Implementation of a European e-Infrastructure for the 21st Century

CERN-OPEN-2013-019
12/07/2013

EIROforum IT Working Group
12th July 2013

This document produced by Members of the EIROforum and is licensed under a Creative Commons Attribution 3.0 Unported
License. Permissions beyond the scope of this license may be available at http://www.eiroforum.org/
Page 1

Implementation of a European e‐
Infrastructure for the 21st
Century
Executive Summary

This document proposes an implementation plan for the vision of an e‐infrastructure as
described in “A Vision for a European e‐Infrastructure for the 21st Century”. The objective
of the implementation plan is to put in place the e‐infrastructure commons that will
enable digital science by introducing IT as a service to the public research sector in
Europe.
The rationale calls for a hybrid model that brings together public and commercial
service suppliers to build a network of Centres of Excellence offering a range of services
to a wide user base. The platform will make use of and cooperate with existing European
e‐infrastructures by jointly offering integrated services to the end‐user. This hybrid
model represents a significant change from the status‐quo and will bring benefits for the
stakeholders: end‐users, research organisations, service providers (public and
commercial) and funding agencies. Centres of Excellence can be owned and operated by
a mixture of commercial companies and public organisations. Their portfolio of services,
starting with those listed by eIRG and the High Level Expert Group on Scientific Data,
will be made available under a set of terms and conditions that are compliant with
European jurisdiction and legislation with service definitions implementing recognised
policies for trust, security and privacy notably for data protection. A funding model
engaging all stakeholder groups is described.
The ability to fully exploit the potential for knowledge and job creation that is locked‐up
in the datasets and algorithms to be hosted by the Centres of Excellence will require the
nurturing of a new generation of data scientists with a core set of ICT skills.
A management board where all the Centres of Excellence operating organisations are
represented will provide strategic and financial oversight is coupled with a user forum,
through which the end‐users themselves, in a cross‐disciplinary body collaborate to
define requirements and policies for the services.
A pilot service is proposed that can be rapidly established by building on the existing
investments. The pilot service will demonstrate the feasibility of the e‐infrastructure
Centres of Excellence model for a range of scientific disciplines and evaluate the
suitability for the ESFRI Research Infrastructures, that are currently under‐development
and represent Europe’s future “big data factories”. Implementation will start in 2014,
initially offering a limited set of services at a prototype Centre of Excellence.

12th July 2013

Page 2

Contents
Executive Summary ............................................................................................................................ 1
Introduction .......................................................................................................................................... 3
Rationale ................................................................................................................................................. 4
Benefits of the hybrid model ..................................................................................................... 6
Pilot Service ........................................................................................................................................... 7
Portfolio of Services ...................................................................................................................... 7
Training and Education .................................................................................................................... 7
Integration of public and commercial service suppliers into a hybrid model .......... 8
Governance Model .............................................................................................................................. 8
Relationship to existing European e‐Infrastructures .......................................................... 9
GEANT ................................................................................................................................................. 9
PRACE .................................................................................................................................................. 9
EGI ...................................................................................................................................................... 10
Engagement of funding agencies ............................................................................................... 10
A Prototype Centre of Excellence .............................................................................................. 13
Funding Model for the Prototype Centre of Excellence .............................................. 15
Timeline ............................................................................................................................................... 17

This document has been prepared by CERN’s IT department on behalf of the EIROforum
IT working group.

12th July 2013

Page 3

Introduction

This document proposes an implementation plan for the vision of an e‐infrastructure as
described in a separate document1 “A Vision for a European e‐Infrastructure for the 21st
Century”. The objective of the implementation plan is to put in place the e‐infrastructure
commons that will enable digital science by introducing IT as a service to the public
research sector in Europe.
The EC, in its Communication “High Performance Computing: Europe’s place in a Global
Race”2 called for the creation of a network of HPC Centres of Excellence and this
implementation plan is a means of addressing this need.
EIROforum3 partners are intergovernmental research organisations – CERN, ESA, EMBL,
EFDA, ESO, European XFEL, ILL and ESRF – covering disciplines ranging from particle
physics, space science and biology to fusion research, astronomy, and neutron and
photon sciences. The partner organisations have a truly European governance, funding
and remit, and in many cases share a global engagement. They are world leaders in basic
research, as well as in managing and operating large research infrastructures and
facilities. The EIROforum collaboration is helping European science reach its full
potential through exploiting its unparalleled resources, facilities and expertise. By
combining international facilities and human resources, EIROforum exceeds the
research potential of the individual organisations, achieving world‐class scientific and
technological excellence in interdisciplinary fields. EIROforum works closely with
industry to foster innovation and to stimulate the transfer of technology.

