2014 COMPENDIUM Edition of National Research and Education Networks in Europe

2014 Edition
of National Research and
Education Networks in Europe
GÉANT Association
COMPENDIUM
compendium.terena.org

GÉANT backbone topology as at January 2014.
What is the GÉANT project?
The GÉANT network is the pan-European research and education network that
interconnects Europe’s National Research and Education Networks (NRENs).
Together, these networks connect over 50 million users at 10 000 institutions
across Europe, delivering a range of networking services for institutions,
projects and researchers.
The GÉANT network, associated services and innovation programme comprise
the GÉANT project (currently administratively known as GN3plus). Co-funded
by the European Commission and Europe’s NRENs, the GÉANT project lies at
the heart of the EU’s e-Infrastructure strategy. It seeks to promote the free,
unimpeded movement of scientific data and knowledge, by connecting and
empowering research and education communities within Europe and other
world regions.
The GÉANT project is a collaboration between 39 partners: 37 European NRENs,
the GÉANT Association’s Cambridge office (formerly DANTE) and Amsterdam
office (formerly TERENA), NORDUnet (representing the 5 Nordic countries). Open
calls allow participation by more partners. The project is managed and run by
the GÉANT Association on behalf of the partners.

ISSN: 1569-4496
2014 Edition
of National Research and
Education Networks in Europe
GÉANT Association
COMPENDIUM

© GÉANT Association 2014
All rights reserved
Parts of this report may be freely copied, unaltered, provided that the
original source is acknowledged and copyright is preserved.
Editor: Bert van Pinxteren
Text, tables and graphs: Bert van Pinxteren, Alper Topçuoğlu
Database and website: Christian Gijtenbeek, Johanna Saladas
Proofreading and correction: Rob Stuart, LocuMotio.nl
Design and production: Omdesign, Omdesign.nl
Printing: Druckhaus Cramer
Paper: Recy Satin
Price: €25 (excl. postage and packing)
For further information or to place an order, please contact:
GÉANT Association
Amsterdam office
Singel 468 D
1017 AW Amsterdam
Netherlands
Email: info@geant.org
The research leading to these results has received funding from the
European Community’s Seventh Framework Programme (FP7 2007–
2013) under Grant Agreement No. 605243 (GN3plus). relating to the project
Multi-Gigabit European Research and Education Network and Associated
Services (GN3plus).
The GÉANT Association is solely responsible for this publication, which does
not represent the opinion of the European Union, nor is the European Union
responsible for any use that may be made of the data appearing herein.

3
GÉANT Compendium of National Research and Education Networks In Europe / Contents
CONTENTS
Introduction 5
Key findings: a brief overview 6
1 Organisational information 9
1.1 European NREN respondents 9
1.2 NRENs in other regions and continents 11
1.3 Legal form of NRENs 14
1.4 Major changes in NRENs 15
1.5 NREN environmental policies 16
2 Client institutions 17
2.1 Overview 17
2.2 Approximate market shares 17
2.3 Numbers of users 21
2.4 Bandwidths 23
2.5 For-profit organisations and international (virtual) research
organisations 25
3 Network and connectivity services 29
3.1 Overview 29
3.2 Core capacity on the routed network 30
3.3 External connectivity: total external links 31
3.4 Dark fibre 32
3.5 Cross-border dark fibre 34
3.6 Network topologies 34
3.7 Major expected network developments 37
3.8 Software-Defined Networking 39
3.9 Network as a Service 40
4 Traffic 41
4.1 Overview 41
4.2 Traffic in 2013 43
4.3 Traffic trends, 2004 -2013 44
4.4 NREN traffic per inhabitant 45
4.5 Congestion 48
4.6 IPv6 deployment 48
4.7 Lambda and point-to-point fixed bandwidth traffic 49
5 Middleware services 53
5.1 Overview 53
5.2 Security services 53
5.3 Authentication and mobility services 58
5.3.1 Identity federations 58
5.3.2 Digital Certificates 61
5.3.3 eduroam 62
6 Collaboration support services 63
6.1 Overview 63
6.2 Network collaboration tools 63
6.3 Networked e-Science resources 65
6.3.1 Cloud services 65
6.4 e-Learning 69
6.5 Clients interaction, and knowledge dissemination 69
6.6 Broker and professional services 70
6.7 Software development 75
7 Funding and staffing 79
7.1 Overview 79
7.2 Staffing 79
7.3 Total budgets, 2009 -2011 and 2012 -2014 82
7.4 Income sources 86
7.5 Expenditure by category 87
7.6 Expenditure by network level 88
APPENDICES 89
1 Alphabetical list of NRENs 89
2 Glossary of terms 91

5
INTRODUCTION
In the more than ten years since its inception, the Compendium has grown into
a much sought-after and authoritative reference source for researchers and
organisations that are interested in the development of research and education
networking. With each successive edition, the information included in the
Compendium has become increasingly varied and dependable, although, as
always, the data should be interpreted with the necessary caution. This year’s
edition of the Compendium – the fourteenth in the series – is produced by the
GÉANT Association, which resulted from DANTE and TERENA joining forces in
October 2014.1
As in previous years, this edition of the Compendium is part of the GÉANT
(GN3plus) project. It has been enhanced with input from activity leaders working
within the scope of that project. Most of the tables and graphs included below
present information compiled from input provided by all GÉANT partner NRENs.
As in previous years, we have examined and attempted to explain multi-year (i.e.
‘longitudinal’) trends. Each section of the Compendium starts with an overview,
which summarises and analyses the most important information presented
in that section. The section on key findings, which follows this introduction,
provides a more general analysis of recent developments in research and
education networking.
This year, for the second time, the Compendium survey was based on a Common
NREN Information Model, which was developed in a joint effort by a worldwide
team of experts: Markus Buchhorn (APAN), Salem Alagtash (ASREN), Thomas
Tam (Canarie), Helmut Sverenyák (CESNET), Marcela Larenas and Florencio
Utreras (RedCLARA), Claudio Allocchio (GARR), Mirjam Kühne (RIPE NCC), Thomas
Lenggenhager (SWITCH), Katrina Sataki (SigmaNet), Tiwonge Banda (UbuntuNet)
and Omo Oaiya (WACREN).
The GÉANT Association hopes that it will be possible to continue this global
collaboration in future years and that this effort will lead to other regional
publications of this type and even to a joint global compendium publication.
Non-European NRENs – except those in Latin America – were invited to submit
their data for inclusion in this edition of the Compendium. Information on Latin
American NRENs is available on the RedCLARA website (www.redclara.net).
Overall, this year’s responses came from a total of 61 NRENs operating in the
same number of countries (49 in Europe and the Mediterranean region; 12
in other parts of the world). All the NRENs were asked to double-check their
responses and ensure that the information was up to date.
Collecting such data requires contributions from, and careful checking by, several
staff members of each NREN. The GÉANT Association would like to thank all those
in the NREN community who gathered, submitted, clarified and checked the data
included in this publication. Where possible, we have also included data from other
sources, such as that on GÉANT traffic, eduroam and the TERENA Certificate Service.
In general, this edition of the Compendium looks back five years, comparing
2014/2013 with 2010/2009. The Compendium consists of two parts: the entire
body of information submitted by the individual NRENs (published online at
compendium.terena.org) and this printed publication. Over the past year, a new
interface has been developed, allowing better comparison of data from several
years and more on-line reporting options than was previously possible. A number
of tables and maps that, in earlier editions, were available in printed format have
been moved to the online publication and are now only available there.
Most of the tables and graphs included below first list all responses from the
GÉANT partner NRENs and then those from other countries. In most cases, the
data are presented in alphabetical order, sorted on the English name of each
country. All the European NRENs included in this Compendium are listed in
Section 1.1. The respondent NRENs in all other regions of the world (except Latin
America) are listed in Section 1.2. In several tables, the responses received from
the NRENs are edited and abridged. The full responses are always available online
at compendium.terena.org.
GÉANT Compendium of National Research and Education Networks In Europe / Introduction
1
See inside cover for further information.

6
GÉANT Compendium of National Research and Education Networks In Europe /Key findings
1
The TNC (formerly: TERENA Networking Conference) is the largest annual European research
networking conference.
2
See Appendix 2 for a definition of the term‘university’.
KEY FINDINGS: A BRIEF OVERVIEW
‘Connected Communities’
Through the GÉANT network, more than 100 countries around the world are
now interconnected with high-speed links, which are dedicated to research and
education.‘Connected Communities’is the theme chosen for the 2015 TNC1
. It
reflects not only the opportunities that are now open to National Research and
Education Networks (NRENs) but also the challenges that are in sight.
As in 2013, this 2014 edition of the Compendium is the product of a global
collaboration effort that has led to an NREN Common Information Model, which
is now being used by regional NREN associations around the world. Policymakers,
funders, NRENs and their users all need to know where the opportunities and
challenges lie and what the trends are. The Compendium seeks to meet this
need by documenting the work currently being done by NRENs, examining the
different contexts in which they operate and explaining important trends in
research and education networking that concern them.
The primary focus of all NRENs is to connect universities and research institutes.
However, many NRENs go beyond this by also connecting institutes of further
education, as well as libraries and museums. Such connections are generally
of very high quality: for universities2
within the GÉANT region, the typical
connection capacity is now above 1 Gb/s — a tremendous increase compared
with the situation a few years ago. All other categories of users have significantly
lower capacities. Based on the data provided by NRENs themselves and our own
conservative assumptions, we estimate that the NRENs in the GÉANT region
provide services to approximately 82% of all university-level students in the
countries involved; that is, a total of 24 million students.
NRENs offer national backbones and international connectivity of exceptional
quality. In most GÉANT partner countries, the typical core capacity is now
10 Gb/s, though some NRENs have reached 100 Gb/s. Germany is planning to
upgrade to Terabit capacity. Some NRENs deploy several paths in their backbone,
effectively increasing the capacity even further. For the first time since the
Please note that, unless otherwise specified, all data indicate the situation on or
around 31 January 2014.
We hope that this fourteenth edition of the Compendium will prove to be at least
as valuable as the previous ones. You are warmly invited to give feedback, which
is the key to the Compendium's future development!
Bert van Pinxteren, GÉANT Association.

7
0% 40%20% 100%80%60%10% 50%30% 90%70%
No
Yes
Impacted by DDoS attacks last year
Implemented DDoS mitigation measures
Impacted by DNS reﬂection attacks
Protection in place
Impacted by route/IP hijacking
Mitigation in place
Security risks, GÉANT NRENs
inception of the Compendium, NRENs have reported that users are experiencing
no congestion at the backbone level.
Connections to the GÉANT backbone are the largest category for GÉANT NRENs.
Connections to GÉANT, to commercial providers, to neighbouring countries via
cross-border fibre and to Internet exchanges together account for 85% of these
NRENs' external capacity.
Within Europe there are still substantial differences: NREN traffic per inhabitant
in Armenia, Azerbaijan, Belarus, Bulgaria, Cyprus, Latvia, Macedonia, Moldova,
Montenegro and Ukraine remains below 15% of the European average level.
In the past few years, traffic growth seems to have slowed down. However, we
believe this is because, increasingly, traffic from very large users is now routed
via separate lambda or other point-to-point connections, over which traffic is not
measured.
Twenty-one GÉANT partner NRENs now
provide dedicated wavelengths (lambdas)
to their customers. Per NREN, the number
of static point-to-point fixed-bandwidth
connections provisioned in January 2014
varies between three and 511 (HEAnet
of Ireland). Within GÉANT, nearly 1300
wavelength circuits are now in use (up from
875 in the year 2012 and 1200 in 2013).
Five NRENs are currently using dynamic
lambdas, which can be set up and taken
down by the NREN or its clients for short-
term requirements.
Growth in services
Authentication and Authorisation Infrastructures (AAIs) are key to giving users
access to services independently of the physical location of both user and
service. All but four of the 42 GÉANT NRENs currently offer, or are planning to
offer, an AAI service. This has enabled the introduction of new services and the
development of collaborative platforms that were previously impossible.
As service access is becoming increasingly independent of the physical location
of the user and/or the service, the need for trust – that is, security services – is
growing.
This edition of the Compendium shows that NRENs are aware of possible security
risks; indeed, over the past year some of them have been impacted by various
types of attacks. Clearly, more can still be done to put mitigation or protection
measures in place.

8
Most of the GÉANT partner NRENs (and a few non-GÉANT NRENs) have joined
or are planning to join the eduGAIN interfederation service — a development
that holds the promise of worldwide service access across federations.
Over the past year, eduroam has shown remarkable growth. It is now available
not only in all GÉANT countries but also in a growing number of other countries
and territories. The number of locations with eduroam availability grew by
25% and the number of authentications more than tripled between January
2013 and January 2014; by September 2014, it was approaching 150 million
authentications per month.
Cloud services not procured via a commercial vendor are currently offered by 17
of the GÉANT partner NRENs, and 12 others are planning to offer such services.
Eighteen of the GÉANT partner NRENs currently provide an e-learning service,
and five others are planning work in this area.
NRENs function as centres of excellence, in service of their clients. This year’s
Compendium survey has identified a number of services being provided by
NRENs in the general category of ’brokerage‘. This seems to be an area in which
NRENs can achieve considerable savings for their customers and where there is
potential for expansion. NRENs are also undertaking framework procurements
for network and related equipment. Ten GÉANT partner NRENs currently offer
some form of support for procurement processes. The majority (21) indicate that
they would be interested in benefiting from framework agreements put in place
by another purchasing body, for a wide palette of services.
Economic and organisational challenges
Indexed on 2009, the average NREN budget in the GÉANT region rose by 2%, even
though average staff size increased by 31%. Although the costs of transmission
capacity and equipment have decreased overall, new services have necessitated
staff size increases.
Comparing 2014 budget data with those from previous editions of the
Compendium reveals that, overall, NREN budgets decreased with respect to 2013.
At the time of going to press, for some NRENs the budget situation is still unclear
because final budget decisions are made in the course of the year. As reported in
previous editions of the Compendium, the overall trend is that, each year, NRENs
are able to deliver more bandwidth and more services for roughly the same
amount of money as in the previous year.
Finally, although it is impossible to make general recommendations on NREN
funding mechanisms, a model that in some way involves all the various
stakeholders in an NREN would seem to provide the best guarantees for its
continued success. In their respective fields, many NRENs are engaged in
innovations, which are often steered by dedicated funding mechanisms. It is
important for NRENs to use such funds to their advantage wherever they exist.

9
GÉANT Compendium of National Research and Education Networks In Europe /Organisational information
1 ORGANISATIONAL
INFORMATION
The GÉANT Compendium, funded by the GÉANT project, is an authoritative
reference on the development of research and education networking in Europe
and beyond. Below, Section 1.1 presents information on the European National
Research and Education Networks (NRENs) that responded to the questionnaire
distributed in May 2014. Map 1.1.1 shows the European NRENs that responded
and a selection of the services that they provide. Section 1.2 includes a
comprehensive list of non-European NRENs, indicating which of them submitted
responses to the questionnaire. Section 1.3 gives an overview of the legal status
of the European NRENs and their relationship with government. Section 1.4
summarises recent major changes in NRENs, their services and/or their users.
Section 1.5 overviews NREN environmental policies.
1.1 European NREN respondents
There are 541
countries in the main area covered by this 2014 edition of the
Compendium (that is, Europe, as well as Mediterranean countries in the Middle East
and North Africa). In three of those 54 countries (Libya, the Palestinian Territory
and Syria) there is either no NREN or we have no knowledge of NREN work there.
In two other countries (Albania and Lebanon) there are NREN initiatives but no
operational NRENs. Replies or partial responses were received from 49 NRENs or
NREN initiatives. In addition, full or partial responses were received from a further
12 NRENs in the same number of countries outside of Europe2
. Map 1.1.1 and
Tables 1.1.1 and 1.2.2 give an overview of the NRENs that submitted responses.
Please note that in most of the tables and graphs included in this edition of the
Compendium, NRENs are identified by abbreviations of their official English names.
Table 1.1.1, which lists the European and Mediterranean NRENs, is divided into two
categories: GÉANT partner countries (42 in total) and other countries.
Table 1.1.1 - European and Mediterranean NRENs included in this Compendium
(GÉANT Association members are shown in bold)
Country NREN URL
GÉANT partner countries
Armenia ASNET-AM www.asnet.am
Azerbaijan3
AzRENA www..azrena.org
Austria ACOnet www.aco.net
Belgium Belnet www.belnet.be
Belarus BASNET www.basnet.by
Bulgaria BREN www.bren.bg
Croatia CARNet www.carnet.hr
Cyprus CYNET www.cynet.ac.cy
Czech Republic CESNET www.cesnet.cz, www.ces.net
Denmark DeIC4
www.deic.dk
Estonia EENet www.eenet.ee
Finland Funet www.funet.fi (www.csc.fi/funet)
France RENATER www.renater.fr
Georgia GRENA www.grena.ge
Germany DFN www.dfn.de
Greece GRNET S.A. www.grnet.gr
Hungary NIIF/HUNGARNET www.niif.hu
Iceland RHnet www.rhnet.is
Ireland HEAnet www.heanet.ie
Israel IUCC www.iucc.ac.il
Italy GARR www.garr.it
Latvia SigmaNet www.sigmanet.lv
Lithuania LITNET www.litnet.lt
Luxembourg RESTENA www.restena.lu
Responses received
No responses received
NREN planned but not operational
No NREN or NREN status unknown
Legend for Table 1.1.1
1
This number excludes Andorra, Kosovo, Liechtenstein, Monaco, San Marino and the Vatican City
State, which are not included separately in this Compendium.
2
Several more responses were received from non-European countries after the deadline for the start
of our analysis had expired; these are available on the Compendium website:
compendium.terena.org.

10
Austria
Morocco
Slovenia
Italy
Turkey
France
Spain
Portugal
United Kingdom
Ireland
Belgium
Netherlands
Luxembourg
Switzerland
Czech Republic
Poland
Algeria
Slovakia
Hungary
Malta
Greece
Denmark
Norway
Iceland
Sweden Finland
Estonia
Latvia
Lithuania
Belarus
Ukraine
Bulgaria
Moldova
Serbia
Montenegro
Macedonia
Israel
Romania
Azerbaijan
Russian Federation
Georgia
Cyprus
Albania
Croatia
Germany
Bosnia/
Herzegovina
Armenia
Kazakhstan
Tunisia
Egypt
Responses received
No responses received
No NREN or no known NREN work
Multimedia repository
Procurement support
Lambdas
SDN testbed
NREN
SERVICE (Red = planned)
Lebanon
Jordan
Map 1.1.1 - European NRENs and selected services

11
Macedonia, FYRo MARNet www.marnet.mk
Malta UoM/RicerkaNet www.um.edu.mt/itservices/about
Moldova RENAM www.renam.md
Montenegro MREN www.mren.ac.me
Netherlands SURFnet www.surfnet.nl
Norway UNINETT www.uninett.no
Poland PIONIER www.pionier.net.pl
Portugal FCCN www.fccn.pt
Romania RoEduNet www.roedu.net
Serbia AMRES www.amres.ac.rs
Slovakia SANET www.sanet.sk
Slovenia ARNES www.arnes.si
Spain RedIRIS www.rediris.es www.red.es
Sweden SUNET www.sunet.se
Switzerland SWITCH www.switch.ch
Turkey ULAKBIM www.ulakbim.gov.tr
Ukraine URAN www.uran.ua
United Kingdom Janet www.ja.net
Other European and Mediterranean countries
Albania ANA www.rash.al
Algeria ARN www.arn.dz
Azerbaijan AzScienceNet www.ict.az, www.azsciencenet.az/en
Bosnia/Herzegovina SARNET 5
jusarnet.net
Egypt EUN www.eun.eg
Jordan JUNet www.junet.edu.jo
Lebanon CNRS www.cnrs.edu.lb
Libya
Morocco MARWAN www.marwan.ma
Table 1.1.1 - continued
Country NREN URL
3
In Azerbaijan, there are two NRENs: AzRENA, which is a partner in the GÉANT network, and
AzScienceNet, which is not.
4
The five Nordic NRENs (DeIC, Funet, RHnet, SUNET and UNINETT) are represented in the GÉANT
Association through NORDUnet.
5
SARNET is active only in the Republika Srpska entity of Bosnia/Herzegovina.
Palestinian Territory
Russian Federation e-ARENA www.e-arena.ru
Syria HIAST www.hiast.edu.sy
Tunisia RNU www.cck.rnu.tn/?lg=fridm=1id=2
Ukraine UARNet www.uar.net/en
Country NREN URL
Other European and Mediterranean countries
1.2 NRENs in other regions and continents
Table 1.2.1 lists sources of information on NRENs in other regions and continents.
Area Organisation/project URL
Arab states ASREN www.asrenorg.net
Eastern and Southern Africa Ubuntunet Alliance www.ubuntunet.net
West and Central Africa WACREN www.wacren.net
Asia/ Pacific APAN www.apan.net
Central Asia CAREN caren.dante.net
Latin America CLARA www.redclara.net
Caribbean CKLN www.ckln.org
Canada CANARIE www.canarie.ca
USA Internet2 www.internet2.edu
ESnet www.es.net
National Regional
Networks consortium
www.thequilt.net
Tables 1.2.1 - Information on non-European NRENs
Several projects that aim to connect research communities around the globe to
the GÉANT network are listed at www.geant.net/Network/GlobalConnectivity.

