An Effective Approach for Network Management Based on Situation Management and Mashups

Instituto de Informática – UFRGS
An Effective Approach for Network
Management Based on Situation
Management and Mashups
Oscar Mauricio Caicedo Rendón
Supervisor: Lisandro Zambenedetti Granville
Federal University of Rio Grande do Sul – UFRGS
Institute of Informatics - http://inf.ufrgs.br/en/
Computer Networks Group – http://networks.inf.ufrgs.br/
Porto Alegre, RS – Brazil
16-03-20151

Outline
• Introduction
• Mashment ecosystem
• Process for addressing nmsits
• Mashment system architecture
• Evaluation
• Summary and conclusions
Oscar Mauricio Caicedo Rendón 2

Introduction
Contextualization
• Situation Management (SM)
• Mashups

4
Introduction
Contextualization
• What is SM?
– Address situations happening or that might happen in dynamic systems (Jakobson et al., 2005)
– A situation is a composed entity in a particular application domain whose components are other
entities, their attributes, and relationships in a time interval (Kokar et al., 2009)
– Provide solutions that enable analyzing, correlating, and coordinating interactions among
people, information, technologies, and actions targeted to overcome situations (Jakobson, 2014)

Introduction
Contextualization
• SM application domains
– Origin - > militar domain
– Satellite networking (Gopal, 2007)
– Disaster response (George et al., 2010)
– Smart power grid (Magoutas et al., 2011)
– Civil crisis (Hein et al., 2012)
– Aviation (Koelle and Tarter, 2012)
– Public health (Pereira et al., 2013)
– Electric power systems (Krohns-Valimaki et al., 2013)
– Emergency medical assistance (Bruns et al., 2014)
There is few work about the use of
SM in network management

6
Introduction
Contextualization
• What are mashups?
– Mashups are Web applications built up by end-users through the combination of Web
resources available along the Internet (Simmen et al., 2008) (Cappiello et al., 2010)
– Mashups allow end-users, without advanced programming skills, to create their own
and customized applications (Laga et al., 2012)
– Mashups encourage both cooperation and reuse among end-users (Huang et al., 2012)

7
Introduction
Mashups have not been used to face the
complexity and time of tasks performed
by network administrators in their daily
work when coping with situations
Contextualization
• Mashups application domains
– Origin - > music domain
– Project management (Ozkan and Abidin, 2009)
– Water floods (Tosti and Smari, 2010)
– Fire emergencies (Jchrzak and More, 2011)
– Telco services (Gebhardt et al., 2012)
– Data integration (Han et al., 2013)
– Immersive mirror worlds (Stirbu et al., 2013)
– IT Service Management (Santos et al., 2011) (Santos et al., 2013)
– Music (Davies et al., 2014)
– Specific tasks for network management (Bezerra et al., 2010) (Santos et al., 2010)

8
Introduction
Problem Statement
Find the root cause of unexpected
and multiple packet transmission
failures in network slices formed by
OpenFlow networks
Situations Options Shortcomings
Manage the abrupt performance
degradation of one or more
switches/routers on networks that
use diverse virtualization
technologies
Monitor sudden violations in service
level agreements in a near real time
way
Employ non-integrated solutions
such as traceroute, ZenOSS, and
OpenNMS
Develop low-level scripts to integrate
mismatched solutions
Use research solutions based on
policies (e.g. PROCERA), business
processes (e.g. MEICAN), and service
composition (e.g. NetOpen)
It hinders and overloads the work
carried out by network
administrators
Developing scripts is complex for a
network administrator because,
usually, he/she is not an expert
programmer
Writing policies, queries, or
primitives in specific languages
and/or controllers remains
complexity in network management
(Chen et al., 2010) (Kim and Kim, 2011) (Santanna et al., 2012)
(Monsanto et al., 2013) (Kim and Feamster, 2013) (Zaalouk et al., 2014)

9
Introduction
Problem Statement
• In the network management domain, large research efforts (Chen et al., 2010) (Kim
and Kim, 2011) (Santanna et al., 2012) (Monsanto et al., 2013) (Kim and Feamster,
2013) (Zaalouk et al., 2014) have been made to automate and facilitate the
management tasks
• So far, none of these efforts has carried out network management by means of SM
and mashups
How to carry out network management, in an
effective way, by SM and mashups?