This implementation plan establishes a pilot service aimed at addressing the challenges
ahead to capture, manage, and process the vast amounts of data to be generated, not
only by the established fundamental research domains but in a growing range of
scientific disciplines, large and small. This pilot service is seen by the research
organisations as an exploratory activity to be operated and funded in addition to their
baseline plans for the production services they offer to their user communities.

1
http://cds.cern.ch/record/1550136/files/CERN‐OPEN‐2013‐018.pdf
2
COM(2012)45, 15.2.2012. http://eur‐
lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2012:0045:FIN:EN:PDF
3
http://www.eiroforum.org/

12th July 2013

Page 4

Rationale

The rationale for this implementation approach is inspired by the Helix Nebula project’s
publication “Helix Nebula – The Science Cloud: A catalyst for change in Europe”5 from
which extracts have been taken.

Helix Nebula was conceived as a way of bringing coherence to a highly fragmented IT
services industry through the vision of a federated ‘science cloud’ integrated with
publicly‐funded scientific e‐Infrastructures. Two fundamental steps remain if the vision
of Helix Nebula is to become reality. One is to bring together a critical mass of supply‐
side interests and the other is to do the same for the demand‐side.
The work of Helix Nebula on an architecture model6 has produced a number of
scenarios exploring the means by which publicly funded infrastructures can
interoperate with commercial cloud services7. Such hybrid systems are in the interest of
the users of publicly funded infrastructures and funding agencies because they will
provide ‘freedom and choice’ over the source of resources to be consumed and the
manner in which they can be obtained, facilitating the goal of ‘making every researcher
digital’. Hybrid systems serve the interests of suppliers by encouraging the adoption of
cloud services by more researchers and thus creating a larger and more vibrant
marketplace in which they can offer their services.
This integration will allow the public infrastructure users to strike their own balance
between publicly funded resources and commercial cloud services while taking into
account aspects of policy and cost. Private sector service providers can supply
resources/services and additional capacity not available in the public funded
infrastructures. Research and innovation activities that have the potential for
commercial exploitation can work with the private sector to unlock that potential,
demonstrate feasibility and market acceptance. It is expected that services will migrate
from publicly funded infrastructures to commercial suppliers as suitable offers,
addressing cost and policy aspects, become available on the market. In this manner,
Helix Nebula is implementing what the e‐Infrastructrue Reflection Group (eIRG) refers
to as the e‐Infrastructure Commons in its 2012 Roadmap paper8.
This hybrid cloud has the potential to bring the public and private sectors together so
that the public sector has something to offer the private sector more than a simple
demand for cloud services. The creation of a continuum of services across the public
and private sector, instead of a simple supply‐demand model will alleviate the current
situation which is inhibiting the establishment of hybrid cloud infrastructures. It is also
inhibiting the establishment of big‐data services since there are a number of
sensitivities, explained below, that need to be addressed before a model can be envisage
where all datasets are hosted on commercial cloud services.

5
http://cds.cern.ch/record/1537032/files/HelixNebula‐NOTE‐2013‐003.pdf
6
http://cdsweb.cern.ch/record/1478364/files/HelixNebula‐NOTE‐2012‐001.pdf
7
http://cds.cern.ch/record/1548323/files/HelixNebula‐D6_1.pdf
8
http://www.e‐irg.eu/images/stories/publ/e‐irg_roadmap_2012‐final.pdf

12th July 2013

Page 5

Business models need to be developed that ensure the Data Stewards9, which are
frequently publicly funded research centres, retain control over the curation and usage
of the datasets and the associated intellectual property. Researchers also need
guarantees about the long‐term availability of the cloud services and the data. Such a
hybrid model would address these points and allow the progressive migration of
services between public and private providers. Furthermore, it is of upmost importance
that data catalogue services, such as those being developed by the EUDAT project10, are
available to all research communities’ datasets hosted by both public and private cloud
services.
There have been many community specific research data infrastructures established in
recent years, such as The Catalog of Life indexing the world’s known species11,
iMARINE12 and GENESI‐DEC13, which have produced valuable data curation tools and
expertise, along with data sharing policies. Being able to interface these data e‐
infrastructures into a hybrid cloud model will allow a larger user base to exploit the
data, bigger opportunities to contribute scientific data to multi‐disciplinary research,
and provide sustainability models for their continued existence.