12
United States
A T L A N T I C
P A C I F I C
P A C I F I C
I N D I A N
A R C T I C
O C E A N
O C E A N
O C E A N
O C E A N
O C E A N
Canada
Chile
Colombia
Malawi
Mozambique
Nepal
New Zealand
Taiwan
Mexico
Cuba
Guatemala
Argentina
Bolivia
Venezuela
Peru
Ecuador
Paraguay
Uruguay
El Salvador
Brazil
Kenya
Sudan
Somalia
Ethiopia
South Africa
Tanzania
Uganda
Ghana
United Arab
Emirates
India
Pakistan
Iran
Turkmenistan
Uzbekistan
Kazakhstan
Kyrgyzstan
Tajikistan
Bangladesh
China
Mongolia
Vietnam
Hong Kong
Thailand
Singapore
Indonesia
Malaysia
Philippines
Australia
JapanRepublic of
Qatar
Sri Lanka
Egypt
Oman
Korea
Zambia
Togo
Costa Rica
Ivory Coast
Nigeria
Senegal
Malagasy Republic
Trinidad and Tobago
Operational NREN, information
included in the Compendium
Operational NREN, information
not included in the Compendium
No NREN or no known NREN work
in this country
NREN
Map 1.2.1 - NRENs known to be operating in other countries

13
Country NREN URL
Argentina INNOVA|RED www.innova-red.net
Australia AARNet www.aarnet.edu.au
Bangladesh BdREN www.bdren.net.bd
Bolivia ADSIB www.adsib.gob.bo
Brazil RNP www.rnp.br
Canada CANARIE www.canarie.ca
Chile REUNA www.reuna.cl
China CERNET www.edu.cn
China CSTNet www.cstnet.net.cn
China (Hong Kong) HARNET www.harnet.hk
Colombia RENATA www.renata.edu.co
Operational NREN, information included in this Compendium
Operational NREN, information not included in this Compendium
No NREN or NREN status unknown
Table 1.2.2 lists those NRENs and NREN initiatives of which GÉANT is currently
aware beyond the region covered by this Compendium. Note that this list is not
comprehensive: there may be other NRENs of which we have no knowledge.
Also, in some countries the situation may be subject to rapid change. Twelve
NRENs from non-European countries submitted data for this Compendium; they
are highlighted in green. Their full responses are available at
Further information on Latin American NRENs is published in the CLARA
Compendium of Latin American National Research and Education Networks,
available at www.redclara.net/compendio.
Costa Rica CONARE www.cenat.ac.cr/computacion-avanzada/red-
conare
Cuba RedUNIV www.reduniv.mes.edu.cu
Ecuador CEDIA www.cedia.org.ec
El Salvador RAICES www.raices.org.sv
Ethiopia EthERNet
Ghana GARNET www.garnet.edu.gh
Guatemala RAGIE www.ragie.org.gt
India ERNET www.eis.ernet.in
Indonesia INHERENT-DIKTI www.inherent-dikti.net
Iran IRANET/IPM www.iranet.ir
Ivory Coast RITER www.riter.ci
Japan SINET www.sinet.ad.jp
Japan NICT www.jgn.nict.go.jp/english
Kazakhstan KazRENA www.kazrena.kz
Kenya KENET www.kenet.or.ke
Korea, Republic Of KOREN www.koren.kr
Korea, Republic Of KREONET www.kreonet.re.kr/en_main
Kyrgyzstan KRENA-AKNET www.krena.kg
Malagasy Republic iRENALA www.irenala.edu.mg
Malawi MAREN www.malico.mw/maren
Malaysia MYREN www.myren.net.my
Mexico CUDI www.cudi.edu.mx
Mongolia ErdemNET www.erdemnet.mn
Mozambique MoRENet
Nepal NREN www.nren.net.np
New Zealand REANNZ www.reannz.co.nz
Nigeria ngREN ngren.edu.ng
Oman OmREN www.trc.gov.om
Pakistan PERN www.pern.edu.pk
Country NREN URL
Table 1.2.2 - NRENs known to be operating in other countries

14
Paraguay Arandu www.arandu.net.py/cms/index.php
Peru RAAP www.raap.org.pe
Philippines PREGINET www.pregi.net
Qatar Qatar Foundation www.qf.qa
Senegal SnRER snrer.edu.sn
Singapore SingAREN www.singaren.net.sg
Somalia Somaliren www.somaliren.org
South Africa SANReN www.sanren.ac.za
Sri Lanka LEARN www.learn.ac.lk
Sudan SudREN www.suin.edu.sd
Taiwan TWAREN www.nchc.org.tw/en
Tajikistan TARENA www.tarena.tj
Tanzania TERNET www.ternet.or.tz
Thailand ThaiREN www.thairen.net.th
Thailand UniNet www.uni.net.th/UniNet/Eng/index_eng.php
Togo RNERT www.togorer.tg
Trinidad and Tobago TTRENT www.ttrent.edu.tt
Turkmenistan TuRENA www.science.gov.tm/en/turena
Uganda RENU www.renu.ac.ug
United Arab Emirates ANKABUT www.ankabut.ae
United States Internet2 www.internet2.edu
Uruguay RAU www.rau.edu.uy
Uzbekistan UzSciNet www.uzsci.net
Venezuela REACCIUN www3.reacciun.ve
Vietnam NASATI www.vista.gov.vn
Zambia ZAMREN www.zamren.zm
Country NREN URL
1.3 Legal form of NRENs
NRENs have various legal forms. NREN names and their English translations
may be misleading: what is called a‘foundation’in one country may be quite
different from a foundation in another country. The same is true of several other
designations, including‘association’. This section distinguishes two parameters
which, together, help to characterise the legal form of an NREN:
1) Its relationship with government; and
2) Whether it is a separate legal entity.
In some countries, there is a distinction between the name of the physical
network (e.g. PIONIER in Poland or Funet in Finland) and the name of the
organisation that runs it. Thus, Funet is run by an operational unit within CSC,
an organisation that performs several other functions within Finland, such as
high-performance computing. Although the two parameters cited above can
usefully characterise the legal form of an NREN, the classification is not always
straightforward.
Relationship with government
In this Compendium, we distinguish three situations:
a) A NREN that is under direct government control. Such is the case if an NREN is
a government agency or parastatal, or a part thereof.
b) A NREN operating independently of government to a certain extent; for
example, those that are separate legal entities with governing boards at least
half of whose members are government-appointed. Also, some NRENs that
are government agencies enjoy a certain degree of autonomy comparable to
that of NRENs that are separate legal entities.
c) A NREN having no direct government ties, even though, typically, the
majority of their client institutions are largely government-funded.

15
Separate legal entity
Many NRENs operate as separate legal entities.
A combination of the two parameters leads to six categories, as shown in
Map 1.3.1:
Legal entity? Separate legal
entity
Not separate
legal entity
Not under direct government control
Largely government-controlled
Entirely government-controlled
Level of government control
In Portugal, FCCN has recently merged with FCT, bringing it under closer
government control.
It seems self-evident that for an NREN to develop, the commitment of all its
major stakeholders − including funders and users − is required. A governing
model that allows all such stakeholders to participate would seem to be the most
appropriate; such a situation can be achieved in various ways.
NRENs that can operate with a certain degree of independence from their
respective governments may have distinct advantages, such as easier decision-
making processes and the ability to recruit and retain suitably qualified staff,
partly by setting salaries at competitive levels. This may partially explain why this
model is more common in countries where, after many years of development,
research and education networking is well-established.
1.4 Major changes in NRENs
All the NRENs covered by this 2014 edition of the Compendium were requested
to briefly describe any major changes in their mandate or remit, user-base, or
technology and services that occurred in the past year or were expected to occur
in the coming year. The full responses are available at
compendium.terena.org/reports/major_changes.
Several NRENs report that they have further upgraded and extended their
networks. Several also mention the development of cloud storage and other
services. CARNet of Croatia reports a great expansion of its services to the
secondary education sector. In Hungary and Ireland, secondary schools were
integrated into the NREN network.
Map 1.3.1 - Legal form of NRENs

16
7
Available at www.geant.net/Network/Environmental-Impact/Pages/Policy.aspx
1.5 NREN environmental policies
Environmental issues started to feature on NREN agendas a few years ago. NRENs
and their users began to realise that it is important to address such issues, for
example by measuring and reducing energy consumption, and by promoting
green uses of network technology in order to reduce greenhouse gas (GHG)
emissions. Various NRENs have made progress on environmental issues.
As part of the GN3plus project, five NRENs (AMRES, GRNET, HEAnet, NIIF/
HUNGARNET and SURFnet) have audited their GHG emissions. The full report is
available at: www.geant.net/Resources/Media_Library/Documents/D3-3_
DN3-3-1_Environmental-Impact-Report-2014.pdf.
This report shows that although the quantity of emissions differs considerably
from country to country, the overall trend seems to be upward. One explanation
for the differences is that GRNET of Greece has a large data centre which has
expanded over the years, whereas the other NRENs do not. In the case of
SURFnet of the Netherlands, differences can be explained by changes in the
underlying measurement methods, which are still under development. In the
cases of Ireland and Hungary, savings in some areas were more than offset by
expanded networks and activities in other areas, such as storage and computing.
The GÉANT project team aims to develop metrics that are independent of such
developments, such as energy/bit.
Eight GÉANT NRENs (Belgium, Greece, Hungary, Ireland, Netherlands, Portugal,
UK and Ukraine) report that they have an environmental policy in place, whereas
only four did in 2012. Now that the GÉANT project has developed a template
for such policies7
, a further increase is expected. The Irish and UK policies are
available at the respective NREN websites:
www.heanet.ie/about/environmental_policy
www.ja.net/about-janet/environmental-policy

17
2 CLIENT INSTITUTIONS
Section 2.2 indicates the‘market shares’; that is, how many institutions in various
categories are actually connected to the NREN; Section 2.3 estimates the numbers
of connected users that correspond to these market shares; an expanded Section
2.4 documents typical bandwidths; Section 2.5 focuses on dark fibre connections
to client institutions. Finally, Section 2.6 examines connectivity to for-profit
organisations and international (virtual) research organisations.
2.1 Overview
As indicated in previous editions of the Compendium, all the NRENs covered
by this publication are allowed to connect universities and research institutes.
Nearly all may also connect institutes of further education, as well as libraries
and museums. Information on connection policies is not repeated in this year’s
edition: even though NRENs differ greatly in this respect, there have been no
significant changes in the past year.
Even if an NREN is allowed to connect a certain institution, it does not necessarily
do so. In the university sector, NRENs obviously have very high market shares.
Based on data provided by NRENs themselves and our own conservative
assumptions, we estimate that the NRENs in the GÉANT region provide services
to approximately 82% of all university-level students in the countries involved;
that is, a total of 24 million students.
For universities1
within the GÉANT region, the typical connection capacity is
now above 1 Gb/s — a tremendous increase compared with the situation a few
years ago. All other categories of users have significantly lower capacities. As
documented in Section 2.4, differences in bandwidth exist not only between but
also within countries.
Only a handful of NRENs connect for-profit organisations and/or international
(virtual) research organisations. In such cases, the number of organisations
1
See Appendix 2 for a definition of the term‘university’.
GÉANT Compendium of National Research and Education Networks In Europe /Client institutions
connected is very limited; in most cases these connections are made specifically
to access services provided by the research and education sector.
2.2 Approximate market shares
Table 2.2.1 gives an overview of the number of institutions in each user-category
and indicates the percentage of users that are serviced by each NREN. Note that
Compendium respondents submitted only approximate percentages.
Many NRENs operating in a strong hierarchy of Metropolitan or Regional Area
Networks (MAN/RAN) were unable to provide connection figures but did
indicate that they service high percentages of their respective communities. For
additional information on individual NRENs, see the Compendium website:
compendium.terena.org.
For this year’s edition of the Compendium, we asked NRENs to use the new (2011)
ISCED classification2
of educational institutions. More clearly than the previous
scheme, this new classification differentiates tertiary institutions offering courses
below Bachelor level from those offering Bachelor- and higher-level courses (that
is, at ISCED levels 6, 7 and 8). We asked NRENs to provide combined data for all
tertiary education institutions in the latter category. In some countries, including
Ireland, only institutions offering education up to level 8 (the doctoral level) are
considered to be true universities. In the Netherlands, for example, a distinction
is made between universities of applied sciences and other universities.
Unfortunately, because of this new ISCED classification, for universities and
other institutions at ISCED levels 6, 7 and 8 the figures for 2014 are not directly
comparable with those for previous years. In France, a number of secondary
and primary schools are connected, but indirectly and therefore the numbers
connected are not known. Similar situations may exist in other countries as well.
2
www.uis.unesco.org/Education/Pages/international-standard-classification-of-education.aspx

18
Legend NREN-connected institutions
80%
60 - 80%
40 - 60%
20 - 40%
20%
Unknown or not applicable
No answer provided
Country Universities Institutes of
further education
Research institutes Secondary schools Primary schools Libraries,
museums,
archives, cultural
institutions
Hospitals (other
than university
hospitals)
Government
departments
(national, regional,
local)
Armenia 3 35 4 2
Austria 48 27 12 5 23
Belarus 10 59 17 5 8
Belgium 63 5 37 5 0 15 20 48
Bulgaria 22 42
Croatia 318 20 62 424 1670 70 36 134
Cyprus 8 3 4
Czech Republic 35 23 32 77 12 40 43 34
Denmark 8 7 22 0 0 5 0 0
Estonia 22 14 20 43 8 58 0 9
Finland 48 10 7 8
France 409 340 377 19 7 19
Georgia 9 19 4 3 3 1 3
Germany
Greece 40 15
Hungary 25 28 73 1580 3050 560 5 4
Iceland 7 2 8 1 1
Ireland 27 10 12 800 3200 0 0 13
In the past, almost all NRENs responded that they connected all or nearly all
the universities in their country. However, since 2013, several NRENs have
reported for the first time that they connect a significantly lower percentage of
the universities in their country. Apparently, this is not because the number of
connected universities has decreased but because newer institutions providing
education at ISCED level 6 in these countries are not connected to the Internet
via the NREN.
Table 2.2.1 – Approximate market shares

19
Israel 12 0 4 0 0 1 0 0
Italy 99 0 329 104 19 39 56 3
Latvia 11 2 8 3 0 3 0 0
Lithuania 42 53 32 540 37 29 4 31
Luxembourg 7 0 21 61 226 14 0 15
Macedonia 1 1 1 0 0 0 0 0
Malta 1 3
Moldova 7 2 36 1 0 10 10 1
Montenegro 19 1 2 2 1
Netherlands 46 32 33 20 4
Norway 37 20 74 2 4 17 0 0
Poland 199 20 199 244 64 47 157
Portugal 48 0 11 0 0 1 0 9
Romania 50 10 56 350 140 40 30
Serbia 87 9 43 13 5 21 6 2
Slovakia
Slovenia 4 17 53 154 546 212 0 24
Spain 93 6 199 48 57 32
Sweden
Switzerland 44 2 11 2 0 1 0 6
Turkey 162 16 1 13
UK 167 428 36 8 6
Ukraine 80 10 7 3
Other countries
Albania
Algeria 58 36 15 0 6
Australia 58 16 28 61 80 8 6 8
further education
museums,
archives, cultural
institutions
Hospitals (other
than university
hospitals)
Government
departments
local)
Table 2.2.1 – continued

20
Brazil 826 3 167 11 154 53 77
Canada 190 2000 24 62 127
Japan
Kazakhstan
Korea 51 0 70 1 0 70 0 16
Lebanon 0 0 5 0 0 0 0 0
Morocco 94 9 5 0 0 2
Russian Federation 250 140
South Africa 25 5 26 0 2 0 0
Taiwan 90 5 5 1000 1000 5 1 5
USA
further education
museums,
archives, cultural
institutions
Hospitals (other
than university
hospitals)
Government
departments
local)
Other countries

21
2.3 Numbers of users
In the questionnaire for this edition of the Compendium, NRENs were asked
to estimate the numbers of university students they serve. This information
has been combined with enrolment figures derived from UNESCO statistics3
to produce Graphs 2.3.1 and 2.3.2, showing estimated numbers of university
students served by GÉANT NRENs. Note that, in those cases where NRENs were
unable to give such an estimate, we have made conservative assumptions.
For example, PIONEER in Poland is structured as a network connecting the
country’s regional networks, which, in turn, connect all or nearly all of the Polish
universities. We have therefore assumed that 90% of students in Poland benefit
from PIONEER’s services.
Based on the data received and the assumptions explained above, we estimate
that NRENs in the GÉANT region (except Azerbaijan) provide services to
approximately 82% of all university-level students in those 41 countries; that is, a
total of 24 million students4
.
The data above also suggest that (as was the case in 2013) approximately five
million students in the GÉANT region are not serviced by an NREN, the largest
numbers being in the Ukraine and Belarus.
Of course, not only whether a student receives NREN services but also the quality
of service provided must be taken into account for an accurate impression of the
present situation. Map 2.4.3, which indicates the spread of bandwidth between
the universities within each country, shows that for a number of countries the
typical university connection capacity is less than 1 Gb/s.
Section 4.4 provides another indicator, showing much higher traffic per
inhabitant for the UK than for Turkey, for example.
3
In a few cases, Eurostat data were used.
4
In 2012, the GÉANT region was smaller: 36 countries, whose market share was approximately 88%,
a total of around 20.5 million students.
0
4 500 000
3 500 000
5 000 000
2 000 000
Numberofstudents
4 000 000
3 000 000
500 000
1 500 000
2 500 000
France
Germany
Belarus
Italy
Spain
UK
Netherlands
Turkey
Ukraine
Greece
1 000 000
Poland
Romania
Number of students
Students connected
Graph 2.3.1 – Number of students serviced by an NREN, GÉANT region,
for countries with 500 000 students

22
Malta
Iceland
Luxembourg
Cyprus
Slovenia
Moldova
Macedonia
Georgia
Croatia
Montenegro
Estonia
Latvia
Ireland
Slovakia
Lithuania
Norway
Denmark
Serbia
Bulgaria
Finland
Austria
Hungary
Israel
Switzerland
Portugal
CzechRepublic
Sweden
Belgium
Number of students
Students connected
0
450 000
350 000
500 000
200 000
Numberofstudents
400 000
300 000
50 000
150 000
250 000
100 000
Graph 2.3.2 – Number of students serviced by an NREN, GÉANT region,
for countries with 500 000 students

23
2.4 Bandwidths
The 2008 edition of the Compendium stated that:
In 2003, the ‘average’ university was connected at Megabit capacity; by
2008, that had changed to Gigabit capacity.
The typical capacity for universities in GÉANT partner countries now exceeds
1 Gb/s. All other user-categories have lower capacities, although research
institutions are generally close to 1 Gb/s.
0
10%
30%
20%
40%
50%
60%
Institutes of further education
Universities and other
Secondary schools
Research institutes
10 Gb/s100 Mb/s 1 Gb/s =100 Gb/s40 Gb/s 100 Gb/s10 Mb/s
Graph 2.4.1 gives an overview of the spread of bandwidths available to NREN
users. Note that not all NRENs provided information relevant to this overview,
so the set of countries is not exactly the same in each user-category.
Table 2.4.2 shows the information received from non-GÉANT countries.
Country A B C D E F G H
Albania
Algeria
Australia
Brazil
Canada
Hong Kong
Japan
Kazakhstan
Korea
Kyrgyzstan
Lebanon
Morocco
New Zealand
Russian Federation
South Africa
Taiwan
40 Gb/s
10 Gb/s
1 Gb/s
100 Mb/s
10 Mb/s
Legend
A Universities and other
B Institutes of further education
C Research institutes
D Secondary schools
E Primary schools
F Libraries, museums, archives, cultural institutions
G Hospitals (other than university hospitals)
H Government departments (national, regional, local)
Graph 2.4.1 – Typical bandwidth, GÉANT partner countries
Table 2.4.2 - Typical bandwidth, non-GÉANT countries

24
We have also considered the spread within individual countries. It should be
noted that in some countries all or most institutions in a particular category are
connected at similar capacities, whereas in other countries there may be large
capacity differences at national level.
To give an impression of the spread of bandwidths among universities within
individual countries, we have categorised NRENs according to the spread
between the lowest and highest connection categories, using the seven
categories indicated in Table 2.4.2. Those countries with no more than one
category difference between lowest and highest have been designated as being
‘even’, those with two or three categories difference as having‘moderate spread’,
and those with a larger difference as being‘differentiated’. Each category is
subdivided into those countries with a typical university capacity of less than 1
Gb/s and those with a typical capacity of 1 Gb/s or more.
The result is shown in Map 2.4.3. The NRENs with the lowest-capacity
connections are Albania, Belarus, Georgia, Macedonia and Ukraine.
We have examined the technological aspect in greater detail by asking NRENs
what connectivity services they provide to their client institutions that are
connected with dark fibre. The replies are summarised online at compendium.
terena.org/reports/non_ip_connectivity. With a dark fibre infrastructure,
NRENs can offer virtual private networks (VPNs) as a service, often marketed as
LAN extensions. In this way, a single university campus and its LAN can extend
across multiple separate sites, which can bring significant savings in service and
support costs for university IT departments.
Even (typical bandwidth 1 Gb/s)
Even (typical bandwidth) ≥1 Gb/s)
Moderate spread (typical bandwidth 1 Gb/s)
Moderate spread (typical bandwidth ≥ 1 Gb/s)
Differentiated (typical bandwidth ≥ 1 Gb/s)
Legend for Map 2.4.3
Map 2.4.3 – Spread of bandwidth among universities within countries

25
2.5 For-profit organisations and international
(virtual) research organisations
We asked the NRENs whether they connect commercial, for-profit organisations.
Several do so, though in most cases this is specifically to access services provided
by the research and education sector. In all cases, the reported number is quite
limited, the highest being 31 (for Croatia).
We also asked NRENs whether they connect international (virtual) research
organisations. In contrast to 2013, many NRENs reported that they do this, though
the numbers of connected organisations are no higher than five (for the UK). Note
that all NRENs have an Acceptable Use Policy, to which all their clients must adhere.
Country Connects for-profit organisations? Connects
international
(virtual) research
organisations?
Activity related
to virtual
research
communities?
Brief description:
Armenia No Yes Yes ATLAS, ALICE, ENVIROGRIDS
Belarus No Yes Yes
Belgium No Yes Yes * Flemish Supercomputer Center (www.vscentrum.be): supercomputers of five universities are
interconnected/clustered over our backbone via 10 GE lightpaths for shared use by researchers (further
information: https://vscentrum.be/nl/en). A similar initiative has started in the Walloon region (further
information: http://www.ceci-hpc.be/)
Bulgaria Specifically to access services
provided by the research and
education sector
Yes No
Croatia Specifically for offering services to
the research and education sector
Yes No
Cyprus Specifically for offering services to
Yes No
Czech Republic No Yes Yes Auger observatory, EGI, ELIXIR, XiFi
Denmark On the same basis as RE
organisations
Yes Yes Connection between CERN and Niels Bohr Institute, Copenhagen.
Finland Specifically for offering services to
Yes Yes NeIC (http://neic.nordforsk.org/) operates the Nordic WLCG Tier-1 center; ELIXIR
France On the same basis as RE
organisations
Yes Yes We collaborate with communities including LHC and PRACE to understand their needs and adapt the
network infrastructure
Georgia No Yes Yes On Grid site several VOs are registered
Germany Specifically for offering services to
Yes Yes We provide network services (IP, Lambda), security services (PKI, CA) and participate in testbeds
Table 2.5.1 – Connections to for-profit and to international research organisations