Introduction
General Objective
• Investigate the feasibility of using SM and mashups as an effective approach
for network management
• Effective approach in terms of
– The complexity perceived and the time taken (time-consuming) by network
administrators when carrying out tasks for facing situations
– The traffic and the time of response (time-response) of solutions used by network
administrators for coping with situations

11
Introduction
Hypothesis
• The employment of SM and mashups provides an effective approach
for network management

12
Introduction
Research questions
• What is the performance, in terms of the complexity and consuming of time, of
solutions that use SM and mashups for network management?
• What is the performance, in terms of traffic and time of response, of solutions based
on SM and mashups for network management?
• Which mechanisms could be employed to improve the performance of solutions that
use SM and mashups for network management?

13
Introduction
Contributions
• Conceptual
– The network management
situation concept and its data
model
– The mashment concept and its
data model
• Specific
― The mashment ecosystem
― The process for addressing network
management situations
― The mashment system architecture
― The model for measuring complexity and
time-consuming
― The mechanism to automatically recognize
network management situations
― The mechanism to dynamic composition of
mashments

Mashment Ecosystem
OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti Granville. A
Mashup Ecosystem for Network Management Situations. The IEEE Global Communications
Conference (GLOBECOM 2013), 9-13 December 2013, Atlanta, United States
14Oscar Mauricio Caicedo Rendón

15
Mashment ecosystem
What is a nmsit?
• A network management situation (nmsit) is an unexpected, dynamic, and
heterogeneous situation happening or that might happen in the network
management domain
– Manage the abrupt performance degradation of one or more switches/routers on
networks that use diverse virtualization technologies
– Monitor sudden violations in service level agreements in a near real time way
when, for instance, links formed by virtual switches suffer abrupt and
intermittent performance degradation
– Find the root cause of the unexpected traffic overload suffered by a set of
routers/switches

16
Mashment ecosystem
• Mashment
― Tunable mashup
― Diverse types of
resources
― Multiple providers
― SM investigative
aspect
― SM control aspect
What is a mashment?

17
Mashment ecosystem
What is a mashment ecosystem?
Interactions
among
software
entities,
providers, and
end-users
Mashment
Ecosystem
Acting as a single unit
Software
entities
supporting
activities
Activities
performed by
end-users and
providers
Mashments
End-users
creating and
using
mashments
Resources for
network
management
Providers
supplying
resources for
network
management
and software
entities

Mashment ecosystem

Process for adressing nmsits
OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti Granville. An
Approach to Overcome the Complexity of Network Management Situations by Mashments. The 28th
IEEE International Conference on Advanced Information Networking and Applications (AINA 2014),
13-16 May 2014, Victoria, Canada

Process for addressing
nmsits
Overall Process

28 / 16
nmsits
Complexity
Complexity is the difficulty perceived by network administrators to
carry out tasks targeted to overcome nmsits
These complexities are expressed by using per-task metrics defined
for IT Service Management processes (Diao and Keller, 2006)
𝞯 = 𝞯 𝑠𝑒𝑙
𝑖
1
+ 𝞯 𝑐𝑜𝑛 + 𝞯 𝑐𝑜𝑚 + 𝞯𝑙𝑎𝑢
𝑜
1
𝑘
1
𝑗
1

nmsits
Time-consuming
Time-consuming is the time that network administrator takes to carry
out tasks targeted to face nmsits
These times are expressed using the Keystroke-Level Model (Kieras,
2001) (Tian et al., 2011)
𝑻= 𝑻 𝑠𝑒𝑙
𝑖
1 + 𝑻 𝑐𝑜𝑛 + 𝑻 𝑐𝑜𝑚 + 𝑻𝑙𝑎𝑢
𝑜
1
𝑘
1
𝑗
1

38 / 16
Mashment system architecture
OSCAR MAURICIO CAICEDO RENDON, Carlos Raniery P. dos Santos, Arthur Selle Jacobs, Lisandro
Zambenedetti Granville. Monitoring Virtual Nodes Using Mashups. Computer Networks (COMNET), v. 64,
pp. 55-70, May 2014. ISSN 1389-1286
OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti Granville. A
Mashup-based Approach for Virtual SDN Management. The 37th IEEE Annual International Computer
Software & Applications Conference (COMPSAC 2013), 22-26 July 2013, Kyoto, Japan
OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Vinicius Guimarães, Lisandro
Zambenedetti Granville. Rich Dynamic Mashments: An Approach for Network Management Based on
Mashups and Situation Management. Computer Networks (COMNET), 2015, submitted

Managed resources layer

Adaptation layer

Composition layer

43 / 16
Composition layer

44 / 16
Composition layer

Composition layer

Presentation layer

Evaluation

SDN Case Study
- test environment -
challenge: it is needed the fast and ease development of a solution able to
present, in an integrated, visual, and intelligible way, information about Open
vSwitches, links, and flows handled by Beacon, POX, and Floodlight
nmsit: the Network Administrator needs to investigate which are the Open
vSwitches that are causing sudden performance degradation in the
OpenFlow-based networks of the test environment