Initial investigations on the potential impact of cloud services in the research
community suggest that the commercial public cloud services are likely to be adopted
initially for the ‘long tail of science’ conducted by researchers that do not have access to
significant in‐house computing resources and skills. Efforts must be made to simplify
access to commercial cloud services for such groups that generally have straightforward
requirements and frequently do not have sufficient in‐house IT expertise to manage and
operate their own computing resources.
Conversely, it will require further reduction in the costs of commercial services for large
research users with important in‐house computing capacity to find commercial cloud
services as financially attractive as is the case for small scale users. As one industry
representative put it ‘why would I hire a car on a daily basis if I know in advance I will use
it every day for 3 years – it will be cheaper to buy my own’14.
It is recognised that not all publicly funded research centres are in a position to make
accurate estimations of the total cost of ownership of in‐house IT services since some
contributing costs are borne by different departments. But in order for the demand‐side
users to be encouraged to purchase cloud computing services, the services offered must
be economically advantageous compared to other means of procuring IT services.
These alternatives include purchasing and operating IT equipment internally which
requires capital investment and IT expertise but remains economically attractive for IT‐
intensive applications with a sustained and predictable usage.

9
http://en.wikipedia.org/wiki/Data_governance
10
http://www.eudat.eu/
11
http://www.catalogueoflife.org/
12
http://www.i‐marine.eu/Pages/Home.aspx
13
http://www.genesi‐dec.eu/
14
Similar cost arguments are made in this article:
http://www.theregister.co.uk/2013/04/26/cloud_magic_number/

12th July 2013

Page 6

Benefits of the hybrid model

The hybrid model outlined in the rationale above represents a change from the status‐
quo but is necessary to kick‐start new and emerging sciences use of e‐infrastructures.
The identified benefits of this model for the stakeholders include:

Users
 A wide range of integrated services and data resources providing a more
complete working environment supporting the full lifecycle of research activities
 Easy means to publish their own research results and data
 Freedom and choice over which services to use according to preferences for
functionality, policy and budget
 An environment into which their own bespoke services can be integrated
 Readily available training material and support
 Time to focus on their research objectives rather than having to learn about and
dedicate time to operate IT infrastructure
 Keep control over their results and data with the assurance that they will be
available over the long‐term
 Have a voice on how the services they use are governed

Research Organisations
 A means to expose their research products to a wider community
 A platform through which research organisations can be more closely connected
to their users and react more quickly to their changing needs
 Ability to procure additional resources on‐demand
 Grouped negotiation of service procurement with other research organisations
 A ready‐made channel for improving knowledge and technology transfer with
industry

Service providers (public and commercial)
 Creation of a market with an extensive user base in to which providers can offer
their services
 Privileged communication channel with end‐users and purchasing organisations
 A means of estimating the scale and sophistication of services that would be
suitable for the public research sector
 Reduced risk and up‐front investment during the development and introduction
of innovative services

Funding agencies
 A means of encouraging the uptake of cloud services by the public sector
 Creation of an public‐private innovation platform for ICT
 A simple means for implementing open access policies of publications and data
 Democratisation of access to data and compute resources so they are available
to any users, laboratory or project, regardless of size
 A channel for increasing the impact and exploitation of publicly funded research
 A means to monitor the impact of public investments

12th July 2013

Page 7

Pilot Service

In order to verify the feasibility of this vision and rationale, a pilot service is proposed
that can be rapidly established by building on the existing investments to demonstrate
added‐value to the research community and the sustainability of the approach.

Implementation will start in 2014, initially offering a limited set of services at a
prototype Centre of Excellence. This ramp‐up phase will include a roadmap for how to
extend the e‐infrastructure to a group of independently operated and cooperating
Centres of Excellence offering a richer set of layered services and higher total capacity to
serve the European Research Area. Such facilities will be Centres of Excellence providing
capacity style High Performance Computing (HPC) services for data centric applications
and will be federated into a network of such centres across Europe.