26
Greece No Yes Yes GRNET participates in EGI.eu and coordinates the HP-SEE (http://www.hp-see.eu) project. It serves
several Virtual Research Communities including Life Sciences, Biomed, Computational Chemistry and
Computational Physics. GRNET will also participate as computational infrastructure provider in the Greek
national CLARIN center (CLARIN-EL project)
Hungary Under certain circumstances Yes Yes Advanced videoconferencing facilities and service, extended basic investigations in VRE (Virtual Research
Environments) and VRC (Virtual Research Communities), including eduCONF and WebRTC
Iceland Specifically for offering services to
Yes No We host a Mantycore node
Ireland Specifically for offering services to
Yes Yes [1] HEAnet client ICHEC is very active in the PRACE project: http://www.ichec.ie/research/prace_proto.
php [2] HEAnet client Marine Institute is involved in the Smartbay project: http://www.marine.ie/home/
research/ProjectsDatabase/CurrentProjects/SmartBay [3] HEAnet clients WIT and TSSG are involved in
the Finessey and XiFi projects [4] Many Irish Researchers take part in FP7, ESFRI, PRACE, BBMRI, ECRIN,
EUROBIOIMAGING, ISBE, DARIAH, ESS, SHARE, EMSO and EPOS projects using IP connectivity including:
Tyndall, CTVR and Clarity
Israel No Yes Yes
Italy Under certain circumstances Yes No
Latvia No Yes Yes We participate in the National Grid Initiative and the European Grid Initiative
Lithuania Specifically for offering services to
Yes No
Luxembourg Specifically for offering services to
No No
Macedonia No Don’t know No 2 clusters are included in EGI-InSPIRE (http://www.egi.eu/about/egi-inspire/), HP-SEE (http://www.hp-see.
eu/)
Malta Specifically for offering services to
Yes No
Moldova Under certain circumstances Yes Yes We provide connectivity to support various VRCs that are using resources of the National Grid
Infrastructure; we also provide access to VRCs that use regional SEE HPC infrastructure
Netherlands Under certain circumstances Yes Yes Community support, outreach programme, cloud services, middleware support, pilots
Norway No Yes Yes Connected to the worldwide LHC Computing Grid
Poland Specifically for offering services to
Yes Yes For users including AstroGrid-PL, HEPGrid, QosCosGrid, LHCONE, VLBI
Portugal Under certain circumstances Yes Yes We manage the datacentre that houses the Portuguese Tier-2 node for CERN
international
(virtual) research
organisations?
Activity related
to virtual
research
communities?
Brief description:

27
Romania No Yes No
Russian
Federation
Under certain circumstances Yes Yes LHC Computing GRID, IVOA
Serbia No Yes No
Slovakia Specifically for offering services to
No No
Slovenia No Yes Yes Cluster supports Atlas VO, IPv6 testing, CVMFS testing
Spain Specifically for offering services to
Yes Yes For-profit organisation connections only for R+D departments of companies participating in research
projects and only for the duration of the project. Virtual research communities: LHC and DEISA. We are
members of the Spanish Grid Initiative and participate in the EGI-InSPIRE project.
Sweden Specifically for offering services to
Yes Yes DEISA, LHC
Switzerland Specifically for offering services to
Yes Yes Member of the Swiss NGI
Turkey No Yes Yes Atlas, CMS, BIOMED, SEE, BELLEII, EUMED and 4 national VOs (TRGRIDA, TRGRIDB, TRGRIDD, TRGRIDE)
UK Specifically for offering services to
Yes No
Ukraine Specifically for offering services to
Yes No
Other countries
Albania Specifically for offering services to
No
Algeria No Yes Yes We participate in EUMEDGRID infrastructure with EUMED VO use for some of our researchers
Australia Specifically for offering services to
Yes No
Brazil Under certain circumstances Yes No
Canada Specifically for offering services to
Yes Yes We provide connectivity to advanced computing consortia and fund development of middleware to serve
VRCs
Japan Specifically for offering services to
Yes Yes APAN
Kazakhstan No Yes No
international
(virtual) research
organisations?
Activity related
to virtual
research
communities?
Brief description:

28
Korea Specifically for offering services to
Yes Yes LHC Computing Grid, GLIF, GLORIAD, PRAGMA, CINEGRID, APAN
Kyrgyzstan No Yes Yes www.KyrgyzstanVSL.org
Lebanon Specifically for offering services to
Yes No
Morocco No Yes Yes Morocco is a member of the ATLAS collaboration (LHC).
New Zealand Specifically for offering services to
Yes No
South Africa No Yes Yes Via SAGRID we enable: biomed - life sciences, enmr.eu - structural biology, ATLAS collaboration production,
ALICE collaboration production, GEANT4 collaboration, EELA - Latin America Catch-All VO, sagrid.ac.za -
South African Catch-All VO, EUMED - North-African Catch-All VO
Taiwan No Yes Yes We have a direct lightpath connection to the iGENI platform and run OpenFlow collaboration on top of it.
USA Yes Yes
international
(virtual) research
organisations?
Activity related
to virtual
research
communities?
Brief description:

29
GÉANT Compendium of National Research and Education Networks In Europe /Network and connectivity services
3 NETWORK AND
CONNECTIVITY SERVICES
This section provides insights into several important network characteristics.
Section 3.2 presents information on core capacity on the routed network. Section
3.3 examines the capacity of NREN external connections. Section 3.4 documents
recent developments in dark fibre, and Section 3.5 focuses on cross-border
dark-fibre links. Section 3.6 examines network topologies. Section 3.7 includes
an overview of expected major developments in research and education
networking. Sections 3.8 and 3.9 look at Software-Defined Networking (SDN) and
Network as a Service (NaaS), respectively.
3.1 Overview
In most GÉANT partner countries, the typical core capacity is now 10 Gb/s,
though some NRENs have reached 100 Gb/s. Because many NRENs in the GÉANT
region have access to dark fibre (see Section 3.4), which is potentially able to
handle high capacities, they can increase capacity easily and economically
whenever required. Some NRENs allow several paths in their backbone,
effectively increasing the capacity even further.
In other OECD countries, the trend is similar. Australia and the USA, for example,
have both introduced 100 Gb/s capacities.
In several cases, the NREN’s network is structured as a mesh, having redundant
core and access links. Thus, SURFnet indicates that its core capacity is 200 Gb/s,
even though the capacity of the individual links is 100 Gb/s.
In general, connections not only to the European academic backbone network
(i.e. GÉANT) but also to the general Internet are crucial to NRENs. Connections to
the GÉANT backbone are the largest category for GÉANT NRENs. Connections to
GÉANT, to commercial providers, to neighbouring countries via cross-border fibre
and to Internet exchanges together account for 85% of these NRENs' external
capacity
For the purposes of the Compendium, we have been monitoring the uptake of
dark fibre by NRENs since 2005. At that time, only a few networks used dark fibre
in their backbones, and GÉANT was just starting to use dark fibre and light it for
transnational trunks.
This year (2014), the aggregate length of dark fibre used internally by NRENs in
the GÉANT region exceeds 139 000 km, approximately the same as in 2013.
Within an NREN there are many ways to use dark fibre cost-effectively, all of
which are focused on enhanced services to clients and users. NRENs, as a result
of moving from managed network links to their own transmission infrastructure,
have been able to develop new features and services at various levels.
Another continuing development is the implementation of cross-border dark-
fibre links between NRENs. Section 3.5 presents information on new links
implemented since January 2013.
In the past, the logical and physical structures of a network were roughly the
same. This is no longer always the case, for it has become possible to define
logical routes in a network independently of the physical routes. This presents
advantages including lower latency and the possibility of creating logical links
of greater capacity than that of the largest individual link. Section 3.6 gives an
example of this.
Six GÉANT partner NRENs currently have a testbed for Software-Defined
Networking (SDN). This enables the creation of virtual networks using the
underlying infrastructure, though with varying characteristics and topologies,
adapted to user needs. Ten NRENs have plans in this area.
Network as a Service (NaaS) is currently being considered in three countries;
several more have plans in this area.

30
Major expected network developments reported by NRENs include:
• In the OECD countries, dark-fibre networks are already in place and being
upgraded and extended. Several NRENs are preparing for 100 Gb/s or even
multiples thereof; Germany is planning an upgrade to Terabit capacity.
• Many NRENs also report that they have expanded the types of services they
offer to clients.
0
1
2
3
4
5
Corecapacities(Gb/s)
Belarus
Belgium
Bulgaria
Croatia
Cyprus
CzechRepublic
Estonia
France
Georgia
Germany
Greece
Hungary
Ireland
Israel
Italy
Latvia
Lithuania
Luxembourg
Austria
Armenia
Denmark
Finland
Iceland
Malta
Montenegro
Netherlands
Portugal
Serbia
Slovakia
Slovenia
Spain
Switzerland
Turkey
UK
Ukraine
Moldova
Norway
Romania
Poland
Sweden
≥ 10 Gb/s, 40 Gb/s
≥ 1 Gb/s, 10 Gb/s
1 Gb/s
≥ 100 Gb/s
≥ 40 Gb/s, 100 Gb/s
1
2
3
4
5
Graph 3.2.1 – Core capacities, GÉANT partner NRENs
3.2 Core capacity on the routed network
The term‘core usable backbone capacity’means the typical core capacity of the
linked nodes in the core. Some NRENs have dark fibre with a very high theoretical
capacity; in such cases, we requested data on the usable IP capacity.
In most GÉANT partner countries, the typical core capacity is now 10 Gb/s,
though some NRENs have reached 100 Gb/s. The trend is similar in non-GÉANT
countries that submitted data for this year’s Compendium survey. Both Australia
and the USA have introduced capacities of at least 100 Gb/s. Graphs 3.2.1 and
3.2.2. present the data for the individual countries.
As many NRENs in the GÉANT region have access to dark fibre (see Section 3.4),
which is potentially able to handle high capacities, they can increase capacity
easily and economically whenever required. In addition, as indicated in Section
3.4, many NRENs now have several point-to-point circuits and lightpaths, which
offer additional capacity that is not usually included in normal traffic statistics.

31
0
1
2
3
4
5
Corecapacities(Gb/s)
Bosnia/Herzegowina
Brazil
Canada
HongKong
Korea
Lebanon
NewZealand
Sudan
SouthAfrica
Taiwan
USA
Australia
Algeria
Morocco
RussianFederation
= 10 Gb/s, 40 Gb/s
= 1 Gb/s, 10 Gb/s
1 Gb/s
= 100 Gb/s
= 40 Gb/s, 100 Gb/s
Knowing the typical capacity of the links is no longer sufficient. In several cases,
the NREN’s network is structured as a mesh, having redundant core and access
links. For example, GARR (Italy) allows load-balanced traffic between two points
using multiple paths through the mesh, which means that the GARR network’s
effective core capacity is not 60 Gb/s (the typical link capacity) but 600 Gb/s
(the capacity of the mesh as such). Similarly, SURFnet indicates that its core
capacity is 200 Gb/s, even though the capacity of the individual links is 100 Gb/s.
SUNET finds it impossible to state a figure for its core capacity, because it has a
“redundant star-shaped backbone”.
It should be noted that network capacity grows stepwise rather than linearly.
3.3 External connectivity: total external links
The NRENs covered by this edition of the Compendium were asked to list all their
external connections as of January 2014.
It should be noted that the Nordic NRENs (DeIC of Denmark, Funet of Finland,
RHnet of Iceland, SUNET of Sweden and UNINETT of Norway) share their external
connections through NORDUnet.
In general, connections to GÉANT and to other NRENs carry research and
education traffic, whereas peerings and other connections convey traffic to and
from the general Internet. Research and education traffic may consist of highly
specialised data and is often transmitted in huge volumes within very short
time-frames; for example, real-time observational data from a radio telescope,
which must be transmitted over large distances for pre-processing and storage.
As high traffic peaks can be expected on such links, they must be dimensioned
to accommodate them; it is not unusual to see a flow of 1 Gb/s generated by a
single high-end researcher. Thus, the average volume of traffic is not a reliable
indicator of the required capacity of the link.
In contrast, traffic to and from the general Internet tends to be aggregated and
smoothly varying. It comprises a large number of small-to-medium data flows,
which combine to produce a fairly predictable traffic pattern. Therefore, the
required capacity of the link can be reliably related to the average flow of data.
Note that in Graph 3.3.1, which represents the average situation for all GÉANT
partner NRENs, these two distinct categories of connections are combined.
In general, connections not only to the European academic backbone network
(i.e. GÉANT) but also to the general Internet are crucially important to NRENs.
As Graph 3.1.1 illustrates, connections to the GÉANT backbone are the largest
category for GÉANT NRENs. Connections to GÉANT, to commercial providers,
to neighbouring countries via cross-border fibre and to Internet exchanges
together account for 85% of these NRENs' external capacity. Compared to 2013,
total capacity shows continued growth, though the capacity to commercial ISPs
now is smaller (percentage-wise) than last year.
Graph 3.2.2 – Core capacities, other NRENs

32
25%
0%
5%
8%
20%
20%
20%
2%
GÉANT direct
NORDUnet
GÉANT indirect
Other NREN Cross-border fibre
Commercial
Internet exchange
Other
4 000
2 500
0
500
1 000
5 000
3 500
4 500
1 500
2 000
3 000
Totalcapacity(Gb/s)
2011 2013 20142012
Other
Cross-border fibre
Commercial
Internet exchange
NORDUnet
GÉANT indirect
Other NREN
GÉANT direct
Other
Cross-border fibre
Commercial
Internet exchange NORDUnet
GÉANT indirect
Other NREN GÉANT direct
3.4 Dark fibre1
As some NRENs either own, lease or have indefeasible rights of use (IRUs)2
to
dark fibre, they can decide what technology and speeds to use on it. The NRENs
covered by this edition of the Compendium were asked whether they currently
own or have IRUs to dark fibre, or plan to acquire it within the coming two years.
The NRENs were also asked to state approximately what percentage of their
backbone is accounted for by dark fibre.
The Compendium has been monitoring the uptake of dark fibre in NRENs since
2005. At that time, only a few networks used dark fibre in their backbones, and
GÉANT was just beginning to use dark fibre and light it for transnational trunks.
This year (2014), the aggregate length of dark fibre used internally by NRENs
in the GÉANT region exceeds 139 000 km, approximately the same as in 2013.
Cross-border fibres are also an important feature, and in some cases these bi-
lateral links are being integrated into the GÉANT pan-European infrastructure.
Within an NREN there are many ways to use dark fibre cost-effectively, all of them
focused on enhanced services to clients and users. NRENS, as a result of moving
from managed network links to their own transmission infrastructure, have been
able to develop new features and services at various levels, including:
• Network engineering;
• Campus networks at regional and national levels;
• Premium IP networks;
• Hybrid and multi-protocol networks;
• Wavelength or Lambda services;
• Dynamic Lambdas.
Dark fibre can be used to engineer‘future-proof’networks, which are sustainable
against the inevitable demand for greater bandwidth at reduced unit cost.
Although budgets generally have not increased much since 2009 (see Section 7
1
Analysis provided by Mike Norris, formerly of HEAnet.
Graph 3.3.1 – Capacity of NREN external connections, GÉANT partner countries, 2014
Graph 3.3.2 – Capacity of NREN external connections,
GÉANT partner countries, 2011 – 2014

33
on funding and staffing), NRENs have been able to scale up their capacity from
multi-megabit to multi-gigabit per second. A major advantage of using dark fibre
is that clients can upgrade their access capacity quickly and relatively cheaply.
Lambdas (also referred to as‘lightpaths’) are an advanced service that has
become available due to the move to dark fibre and is provided nationally
by several NRENs. It is important to distinguish this service from the NRENs’
use of wavelengths to engineer their networks and to provide different layers
for switched and routed services. The use of lightpaths, unlike the use of
wavelengths, arises from user requirements for specific circuits to interconnect
resources. The responses to this year’s Compendium questionnaire indicate that
21 European NRENs (the same number as in the past two years) are offering
lambdas as a service. Two more are planning to introduce this. For further
information, see Section 4.7.
At the European level, GÉANT provides a lambda service – analogous to the
NRENs’national-level services – which can be extended to its peer networks in
North America. The use cases, though not plentiful, include some of the most
cutting-edge research areas.
In 2013, dark-fibre network expansions of at least 15% occurred in the Czech
Republic and Estonia. Outside Europe, Australia is the country with by far the
largest dark fibre network.
See also the map at compendium.terena.org/reports/darkfibre.
2
Effective long-term leasing (temporary ownership) of a portion of the cable’s capacity.
The distinction between an IRU and a lease is becoming less clear; therefore,
in this section these two categories have been combined.
Country Total length of
dark fibre (km)
Proportion of
total network
length (%)
Percentage
of dark fibre
length added
Percentage of
dark fibre length
decommissioned
Armenia 70 80
Austria 4500 100
Azerbaijan 65 20
Belgium 2068 100 0.1 0
Croatia 131 100 0.011 0.0036
Czech Republic 6000 100 16.3 0
Denmark 2591 54 10.3 0
Estonia 1400 80 22 0
Finland 4290 91 1.2 0
France 13800 92
Georgia 50 10
Germany 10547 99
Greece 9000 100
Hungary 3200 85
Iceland 190 19
Ireland 2600 94
Israel 770 100
Italy 9500 85 9.1 1.1
Lithuania 1350 54 8 0
Luxembourg 480 80 0.14 0
Macedonia 20 95 50
Moldova 165 78 7 0
Montenegro 10.4 5 0 0
Netherlands 11000 100 0.01 0.01
Poland 6479.3 89.46 0 2.56
Portugal 1000 25
Romania 5350 98 1 6
Table 3.4.1 – Dark fibre on NREN backbones, 2014

34
Serbia 4000 95
Slovakia 2360 100 2.6 0
Slovenia 1660 100 0.72 0
Spain 13750 97 0.75 0
Sweden 8000 95
Switzerland 2969 100 13.4 1
Turkey 200 0.3 0.08 0
UK 9500 70
Ukraine 350 10
Other countries
Australia 91000 93 0.04 0
Bosnia/Herzegovina 722 100 0 0
Brazil 1461 5.03 0.32 0
Canada 4900 50 0
Hong Kong ~300*
Kazakhstan 59
Korea 1842 47.5 47.5 0
New Zealand 2205 68
Russian Federation 650 4.4 8 0
Sudan 0
USA 25294 86
Country Total length of
dark fibre (km)
Proportion of
total network
length (%)
Percentage
of dark fibre
length added
Percentage of
dark fibre length
decommissioned
3.5 Cross-border dark fibre
Most European NRENs have already installed or are planning to install cross-
border dark-fibre links to neighbouring NRENs. Cross-border dark fibre is an
(optical) network connection between two NRENs / RE networks, crossing
national borders and / or going outside the usual service area or country of the
network. There has been considerable development in this area over the past five
years, as was documented in the 2013 edition of the Compendium.
In 2013, the following links were added:
• EEnet – CSC/Funet: a 10 Gb/s lambda between Estonia and Finland;
• LITNET – PIONIER: a 40 Gb/s connection between Kaunas University
in Lithuania and PIONIER of Poland.
3.6 Network topologies
In the early days of the Internet, the logical and physical structures of a network
were roughly the same. In other words, the path the bits and bytes took through
the copper wires corresponded to the logical geographical path between two
systems. In modern network designs, the physical backbone structure, which is
based on dark fibres, has a high degree of redundancy. This means that there are
several geographical paths between two systems. The IP network (on layer 3) is
usually simpler – often more centralised – with fewer nodes than the physical
network. This may simplify network management and reduce operational costs.
In addition, it is possible to create logical links that simultaneously use several
physical connections. Such point-to-point shortcuts can reduce the latency
and create a capacity greater than that of the largest individual link. In theory,
that capacity could equal the sum of the capacities of all the links to a specific
node. Furthermore, a modern physical layer with optical switching increases the
stability and flexibility of an IP network.
* Via managed optical network services with usable lambda.