Portfolio of Services
The Centres of Excellence will offer an initial portfolio of services taken from the list
documented by eIRG in its blue paper of 201015 with the technical characteristics
identified by the High Level Expert Group on Scientific Data in their “Riding the Wave”
report from the same year16.
All services will incorporate state‐of‐the‐art security measures, including host‐based
protections, system logging and access traceability. All services will make use of a
federated identity management system addressing authentication and authorization to
ensure seamless and secure access to the services and data. While the Centres of
Excellence will take the role and responsibilities of Data Custodians17, the Stewards of
the data‐sets hosted by the Centres of Excellence will be responsible for defining and
managing the access policy for their data.
Brokering services will be deployed so that the full portfolio of services from all Centres
of Excellence is visible to all users and can be compared and accessed.

Training and Education
The design, creation and operation of e‐infrastructure services are essential tools in the
development of skills and competencies for the European market. The services offered
by the e‐infrastructure Centres of Excellence will provide a focus for the creation of
public‐private teams of skilled personnel across the cooperating research organisations,
service suppliers and users, enabling the sharing of knowledge to develop a co‐design
approach18 to services and applications. The ability to fully exploit the potential for
knowledge and job creation that is locked‐up in the datasets and algorithms to be hosted
by the centre will require the nurturing of a new generation of data scientists with a core
set of ICT skills. To ensure the training and education services impact as large an
audience as possible while controlling the costs of offering such services, massive open

15
http://www.e‐irg.eu/images/stories/eirg_bluepaper2010_final.pdf
16
http://cordis.europa.eu/fp7/ict/e‐infrastructure/docs/hlg‐sdi‐report.pdf
17
http://en.wikipedia.org/wiki/Data_governance
18
http://echallenges.org/e2010/outbox/eChallenges_e2007_ref_195_doc_3562.pdf

12th July 2013

Page 8

online course (MOOC19) techniques will be employed which will be supported by the
Centres of Excellence model itself via the digital library services. The EIROforum
organisations have core competences in training and education20 which can form the
basis of this activity. Since its formation in 2002, EIROforum joint initiatives have
already resulted in an important impact in the areas of Outreach and Education. High‐
profile activities have involved thousands of European science teachers and students,
thus emphatically supporting Europe‐wide efforts to raise interest in science and
technology and to secure a sound recruitment base for European R&D efforts in the
future. Such collaboration in the area of human resources also have a bearing on the
European Research Area: by attracting and retaining world‐class researchers in Europe,
through technological exchange and common development, and by organising multi‐
disciplinary scientific conferences.

Integration of public and commercial service suppliers into a
hybrid model

The investigations within Helix Nebula have revealed that with current pricing of
commercial services many public organisations will continue to find it financially
attractive to provide the services themselves in‐house. Hence this implementation plan
foresees that Centres of Excellence can be owned and operated by a mixture of
commercial companies and public organisations. Each Centre of Excellence will include a
pay‐per‐usage element in its funding model but the operating organisation is free to
define its own pricing strategy and implementation. The portfolio of services will be
made available under a set of terms and conditions that are compliant with European
jurisdiction and legislation with service definitions implementing recognised policies for
trust, security and privacy notably for data protection.
By integrating commercial and publicly operated Centres of Excellence into a hybrid
model it will be possible to migrate workloads to commercial suppliers once acceptable
pricing and terms and conditions have been negotiated. In this way the Centre of
Excellence model will bolster the creation of a market to which the commercial suppliers
can offer their services.

Governance Model
The governance model foreseen for the pilot service will be based on two bodies:
 A management board where all the Centres of Excellence operating
organisations are represented and the strategic and financial aspects of the
Centre of Excellence model are addressed. Policy aspects impacting the
operation of the network of Centres of Excellence will be addressed by the
management board. The introduction of a new Centre of Excellence into the
network will be approved by the management board subject to verification of

19
http://en.wikipedia.org/wiki/Massive_open_online_course
20
http://www.scienceinschool.org/

12th July 2013

Page 9

the policies (see Integration of public and commercial service suppliers) under
which it operates and the quality of the services provided.
 A user forum21, through which the end‐users themselves, in a cross‐
disciplinary body collaborate on requirements and policies concerning the e‐
infrastructure services. The User Forum will also act as a consultation body for
management board.