35
We asked NRENs to provide us with (hyperlinks to) their network maps of the
physical and the logical layers. The maps of RedIRIS (Spain) nicely demonstrate
the differences. Map 3.6.1 shows the logical structure of the RedIRIS IP network.
For example, there is a direct logical connection between CICA, the regional
computer network of Andalusia, and the University of Valencia (UV on the
map). Map 3.6.2 shows the physical network, in which there are multiple paths
between CICA and UV, each with several intermediate nodes.
Map 3.6.1 – RedIRIS (logical) Map 3.6.2 – RedIRIS (physical)

36
Another illustration, more complicated but conceptually similar, is from GARR
in Italy. Map 3.6.3 shows how each of the nodes in the network is connected to
several other nodes. This map illustrates how various logical routes from point
A to point B on the GARR network are possible. Of course, it is the underlying
physical infrastructure, shown in Map 3.6.4, which makes this possible.
ALBANIA
MALTA
FRANCE
AUSTRIA
GREECE
GÉANT GÉANT
NaMeX
Cogent
TIX
VSIXTop-IX
MIX
Level3
Global
Crossing
SWITZERLAND
SLOVENIA
COUNTRY
GARR Network PoP
Fiber optics
operational
National International links
Research links
planned
Internet links
Cross Border Fibers (CBF)
© GARR, October 2013
Map 3.6.3 – GARR (logical)
Map 3.6.4 – GARR (physical)

37
Germany Upgrade to TB/s 4 years Likely
Greece Upgrade of optical transmission equipment 6 months Quite certain
Hungary Primary running task is the modernisation of the
facilities and services of the school network and
improvement of backbone flexibility
until late
2015
Quite certain
Iceland Replacement of leased capacity from other
providers with fibre
1-3 years Uncertain
Ireland Enhanced Optical Resilience 2014 Quite certain
Replacement of Layer 2 network 2015/2016 Likely
Italy GARR-X Progress aims to strengthen the optical network
dedicated to the research and education community in the
four regions of convergence, to include, by the end of the
project, about 3900 km backbone, 24 Points of Presence
(PoP) and over 2200 km of user access links. The transmission
and IP/MPLS levels will be implemented with leading-edge
equipment, characterized by high degrees of scalability,
speed and flexibility, which will deliver capacities in the order
of 100 Gb/s and multiples thereof. User access links will also
be upgraded, thanks to the implementation of symmetric
fibre links with capacities of at least 100 Mb/s. Overall,
the project includes a minimum of 120 sites belonging to
universities and research institutions (which could increase
to 400 as the project progresses), and about 380 secondary
schools located in main urban areas, for more than 2000 km
of access fibre links to be rolled out by the project.
2014-2015 Quite certain
Latvia None Quite certain
Lithuania WiFi as a service 2016 Likely
VPN as a service 2016 Likely
Firewall on demand 2016 Likely
WiFi for schools 2016 Quite certain
Cloud services 2017 Quite certain
Luxembourg Full 10G backbone coverage 1 Quite certain
Extend DWDM network coverage 2 Quite certain
Planning IRU contracts replacement 4 Quite certain
Malta More use of dark fibre 2015-2017 2015-2017 Likely
3.7 Major expected network developments
The NRENs covered by this edition of the Compendium were asked to outline
major initiatives related to development of their underlying network that they
expect to realise within the next two to five years. Several NRENs that did not
respond to this question did provide information on major changes in their
organisations (see Appendix 1). Table 3.7.1 provides general insight into major
developments in research and education networking that are expected by NRENs
in the various countries of Europe and other continents, including:
• In the OECD countries, dark-fibre networks are already in place and being
upgraded and extended. Several NRENs are preparing for 100 Gb/s or even
multiples thereof; Germany is planning an upgrade to Terabit capacity.
• Many NRENs also report that they have expanded the types of service they
offer to clients.
Belarus Increase the capacity of the link to PIONIER to 10
Gb/s
2014-2015 Likely
Belgium Implementation of a new, state-of-the-art optical
network
Mid 2014 Quite certain
Implementation of a new, state-of-the-art IP
network
Mid 2015 Quite certain
Extension of the Belnet fibre footprint Continue Quite certain
Testing implementing SDN/NFV technology 2016-2017 Likely
Bulgaria 100 Gbit/s 2-5 year Likely
Cyprus Dark fibre deployment 18 months Quite certain
Denmark DWDM upgrade 2016 Quite certain
Faroe Islands connection 2014 Likely
Greenland connection 2015 Uncertain
Estonia 100G 2016 Likely
Finland 100G upgrade 2 years Quite certain
Introduce dynamic lightpath services 2 years Quite certain
Country Developments Timeframe
(years)
Confidence
(years)
Confidence
Table 3.7.1 – Major expected network developments

38
Moldova EPIC (EaP@interconnect) 2016 Quite certain
Netherlands SURFnet8 - renewal service layer 2017 Likely
SURFnet8 - renewal photonic layer 2017 Likely
Norway Dark fibre from Longyearbyen to Ny-Ålesund
(Spitsbergen)
2014 Quite certain
Poland Deploy 100G links 2015 Quite certain
Romania Deploying 100 Gb/s core with new routing
equipment
1 year Quite certain
Serbia SEELight project - full realization – to connect ~110
new institutions, increase backbone links capacity to
10Gb/s and implementation of DWDM technology
on backbone links
3 Likely
Educational network: plans for connecting ~2900
schools on AMRES network by ADSL and 3G
technology
3 Likely
Slovakia Capacity upgrade 2 - 3 years Quite certain
New POPs installation Future Likely
Slovenia Additional CBF between Nova Gorica, Slovenia and
Gorizia, Italy
Q4/2014 Quite certain
Spain New fibre to Baleares 2015 Quite certain
Extension of the network (Donana, some
technological parks in Seville, Santander and
Madrid, etc.)
2015 Quite certain
Sweden New nationwide backbone 2016 Quite certain
Switzerland New DWDM system 1 year Quite certain
First 100G backbone links 1 month Quite certain
First 100G client connection 3 months Quite certain
UK Insourcing of regional networks Within next
two years
Quite certain
Albania 10 Gb/s backbone Q4, 2014 Quite certain
1 Gb/s link to most universities and research centres Q4, 2014 Quite certain
Algeria Upgrade GEANT link 1 Quite certain
Brazil Introduction of SDN pilot 1 Quite certain
Deployment of some 100G links both at national
and international levels
2 Quite certain
Extension of performance monitoring to include
user access links
2 Quite certain
Canada Fibre builds mid 2015 Quite certain
100G Core IP Upgrade end of 2014 Quite certain
SDN development 2015 Likely
Hong Kong Enable IPv6 at all participating institution 2-3 years Likely
Experiment mutual backup/DR sites within
members’campuses
1-2 years Likely
Experiment NRENum among participating institutions 1-2 years Likely
Cloud Services 2-3 years Uncertain
Kyrgyzstan eduroam
e-library
Hosted LMS services for universities
Lebanon Switch to fibre optic 1 year Quite certain
Morocco Marwan 4 2015 Quite certain
Russian
Federation
Russian Ring 10Gbit/s 2014 Quite certain
DWDM Moscow - St. Petersburg 2015 Likely
SDN Testbed 2015 Likely
Sudan Dark fibre 2016 Uncertain
Increase bandwidth 2016 Uncertain
Increase members 2016 Uncertain
South Africa Backbone upgrade 2015-2017 Quite certain
International capacity upgrade 2015-2018 Quite certain
Addition of new sites 2014-2018 Quite certain
Backbone expansion 2014-2018 Quite certain
Taiwan Upgrade to 100G backbone Dec, 2014 Quite certain
(years)
Confidence
(years)
Confidence
Other countries
Table 3.7.1 – continued Table 3.7.1 – continued

39
Czech
Republic
Planned First to implement and test it
locally, then in cooperation with
other partners
Yes 2015
Denmark Planned Add BoD as a service Yes We have
participated
in several
projects
Germany Yes Currently prepared in GN3plus/
SA2
Yes After stabilisation of test
services
Greece Planned Yes 2014 SDN capabilities are
being assessed for the
next generation of Data
Centers hosting our cloud
infrastructure
Hungary Planned Provide testbed interface. Yes next few years
Ireland Yes Yes
Italy Yes National testbed available for the
researchers. Available by 2014 Q3
Don't know 2015 Q2
Lithuania Yes Don't know During 2013
Moldova Planned Participate in GN3plus GN4
TaaS activities
Yes 2015-2016
Netherlands Planned Collaboration with other NREN
and universities
Yes 2014 2015-2018
Norway Yes Add Lambdas Yes 2013 No
Poland Yes Upgrade to higher versions of
OpenFlow protocol
Yes 2014
Russian
Federation
Planned Yes 2015 Yes
Slovenia Planned Don't know
Spain Planned Yes 2015/2016
Switzerland Planned Don't know According to
user demand;
also, willing to
connect sites
to GÉANT SDN
testbed.
No
3.8 Software-Defined
Networking
Software-Defined Networking (SDN) enables the
creation of virtual networks using the underlying
infrastructure, though with varying characteristics
and topologies, adapted to user needs4
.
For this year’s Compendium, we asked NRENs
whether they have an SDN testbed, how it is
managed and for what purposes it is being, or
will be, used. The responses are shown in Table
3.7.1. Six GÉANT partner NRENs currently have
such a testbed, up from four in 2013. Ten more are
planning to deploy one. In all cases, the setup is
still experimental. The exception is Brazil, which has
offered the testbed to researchers.
Countries Testbed?
How will it be
extended in future?
Interest
in multi-
domain SDN
capabilities
Plan to
implement an
SDN-based
testbed
Introducing SDN-based
capabilities in the
production network?
4
For more information, see for example
en.wikipedia.org/wiki/Software-defined_networking
Table 3.8.1 – Software-defined networking

40
Testbed?
How will it be
extended in future?
Interest
in multi-
domain SDN
capabilities
Plan to
implement an
SDN-based
testbed
Introducing SDN-based
capabilities in the
production network?
Other countries
Algeria Planned 2014 Third
quarter
Australia Yes By exploiting the MPLS capability
of AARNet4. It is planned to
make it compatible with (for
example) GENI, with initial
international deployments in
response to researcher demand,
e.g. in NZ, USA and Europe.
Yes Employing MPLS, e.g. for
DaShNet (https://www.rdsi.
edu.au/dashnet). A testbed
is about to be introduced
for researchers needing it,
determined by demand for
which countries they want to
connect to.
Brazil Yes The present FIBRE testbed
currently reaches 10 institutions
in Brazil and 3 in Europe (http://
www.fibre.org.br). Extensions
are planned by including more
institutions in Brazil, and by
federating with testbeds in other
countries, especially EU and US.
Yes The main purpose is to
implement a virtualized
backbone resource in such
way that it could be controlled
by software and isolated. In
this way we can distribute the
network resource based in a
Business Model.
Canada Yes Extending the testbed access
to researchers and regional
providers.
Current year
Kazakhstan Planned 2014/2015
Korea Yes Extension to international
domain like Hong Kong, Seattle,
Chicago and Daejeon.
Yes
Morocco Planned Yes 2015 No
New Zealand Planned
South Africa Planned Don’t know 2015-2016
Taiwan Yes We plan to connect more
universities and commercial
companies to facilitate
OpenFlow advancement.
Yes SDN switches are planned to be
deployed on the backbone and
interconnected by VPLS VPN
to provide the SDN testbed
services to the researchers.
3.9 Network as a Service
Network as a Service (NaaS) gives cloud users the
capacity to reserve bandwidth on the network
dynamically5
.
Currently, only three NRENs have deployed such
capabilities on their network infrastructure (Israel
was the latest country to implement this, following
Ireland and the Netherlands last year).Seven
GÉANT NRENs have plans in this area; a few non-
European NRENs including Brazil and Korea also
have such plans. For SURFnet, the purpose is to
provide researchers with a flexible Bandwidth-on-
Demand (BoD) service. Lithuania would like to serve
smaller institutions who do not have the necessary
capabilities to construct the network themselves.
Italy is planning to use NaaS for inter- and intra-
data-centre connectivity for distributed clouds.
Brazil aims to dynamically provision a set of services
that integrate network (VPN, BoD, VLAN), storage
and computer processing.
5
For more information, see for example en.wikipedia.org/wiki/
Network_as_a_service

41
GÉANT Compendium of National Research and Education Networks In Europe /Traffic
4 TRAFFIC
As in previous years, the NRENs covered by this edition of the Compendium were
requested to report their total annual traffic flows at the boundaries of their
networks. The four flows they were asked to specify are defined in Diagram 4.0.1.
Customer
connections
All external
networks
peerings
NREN network
T1
T2
T3
T4
T1 All IP traffic from customer sites and the NREN.
T2 All IP traffic to customer sites and the NREN.
T3 All IP traffic leaving the NREN
T4 All IP traffic entering the NREN
External traffic = all traffic to
the NREN’s regional backbone,
the commercial Internet,
Internet Exchanges, etc.
(consisting of T3 + T4 above).
Note: all IP traffic includes both
IPv4 and IPv6 traffic.
What this diagram does not show is that, in some cases, the reported traffic may
include traffic flowing between separate customers of the NREN. In other cases,
such traffic may be confined to separate MANs/RANs and not appear on the
NREN backbone.
In certain cases, non-routed traffic within a customer network may be switched
via NREN point-to-point circuits but not be reported in traffic statistics, as it does
not appear as IP traffic on the NREN network.
Similar distinctions apply to external traffic. IP traffic is exchanged with external
networks via neutral exchange points, via peerings with commercial Internet
service providers, via GÉANT with other NRENs in Europe and around the world,
and in some cases via cross-border fibre with neighbouring NRENs. In addition,
some NRENs use direct point-to-point circuits to interconnect, for example,
research facilities with their users. Because traffic on such links does not go
through the routed IP network, it is not reported here.
Below, Section 4.1 gives an overview of the NRENs’responses, as well as this
year’s traffic trends. Section 4.2 considers traffic in 2013, whereas Section 4.3
analyses traffic trends over the past ten years. Section 4.4 gives information on
NREN traffic per inhabitant. Section 4.5 considers the issue of congestion. Section
4.6 examines deployment of IPv6. Finally, Section 4.7 focuses on lambda and
point-to-point fixed bandwidth traffic.
4.1 Overview
Most of the NRENs that responded to the questionnaire for this 2014
Compendium reported annual IP traffic flows at the points where they exchange
traffic with external networks (T3 T4); most NRENs also reported annual traffic
flows between their connected sites and their backbone network (T1 T2). The
T3 and T4 traffic levels are relatively easy to measure and record, as there are
only a few points on the network to monitor. Below, Graphs 4.2.1 (2013 traffic,
T3 10 000 TB) and 4.2.2 (2013 traffic, T3 10 000 TB) represent all the national
responses submitted in 2014. Comparison with data from previous years reveals
that IP traffic continues to grow. Over the past ten years, the annual rate of
growth has fluctuated around an average of almost 30% (but always remained
positive). In recent years, growth seems to have slowed down. Traffic on the
GÉANT IP backbone has decreased, likely because increasing volumes of traffic
from very large users are now routed via separate lambda connections, over
which traffic is not measured.
It should be noted that traffic growth patterns differ from country to country and
from year to year.
Diagram 4.0.1 – Types of traffic flow

42
0
16
6
2
4
8
10
12
14
Ukraine
Azerbaijan
Bulgaria
Belarus
Macedonia
Cyprus
Latvia
Armenia
Moldova
Montenegro
2010
2009
2012
2011
2013
Analysing the available traffic data reveals substantial differences within Europe: NREN traffic per inhabitant in
Armenia, Azerbaijan, Belarus, Bulgaria, Cyprus, Latvia, Macedonia, Moldova, Montenegro and Ukraine remains
below 15% of the European average level. Compared to 2012, traffic per inhabitant increased in Serbia and
Turkey, bringing them above the 15% level. (See Graph 4.4.2.)
For the GÉANT partner countries, average estimated
congestion levels have increased at the campus and
access-network level but have almost disappeared
at the levels of backbone and external connections
levels.
The vast majority of NRENs provide some or all of
their clients with both IPv4 and IPv6 connectivity.
Fourteen of the thirty GÉANT partner NRENs
that provided information for this section of the
Compendium have so far allocated less than 1% of
their IPv6 address space to client institutions. NRENs
do not generally engage in carrier-grade NAT or
other forms of address sharing, though there are a
few exceptions.
Traditionally, NRENs have provided connected
institutions with both the physical connections and
the routing that is necessary to transport IP packets
from their place of origin to their destination. Traffic
is measured by counting the IP packets that are
routed by the NREN. However, with the advent
of optical networking, it has become possible to
directly interlink high-end users by means of a
dedicated physical connection in the form of a
wavelength or‘lambda’on an optical fibre cable. In
principle, the users that are thus connected decide
how to use that wavelength; the routing is no
longer the automatic or standard responsibility of
the NREN. (In theory, over such a wavelength it is
possible to transport data using methods other than
standard IP.) Therefore, it is no longer feasible for
Graph 4.1.1 – Traffic per inhabitant, 15% of European average

43
the NREN in question to measure the traffic on such
links. Given that such traffic volumes are normally
supplied only to high-end users, it is reasonable to
assume that they are substantial and growing.
Twenty-one GÉANT partner NRENs now provide
dedicated wavelengths (lambdas) to their clients.
Per NREN, the number of static point-to-point fixed
bandwidth connections provisioned in January
2014 varies between three and 511 (HEAnet of
Ireland). At present, the number of circuits seems
to be the only measurable parameter that can be
used to document the evolution of lambda traffic.
Nearly 1300 wavelengths circuits are now in use by
the GÉANT NRENs for high-bandwidth, low-jitter
transport (up from 1200 in the year 2013 and 875
in 2012). Nearly all lambdas reported by the 21
NRENs concerned are static or fixed. Five NRENs are
currently using dynamic lambdas, which can be set
up and taken down (by the NREN or its clients) for
short-term requirements.
4.2 Traffic in 2013
Graph 4.2.1 represents the data submitted by those NRENs whose T3 traffic exceeds 10 000 terabytes per year,
whereas Graph 4.2.2 represents the data submitted by NRENs with lower levels of T3 traffic. (In both graphs,
the countries have been sorted on the amount of T3 traffic.) These graphs clearly show how the distribution
of total traffic between the four categories (T1 to T4) differs from NREN to NREN. Note that not all respondent
NRENs provided all four traffic values.
For most NRENs, the reported traffic sent into their backbone (T1+T4) is equal or nearly equal to the reported
traffic sent out of their backbone (T2+T3). The few NRENs for which this is not the case cite, as main reasons
for the discrepancy, the effects of traffic-monitoring efforts, hosting of Content Delivery Networks such as
Akamai, and difficulties in separating out the various traffic types.
Graph 4.2.1 – 2013 traffic, T3 10 000 TB
0
50 000
250 000
300 000
200 000
100 000
T1
T4
T2
T3
150 000
+
+
CzechRepublic
Finland
Germany
Greece
Hungary
France
Italy
Netherlands
Norway
Poland
Romania
Slovenia
Spain
Sweden
Switzerland
Turkey
UK
350 000
400 000

44
The annual average growth rate was almost 30%. Graph 4.3.2, which shows the
growth rate variation over the same period, reveals that, overall, growth has
slowed down somewhat in recent years. Figures from Cisco1
show global IP traffic
growth slowing down to 25% in 2013. However, this is still considerably higher
than the NREN traffic growth of 3%.
Using data from GÉANT service reports, GÉANT IP traffic growth has been plotted
in Graph 4.3.3. This shows a traffic decrease over the past year, probably because
traffic flows from very large international users increasingly use dedicated
bandwidth services. Traffic on these services, although substantial, is not
measured together with normal IP traffic. The effects of this development, which
started a few years ago, are starting to become evident on the GÉANT backbone
network. (See also Section 4.7.)
4.3 Traffic trends, 2004-2013
As in the 2013 edition of the Compendium, Graph 4.3.1 shows T3+T4 values for a
subset of 23 NRENs that have consistently submitted complete data. Clearly, over
this ten-year period (2004-2013) NREN traffic has continued to grow steadily,
even though growth in 2012/2013 was lower than in earlier years.
0
200
1 400
1 200
400
2006 20092004
1 000
2008
Traffic(PB/year)
600
800
2010 2011 201220072005 2013
1
See www.cisco.com/web/solutions/sp/vni/vni_forecast_highlights/index.html.
Graph 4.3.1 – NREN annual traffic flows (T3+T4), 2004-2013, n = 23 NRENsGraph 4.2.2 – 2013 traffic, T3 10 000 TB
0
20 000
10 000
15 000
25 000
T1
T4
T2
T3
+
+
Luxembourg
Latvia
Bosnia/Herzegovina
Belarus
Georgia
Iceland
Moldova
Morocco
Montenegro
Ireland
Brasil
Australia
5 000

45
Note that the traffic growth pattern is somewhat erratic: it differs from country
to country and from year to year. Growth is influenced by many factors, such as
backbone and access-network capacity upgrades, connections to large users,
and changes in policy and charging models.
4.4 NREN traffic per inhabitant
In 2009, we attempted to identify an indicator that would enable NRENs to be
compared in terms of traffic. After considering several alternatives, the simplest
indicator – traffic-per-inhabitant – was found to be the most reliable, because
there is a roughly proportional relationship between a country’s total population
and the size of its education and research community. We have found this
relationship to be valid for most countries surveyed and, therefore, no other
assumptions or data convolutions need to be made. However, as with other
high-level indicators, these statistics should be treated with caution, as large
differences between countries do exist and any comparison of this type is a
simplification of reality.
Graphs 4.4.1 and 4.4.2 show NREN annual T3+T4 traffic in 34 countries over the
period 2009-2013, normalised according to the total national population in each
corresponding year. Note that this figure should not be taken as an indicator of
the network traffic generated by a typical NREN user.
In these 34 countries, the average traffic per inhabitant grew from 166 MB/month
in 2009 to 342 MB/month in 2013, with an average annual growth rate of 19.7%.
0
20
100
40
10
30
2006 2009
80
2008
Growth(%)
60
50
70
90
2010 201320112004 20072005 2012
0
10
20
30
40
70
50
2006 2007 20122005
60
2008 2009 2010
Traffic(PB/quarter)
20142011 2013
Graph 4.3.2 – Growth rates of NREN and global IP traffic, 2004-2013 (in %, T3+T4)
Graph 4.3.3 – GÉANT IP traffic 2005-2014 (quarterly data)

46
0
200
1 600
1 200
400
600
800
1 000
AVERAGE
Denmark
Hungary
Sweden
CzechRepublic
Lithuania
Netherlands
Finland
Switzerland
Norway
Iceland
Slovenia
2010
2009
2012
2011
2013
1 400
UK
Austria
Slovenia has had consistently high nominal traffic (per inhabitant) over the
five-year period (2009-2013). Although Slovenia’s population is quite small,
the country has a relatively high proportion of traffic generated by primary
and secondary schools: indeed, nearly 700 schools and a few universities are
connected to the ARNES backbone, and some of the schools are connected with
gigabit capacities. Therefore, the proportion of the population that is connected
by the NREN is relatively high, and Slovenia’s external traffic is higher than that of
the other European countries shown.
Graph 4.4.1 – Nominal external traffic (T3+T4) divided by total national population: greater than European average

47
0
50
350
100
150
250
300
200
Bulgaria
Moldova
Montenegro
Serbia
Turkey
Latvia
Belarus
Armenia
Italy
Spain
Portugal
Romania
Cyprus
Poland
Germany
Belgium
Ireland
Luxembourg
Greece
Georgia
AVERAGE
2010
2009
2012
2011
2013
Israel
Note that the vertical scale of Graph 4.4.2 is much smaller than that of Graph
4.4.1. Clearly, there remain substantial differences within Europe: Armenia,
Bulgaria, Cyprus, Georgia, Latvia, Montenegro, Moldova, Serbia and Turkey show
much lower values than the rest of Europe (despite the traffic growth in Latvia,
Serbia and Turkey).
Like Slovenia, many countries with quite a small population have relatively high
traffic per inhabitant. Simply having an NREN in itself generates a certain amount
of traffic (including mirroring services, news groups and library databases) — at
least, once the NREN has attained a certain minimum level of development and
connectivity. For larger countries, that‘base-load traffic’does not significantly
affect the traffic statistics.
Graph 4.4.2 – Nominal external traffic (T3+T4) divided by total national population: lower than European average

48
2
This index was developed for the Compendium by Mike Norris, formerly of HEAnet. The index was
modified in 2009 to set the minimum value at 0 rather than 5.
3
For information on IPv4 exhaustion, see for example:
www.ripe.net/internet-coordination/ipv4-exhaustion.
4.5 Congestion
The NRENs covered by this edition of the Compendium were asked to estimate
the percentage of institutions connected to their networks that experience none-
to-little, some-to-moderate, or serious congestion at the various network levels.
From the subjective levels reported by NRENs, a metric was derived for the level
of congestion in each network element, using the following formula2
:
congestion index = (0.05*little + 0.2*some + 0.5*serious) – 5
Note that the data for MANs and access networks were combined. Applied to all
the submitted data on congestion, this formula provides a single uniform metric.
As shown by Graph 4.5.1, for the GÉANT partner countries, the average estimated
congestion levels have increased at the campus and access-network levels. For
the first time this year, no congestion is reported at the backbone level, whereas
at the level of external connections, the congestion has dropped to minimal
levels. However, it seems that recent investments in increased capacity at the
external and backbone levels are causing some bottlenecks at the access-
network level.
4.6 IPv6 deployment
There have been numerous reports concerning the impending shortage of IPv4
address space3
and the need to start deploying IPv6. As indicated in previous
editions of the Compendium, because most European NRENs have already
deployed IPv6 on their backbones, client institutions can start to use this new
protocol as and when the need arises. The ratio of IPv6 to IPv4 traffic is currently
very low and has not risen much over the last few years.
Nine GÉANT NRENs do not yet provide IPv6 connectivity to any of their clients.
However, the vast majority of NRENs do provide some of their clients with both
IPv4 and IPv6 connectivity. Four NRENs (one more than in 2013) provide this to
all their clients. The GÉANT network itself has been carrying IPv6 traffic, along
with IPv4, since 2000.
0
2
Campus Access/Metro Backbone External
2011
2010
2013
2012
4
1
3
2.5
0.5
1.5
3.5
2014
Graph 4.5.1 – Congestion index, GÉANT partner countries, n = 31