The pilot service will demonstrate the feasibility of the e‐infrastructure Centres of
Excellence model for a range of scientific disciplines. The pilot service will also provide a
means for the ESFRI Research Infrastructures, that are currently under‐development
and represent Europe’s future “big data factories”, to evaluate the suitability of the
services provided by Centres of Excellence. The integrated service portfolio offered by
the network of Centres of Excellence will enable multi‐disciplinary research across the
Research Infrastructures and contribute to ensuring excellence in science.

Relationship to existing European e‐Infrastructures
The Centres of Excellence will make use of and cooperate with existing European e‐
infrastructures to build the e‐infrastructure commons by jointly offering integrated
services to the end‐user. The relationship with these existing e‐infrastructures will be
defined via written agreements and is described below.

GEANT
The Centres of Excellence will be connected to the GEANT network in order to provide
high performance access for the whole European Research Area and ensure integration
with other regions of the world so Europe can be a leading participant in global research
challenges. Building on the experience gathered through the Helix Nebula project which
is extending GEANT connectivity to commercial data centres will greatly simplify the
establishment of a hybrid model engaging both public and commercial service
providers.

PRACE
The capacity style HPC services for data centric applications offered by the Centres of
Excellence will be complimentary to the capability style HPC services provided by
PRACE. The expertise developed by PRACE and related projects in efficient parallel
programming paradigms and optimising software for a range of architectures is directly
relevant to the e‐Infrastructure Centres of Excellence and application/service
developers. The European Technology Platform for High Performance Computing
project22 recently published a Strategic Research Agenda for achieving HPC leadership

21
David Foster, Bob Jones, “Science Strategy and Sustainable Solutions; A Collaboration on the
Directions of e-Infrastructure for Science”, CERN-OPEN-2013-017,
http://cds.cern.ch/record/1545615/files/CERN-OPEN-2013-017.pdf
22
http://www.etp4hpc.eu/

12th July 2013

Page 10

in Europe23 which specifically highlights the upcoming big‐data challenges for leading
research activities and the relevance of cloud services:

“Europe is in a unique position to excel in the area of HPC Usage and Big Data
owing to the experience level of current and potential users (and the recognition of
the importance of data by such users as CERN, ESA, and biological data banks) and
the presence of leading ISVs for large‐scale business applications. Europe should
exploit that knowledge to create competitive solutions for big‐data business
applications, by providing easier access to data and to leading‐edge HPC
platforms, by broaden the user base (e.g., through Cloud Computing and Software
as a Service (SaaS)), and by responding to new and challenging technologies.”

So as well as sharing expertise, the services offered by the Centres of Excellence and the
PRACE Tier‐0 and Tier‐1 centres should be integrated to form part of the overall e‐
infrastructure ecosystem. This will require the PRACE HPC centres to participate in the
federated identity management scheme and data sharing services if the PRACE centres
are to be fully integrated as service providers within this model.

EGI
The experience gathered by EGI in managing a federated grid infrastructure will be
directly relevant to the Centres of Excellence model. The Centres of Excellence will be
integrated with the EGI federated cloud24. Integrating the Centres of Excellence with the
EGI distributed computing infrastructure provides a clear direction for how EGI can
contribute its experience and make the full portfolio of services accessible to its existing
user‐base. The pilot service is intended to be technology agnostic allowing the e‐
infrastructure commons to profit from the long‐standing experience EGI has gained with
grid computing combined with the innovation potential created by the uptake of cloud
computing in research and business sectors.

It is expected that data services currently under development by various projects,
notably EUDAT, will provide candidates for future services and potentially additional e‐
infrastructure Centres of Excellence. A goal will be to introduce services that can profit
from the co‐location of data and compute services to support multi‐disciplinary
research. Metadata and indexing facilities across the set of services in all the e‐
infrastructure Centres of Excellence are seen as being particularly relevant. It is essential
that new services are fully integrated with existing services to preserve the data and
compute continuum of the platform and support the e‐infrastructure commons.
Engagement of funding agencies
The implementation plan and associated funding model have been designed so that the
e‐infrastructure Centres of Excellence can be sustained by their operating organisations
according to a continuum of financial models ranging from sponsored resources for
peer‐reviewed scientific cases to communities who would pay for the services they

23
http://www.etp4hpc.eu/wp-content/uploads/2013/06/ETP4HPC_book_singlePage.pdf
24
https://wiki.egi.eu/wiki/Fedcloud-tf:UserCommunities

12th July 2013

Page 11

receive. Additional resources will be required in order for these services to be expanded
and to serve a wider range of users. The European Commission and national funding
agencies will be invited to become stakeholders and contribute to the expansion of the
e‐infrastructure Centre of Excellence model. The guiding principle is that funding from
such stakeholders will be focused on innovation of services and uptake by new user
communities and business actors while the operational costs will be borne by the
operating organisations and the users themselves. Below is a non–exhaustive list of
areas where funding agencies may contribute to the expansion of the Centre of
Excellence model:

 Development of new services to be deployed on the e‐infrastructure. Significant
effort will be required to co‐develop scalable services that can operate in a
distributed environment and serve a wide range of users.

 Financial incentive scheme to increase adoption of services (both public and
private) by users including ‘long‐tail of science’ research groups and SMEs.

 Engaging the use of the services by new research communities (e.g. curation of
data‐sets, connection of identity federations, deployment of community specific
services, training for new users, etc.)

 Develop training and educational activities building on the e‐infrastructure
services to maximise their impact. This can also include expansion of services to
support for volunteer computing so that researchers can build citizen‐
cyberscience communities and further engage the general public in science.

 Organisation of user forum events as well as outreach and dissemination to a
range of audiences and production of material for policy related activities.

 International collaboration (beyond Europe) through interoperation with
equivalent structures in other regions of the world.

 Expansion of the network of the Centres of Excellence across the European
member states to address national and thematic needs.

 Many research organisations that operate research infrastructures do not have
the mandate to provide e‐infrastructure services to their users for the
management and processing of their experimental data. This represents a gap in
the scientific lifecycle and a missed opportunity to highlight the results and
impact of public funded research. These research organisations will require
assistance to bridge this gap by supporting their users so they can make use of e‐
infrastructures services to manage and process their experimental data.

The recently completed eInfraNet project25 brought together national funding agencies26
from across Europe to explore the economics of e‐Infrastructures and align European

25
http://e-infranet.eu/
26
http://e-infranet.eu/partners/

12th July 2013

Page 12

policies with the needs of the member states. In its final report27 the project
recommended:

“The use of standards‐based private cloud infrastructures may also provide the
needed flexibility in case of jobs that exceed available resources through cloud
bursting (the use of external public clouds for the exceeding capacity) or by
pooling individual institutional private clouds into a federated cloud that can
provide the necessary resources in a federated way, following the extremely
successful approach already in use for GRIDs.”

The model for a network of Centres of Excellence described in this document can serve as
a blueprint to be expanded across Europe. The establishment of Centres of Excellence
could be supported by national funding agencies to address local or thematic needs and
provide a focus for innovation in e‐Infrastructures. Such national or thematic Centres of
Excellence should be integrated with existing European e‐Infrastructures (see
Relationship to existing European e‐Infrastructures) and will be in a position to host
data, information and services that for policy reasons need to remain within a national
context. As these Centres of Excellence come online they must be federated together to
provide a European dimension to the e‐infrastructure commons.

27
Cloud Computing Economics: An evidence-based approach for Research Applications http://e-
infranet.eu/wp-content/uploads/2013/02/Cloud-Computing-Economics-An-evidence-based-approach-
for-Research-Applications1.pdf

12th July 2013

Page 13

A Prototype Centre of Excellence

A prototype Centre of Excellence is planned in order to act as “pathfinder” for the overall
model and pilot service. It is proposed to use the resources installed by CERN at the
Wigner Research Centre for Physics in Budapest, Hungary29 to develop a prototype
Centre of Excellence for the pilot service. The Wigner centre was contracted by CERN
following a competitive call to tender to which approximately 20 sites around Europe
responded. Wigner is operated as an extension to the CERN Data Centre (hosted on the
CERN site in Switzerland) and consequently CERN can assure the business continuity for
the critical systems in case of major problems. The Wigner data centre is connected to
CERN via multiple 100Gbs network links. CERN is integrated with many of the world
leaders in international networking, including hosting a GEANT Point of Presence (PoP)
as well as being a major center for commercial networking for both telecoms suppliers
as well as Internet Service Providers (ISPs). Additional network options provided by
public and private bodies are welcome according to the needs expressed by the users,
ranging from commodity IP services to dedicated circuits directly to Wigner or via
CERN.