49
4.7 Lambda and point-to-point fixed
bandwidth traffic
As in the past two years, 21 of the GÉANT partner NRENs who responded to
our questionnaire this year currently offer dedicated wavelengths (lambdas)
to their clients. As of 31 January 2014, five NRENs (up from three a year earlier)
had dynamic lambdas. Three NRENs offered only one or two ports for this; the
Netherlands and Poland offered 35 and 21 ports, respectively.
The number of static point-to-point fixed bandwidth connections that were
live at the end of January 2014 ranges from three to 511 (HEAnet in Ireland).
The provisioning time ranges from a few minutes to three months. Table 4.7.1
shows that at the end of January 2014 almost 1300 static point-to-point lambdas
were live, while there were over 1400 static point-to-point fixed bandwidth
connections.
As originally indicated by the 2010 Compendium, measurement of the traffic
on such lambdas is sometimes complex and not added to the traffic on the rest
of the network. Table 4.7.1 gives an overview of the large number number of
lambdas provisioned as of 31 January 2014 and it is an indication of the large
amount of traffic carried over separate paths.
Many NRENs do not measure this type of traffic at all, whereas others are only
able to measure it via their own routers and/or IP-based traffic. An alternative
method to measure lambda take-up and traffic still needs to be found. One
proposal is to measure the number of circuits, rather than the traffic itself. Table
4.7.1 at least provides an overview of the number of lambdas provisioned as of
31 January 2014.
Four of the 30 GÉANT NRENs that provided information for this section of the
Compendium have not yet allocated any IPv6 address space to their client
institutions. Ten have allocated less than 1% (down from 17 last year), eight have
allocated more than 1%, and one has allocated 100%.
We asked NRENs if they have taken any measures in response to the growing
deployment of carrier-grade NAT or other forms of address sharing. The only
NRENs who report doing anything of this kind are those in Bosnia/Herzegovina
and Hong Kong. A few more NRENs or NREN customers are deploying carrier-
grade NAT themselves. However, most NRENs are not doing this and are
discouraging their customers from doing so.
Country NREN or its customers deploying or considering the use of carrier-grade
NAT or similar address sharing technologies?
Belgium Belnet: no. We discourage our customers from doing this.
Greece Not yet. We had an issue with resource exhaustion in the cloud infrastructure,
but we decided not to invest in CGN.
Hungary Yes
Lithuania We advise our customers not to do this.
Serbia There are plans for connecting schools, and we estimate that the entire
network will need to use private IPv4 addresses with a carrier grade NAT solution,
because of the lack of IPv4 address space.
Sweden Yes
Other countries
Australia AARNet: no; customers: yes
Bosnia/
Herzegovina
Yes
Brazil Not that we are aware of.
Hong Kong Yes
Table 4.6.1 – Carrier-grade NAT deployment

50
Country Lambdas? Number
static
Static
point to-point
fixed-bandwidth
connections
Number
dynamic
Charge? Provisioning time Number leaving
country or NREN
Belgium Yes 123 132 0 Yes 24 hours 2
Czech Republic Yes 15 4 1 No, but planned 3 months 5
Denmark Yes 142 20 2 Yes 3 weeks We use Bandwidth-
on-Demand for this
purpose
Estonia Planned
Finland Yes 117 234 0 Yes. For example, fixed yearly price
for 1 Gb/s and 10 Gb/s lightpaths.
1 Gb/s: typically 1 week if fibre infrastructure exists. 10 Gb/s:
8-12 weeks.
3
France Yes 132 No 3 months 8
Germany Yes 43 43 0 Yes On average, 10 days 16, Only CBF
Greece Yes 0 No 2 days to provision a new lambda on existing link None
Hungary Yes 70 10 0 No Depends on availability of transponders or ports 2
Ireland Yes 64 511 1 Yes, they are charged as per a Normal
circuit.
4-6 weeks 0
Italy Yes 110 30 0 Yes 30 days 5
Lithuania Yes 13 8 0 No, but in practice users speed up
the process by buying their own
coloured interfaces.
Aliens add up easily, but we do not possess surplus
transponders so new ones should be purchased by means of
public procurement
7
Luxembourg Yes 24 34 0 No 1 month 1
Moldova Planned 4
Netherlands Yes 210 242 35 Yes 3 weeks excluding new fibre orders 18
Norway Yes 3 4 0 Yes 4-8 weeks 0
Poland Yes 80 0 21 No 1 day 12
Portugal Yes 25 147 0 Historically no costs were charged,
though the official policy states that
costs could be charged.
If no purchases are needed, one week.
If purchases are needed, one month
3
Slovenia Yes 3 3 0 No Approx. one week, excluding all the meetings with the
institution and waiting time for the equipment that the
institution needs to provide to be delivered
2
Spain Yes 176 No Minutes, if hardware (TRBDs) is available on equipment 8
Table 4.7.1 – Lambda provisioning

51
Sweden Yes Less than 10 Less than 10 0 Yes Between 10 min and 12 weeks 8
Switzerland Yes 6 14 0 Yes 1 month 3
UK Yes 34 0 0 Yes, over 1 Gb 45 working days 17
Other countries
Brazil Yes 0 0 0 No Two days 0
Canada Yes 10 65 0 No 10 days 14
Hong Kong Yes
Japan Yes No Unknown
Korea Yes 10 30 - No Within 1 minute 10
New Zealand Planned
Russian Federation Yes 4 12 Yes 0 4
USA Yes Yes Pre-deployed services, such as pre-placed transceivers, may
be turned up within hours but generally require 72 hours
notice to ensure proper notification of non-service-affecting
maintenance. Due to the diverse and expansive geography
of the United States, shipping times may dominate turn-up
speed for services that are not pre-deployed.
Several circuits,
such as the
ANA 100G, exit
Internet2, however
no lambdas egress
and become alien
wavelengths in
another optical
domain.
Country Lambdas? Number
static
Static
point to-point
fixed-bandwidth
connections
Number
dynamic
Charge? Provisioning time Number leaving
country or NREN

52

53
GÉANT Compendium of National Research and Education Networks In Europe /Middleware services
5 MIDDLEWARE SERVICES
Middleware services provide an essential bridge between the network
infrastructure and its uses. Below, Section 5.2 considers key security services,
including anti-spam measures and mechanisms for sharing information on
security issues. Section 5.3 examines various authentication and mobility
services: specifically, identity federations (5.3.1), digital certificates (5.3.2) and
eduroam take-up across Europe (5.3.3).
5.1 Overview
Service access is becoming increasingly independent of the physical location
either of the user or of the service. As a result, there is a growing need for
security services, identity federations and certification services; in all these
areas, which involve secure access by remote users, developments are rapid.
Important new developments in security include the adoption of structured
formats for exchanging information about computer incidents, and the use of
network devices for addressing security threats. Many NRENs are also active in
the related area of spam filtering.
Twenty-seven GÉANT partner NRENs currently provide an Authentication and
Authorisation Infrastructure (AAI); eleven are planning to introduce one. In
most cases, the web single-sign-on federation is operated by the NREN. Most
of the GÉANT partner NRENs (and a few non-GÉANT NRENs) have joined or are
planning to join the eduGAIN interfederation service — a development that
holds the promise of service access across federations.
By 31 December 2014, approximately 150 000 valid outstanding server
certificates were being used by GÉANT partner NRENs. Of those, around two-
thirds had been issued under the TERENA Certificate Service.
eduroam has shown remarkable growth over the past year: the number of
locations with eduroam availability grew by 25%. The number of authentications
more than tripled. By September 2014, it was approaching 150 million
authentications per month.
5.2 Security services
As security services are becoming increasingly important to NRENs, the
questionnaire for this year’s Compendium included several additional questions
related to security. The responses are summarized in Graph 5.2.1. In addition, the
answers to some of the new questions are shown in Table 5.2.2.
This information shows that 90% of the GÉANT NRENs take measures to reduce
spam activities. Although nearly two-thirds of the GÉANT NRENs were impacted
by DDoS1
attacks last year, less than 10% had previously implemented mitigation
measures. By contrast, only 5% of the GÉANT NRENs were impacted by route/
IP hijacking, whereas 36% had mitigation measures in place. Thirty-nine percent
of the NRENs have a Chief Information Security Officer. A similar percentage of
NRENs work according to a formal security policy, even though only 14% follow
formal security standards (such as ISO 27001). Twenty-five percent of GÉANT
NRENs are planning to introduce such standards.
For further information on how NRENs have recently been collaborating in the
area of security, see for example the TF-CSIRT2
web pages at
www.terena.org/activities/tf-csirt.
1
Distributed Denial of Service Attack.
2
TF-CSIRT is the GÉANT Association Task Force on Computer Security Incident Response Teams.

54
0% 40%20% 100%80%60%10% 50%30% 90%70%
No
Planned
Yes
Impacted by DDoS attacks last year
Implemented DDoS mitigation measures
Impacted by DNS reflection attacks
Protection in place?
Measures to reduce spam activities
Network devices for addressing security threats
Anti-spoofing measures?
Structured format for the exchange of computer incident information
CISO (Chief Security Officer)
Formal security policy
Formal security standards
Impacted by route/IP hijacking
Mitigation in place
Graph 5.2.1 – Security services, GÉANT NRENs

55
Country Negatively
impacted by
DDoS attacks
DDoS mitigation
measures
Negatively impacted
by DNS Reflection /
Amplification attacks?
Protection
in place?
Negatively
impacted
by route/IP
hijacking
Mitigation measures Designated
CISO
Formal
security
policy
Formal
security
standards
Armenia Yes No Yes Yes No Yes Yes
Austria No NetFlow Data No No No No No No No
Belarus Yes No No No No Monitor routing changes No Yes No
Belgium Yes Filters on the core routers +
remote triggered blackhole
filtering towards upstream
providers
Occasionally additional
load on DNS server (but
not problematic)
Yes No No No No Planned
Bulgaria No No No Yes No No No Yes No
Croatia Yes No Yes Yes No Yes Yes Yes Planned
Cyprus Yes No No No No No No No No
Czech Republic Yes NetFlow data. Yes Yes No Monitoring No No No
Denmark No No No No No No No No Planned
Estonia Yes No Yes No No No No No No
Finland Yes We have implemented BGP
blackhole communities
that can be used to discard
traffic to affected hosts. We
also have a self-made tool
for adding stateless firewall
filters on demand. However,
both of the tools require
manual intervention.
Yes Yes No We use RIPE RIS alert service
in order to detect whether
someone announces our
prefixes or more specifics.
No Yes Yes
France Yes Blackholing at several levels
already in place. DDOS
mitigation will be deployed
during 2015.
Yes Yes No No Yes Planned Planned
Georgia Several times Depends on case No No No Monitor routing changes No Yes No
Germany No NetFlow-Data and SNMP-
Data
No Yes No Yes Yes Yes Yes
Greece Yes Firewall on Demand service
to currently block (could
also mitigate) the attack
Yes Yes No No Yes No No
Table 5.2.2 – NREN security services

56
Hungary Yes blackholing outside NIIF
network after detection of
the sources
Yes Yes No No No Yes Yes
Iceland Yes Blackhole routing. Yes No No No Yes No No
Ireland Yes Filtering Yes Yes Yes Monitor routing changes. Yes No No
Israel No
Italy Yes Yes No No
Latvia Yes Remote triggered
blackhole
Yes No No No No No No
Lithuania Yes No. Transition to IPv6 is
imminent.
Yes No No No No Yes No
Luxembourg No No No Yes No No No No No
Moldova No No No No No Yes No Planned
Montenegro No ASA IPF No No No Yes No Yes Planned
Netherlands No Filtering Yes Yes No RPKI Yes Yes Planned
Norway Yes Rate limiting. Yes Yes No Border filtering and customer
filter. In addition, strict
filtering on BGP towards
customers.
Yes Yes Yes
Poland Yes Blackholing and more
sophisticated methods
depending on DDoS
targets. DNS servers
auditing and patching —
implementation of BCP
140. Implementation of
BCP 38 was possible.
Yes Yes No No No No No
Portugal Yes No No No No RPKI Yes No No
Romania Yes RTBF Yes No No We’re using BGPmon for
route changes.
No Yes No
Russian
Federation
Yes No Yes No No No No No Planned
Country Negatively
impacted by
DDoS attacks
DDoS mitigation
measures
Negatively impacted
by DNS Reflection /
Protection
in place?
Negatively
impacted
by route/IP
hijacking
CISO
Formal
security
policy
Formal
security
standards

57
Serbia No No No No No No Yes No No
Slovakia Not significantly No No No No No No No No
Slovenia No Routes to blackhole with
BGP, rate limiting.
No No No No No No No
Spain No Mitigation system at the
backbone using BGP.
No Yes No No No No Planned
Sweden No No No No Yes Yes No
Switzerland Yes No Yes No No No Yes No No
Turkey Some of our
users (universities
and research
organizations)
were affected.
Some of our users
(universities and research
organizations) were
affected.
No No No No Yes No
UK Yes Filtering of incoming UDP
traffic using traffic policing
rules at network ingress
points.
Yes No No No Yes Yes Planned
Other countries
Algeria Yes IPS for Data Center No Yes No No Yes No No
Australia Yes RTBH, packet filtering via
ACLs.
Yes Yes Yes We monitor routing changes
(usually post hoc); not using
RPKI.
No Yes No
Bosnia/
Herzegovina
No No No No No No No No No
Brazil Yes DDoS mitigation through
traffic monitoring using
Arbor Peakflow solution.
Flowspec (RFC 5575)
and a proprietary threat
management system.
Yes Yes Yes No No Yes Planned
Canada No No No No No No No No No
Hong Kong No No No Yes No No No No No
Country Negatively
impacted by
DDoS attacks
DDoS mitigation
measures
Negatively impacted
by DNS Reflection /
Protection
in place?
Negatively
impacted
by route/IP
hijacking
CISO
Formal
security
policy
Formal
security
standards

58
Country Negatively
impacted by
DDoS attacks
DDoS mitigation
measures
Negatively impacted
by DNS Reflection /
Protection
in place?
Negatively
impacted
by route/IP
hijacking
CISO
Formal
security
policy
Formal
security
standards
Other countries
Kazakhstan No No Yes No No No No Planned
Korea No No No Yes Yes Planned
Lebanon No No No No No Yes No No
Morocco No No No Yes No No
Sudan No No No No No No No
South Africa Yes No No No No No No No
Taiwan Yes. DDoS attacks
originated from
outside and inside
were both often
seen.
No Yes, mainly originated
from the outside.
No No No No Yes Yes
5.3 Authentication and mobility services
State-of-the-art services give specific user groups access based on each user’s
identity and role (faculty, staff, student, etc.) as provided and managed by the
users’home institution. As a result, access to services is becoming less dependent
on the physical location either of the user or of the service. The research and
education community is at the forefront of this development. Security is a key
issue in this area: it is essential to know who wants to access a particular service
and who is entitled to do what. This means that authentication and mobility
services go hand in hand and that the development of these services can either
constrain or stimulate the way other services are developed and delivered to users.
A pioneering mobility service is eduroam®3
, which was developed by the
European NRENs and has grown into a secure roaming access service for the
research and education community internationally. This service is currently
offered by all GÉANT partner NRENs except Azerbaijan, Georgia and Ukraine,
and by NRENs in a growing number of countries in other regions (for further
information, see Section 5.3.3).
eduroam offers general Internet access. Access to more specific services belongs
to the domain of identity federations.
5.3.1 Identity federations
An identity federation enables a user registered in the identity management
system of his university or home institution to access services provided either
by his university or by other institutions participating in the identity federation.
Federated authentication across institutional boundaries originated in the NREN
community. Like NRENs, federations have a variety of organisational forms (e.g.
3
A registered trademark of the GÉANT Association.

59
a project within an NREN, a consortium, a separate entity, a collaboration with
primary education, etc.). Most countries have a single federation for higher
education and research. NRENs either operate the research and education
federation themselves or have close organisational ties with the federation
in their country. These federations have implemented data protection in
accordance with national and EU Data Protection Acts and actively work to
preserve privacy while enabling sharing of user-related information.
Identity federations provide access to a variety of services, which may include:
• library resources;
• catalogue systems and document delivery;
• collaboration tools such as wikis;
• web-conferencing and mailing-list subscription services;
• e-learning tools and web portals.
In addition, there are services such as:
• video- and web-conferencing;
• MCU booking systems;
• streaming video portals;
• software licensing;
• webshops for a range of academic services.
Service providers can use federated access to identify and authorise a particular
set of users; for example, students who may be entitled to special terms for travel,
mobile phones, etc.
As reported in the Compendium since 2006, the number of identity federations
has been growing continually. In order to foster collaboration in this area,
TERENA has facilitated the formation of REFEDS (Research and Education
Federations), in which most federations collaborate. For further information, see
www.refeds.org.
With the growth of identity federations and federated services, the extra
advantages of interfederating them has been recognised. The oldest operational
interfederation activity is Kalmar2, which links the Nordic federations. The
eduGAIN interfederation service started operating in 2011. For further
information on eduGAIN, see www.eduGAIN.org. Both eduroam and eduGAIN
are supported by the EU through the GN3plus project.
Table 5.3.1.1 shows the current situation. Differences with respect to 2013 are
highlighted in colour. Almost all GÉANT partner NRENs provide an AAI or are
planning to do so. There are only four exceptions (BREN of Bulgaria, CYNET of
Cyprus, SANET of Slovakia and URAN of Ukraine).
Since 2013, five more countries (Austria, Ireland, Israel, Slovenia and the UK) have
joined eduGAIN, and more are planning to do so.
Country AAI provided? Federation NREN
operated?
Interfederate?
Armenia Yes No
Austria Yes Yes Yes Yes: eduGAIN
Azerbaijan No No
Belarus Yes No No
Belgium Yes Yes Yes Yes: eduGAIN
Bulgaria No No Yes Yes: eduGAIN
Croatia Yes Yes No Yes: eduGAIN
Cyprus No No Yes No
Czech Republic Yes Yes Yes Yes: eduGAIN
Denmark Yes Yes Yes Yes: Kalmar, eduGAIN
Estonia Yes Yes Yes No, but plan for 2014
Finland Yes Yes Yes Yes: Kalmar, eduGAIN
France Yes Yes Yes Yes: eduGAIN
Table 5.3.1.1 – AAI (yellow highlights denote differences with respect to 2013)

60
Georgia Plan No No
Germany Yes Yes Yes Yes: eduGAIN
Greece Yes Yes Yes Yes: eduGAIN
Hungary Yes Yes Yes Yes: eduGAIN
Iceland No No No
Ireland Yes Yes Yes Yes: eduGAIN
Israel Yes Yes Yes Yes: eduGAIN
Italy Yes Yes Yes Yes: eduGAIN
Latvia Plan No No, but plan
Lithuania Yes Yes Yes No, but plan for 2014
Luxembourg Yes Yes Yes No, but plan for 2014
Macedonia Plan No No
Malta Yes No No No, but plan
Moldova Plan No No
Montenegro Yes No No, but plan
Netherlands Yes Yes Yes Yes: eduGAIN
Norway Yes Yes Yes Yes: Kalmar, eduGAIN
Poland Yes Yes Yes Yes: eduGAIN
Portugal Yes Yes Yes No, but plan for 2014
Romania Plan No No, but plan
Serbia Yes Yes Yes No, but plan for 2014
Slovakia No No No
Slovenia Yes Yes Yes Yes: eduGAIN
Spain Yes Yes Yes Yes: eduGAIN
Sweden Yes Yes Yes Yes: eduGAIN
Switzerland Yes Yes Yes Yes: eduGAIN
Turkey Yes Yes Yes Plan for 2014
UK Yes Yes Yes Yes: eduGAIN
Ukraine No No No
Albania No Yes No
Algeria Yes No No, but plan
Australia No Yes No Yes: eduGAIN
Bosnia/
Herzegovina
Plan No
Brazil Yes Yes Yes Yes: eduGAIN
Canada Yes Yes Yes Yes: eduGAIN
Japan Plan No Yes: other
Lebanon Plan No
Morocco Yes Yes Yes No, but plan for 2014
Kazakhstan Yes Yes No Yes: other
Korea Plan No No, but plan
Kyrgyzstan Plan No No, but plan
Hong Kong No No
New Zealand Plan Yes No Yes: other
South Africa Plan No No No
Russian
Federation
Plan No No, but plan
Sudan Plan No No, but plan
Taiwan No No No
US Yes Yes Yes Yes: other
operated?
Interfederate?
operated?
Interfederate?
Other countries
Table 5.3.1.1 – continued Table 5.3.1.1 – continued

61
5.3.2 Digital Certificates
Certification Authorities (CAs) issue digital certificates that certify that servers or
users are who they claim to be. As such, they help to establish secure and reliable
communication between servers, between users, or between a user and a server.
Examples of such use include:
• a user with his browser connecting securely to a web server (server certificate);
• a user authenticating securely with a server (personal and server certificate);
• two users exchanging encrypted emails (personal certificates).
The Grid community requires secure authentication for users to log in to Grid
resources; this requirement is met by using personal certificates. At present,
server certificates are more widely used than client/personal certificates, as the
former are required whenever a secure connection is needed between servers, or
between a client and server.
In order to support the user community in deploying services securely (for
example, in eScience), many NRENs run a certification authority (CA). However,
certificates issued by these authorities are not automatically trusted outside the
NREN’s own domain. Therefore, a few years ago NRENs requested that TERENA
offer what subsequently became known as the TERENA Certificate Service (TCS,
see www.terena.org/activities/tcs). By December 2013, the TCS supported
28 NRENs for server certificates, 18 for personal certificates and eight for code-
signing certificates.
Many NRENs do not rely solely on the TCS for issuing certificates. Currently,
29 GÉANT partner NRENs operate certification authorities in addition to, or
independent of, the TCS (six more than last year). By far the largest of these is in
Germany. In seventeen cases, the CA is operated by the NREN itself. Twelve NRENs
use a commercial certification authority. Beyond the GÉANT area, many NRENs
also operate their own CA. These CAs issue server certificates; most issue personal
certificates as well. In recent years, several initiatives have been set up to create a
trust fabric within the academic community among academic CAs. One example
is TACAR (www.tacar.org), a repository set up by TERENA for safe storage and
distribution of root CA certificates. Another, more far-reaching example is the set
of Policy Management Authorities set up within the international Grid community.
The European body is the EUGridPMA (www.eugridpma.org); worldwide
collaboration is realised through the International Grid Trust Federation (IGTF,
www.igtf.net). Many NREN CAs are affiliated to the IGTF.
Although there were some problems with the underlying dataset, we estimate
that, as of 31 December 2013, there were around 150 000 valid outstanding
server certificates in the GÉANT region (10 000 more than one year earlier).
Spain, 4%
Sweden, 3%
Netherlands, 7%
Italy, 3%
Switzerland, 2%
UK, 19%
Austria, 3%
Belgium, 3%
Czech Republic, 3%
Denmark, 2%
France, 10%
Finland, 1%
Germany, 30%
Ireland, 1% Hungary, 1%
Greece, 1%
Norway, 2%
Poland, 2%
Portugal, 1%
Graph 5.3.2.1 – Division of server certificates, GÉANT partner countries

62
Around two-thirds of these had been issued under the TCS. How the server
certificates are divided among the countries is presented approximately in Graph
5.3.2.1.
In addition to server certificates, there are several other categories of certificates.
In terms of numbers issued, personal certificates are the second most important
category after server certificates. Germany has issued the vast majority of these:
almost 400 000 (compared to around 10 000 issued under the TCS).
5.3.3 eduroam
There is a secure, world-wide roaming access service that has been developed for
the international research and education community: eduroam®.
This service allows students, researchers and staff at participating institutions to
obtain Internet connectivity campus-wide and when visiting other participating
institutions, simply by opening their laptop and connecting to the eduroam WiFi
network.
As of June 2014, eduroam is available in 69 countries and territories worldwide,
up from 66 in 2013. There are 39 countries in the European eduroam
confederation. Up-to-date information on eduroam is available at
www.eduroam.org.
Even though eduroam may be available in a certain country, that does not mean
that all institutions or students in that country have access to it. Of course, the
fact that a student has eduroam credentials does not necessarily mean that he
or she actually uses the eduroam service; it has to be available at appropriate
locations. In October 2014, eduroam was available at more than 10 000 locations4
in the GÉANT area, 25% more than in 2013; the highest numbers of locations
were in the UK and Germany5
.
Take-up of eduroam has grown strongly in recent years, as illustrated by the
number of successful authentications in the GÉANT area over the past two years.
Between January 2013 and January 2014, the number of authentications more
than tripled and is now approaching the 150 million per month. The growth in
international authentications slightly outpaced the rate of national growth, as
shown in Graph 5.3.3.1.
120
0
60
160
140
20
100
80
40
20%
0%
5%
30%
25%
15%
10%
International
National
Millionsofauthentications
%ofinternationalauthentications
Jan. 2012 Sep. 2012 Jan. 2013 Sep. 2013 Jan. 2014 Sep. 2014
4
‘Location’here means a hotspot, a building or a set of buildings where an eduroam service is offered
by a single organisation.
5
Statistics provided by the eduroam operations team and collated by Miroslav Milinović, SRCE.
6
Data from the ETRL (European top level radius proxies logs) – not publicly available.
Graph 5.3.3.1 – Successful eduroam authentications, GÉANT area6

63
GÉANT Compendium of National Research and Education Networks In Europe /Collaboration support services
6 COLLABORATION SUPPORT
SERVICES
It is becoming easier to introduce collaboration support services, because
middleware services (see Section 5) are increasingly widespread. Below, Section
6.1 gives an overview of recent developments. Section 6.2 considers network
collaboration tools. Section 6.3 documents NREN involvement in the provision of
networked e-Science resources, including cloud resources. Section 6.4 examines
e-learning and Section 6.5 interaction with NREN clients. Section 6.6 explores
the relatively new areas of broker services and professional development. Finally,
Section 6.7 examines the area of software development.
6.1 Overview
Seventeen GÉANT partner NRENs currently offer a multimedia content
repository; ten others are planning to establish one.
We asked NRENs to provide information on three related resources in the area of
networked e-Science resources:
• Grid middleware;
• Computing power;
• Storage facilities.
In some countries, these three related resources are provided by the same
institution, but in many other countries the situation is mixed. In most countries,
the NREN has some involvement in providing these resources.
Seventeen of the GÉANT partner NRENs currently offer cloud services that are not
procured via a commercial vendor, and twelve others are planning to offer such
services. NRENs that offer a commercial storage service or that host commercial
content on their networks are far less common, though ten of the GÉANT partner
NRENs (and several NRENs outside the GÉANT region) already do, or are planning
to do, this. This situation has not changed much since last year.
Eighteen of the GÉANT partner NRENs currently provide an e-learning service,
and five others are planning work in this area.
Twenty-eight of the GÉANT partner NRENs have separate customer-support
departments. Sixteen GÉANT partner NRENs offer some form of best-practice
guidance to client institutions, mostly through individual consultation and by
publishing technical guides in various areas. Twenty-six GÉANT NRENs have
dedicated marketing and communications staff, though in most cases no more
than one or two staff members.
NRENs in the general category of ’brokerage‘. This seems to be an area in which
NRENs can achieve considerable savings for their customers and where there is
potential for expansion. NRENs are also undertaking framework procurements for
network and related equipment. Ten GÉANT NRENs offer some form of support
for procurement processes. The majority (21 NRENs) are interested in benefiting
from framework agreements put in place by another purchasing body, for a wide
palette of services.
Eleven of the GÉANT NRENs own intellectual property rights to network software.
Twenty-one are involved in open-source software development.
6.2 Network collaboration tools
Over the past ten years, NREN collaboration infrastructures and related services
have become the cornerstone of European and worldwide collaboration among
researchers and higher-education providers. Although collaboration hardware
and software has not changed profoundly in the past few years, its significantly
increased quality and reduced prices have made network-based virtual meetings

64
1
Available via the services map at compendium.terena.org/reports/nrens_services.
more appealing than ever. In research and education, collaboration technology
are playing a key role in making project, research and administration work more
effective, by virtually connecting remotely located personnel. Such remote
collaboration helps to optimise how time is used, to reduce travel costs and to
lower the environmental impacts of travelling.
Four pillars of the NREN collaboration infrastructure are:
1. Numbering schemes and Voice over IP (VoIP) to connect institutional IP
telephony deployments or individual end-users.
2. Video- and web-conferencing to provide a high-quality audio/video-based
collaboration environment, often enhanced by other tools enabling joint
work.
NB: This year’s Compendium does not report on elements 1 and 2.
3. Group collaboration services: i.e. the bundling of services that allow
collaborative groups to form and work together easily, independent of their
location.
NB: Information on element 3 is available online via the NREN Services map on
the Compendium website1
.
4. Multimedia content repositories for online presentation of materials recorded
by higher education and research organisations to complement remote
teaching/learning and science dissemination.
Element 4 is reported on below.
As in previous years, we asked NRENs about their use of multimedia repositories
(i.e. audio/video archives) and the streaming services they offer. As shown in Map
6.2.1 , a total of 16 GÉANT partner NRENs currently offer a multimedia content
repository, and 10 others are planning to establish one (no change since last
year). An online table1
gives additional information, including links to the various
repositories.
Deployed
Planned
Not planned
Map 6.2.1 – Multimedia content repositories

65
6.3 Networked e-Science resources
Because resources needed by scientists are increasingly being made available
via the Internet, this area has become important for NRENs. In many cases, the
NRENs already provide the networking infrastructure for such services and are
offering additional services to the Grid community. For this year’s Compendium,
we asked NRENs to provide information on three related resources in this area:
• Grid middleware;
• Computing power;
• Storage facilities.
In some countries, these three related resources are provided by the same
institution, but in many other countries the situation is mixed. In most countries,
the NREN has some involvement in providing these resources. In some countries,
not all of them are available.
Graph 6.3.1 gives an overview of the wide variety of existing models for provision
of such resources, in which we were unable to distinguish any geographic
pattern.
Unknown or no answer, 13%
Mix: NREN + other bodies, 26%
Individual institutions, 13%
Other body, 8%
Mixed institutions, not the NREN, 2%
NREN, 30%
Mix: NREN + individual institutions, 4% NREN + another body, 4%
6.3.1 Cloud services
Seventeen of the GÉANT partner NRENs currently offer cloud services that are not
procured via a commercial vendor (up from 16 last year), and twelve others are
planning to offer such services. This year, five NRENs brokered agreements with
cloud service providers (up from one last year). Several other NRENs are considering
doing this. Table 6.3.1.1 gives details for those NRENs that have either existing
activities or plans in this area. Changes with respect to 2013 are highlighted.
NRENs offering non-commercial services are far more common than those
offering commercial storage services or hosting commercial content on their
networks. However, a number of NRENs are already involved, or are planning to
become involved, in such commercial activities. Many NRENs, including 19 of
the GÉANT partners NRENs (the same number as last year), mirror content from
outside the NREN network. Because little has changed in this area since last year,
the full information is not repeated in this year’s Compendium.
Graph 6.3.1 – Provision of networked e-Science resources

66
Country Cloud services not
obtained through
a vendor?
Virtual machines? Details Which other services do you offer or are planning
to offer?
Does your NREN broker
agreements with cloud
service providers?
GÉANT countries
Armenia Yes Yes No
Belarus Yes Yes Standard VMs, small resource pool available. No
Belgium Yes Planned IaaS Storage and computing. No
Cyprus Planned Planned Back-Up MX, Primary DNC, etc. No
Czech Republic Yes Yes Distributed storage infrastructure with 22 PB raw
capacity.
IaaS HPC cloud, PaaS for scientific users planned, SaaS
and storage services under discussion.
No
Denmark Planned Yes KVM Storage and computing. Yes: IPNETT, BOX: Box.com
+ XaaS
Estonia Yes Yes VMs SaaS No
France Planned
Georgia Planned Planned No
Germany Yes Planned Yes DFN-members, non-
commercial: storage
Greece Yes Yes We offer two IaaS services addressed to different
end-users (i.e. ViMa is offered to NOCs and ~okeanos
to end-users). The characteristics of the VMs vary.
Maximum allowed values are 4vCPUs, 4GigRAM,
100Gig Disk. More resources may be provided upon
request and justification.
Computing and Storage services.
~okeanos (https://okeanos.grnet.gr/home/) is a
brand new IaaS Service based on the GRNET cloud
computing infrastructure. Cyclades is the Virtual
Compute and Network service of ~okeanos. With
Cyclades you can build your own virtual machines,
always connected to the Internet, manage them,
destroy them, connect to them and take a handful
of other actions, all from inside your favourite web
browser. Pithos+ is the Virtual Storage service of
~okeanos. Pithos+ enables you to store your files
online, share them with friends and access them
anytime, from anywhere in the world.
No
Hungary Yes Yes KVM-based IaaS cloud, virtual networks, additional
geographically distributed huge storage, multi-site
deployment.
IaaS virtual machines, virtual interconnects. No
Ireland Yes Planned EduStorage (Cloud storage). Planning VMs as a service, planning cloud computing
(virtual servers).
Yes: Microsoft, Blackboard,
Moodlerooms: storage,
computing, LMS
Israel Planned Yes Planned
Table 6.3.1.1 – Cloud services

67
Italy Planned Yes Aiming to provide up to a few thousands VMs, with
different configurations: from single-CPU to many.
IaaS, Storage-aaS (personal, initially), PaaS ( similar to
Identity Provider as a Service).
Planned: ownCloud,
Google, Box.
Integration of commercial
and NREN storage provision
in a way that is transparent
to end-users.
Latvia Yes Yes No
Lithuania Planned Planned Currently in the planning stage. Backup, storage, webspace, email. No
Macedonia Yes Part of health information system (Ministry of health).
Moldova Planned
Netherlands Yes Yes Commercial (cloud) service agreements are provided
via SURFmarket, HPC services via SURFsara.
HPC services of SURFsara among others, for an
overview see https://surfsara.nl/SURFsara-services.
Overview of all the cloud services from commercial
service providers and member institutes is available
at http://www.surf.nl/diensten-en-producten/
surfconext/op-surfconext-aangesloten-diensten/
index.html.
Yes: see http://www.surf.
nl/diensten-en-producten/
surfconext/op-surfconext-
aangesloten-diensten/
index.html. Current focus
is on Microsoft, IaaS
providers, PaaS providers,
Identity-as-a-Service
providers, global services
(and others).
Norway Planned Planned IaaS, PaaS, SaaS. Yes: Box.com. Personal
storage and collaboration
Poland Yes Yes Virtual machines, applications on demand, cloud
storage.
On-Demand computing (IaaS, PaaS, SaaS), cloud
storage: remote distributed file system (National Data
Storage, OpenStack Storage (for several projects).
Planned: Microsoft Office
365.
Portugal Planned
Romania Planned No N/A N/A No
Slovenia Yes Yes Minimum setup: 4x vCPU, 8GB RAM, 160GB disk. IaaS No
Spain Planned Planned We offer cloud spam filtering. We are planning to
offer more services – main candidates are mail, VoIP,
electronic signature, IaaS and webconferencing.
Planned: Spamina
Sweden Not planned Planned, will hopefully be
launched end of this year.
obtained through
a vendor?
to offer?
service providers?
GÉANT countries
Table 6.3.1.1 - continued

68
Switzerland Yes Yes VMs based on OpenStack, offered to all universities
who opt in.
SWITCHdrive – (‘sync share’) service based on
OwnCloud.
No
Turkey Yes Yes OpenStack with KVM. Production level IaaS and Hadoop services are offered. No
UK Not planned We have framework agreements for File Sync and
Share and Telephony.
Yes: Arkiving, Google,
Microsoft. Archive tape
storage; Google Apps for
Education; Microsoft 365.
Ukraine Yes Yes
Other countries
Algeria Planned Yes No
Australia Yes No Storage, videoconferencing, VPN. AARNet needs to
work in harmony with government programmes
NeCTAR and RDSI to avoid duplication of cloud and
related services.
Yes: Box, Zoom. Cloud
storage, videoconferencing.
Bosnia/
Herzegovina
Planned No No
Brazil Planned Planned Storage Planned
Kazakhstan Planned No No
Korea (South) Planned Yes VMWARE DTN (Data Transfer Node): maximizing data transfer
performance.
No
Morocco Planned
New Zealand Yes Net+ - http://www.internet2.edu/netplus/.
Russian Federation Yes Yes Cloud services on Openstack + ClasterFS platform. No
Sudan Planned Planned Web hosting, DNS, Mail. No
obtained through
a vendor?
to offer?
service providers?
GÉANT countries
Table 6.3.1.1 - continued

69
Deployed
Planned
Not planned
6.4 e-Learning
As shown by Map 6.4.1, 18 of the GÉANT partner NRENs currently provide an
e-learning service, up from 17 in 2013. Five others (highlighted in yellow) are
planning work in this area, up from four last year. Turkey has started a new
e-learning service. In many cases, further information can be found on the NREN
websites. Several NRENs beyond the GÉANT region are also active in this area.
6.5 Clients interaction, and knowledge
dissemination
As in previous years, almost all NRENs provide some form of training courses to
their users, and most organise national user-conferences. Compared to previous
years, the situation has not changed much. For further information about such
activities and the associated resources that NRENs make available to their users,
see the separate TF-CPR Compendium:
www.terena.org/activities/tf-cpr/compendium.
Twenty-eight of the GÉANT partner NRENs have separate customer-support
departments. Twenty-seven GÉANT NRENs publish on their website lists of
the services they offer. Sixteen GÉANT partner NRENs offer some form of best-
practice guidance to client institutions. As clearly shown by Table 6.5.1, such
guidance can take many forms; most common are individual consultation with
member institutions and publication of technical guides in various areas. In
addition, campus best-practice guides are developed through the GN3plus
project: more information is available at:
http://services.geant.net/cbp.
Map 6.4.1 – NREN e-Learning services

70
0
2
10
4
1
3
1 or 2
8
5
NumberofGÉANTNRENs
Number of staﬀ
6
5
7
9
3 - 5 1
Graph 6.5.2 – NREN marketing/
communications staff
Table 6.5.1 – NREN best-practice guidance
Belgium We offer technical advice (if needed on site, free of charge) for all NREN-related
services to our connected institutions and develop best practices for our
connected institutions.
Czech Republic Mainly by means of personal consultations for individual cases.
Estonia See: http://kuutorvaja.eenet.ee/wiki/Kuutõrvaja.
Finland Best-practice documents are available in a wiki: https://info.funet.fi.
Germany Video services, security services , legislation, customer contracts.
Ireland Security, networking, eduroam, federation.
Lithuania Consultations and advisory upon request.
Netherlands Many different areas, including network, collaboration infrastructure, etc. See:
http://www.surf.nl.
Norway We publish UFS (whitepapers on campus best practices) and consultancy on
demand.
Romania On best-effort basis.
Serbia AMRES is working on best-practice documents for client institutions; the results
are available at http://cbp.rcub.bg.ac.rs/.We are also participating in the GEANT
CBP task.
Slovenia Recommendations on network setup and management, as well as IPv6
transition. We also introduce Universities to campus best practice results.
Switzerland AAI guides.
Turkey E-mail list, blog, documents, presentations on national user-conference.
Other countries
Australia Consultancy services: see
http://www.aarnet.edu.au/services/enterprise-services-consulting.
Hong Kong See: http://www.jucc.edu.hk.
Korea (South) ScienceDMZ, PerfSONAR, network architecture of campus network, security, etc.
Sudan Through workshops provided to members, meetings and consulting.
This year, for the first time, we
asked the NRENs whether they
have dedicated marketing/
communications staff and if so, how
many staff members they have in
this area. As summarized in Graph
6.5.2, 26 GÉANT NRENs have some
staff in this area, although for most
of them the staff allocation is no
more than two and in many cases
less than one staff member. The
largest departments, with more than
five staff members, are in Belgium,
the Netherlands, Switzerland and the
UK. Many of the non-GÉANT NRENs
also have staff in this area; RNP of
Brazil reports the largest staff size.
6.6 Broker and professional services
NRENs in the general category of ’brokerage‘; that is, a NREN using its expertise
and knowledge to engage with the market on behalf of its clients. A prime
example of such brokerage is software licensing, whereby NRENs can negotiate
bulk deals at the national level for generic, e-learning and other applications. This
seems to be an area in which NRENs can achieve considerable savings for their
clients and where there is potential for expansion.
NRENs are also undertaking framework procurements for network and
related equipment. Such procurements are often directed primarily at NREN

71
requirements, though client institutions can use the negotiated terms to their
advantage by purchasing equipment for their own networks. Maintenance and
support contracts are often part of such frameworks; in some cases, there is a
demand for the NRENs to manage these contracts as well.
This year, we asked NRENs whether they provide support for procurement
processes or whether they can use a centralised purchasing body. We also asked
whether NRENs provide premium professional services or are involved in any
e-government public services such as e-voting, e-ID issuing or e-ID related
applications. The results, shown in Table 6.6.1, are quite similar to those of 2013.
A: Type of procurement
process support:
B: Can you make use of the
services of a centralised
purchasing body?
C: Do you provide premium
professional services?
1 Joint procurement of
software and software
licenses.
Yes, we have to use such a
body.
Consultancy services.
2 Joint licensing for digital
libraries.
Yes, we are free to use this
body if we want to, but only
for our own needs.
Security audits.
3 Joint procurement of
equipment or negotiation of
preferential rates for users or
end users.
Yes, and we can also use this
for our client institutions.
Support on implementing
NREN services (eduroam, IPv6,
BoD).
4 Other. Other services.
Country A: Type of procurement process support: B: Can you make use of the services
of a centralised purchasing body?
C: Do you provide premium professional services? Involvement in e-government
services
Armenia 3: equipment procurement 1: consultancy No
Austria No 1: obligatory 3 No
Belgium Yes 1: consultancy Yes
Croatia No No 2: security audits Yes
Cyprus 4: other 1: e-procurement from government 3 No
Czech Republic 1: software procurement 2: optional 1: consultancy No
Denmark No 3: also for clients: both state
procurement agency and Nordunet
2: security audits No
Estonia No No 1: consultancy No
Finland 1: software procurement 2: optional 1: consultancy No
France No 1: consultancy No
Georgia No No 1: consultancy No
Germany No No 2: security audits No
Greece No No 3 Yes
Hungary No No 4: other No
Table 6.6.1 – Procurement, broker and professional services

72
Ireland 2: digital libraries, 3: equipment
procurement
2: optional 2: security audits Yes
Israel 1: software procurement 1: obligatory 3 No
Italy No 2: security audits No
Latvia No 2: optional 3 No
Lithuania No 1: consultancy Yes
Luxembourg No 2: optional 3 No
Macedonia No 3 Yes
Moldova Planned 3: also for clients 3 No
Montenegro No No 3 No
Netherlands 3: equipment procurement 1: obligatory 1: consultancy Yes
Norway 1: software procurement 3: also for clients 1: consultancy No
Poland 3: equipment procurement No 1: consultancy Yes
Portugal No 3 Yes
Romania No 1: obligatory No
Serbia No No 3 No
Slovakia No No 3 No
Slovenia No No 1: consultancy No
Spain Yes: dark fibre networks for regional RE
networks
3: also for clients No Yes
Sweden No 2: optional No Yes
Switzerland 1: software procurement No 3 Yes
Turkey No No 1: consultancy Yes
Ukraine No 2: optional 3 Yes
Other countries
Algeria 2: digital libraries 3 No
Australia Consulting on an ad hoc basis, see: http://
www.aarnet.edu.au/services/enterprise-
services-consulting
2: optional 1: consultancy No
services

73
services
Other countries
Brazil Planned 3: also for clients 3 Yes
Canada 3: equipment procurement 3 No
Hong Kong 1: software procurement 3 No
Korea (South) No 1: consultancy No
Lebanon 3: equipment procurement 2: optional No No
Morocco No No 1: consultancy No
New Zealand 3: equipment procurement 1: consultancy No
Russian Federation Planned 2: optional Yes, consultancy services Yes
Sudan No No 3 No
We also asked the NRENs whether they are or would be interested in providing
services to other NRENs. Five GÉANT partners already do this, and six others
are interested in doing so. We also asked the NRENs whether they would be
interested in benefiting from framework agreements put in place by another
Country Services to other NRENs Interested in framework agreements put in place by another contracting authority/
centralised purchasing body?
Belgium No, but we are interested in doing this. Yes, how is not clear yet.
Cyprus No Yes: equipment (routers, switches, etc.).
Denmark No Yes: infrastructure equipment and cloud vendors.
Finland No Yes: cloud services, data security services.
Georgia No Yes: would be good to have centralised purchasing for South Caucasus countries for GEANT
connectivity channel.
Germany No Yes
Greece No, but we are interested in doing this: cross-border link to enable diversity of
physical routes. Cloud services.
Yes
Hungary No, but we are interested in doing this: any services another NREN can reasonably
exploit (and vice versa).
Yes: practically everything related to our e-Infrastructure activities. However, public procurement
rules and regulations may be different in the various countries.
purchasing body. The majority (21) of the GÉANT NRENs is interested in this,
for a wide palette of services ranging from communications and equipment to
software licenses and library periodicals. The full results are shown in Table 6.6.2.
Table 6.6.2 – NREN to NREN services

74
Iceland No Yes: comms eq/IRUs
Ireland Any of our existing services with cost-recovery charging. Yes: software licensing, information systems, cloud services, data transmission.
Israel No Yes
Lithuania No, but we are interested in doing this: connectivity, transit and expertise. LITNET
will provide 10G lambda to GEANT POP, and a peering connection, and fibre to
the border with Belarus if BASNET wishes to connect Minks-Vilnius.
Luxembourg No Yes: network equipment.
Malta Yes: software/support licencing; library periodicals.
Moldova No, but we are interested in doing this. Yes: equipment , license to scientific journals.
Netherlands No, but we are interested in doing this. This subject is too complex to be shortly
described (i.e., within the scope of this study) - subject for TF MSP, GEANT etc.).
Yes: pan-European NREN services, supporting international collaboration (international, cross-
institutional cross-sectoral).
Norway Yes: marketing communication services via NORDUNet. Yes: cloud services, software licenses etc.
Poland Yes: connectivity services for neighbouring NRENs. Yes
Romania Yes: RENAM is connected to GEANT Bucharest POP and Internet using our
infrastructure.
Yes
Slovakia Yes: cross border optical connection to AT, CZ and PL. No
Spain No Yes
Switzerland No Yes: SW and SaaS products.
UK Yes: video Conferencing to HEAnet and Belnet. No
Ukraine Yes
Other countries
Australia No, but we are interested in doing this: advice and guidance (via APAN) – not
charged; FileSender CloudStor+ – not charged; negotiations with vendors (e.g.
Box) on behalf of NRENs operating under similar legal frameworks – not charged.
Yes: network hardware software, cloud services, etc.
Brazil No, but we are interested in doing this. Yes: cloud providers, applications, links.
Canada Yes: wavelength and lightpath services. Yes: leveraging bulk buying of wavelength services to reduce costs.
Korea (South) No, but we are interested in doing this. No
Russian Federation No, but we are interested in doing this. No
Sudan No, but we are interested in doing this: libraries, video conferences, access to
network and researches.
Country Services to other NRENs Interested in framework agreements put in place by another contracting authority/
centralised purchasing body?