The following initial portfolio of services will be provided by the prototype Centre of
Excellence:
 A virtual multi‐tenant compute environment to provision and manage networks
of virtual machines on‐demand.
 A ‘dropbox’ style service for secure file sharing over the internet.
 A point‐to‐point reliable, automated file transfer service for bulk data
transfers30.
 An openAIRE compliant open access repository 31 for publications and
supporting data allowing users to create and control their own digital libraries.
Persistent digital object identifiers will be assigned to all publically available
uploads so as to make them citable and permit the creation of a digital data
continuum spanning from experimental data through to publications including
links to commercial publishers.
 A long‐term archiving service32.
 Integrated Digital Conferencing 33 tools allowing users to manage their
conferences, workshops and meetings.
 Training services: Online training material34 will be made available and re‐
enforced with advanced training via brief residential programmes and

29
http://press.web.cern.ch/press-releases/2013/06/cern-and-wigner-research-centre-physics-inaugurate-
cern-data-centres
30
http://information-technology.web.cern.ch/services/file-transfer
31
http://www.zenodo.org/features
32
http://castor.web.cern.ch/
33
http://indico-software.org/
34
http://indico.cern.ch/categoryDisplay.py?categId=88

12th July 2013

Page 14

secondments, such as CERN’s openlab summer student programme35, which has
recently been expanded with the engagement of EIROforum organisations and
more companies.

These services will be implemented using the OpenStack36 open source software suite
accessible via the most popular interfaces (including EC2, S3 and potentially OCCI).
OpenStack has been selected as the framework for the e‐infrastructure services offered
by the prototype Centre of Excellence because it has emerged as a clear leader in the
vibrant cloud management market. CERN has extensively tested the software and has
adopted it for the management of its production resources at its data centre in Geneva.
OpenStack is fully open source and distributed under the Apache 2 license supported by
a growing consortium of more than 180 private companies and public organisations
with a well‐defined and transparent governance structure representing a global
community of more than 9,000 people across 87 countries. More than 800 developers
are contributing to the software. All of these attributes means OpenStack does not
require public funds to be maintained and offers one of the best environments for
innovation.

Recent developments lead CERN to conclude that the open source software technology
will shortly be available to make this federation of a network of Centres of Excellence
technically feasible. Federation is stated as an objective of the OpenStack Foundation37:

“The vision of federation across deployed OpenStack clouds is critical to the
OpenStack vision. At the core OpenStack is a framework for managing and
provisioning compute, network and storage. From that core many things are
possible. But, we hope and will push for the OpenStack core to remain robust,
consistent and stable. Yes, every deployment will have its own configuration and
technical specifications, but as long as the core concepts and native APIs are robust
and consistent, the possibility to drive real interoperability will exist and the value
from this is a huge advantage of the open cloud concept”.

Long term sustainability, governance and collaboration with industry are achieved by
building the prototype Centers of Excellence on the existing open source community and
means national and European funding can be focused on innovation.

The rationalisation offered by the prototype Centre of Excellence and its integration into
a hybrid model will provide a number of advantages:
 Provide a clear example of the scale and nature of the services that would be
required from commercial suppliers.
 Offer “user aggregation” of procurement since the prototype Centre of Excellence
will be in a position to negotiate with suppliers on behalf of all the partner
research organisations. This will ensure that the most cost effective means of
providing a service will be available at any moment in time. This will also

35
http://openlab.web.cern.ch/news/cern-openlab-summer-student-programme-invites-2013-
applications
36
http://www.openstack.org/
37
http://www.rackspace.com/blog/an-open-letter-to-the-openstack-community/

12th July 2013

Page 15

simplify the contractual and administrative aspects of procuring services from
commercial suppliers for the cooperating research organisations.
 The procurement and deployment cycle for the centre will take several months
and so the integration with commercial cloud service providers will be used to
provide additional capacity in the interim. This will bring added value to the
cooperating research organisations and their users by making it simpler for
them to use commercial cloud services in ‘bursts’ to access additional resources
to fulfil any short‐term requirements.
 The technical aspects of migrating user workloads to commercial services will
be simplified and accelerated.
 The prototype Centre of Excellence will make it possible to host data‐sets on a
cost‐neutral basis while offering open access so they can be exploited via
commercial services. This will remove the need for commercial suppliers to
make investments to host data‐sets without a clear estimation of how much
business they will generate.
 The archiving service would offer a means of ensuring that data‐sets have a safe
back‐up copy.