75
6.7 Software development
Eleven of the GÉANT NRENs (up from nine in 2013) own intellectual property rights
to certain software. Twenty-one are involved in open-source software development
(up from 19 in 2013). NRENs most frequently cite financial and human resources
Country Owner of IPR for
software?
Categories of software products owned Involved in open-source
software development for
network services? If not, what
are the obstacles?
URL to software repository
Armenia No Yes
Belgium No No: lack of resources
Croatia No Yes Microsoft Download Center,
http://www.carnet.hr/internet_services/MSDC
Czech Republic Yes Monitoring tools, security, multimedia and middleware. Yes Redmine (https://homeproj.cesnet.cz/), GIT
Denmark No Yes
Estonia Yes Computing resource management, traffic measurement
software, server factory.
Yes
Finland Yes Small-scale monitoring and management tools. Yes
France No Yes http://www.sympa.org
Georgia No No: lack of resources
Germany Yes Intrusion detection (NeMo). Yes NeMo
Greece Yes All home-grown services are open-source (e.g. virtual
machines provision, firewall on demand, e-voting
software).
Yes https://code.grnet.gr/ and https://github.com/grnet/ Mercurial,
Git Subversion (only for backwards compatibility) repositories
are supported for hosting code. By default, Git is used and Git
repositories are automatically created upon creation of a project,
typically after a 2-minute period.
Hungary No Yes Major SW repositories at NIIFI: https://repo.niif.hu/gitweb/ (open)
and https://repo.niif.hu/trac (Shibboleth protected/internal
usage).
Iceland No No: no human resources
Ireland No Yes We have no public repository. Instead, all software we write is
developed under the BSD licence model, and is available for free
for others to use.
Italy Yes Network Monitoring tools (GINS),Identity Provider in the
Cloud.
Yes Not yet deployed.
as obstacles to becoming involved in this area, although for some NRENs software
development is not in their remit. Table 6.7.1 provides an overview.
Table 6.7.1 – Open-source software

76
Lithuania No Yes https://github.com/fln/addrwatch, https://github.com/fln/nat64
Luxembourg No Yes http://downloads.geant.net/ -‘Roaming’ -‘eduroam CAT’
Moldova No No: Currently, main network
services are implemented by using
open-source software, but NREN
personnel is only adopting existing
open-source products and is not
yet involved in developing such
products.
Netherlands Yes MediaMosa, OpenConext (middleware and
applications), Tiqr, FileSender, Apache RAVE, Mujina and
APIS (most of these open source software is available
via Github) .
Yes Github (several)
Norway Yes Tools for Network Monitoring (http://software.uninett.
no/).
Yes http://software.uninett.no/, http://www.filesender.org,
https://nav.uninett.no/
Poland Yes Digital libraries, network management, HD video. Yes No repository yet.
Portugal No No: lack of resources.
Romania No No: no human resources.
Serbia No No: AMRES has no software
development department.
Slovakia No Yes
Slovenia Yes VoD portal, Portal for Adobe Connect, MCU portal. Yes Filesender: https://filesender.arnes.si/
Spain No No https://forja.rediris.es/
Switzerland Yes AAI middleware, video management. Yes
Turkey No Yes https://github.com/ULAKBIM/NfQuery
Ukraine No No: not in current remit.
Other countries
Australia Yes FileSender (co-owned with UNINETT, HEANet and
SURFnet); CloudStor+ based on ownCloud. See: http://
www.aarnet.edu.au/services/netplus/catalogue.
Yes http://www.filesender.org
Bosnia/Herzegovina No No: no human resources.
software?
are the obstacles?

77
software?
are the obstacles?
Other countries
Brazil Yes Hardware Security Module (HSM) Yes https://github.com/
Canada No No: we don’t do software
development.
Korea (South) Yes Yes
Lebanon No Yes OpenStack
Morocco No Yes
New Zealand No Yes

78

79
7 FUNDING AND STAFFING
Some NRENs provide services only to their country’s research and/or education
community. Others also provide services beyond this community; for
example, they administer the country-code top-level domain, or they connect
companies and/or institutions outside the research or education community.
To enable comparison, we asked the NRENs covered by this 2014 edition of
the Compendium to provide information only about their activities for national
research and/or education communities. We refer to such activities simply as
‘NREN activities’.
Section 7.1 gives an overview of recent developments in funding and staffing.
Section 7.2 details the considerable differences in the number and types of staff
that NRENs employ and attempts to explain some of these differences. Section 7.3
provides information on, and explains the variety of, NREN budgets. Sections 7.4
and 7.5 give further information on income sources and expenditure categories,
respectively. Finally, Section 7.6 details NREN expenditure by network level.
7.1 Overview
It is no easy task to compare NRENs by staff or budget size, because their budgets
are variously structured, depending on their tasks, while their funding also differs
greatly. Compared to 2009, overall budgets in the GÉANT area have risen by 2%,
while staff size has increased by 31%. Transmission capacity and equipment costs
have decreased, but new services have necessitated staff size increases.
Comparing 2014 budget data with those from previous editions of the
Compendium reveals that, generally, NREN budgets decreased with respect to
2013, though this is not necessarily the case for individual NRENs. For some NRENs,
the budget situation is unclear because final budget decisions are made in the
course of the year. Obviously, this is not to be recommended, as it makes forward
planning very difficult for the NRENs in question. The general trend is that, each
year, NRENs are able to deliver more bandwidth and more services for roughly
GÉANT Compendium of National Research and Education Networks In Europe / Funding and staffing
the same amount of money as in the previous year. The Czech Republic, Estonia,
Lithuania and Serbia all suffered budget cuts of more than 15%.
Over the past few years, infrastructural investments have led to savings in
transmission costs. In addition, the resulting infrastructural improvements,
coupled with innovations in the area of authentication and authorisation, have
enabled a new generation of networked services, which have required some
increases in staff size.
Although it is impossible to make general recommendations on NREN funding
mechanisms, a model that in some way involves all the various stakeholders in an
NREN would seem to provide the best guarantees for its continued success. In their
respective fields, many NRENs are engaged in innovations, which are often steered
by dedicated funding mechanisms. It is important for NRENs to use such funds to
their advantage wherever they exist.
7.2 Staffing
Graph 7.2.1 gives an overview of all the staff who are directly employed in NREN
activities, as well as subcontracted staff, in full-time equivalents (FTE). Graph 7.2.2
provides similar information specifically for technical staff. The data is presented in
this way because many NRENs use subcontractors; as a result, staff size alone is not
a reliable indicator of the total person-power available to an NREN.
As in previous years, considerable differences from NREN to NREN are evident, not
only in the number of staff employed but also in their set of skills. One explanation
for these differences is that, in some NRENs, the research network is a service
provided by a parent organisation; therefore, it is not possible for all such NRENs
to specifically estimate the non-technical staff time (e.g. in accounting and human
resources, etc.) allocated to NREN activities. This partially explains why some
NRENs have a higher ratio of technical to total staff than others.

80
NRENs differ considerably in the tasks they perform: for example, some provide
connections to metropolitan area networks (MANs) or to access networks, which
in turn connect institutions. Other NRENs connect institutions directly, and some
manage MANs themselves. The connection policies of NRENs also differ with
respect to secondary and primary schools, for example. This affects the remit
of the NRENs and explains some of the differences in staff numbers that are
apparent in Graphs 7.2.1 and 7.2.2.
Finally, some NRENs provide extensive support to individual end-users at
institutions, some provide limited customer support, and many have service
levels that are somewhere in between. This factor can have a significant effect on
staff levels.
0
20
40
60
120
80
200
180
100
160
140
Directly employedBelarus
Belgium
Bulgaria
Croatia
Cyprus
CzechRepublic
Estonia
France
Georgia
Germany
Greece
Hungary
Ireland
Israel
Italy
Latvia
Lithuania
Luxembourg
Macedonia
Montenegro
Netherlands
Portugal
Serbia
Slovakia
Australia
Algeria
HongKong
Kazakhstan
Kyrgyzstan
Austria
Denmark
Finland
Iceland
Korea
Moldova
Morocco
NewZealand
Norway
Romania
RussianFederation
Lebanon
Boznia/Herzegovina
Canada
Slovenia
Spain
Switzerland
Turkey
UK
Sudan
Sweden
External / outsourced
Graph 7.2.1 – Total NREN staff in FTE

81
0
20
60
40
120
100
80
Iceland
Australia
Cyprus
Macedonia
Morocco
Lebanon
Georgia
Bosnia/Herzegovin
Austria
Israel
Montenegro
NewZealand
Latvia
Luxembourg
Sweden
Algeria
Canada
Romania
Belarus
Moldova
Sudan
Armenia
Turkey
Estonia
RussianFederation
Finland
Bulgaria
Belgium
Spain
Germany
Lithuania
Denmark
Serbia
Slovenia
Italy
Slovakia
Hungary
Ireland
Switzerland
France
Norway
Croatia
Greece
UK
Netherlands
Portugal
CzechRepublic
Korea
Sudan
Kazakhstan
Ukraine
Kyrgyzstan
10
50
30
110
90
70
Graph 7.2.2 – NREN technical staff in FTE

82
RNP of Brazil has more staff than any European NREN: in June 2014, it had 240
employees and 182 outsourced staff.
Graph 7.2.3, which compares budget with staff size for the entire GÉANT region,
indexed on 2009 (=100), illustrates the key points made above in this section.
Clearly, budgets grew only moderately until 2012, decreasing again afterwards.
Nevertheless, staff size has grown considerably over the entire period. As
documented in Section 7.5, the costs of transmission capacity and hardware have
decreased. Over the same period, improvements in infrastructure, coupled with
innovations in authentication and authorisation, have enabled a new generation
of networked services, and where these services are administered by NRENs, they
require more staff.
90
125
100
20122010
95
2009
105
110
130
135
2011 2013 2014
Budget
120
115
Graph 7.2.3 – Total NREN budget and staff size in the GÉANT partner countries,
2009–2014, indexed on 2009 (=100)
7.3 Total budgets, 2009-2011 and 2012-2014
NREN budgets may fluctuate as investment levels vary from year to year. In order
to filter out as much of this effect as possible, in Graphs 7.3.2 and 7.3.3 we have
compared the total NREN budgets averaged over two three-year periods: 2009-
2011 and 2012-2014. Note that for Janet (UK), the financial year is from August to
July; therefore, its 2014 budget is actually its 2013/2014 figure.
How total annual budgets have varied over the period 2009 to 2014 is shown in
Graph 7.3.4, together with the growth in GÉANT traffic1
.
For some NRENs, the budget situation is unclear because funders make their final
budget decisions in the course of the year rather than at the beginning. Obviously,
for the NRENs in this category this makes forward planning more difficult, if
not virtually impossible. This year, we asked the NRENs that face such budget
uncertainties to give greater details of them than in previous years. The responses
are shown in Table 7.3.1.
1
Note, though, that traffic through the GÉANT network is only one of many components of a NREN's
traffic.

83
Country Uncertainties
Belgium The standard process is as follows:
1. Initial budget provided at the beginning of summer of the previous year ;
2. Budget adjustment occurs at the end of the first quarter of the current year.
Several other budget adjustments may also be made at the Government’s
discretion.
Croatia Our budget falls under the state budget and is corrected during the year.
Cyprus Due to the unstable financial situation and the freeze on Government funding,
CYNET has difficulties implementing projects/services that are essential to the
members. Therefore, discussions with members on co-funding are ongoing.
Georgia Our budget depends on payments from users and projects.
Hungary Every year there is some uncertainty in the budget, and 2014 seems to be very
similar to 2013 in this sense.
Macedonia MARnet financing is covered partially by the state budget. It is usually determined at
the end of the previous year for the next year.
Other internal sources should also be estimated as part of the predefined budget,
together with the state budget.
Moldova We have already indicated the 2014 budget estimate. We can only recalculate the
real sum at the end of the reporting year.
Slovakia The basic budget of SANET running cost is included in the chapter‘Education and
Research Infrastructure’of the budget set by the Ministry of Education and Research
of Slovakia, according to the law.
Turkey An additional budget will be requested for 2014.
UK The budget supplied is based on the current prediction; however, this level of
funding has yet to be finalised. In addition, we may secure further grants for projects
in 2014/15 which we are currently not aware of. Our income is partially government
funding and partially from sales to the sector. Therefore, our expenditure is
dependent on achieving the forecast sales level.
Ukraine A great number of payments have been delayed or cancelled.
Territories were annexed or invaded.
Other countries
Bosnia/
Herzegovina
Our budget is set according to the Law on Budget Execution, but the actual financial
situation may lead to a certain decrease.
Brazil The majority of expected revenue is confirmed during the year and is subject to
negotiation of Addenda to the multi-annual Management Contract.
Morocco Our NREN is in this category. Our budget comes from the Ministry of Higher
Education and we have to renegotiate it each year.
Sudan SudREN falls into this category.
For several reasons (see bulleted list) it remains difficult to directly compare
budgets. We asked the NRENs whether their submitted budget figures include the
EU grant for GÉANT activity. For some NRENs, this is the case; for others, this grant
is shown not as part of the budget but as a reduced cost. In Graphs 7.3.2 and 7.3.3,
the NRENs that include the GÉANT subsidy in their budget figure are marked with
an asterisk*. As shown in Section 7.4, the proportion of funds received from the
EU (though not always exclusively for GÉANT) differs considerably. There are other
reasons why comparison is difficult:
Graph 7.3.2 – Total budgets, 2009-2014 averages,
GÉANT partner countries with annual budgets 10 M€
0
10
30
90
80
M€
20
50
40
60
70
Average 2009-2011
Average 2012-2014
AVERAGE
Belgium*
Croatia*
Switzerland*
Sweden
Germany*
France*
Ireland*
Italy*
Netherlands*
Norway*
Spain*
CzechRepublic*
Turkey*
UK*
Greece*
Table 7.3.1 – Budget uncertainties

84
A few NRENs, including GRENA of Georgia, do
not have budget certainty until quite late in the
year. This may be because there is uncertainty
about projects, payments from client institutions
or government funding. In some cases, salary
payments to NREN staff are relatively stable, whereas
additional budgeting for investments is only known
later in the year. Obviously, a NREN that is to be at
the forefront of Internet developments in its country
requires reliable budgeting.
Graphs 7.3.2 and 7.3.3 show a mixed situation:
some NREN budgets have increased, others have
decreased, but the overall average has remained
virtually unchanged. Comparing the three-year
averages reveals quite large budget decreases in
Bulgaria, Romania, Portugal, Spain and Greece.
In part, these can be explained by investments in
network upgrades in the earlier period that led to
savings in the later period. Budgets increased in
Belgium, the Czech Republic and Norway.
Map 7.3.3, which compares 2013 and 2014, reveals
the relatively large budget decreases in the Czech
Republic, Estonia, Lithuania and Serbia. For Lithuania,
this was the second budget decrease in two
consecutive years. There were relatively large budget
increases in Belarus, Belgium, Finland, Israel, Latvia,
Luxembourg, Moldova and Switzerland. The reasons
for these changes are different for every NREN.
• Regional and/or metropolitan area networks (RANs/MANs) are funded differently in different countries;
• In some countries, clients pay for their link to the nearest NREN point of presence; in others, the NREN pays for this;
• Some NRENs spend a large part of their budget on connecting primary and secondary schools; others do not or may
take this separately into account;
• There are large differences in how staff are paid. In the GÉANT area, one NREN spends only 2% of its budget on staff,
whereas another spends 59% of its budget on this aspect. In this context, it should be noted that some NRENs have
staff who are not paid from the NREN budget. Similar differences also exist in other expenditure categories.
0
2
8
12
M€
4
Austria*
Bulgaria*
Cyprus*
Denmark
Estonia*
Finland*
Hungary*
Iceland
Latvia*
Lithuania
Luxembourg*
Slovakia*
Serbia
Slovenia
10
Israel*
Macedonia
Georgia
Portugal*
Romania*
AVERAGE
Average 2009-2011
Average 2012-2014
6
Montenegro
Armenia
Moldova*
Belarus
Graph 7.3.3 – Total budgets, 2009-2014 averages, GÉANT partner countries with annual budgets 10 M€

85
Budget increase 5%
Budget increase 5%
Budget decrease 5 - 15%
Budget decrease 15%
No data received or
decrease 5%
Budget decrease 5%
0
350
100
20122010
50
2009
150
200
400
2011 2013 2014
300
250
NREN Budgets
Graph 7.3.4 – Total NREN budgets and traffic, 2009-2014,
GÉANT partner countries (indexed on 2009 = 100)
Traffic has more than tripled since 2009, as illustrated by the figures for traffic on
the GÉANT backbone, which are plotted in Graph 7.3.43
:
As Graph 7.3.4 shows, the investments in infrastructure upgrades that many
countries have made in recent years have enabled steady traffic growth for
roughly the same amounts of money each year (see also Section 4.3). This has also
enabled growth in both diversity and number of services offered on the network.
The data obtained from the non-GÉANT countries are not sufficiently time-
consistent to allow them to be presented in the same form as those from the
GÉANT partner countries.
The current funding levels of several of the non-GÉANT countries that provided
information in the survey are unlikely to be sufficient for them to reach the
standards now common in the GÉANT area, despite the falling prices of
connectivity in recent years.
3
The 2014 traffic data are an estimate, based on extrapolating the figures for the first six months of
2014. Note that this excludes traffic on lightpaths (lambdas) that are not monitored by GÉANT.
Map 7.3.3 – Budget increases and decreases in the GÉANT area, from 2013 to 2014

86
7.4 Income sources
NRENs are funded in various ways: some receive all their funding directly from
national government; others are funded entirely by their users (who may, in turn,
be government-funded to some extent). Between those extremes there are many
variants. Graphs 7.4.1 and 7.4.2 indicate what percentage of NREN funding comes
from which source. Note that in many cases (see also Graphs 7.3.2 and 7.3.3) the
amount of funding received from the EU is not included.
Although it is impossible to make general recommendations on NREN funding
mechanisms, a model that in some way involves the various stakeholders in
an NREN would seem to provide the best guarantees for its continued success.
It should be noted that, in their respective fields, many NRENs are engaged in
innovations, which are often steered by dedicated funding mechanisms. NRENs
may consider applying for such funds wherever they are available.
As in previous years, NRENs were asked whether they can make use of multi-annual
budgeting. Of the GÉANT partner NRENs, 45% (up from 39% last year) confirmed
that they can use such budgeting.
EU
Other sources
Users/clients
National governments
and public sources
Bulgaria
Greece
Lithuania
Romania
Serbia
Slovenia
Spain
Turkey
Croatia
France
Portugal
Slovakia
Luxembourg
Ireland
Macedonia
CzechRepublic
Estonia
UK
Belgium
Hungary
Montenegro
Norway
Belarus
Georgia
Cyprus
Switzerland
Finland
Netherlands
Latvia
Sweden
Moldova
Israel
Italy
Germany
Austria
Denmark
Iceland
0
10
20
30
40
50
60
70
80
90
100
%oftotalincome
Ukraine
Graph 7.4.1 – Income sources, GÉANT partner countries

87
7.5 Expenditure by category
Graphs 7.5.1 and 7.5.2 show the average percentage of NREN income spent on
various cost categories. To enable comparison, the graphs show the situation in the
same set of (current) EU/EFTA countries for 2009 and 20144
.
Since 2009, the proportion of transmission capacity costs in the EU/EFTA countries
has decreased from 47% of total expenditure to 43%. The proportion of salary costs
has increased (in line with the rise in the number of staff, as documented in Section
7.2) from 19 to 25%. The proportion of equipment costs has declined slightly.
%oftotalincome
0
10
20
30
40
50
60
70
80
90
100
Other sources
EU
and public sources
Users / clients
Bosnia/Herzegovina
Brazil
Canada
Algeria
NewZealand
Morocco
RussianFederation
Lebanon
Sudan
Australia
HongKong
Kyrgyzstan
USA
Other sources
EU
and public sources
Users / clients
Sudan
Australia
HongKong
Kyrgyzstan
USA
4
This includes those of the current 28 EU countries that submitted data, plus Iceland,
Norway and Switzerland.
Salary, 25%
Other costs, 16%
Transmission
capacity costs,
43%
Equipment
costs, 16%
EU/EFTA 2014
Salary, 19%Other costs, 15%
Transmission
capacity costs,
47%
Equipment
costs, 19%
EU/EFTA 2009
Graph 7.4.2 – Income sources, other countries
Graph 7.5.1 – Expenditure by category, EU/EFTA countries, 2009
Graph 7.5.2 – Expenditure by category, EU/EFTA countries, 2014

88
7.6 Expenditure by network level
GÉANT partner NRENs differ widely in terms of which network levels they specifically
fund from their budget and how they account for those levels.This makes it difficult to
effectively compare NREN budgets.
Although most NRENs pay for their external connections, the budget proportion
actually spent on this network level differs widely from country to country. On average,
the GÉANT partner NRENs spend 24% of their annual budget on external connectivity
and pay for 80% of its total cost.These figures are very similar to those of 2013.
NRENs also differ in how they account for their expenditure. HEAnet (Ireland)
reports that it spends 6% of its budget on external connections.That this figure
seems unusually low is explained by the fact that HEAnet’s most expensive external
connection – i.e. to GÉANT – is regarded as part of the core infrastructure, not as part of
the external connections. In the case of UNINETT (Norway), the external connections are
not funded through the NREN budget.
Expenditure on other network levels also differs widely. In some GÉANT partner
countries, metropolitan area networks (MANs) and regional area networks (RANs) are
financed under the NREN budget. In other countries, this is not the case.
These disparities in expenditure and accounting methods highlight the complexity
of comparing NREN budgets, as illustrated by Graph 7.6.1. Note that the combined
expenditures on the five network levels do not necessarily constitute 100% of the
NREN’s budget.This can be due to
differences in the NRENs’accounting
methods or because the NRENs also
spend money on items that are not
directly network-related.
0
20
60
40
100
80
Access Network
Backbone
External connections
Campus LAN
MAN/RAN
UK
Austria
Belgium
Belarus
Bulgaria
CzechRepublic
Estonia
Finland
Greece
Georgia
Hungary
Iceland
Ireland
Israel
Italy
Latvia
Luxembourg
Lithuania
Macedonia
Montenegro
Moldova
Netherlands
Norway
Portugal
Romania
Slovakia
Serbia
Slovenia
Spain
Sweden
Turkey
%ofexpenditure
Denmark
Cyprus
90
30
70
50
10
Graph 7.6.1 – Expenditure by network level, GÉANT countries

89
Appendix 1 Alphabetical list of NRENs
N.B.: For additional information on these NRENs, see the country entries at
GÉANT Compendium of National Research and Education Networks In Europe /Appendices
NREN acronym NREN name Country
AARNet Australia’s Academic and Research Network Australia
ACOnet Österreichisches akademisches Computernetz Austria
ADSIB Agencia para el Desarrollo de la Sociedad de la
Información en Bolivia
Bolivia
AMRES Akademska Mreza Srbije Serbia
ANA Rrjeti Akademik Shqiptar Albania
Ankabut United Arab Emirates
Arandu Paraguay
ARENA Armenian Research and Education Networking
Association (ARENA) Foundation
Armenia
ARNES Akademska in raziskovalna mreža Slovenije Slovenia
ASNET-AM Academic Scientific Research Computer Network of
Armenia
Armenia
AzRena Azәrbaycan Elmi-Tәdqiqat ve Tәhsil Şәbәkәlәri
Assosiasiyası
Azerbaijan
AzScienceNet Azәrbaycan Milli Elmlәr Akademiyası Şәbәkәsi Azerbaijan
BASNET Setka Natsianalnai Akademii Nauk Belarusi Belarus
BdREN Bangladesh Education and Research Network Bangladesh
Belnet (NL): Het Belgische telematicaonderzoeksnetwerk,
Belnet. (FR): Belnet, Réseau télématique belge de la
recherche.
Belgium
BREN Sdruzhenie Bulgarska Izsledovatelska i
Obrazovatelna Mrezha
Bulgaria
CANARIE CANARIE Inc. Canada
CARNet Hrvatska akademska i istraživačka mreža - CARNet Croatia
CEDIA Consorcio Ecuatoriano para el Desarrollo de Internet
Avanzado
Ecuador
CERIST Centre de Recherche sur l’Information Scientifique et
Technique
Algeria
CERNET China Education and Research Network - Zhōngguó
jiàoyù hé kēyán jìsuànjī wang
China
CESNET CESNET, zájmové sdružení právnických osob Czech Republic
CKLN Caribbean Knowledge and Learning Network
CONARE Consejo Nacional de Rectores Costa Rica
CSC / Funet CSC - IT Center for Science Ltd / Funet Finland
CSTNet China Science and Technology Network - Zhōngguó
kējī wăng
China
CUDI Corporación Universitaria para el desarrollo de
Internet
Mexico
CYNET Kypriako Erevnitiko Kai Akadimaiko Diktio Cyprus
DeiC Danish e-infrastructure Cooperation Denmark
DFN Deutsche Forschungsnetz Germany
e-ARENA Nacionalnaia Associacia issledovatelskih i nauchno-
obrazovatelnih electronnih infrastructur“e-ARENA”
Russian Federation
EENet / HITSA Eesti Hariduse ja Teaduse Andmesidevork /
Hariduse Infotehnoloogia Sihtasutus – Information
Technology Foundation for Education
Estonia
ERNET Education and Research Network India
EUN Shabaket El Gamaat ElMasria Egypt
FCCN / FCT Fundação para a Computação Científica Nacional /
Fundação para a Ciência e a Tecnologia
Portugal
GARNET Ghana
GARR Consortium GARR (Gestione Ampliamento Rete
Ricerca)
Italy
GRENA Saqartvelos samecniero-saganmanatleblo
kompiuteruli qselebis asociacia
Georgia
GRNET S.A. Ethniko Diktio Ereynas Technologias Greece
HARNET Hong Kong Academic and Research NETwork Hong Kong
HEAnet HEAnet Ltd. Ireland
HIAST Syria
INNOVA|RED Argentina
INHERENT-DIKTI Indonesia
Internet2 Internet2 United States
IRANET/IPM Iran (Islamic Republic
of)
iRENALA Research and Education Network for Academic and
Learning Activities
Malagasy Republic
IUCC Merkaz Hachishuvim haBain Universitai Israel

90
Janet The JNT Association trading as Janet United Kingdom
JUNet Shabakat Aljamiat Al Urduniyeh Jordan
KazRENA Qazaqstannyn’bilim beru zhane gylymi kompyuter
zhelisin koldanushylar kauymdastygy / Asociaciya
polzovateley nauchno obrazovatrlnoi kompyuternoi
seti Kazakhstana
Kazakhstan
KENET Kenya Education Network Trust Kenya
KOREN Korea Advanced Research Network Korea, Republic of
KRENA-AKNET Kyrgyzskaya Nauchnaya i Obrazovatel’naya
Kompyuternaya Set-AKNET
Kyrgyzstan
KREONET Korea Research Environment Open NETwork Korea, Republic of
LEARN Lanka Education and Research Network Sri Lanka
LITNET Lietuvos mokslo ir studiju instituciju kompiuteriu
tinklas
Lithuania
MAREN Malawi Research and Education Network Malawi
MARNet Makedonska akademska nauchno-istrazhuvachka
mrezha
Macedonia, Former
Yugoslav Republic of
MARWAN MARWAN- Réseau informatique national pour l’
éducation, la formation et la recherche
Morocco
MREN Crnogorska mreza za razvoj i nauku Montenegro
MYREN Rangkaian Pendidikan Penyelidikan Malaysia Malaysia
NASATI National Agency for Science and Technology
Information - Cuc Thông tin Khoa hoc và Công nghê
Quôc gia
Vietnam
NCHC National Center for High-performance Computing Taiwan
ngNREN Nigeria
NiCT Dokuritu Gyousei Houjin Jyouhou Tuusin Kenkyuu
Kikou
Japan
NII National Institute of Informatics Japan
NIIF/HUNGARNET Nemzeti Informacios Infrastruktura Fejlesztesi Intezet
/ Magyar Kutatasi es Oktatasi Halozati Egyesulet
Hungary
NREN Nepal Research and Education Network Nepal
OMREN Oman
PERN Pakistan Education Research Network Pakistan
PIONIER Polski Internet Optyczny - Konsorcjum Akademickich
Sieci Komputerowych i Centrów Komputerów Dużej
Mocy
Poland
PREGINET Philippine Research, Education, and Government
Information Network
Philippines
Qatar Foundation Qatar
RAAP Red Académica Peruana Peru
RAGIE Red Avanzada Guatemalteca para la Investigación y
Educación
Guatemala
RAICES Red Avanzada de Investigación, Ciencia y Educación
Salvadoreña
El Salvador
RAU Red Académica Uruguaya Uruguay
REACCIUN REACCIUN: Red Académica de Centros de
Investigación y Universidades Nacionales
Venezuela
REANNZ Research and Education Advanced Network New
Zealand Limited
New Zealand
RedCyT Red Científica y Tecnológica - Panamá Panama
RedIRIS RedIRIS Spain
RedUNIV Cuba
RENAM Asociatia Obsteasca RENAM Moldova, Republic Of
RENATA Corporación Red Nacional Académica de Tecnología
Avanzada - RENATA
Colombia
RENATER Réseau national de télécommunications pour la
technologie, l’enseignement et la recherche
France
RENU Research and Education Network of Uganda Uganda
RESTENA Fondation RESTENA, Réseau Téléinformatique de
l’Education Nationale et de la Recherche
Luxembourg
REUNA Red Universitaria Nacional Chile
RHnet Rannsókna- og háskólanet Íslands hf (RHnet) Iceland
RNP Rede Nactional de Ensino e Pesquisa Brazil
RNU Réseau National Universitaire Tunisien Tunisia
RoEduNet Agentia de Administrare a Retelei Nationale
de Informatica pentru Educatie si Cercetare—
‘RoEduNet’
Romania
SANET Združenie používateľov slovenskej akademickej
dátovej siete — SANET
Slovakia (Slovak
Republic)
SANReN South African National Research Network South Africa
SARNET Academic and Research Network of the Republic of
Srpska
Bosnia And
Herzegovina
NREN acronym NREN name Country NREN acronym NREN name Country

91
Appendix 2 Glossary of terms
Terms not listed in this glossary are either explained in the main text or
presumed to be commonly understood.
SigmaNet SigmaNet, Latvijas Universitātes Matemātikas un
Informātikas institūta Akadēmiskā tīkla laboratorija
Latvia
SINET Japan
SingAREN Singapore Advanced Research and Education
Network (SingAREN)
Singapore
SnRER Senegal
Somaliren Somalia
SudREN Sudan
SUNET Det svenska universitetsdatornätet SUNET Sweden
SURFnet Netherlands
SWITCH SWITCH Switzerland
TARENA Tajik Academic, Research and Educational Network
Association
Tajikistan
TENET Tertiary Education and Research Network of South
Africa
South Africa
TERNET Tanzania, United
Republic Of
ThaiREN Thailand
TuRENA Türkmenistanyň milli ylym-bilim tory Turkmenistan
TTRENT Trinidad and Tobago Research and Education
Network
Trinidad and Tobago
TWAREN TaiWan Advanced Research Education Network Taiwan
UARNet Derzavne pidpryemstvo naukovo-telekomunikacijnyj
centr“Ukrainska akademichna i doslidnytska mereza”
IFKS NAN Ukrainy
Ukraine
ULAKBIM Ulusal Akademik Ag ve Bilgi Merkezi Turkey
UniNet Thailand
UNINETT UNINETT AS Norway
UoM/RicerkaNet Is-Servizzi tal-IT, L-Università ta’Malta/RiċerkaNet Malta
URAN Asociacija Korystuvachiv Ukrainskoji Naukovo-
Osvitnioji Telekomunikacijnoji Merezhi
Ukraine
UzSciNet O’zbek ilmiy va o’quv tamog’i Uzbekistan
WACREN West and Central African Research and Education
Network
ZAMREN Zambia
AAI Authentication and Authorisation Infrastructure: a term used for systems
that support the process of determining both (1) whether users are who
they declare themselves to be (authentication) and (2) that they have the
appropriate rights or privileges necessary to access a resource (authorisation).
APAN Asia-Pacific Advanced Network: a non-profit international consortium
established on 3 June 1997. APAN is designed to be a high-performance
network for research and development on advanced next-generation
applications and services. APAN provides an advanced networking
environment for the research and education community in the Asia-Pacific
region and promotes global collaboration. For further information, see
http://www.apan.net/.
APN Access Point Name: a computer protocol that typically allows a user’s
computer to access the Internet using the mobile phone network.
ASREN Arab States Research and Education Network.
ASPIRE A Study on the Prospects of the Internet for Research and Education: a
foresight study completed in 2012.
AUP Acceptable Use Policy.
bit or b Binary digit: the smallest unit of data in a computer. In this Compendium:
kilobit (kb), Megabit (Mb), Gigabit (Gb), Terabit (Tb), Pb (Petabit).
Bandwidth on
Demand (BoD)
A data communication technique for providing additional capacity on a link,
as necessary, to accommodate bursts in data traffic, a videoconference, or
other special requirements.
Byte or B 8 bits. In this Compendium: kB (kilobyte) MB (Megabyte), GB (Gigabyte), TB
(Terabyte), PB (Petabyte).
CA Certification/Certificate Authority.
CLARA Cooperación Latino Americana de Redes Avanzadas (= Latin American
Cooperation of Advanced Networks): an international organisation whose
aim is to interconnect Latin America’s academic computer networks. For more
information, see http://www.redclara.net/.
confederation A federation formed by multiple independent federations with a common
purpose. An example in the NREN community is the European eduroam
Confederation, which unites country-level eduroam federations.
congestion index A measure of congestion at different levels of network access. Developed by
Mike Norris, formerly of HEAnet.
CRM Customer Relations Management.

92
ccTLD Country-code Top-Level Domains: Internet Top-Level Domains (TLDs) are
geographically specific and can be assigned to a dependent territory in
addition to a country.
CSIRT Computer Security Incident Response Team.
Dark Fibre Optic fibre cable that is not connected to transmission equipment by the
vendor or owner of the cable and therefore has to be connected (i.e.‘lit’) by
the NREN or the client institution.
DNSBL A DNSBL (DNS-based Blackhole List, Block List or Blacklist) is a list of IP
addresses published through the Internet Domain Name Service (DNS).
DNSBLs are most often used to publish the addresses of computers or
networks linked to spamming; most mail server software can be configured
to reject or flag messages which have been sent from a site listed on one or
more such lists.
DNSSEC The Domain Name System Security Extensions (DNSSEC) is a set of extensions
to DNS which provide to DNS clients (resolvers) origin authentication of DNS
data, authenticated denial of existence, and data integrity, but not availability
or confidentiality.
DWDM Dense Wavelength-Division Multiplexing: in fibre-optic communications, a
technology that uses multiple wavelengths of light to multiplex signals in a
single optical fibre.
E.164 The ITU recommendation that defines the international public
telecommunication numbering plan used in the PSTN and some other data
networks.
eduroam® education roaming service: provides a secure international roaming service
to users in the international research and education community. It allows a
user visiting another institution that is connected to eduroam to log on to the
WLAN using the same credentials he/she would use if he/she were at his/her
home institution.
EC European Commission.
eduGAIN The eduGAIN service is intended to enable the trustworthy exchange of
information related to identity, authentication and authorisation between
federations (in the GÉANT area and beyond).
ENUM E.164 NUmber Mapping, a suite of protocols to unify the telephone system
with the Internet.
EU European Union.
EUGridPMA The international organisation that coordinates the trust fabric for e-Science
grid authentication in Europe.
FTE Full-Time Equivalent.
GDS Global Dialling Scheme: a hierarchy of video-conference gatekeepers that
support the mapping of a telephone number format to access MCUs and VC
end-points worldwide.
GÉANT a) GÉANT Association: association consisting mainly of European NRENs, with
offices in Amsterdam (formerly known as TERENA) and Cambridge (GÉANT
Limited, formerly known as DANTE).
b) GÉANT project: a project mainly to develop the multi-gigabit pan-
European data communications network‘GÉANT’, which is used specifically
for research and education.
GN3plus The Multi-Gigabit European Research and Education Network and Associated
Services (GN3plus) project of the European Community’s Seventh Framework
Programme (FP7). It succeeds the GN3 project. It operates and develops the
GÉANT network.
Grid computing Applying the resources of many computers in a network to a single problem.
Honeypots A honeypot is a trap set to detect, deflect, or in some manner counteract
attempts at unauthorized use of information systems. Generally, it consists of
a computer, data or a network site that appears to be part of a network but is
actually isolated and monitored, and which seems to contain information or a
resource of value to attackers.
Identity
Management
System
ldM: a system that combines technologies and policies to allow institutions
to store users’personal information and keep it up to date. An ldM is the first
step to providing AAI for a local or federated environment.
interfederate Exchanging of metadata by two or more federations to allow members
within different federations to connect via a federated access management
exchange.
IP Internet Protocol: the method whereby data, in the form of packets, is
transmitted over a network.
IPv4 Internet Protocol version 4: the fourth iteration and first widely deployed
implementation of the Internet Protocol. IPv4 supports 32-bit addressing and
is the dominant Internet-layer protocol.
IPv6 The latest generation of the Internet Protocol (designated as the successor to
IPv4) with 128-bit addressing as its most significant feature.
IPR Intellectual property rights.
IRU Indefeasible Right to Use: the granting of temporary ownership of a fibre-
optic cable, allowing the unencumbered use of DWDM technology to
maximize the capacity of the link.
Kalmar The Kalmar e-identity Union, which is building an infrastructure for
exchanging personal information across Nordic borders.
Lightpath A dedicated point-to-point optical connection created through the use of
wavelengths in an optical network, to provide guaranteed service levels for
demanding applications bypassing the shared IP network.
MAN Metropolitan Area Network: covers a geographical region such as a city. This
term is often used interchangeably with Regional Area Network (RAN), which
generally covers a wider geographic area.

93
MCU Multi-point Conferencing Unit: used to interconnect multiple video-
conferencing (VC) end-points. An MCU is also able to translate between
different video formats, including SD (standard definition) and HD (high
definition), in order to provide an optimized viewing experience for each VC
unit connected.
NaaS Network as a Service: creates dynamic bandwidth reservation capabilities for
cloud users.
NAT Network Address Translation.
NOC Network Operations Centre: a place from which a network is supervised,
monitored, and maintained.
NORDUnet An international collaboration between the Nordic NRENs. It interconnects
those networks with the worldwide network for research and education, as
well as with the general purpose Internet. NORDUnet is a member of the
GÉANT Association.
NREN National Research and Education Network (May also refer to the operator of
such a network).
PKI Public Key Infrastructure: enables the use of encryption and digital signature
services across a wide variety of applications.
PoP Point of Presence: the location of an Internet access point.
PSTN Public Switched Telephone Network: the traditional circuit-switched
telephony service using dedicated circuits for the duration of a call.
RAN Regional Area Network: covers a wider geographic area than a Metropolitan
Area Network (MAN).
SAML Security Assertion Markup Language: a fundamental component of federated
identity and access management systems.
SDN Software-Defined Networking: a method that allows the creation of virtual
networks.
SIP Session Initiation Protocol: an IETF-defined signalling protocol widely used
for controlling communication sessions such as voice and video calls over
Internet Protocol (IP).
SPF Sender Policy Framework: an email validation system designed to prevent
email spam by detecting email spoofing, a common vulnerability, by verifying
sender IP addresses. SPF allows administrators to specify which hosts are
allowed to send mail from a given domain. Mail exchangers use the DNS to
check that mail from a given domain is being sent by a host sanctioned by
that domain’s administrators.
TCP Transmission Control Protocol: one of the core protocols of the Internet
Protocol suite.
TCS TERENA Certificate Service: offers a variety of digital certificates for server,
personal and e-Science use at research and educational institutions served by
participating National Research and Education Networks (NRENs).
UbuntuNet
Alliance
A not-for-profit association of NRENs that aims to provide a research and
education backbone network for Africa.
University Institution providing courses of education equivalent to ISCED levels 5
and 6.‘Higher/further education’is equivalent to ISCED level 4;‘secondary
education’corresponds to ISCED levels 2 and 3, and‘primary education’to
ISCED level 1. For more information on ISCED levels, see http://www.uis.
unesco.org
VoIP Voice-over-Internet Protocol: a protocol for transmitting voice via the Internet
or other packet-switched networks. VoIP is often used to refer to the actual
transmission of voice (rather than the protocol implementing it). This concept
is also referred to as IP telephony, Internet telephony, voice over broadband,
broadband telephony, or broadband phone.
VPN Virtual Private Network: a network that uses a public infrastructure such as
the Internet to provide remote offices or individual users with secure access
to their organisation’s network. A virtual private network can be contrasted
with an expensive system of owned or leased lines that can only be used by
one organization. The goal of a VPN is to provide the organization with the
same capabilities, but at a much lower cost.
WACREN West- and Central African Research and Education Networks.
X-ARF Network Abuse Reporting Format: an email format for reporting network
abuse.

What is the GÉANT Association?
The GÉANT Association is the leading collaboration on network and related
infrastructure and services for the benefit of research and education.
The GÉANT Association is made up of its members. This includes 36 National
Members, which are European national research and education network
(NREN) organisations, and one Representative Member - NORDUnet - which
participates on behalf of five Nordic NRENs. In addition, there are a number of
Associates , including commercial organisations and multi-national research
infrastructures.
The GÉANT Association:
• provides practical support for members, educators, researchers and other
partners to collaborate, innovate, share knowledge and agree on policies and
strategies;
• plans, procures, builds and operates large-scale, advanced international high-
speed networks, including the 500 Gb/s pan-European GÉANT network;
• organises events such as workshops, meetings and conferences, including
TNC – Europe’s largest networking conference for research and education;
• develops, operates and supports services relating to such areas as trust and
identity, security and certification, mobility and access, and media and real-
time communications;
• provides staff expertise in procurement, project management, community
engagement, network operations, and outreach including dissemination and
training;
• liaises with other e-infrastructure organisations, user communities, industry
and with the European Union.
The GÉANT Association was formed on 7 October 2014, when TERENA and
DANTE joined forces. The association has two offices, one in Cambridge
(formerly DANTE) and one in Amsterdam (formerly TERENA).

2014 COMPENDIUM Edition of National Research and Education Networks in Europe

2014 COMPENDIUM Edition of National Research and Education Networks in Europe

More Related Content

Viewers also liked (10)

Similar to 2014 COMPENDIUM Edition of National Research and Education Networks in Europe (20)

More from Ed Dodds (20)

Recently uploaded (20)

2014 COMPENDIUM Edition of National Research and Education Networks in Europe