Funding Model for the Prototype Centre of Excellence

To ensure the sustainability of the prototype Centre of Excellence a funding model is
envisaged where all the stakeholders participate. Sustainability is seen as a process
rather than a fixed solution and hence the funding model will evolve over time.

The services provided by the prototype Centre of Excellence will not be made available
on a commercial basis. CERN rents space in the Wigner data centre and commercial
suppliers can discuss directly with the local authorities if they are interested in locating
their own services in Wigner. This co‐location model is already used in a number of
centres around Europe offering services to the public research sector.

The proposed funding model for the pilot service is that CERN will enter into formal
agreements with partner research organisations that wish to jointly develop and use the
services offered by the prototype Centre of Excellence. CERN will take responsibility for
operating the services at the Wigner data centre and the cost during the first year of
operation will be covered by CERN. Beyond this first year, each partner research
organisation engages to fund the cost of the services their registered users consume
according to a pay‐per‐usage model where the detailed cost model will be jointly‐
developed during the first year. The costs of operating the Wigner site at the capacity
foreseen for CERN’s own usage will continue to be paid by CERN itself. CERN will charge
to the partner research organisations the additional costs which are incurred when
increasing the site’s capacity for the purpose of the Centre of Excellence. These additional
costs will include processors, storage, network switches, cabling, energy, software
licensing costs and operational staff. CERN will operate the services on the same
hardware configurations38 and at the same level of quality as those offered to its high

38
Technical Description Servers and Storage for Physics Data Processing, Acquisition and Control,
CERN, MS-3903/IT, http://cds.cern.ch/record/1472671/files/MS-3903_Technical_Description.pdf

12th July 2013

Page 16

energy physics user community. Tools for monitoring the resources consumed and the
quality of services provided will be made available by CERN. The services will be
integrated in a bottom‐up manner, starting with basic IaaS services and allowing each
partner research organisation to integrate and manage their own higher‐level services
and portals. Any new proposed service will only be introduced if there is at least one
identified partner research organisation that is prepared to contribute to the costs of
operating the service. Support for the cloud services will be integrated into CERN’s
standard service portal39 . Effort will be required from each partner research
organisation to curate their data‐sets, connect their identity federations, deploy their
community specific services and portals as well as manage the interaction with their
registered users and support activities.

39
https://cern.service-now.com/service-portal/

12th July 2013

Page 17

Timeline
The timeline below shows the major milestones foreseen from conception through to
the establishment of the pilot service.

Summer 2013:
 Share vision and implementation with EC and national funding agencies and
discuss ways in which they could contribute

September 2013:
 Discuss engagement of existing e‐Infrastructures during the EGI Technical
Forum in Madrid (16‐20 September)
 Discuss implementation plan with commercial suppliers during Helix Nebula
general assembly at EMBL in Heidelberg (23‐25 September)

October 2013:
 Discuss implementation plan with EUDAT project during their conference in
Rome (28‐30 October)

November 2013:
 Implementation plan for the prototype Centre of Excellence refined with further
details of each service offered, SLA, terms and conditions of service

December 2013:
 First meeting of management board
 First meeting of the user forum

January 2014:
 Initial portfolio of services (see Portfolio of Services) made available for testing
by a small number of research partner organisation(s) via the prototype Centre
of Excellence

June 2014:
 Management board meets to approve opening of the pilot service

July 2014:
 Start of pilot service

Implementation of a European e-Infrastructure for the 21st Century

More Related Content

What's hot (19)

Viewers also liked (15)

Similar to Implementation of a European e-Infrastructure for the 21st Century (20)

More from Ed Dodds (16)

Recently uploaded (20)

Implementation of a European e-Infrastructure for the 21st Century