SDN Case Study
• The Network Administrator creates, launches, and uses different and mismatched
Web-based solutions (e.g., Beacon Web Tool, Floodlight Web Tool, and POX Web Tool)
• The Network Administrator creates, launches, and uses a Situational Script that
integrates such mismatched solutions

SDN Case Study
• The Network Administrator develops and launches PMM by the Mashment Maker

SDN Case Study
• The Network Administrator uses PMM to retrieve information about switches, links, and
flows handled by Beacon, POX, and Floodlight

SDN Case Study
- complexity -
The mashment-based
approach decreases the
complexity (approx
37%)

SDN Case Study
- time-consuming -
The mashment-based
time-consuming
(approx 75%)

SDN Case Study
- time-response -
The time-response of PMM is
always ranked as good
Time-response (Joines et al., 2002)
optimal (r < = 100)
good (100 < r < = 1000)
admissible (1000 < r < = 10000)
deficient (r > 10000)

SDN Case Study
- traffic -
PMM - additional traffic
less than 10%

Virtual Nodes Case Study
challenge: it is needed a situational solution that presents, in an integrated, visual, and intelligible way,
information about Xen, VirtualBox, guests, and Open vSwitches (including links, flows, ports, and traffic)
nmsit: the Network Administrator needs to investigate the Host Computer Systems and guests suffering
unexpectedly overload in processor, memory, and/or network as well as the Open vSwitches that are causing
sudden performance degradation in an OpenFlow-based network handled by Floodlight

• The Network Administrator creates, launches, and uses diverse and disparate Web-based solutions
(e.g., oVirt and Floodlight Web Tool)
• The Network Administrator creates, launches, and uses a Monitoring Script that integrates such
disparate solutions

• The Network Administrator develops and launches VNMM by the Mashment Maker

65 / 16
• The Network Administrator uses VNMM to retrieve information about guests, switches, and links
located at a virtual node

66 / 16
- complexity -
The mashment-based
complexity (about 20%)

- time-consuming -
The mashment-based
time-consuming (about
75%)

- time-response -
The time-response of
VNMM is always ranked
as good

- traffic -
VNMM – additional traffic less than 10%

72 / 16
Dynamic Mashments Case Study
• Recognize several nmsits by nmsit patterns
• Compose mashments by templates

73 / 16

74 / 16

75 / 16
- time-recognition -
The nmsits are recognized
in a short time

76 / 16
- time-composition -
The time-composition is
short (Ordonez et al., 2014)

77 / 16
- time-consuming -
The mechanisms for
automatically
recognizing nmsits and
dynamically composing
mashments decrease
the time-consuming
(about 92%)

Summary and conclusions
Final remarks
• The network management is driven towards an environment focused on
situations, composite situational solutions, and network administrators
• The mashup foundations are brought up to the SM discipline to carry out its
investigative and control aspects
• The mashup technology is led to a novel application domain located in the
intersection of SM and the network management

• Flexibility
― Mashments provide a flexible and easy-to-use way for user-centric service
composition
― Network technologies and situational management operations are
represented as mashup-able components
― Network Administrators do not require a lot of Web programming skills to
create situational management capabilities
― Network Administrators can customize and improve their workspace by
themselves
Final remarks

• Extensibility
― Network Administrators can addressing nmsits by conducting a simple
process for creating/using mashments
― Network Administrators who develop mashments are able to extend the
Mashment Ecosystem and, therefore, enhance/improve their workspace
― Building up a mashment from existing applications is easier than
developing a composite application from scratch
Final remarks
Conclusions and future work

82 / 16
Answering the research questions
What is the performance, in terms of
the complexity and consuming of
time, of solutions that use SM and
mashups for network management?
The evaluations indicate that the mashment-based
approach can be less complex and consuming of time
than proprietary and incompatible CLIs and GUIs used
currently for coping with nmsits
What is the performance, in terms of
traffic and time of response, of
solutions based on SM and mashups
for network management?
Which mechanisms could be
employed to improve the
performance of solutions that use SM
and mashups for network
management?
The evaluations indicate that the time of response of
mashments can be ranked as optimal and/or good.
Their behavior in traffic is similar to the current
solutions used for coping with nmsits
The evaluations indicate that mechanisms for nmsits
recognition and dynamic mashment composition can
be used to improve the performance, in terms of
complexity and time-consuming, of the mashment-
based approach

Future work
• Distribute the mashment-based approach
• Create the deployment cost model of the proposed approach
• Explore the use of Big Data and Artificial Intelligence on mashments

Publications
• OSCAR MAURICIO CAICEDO RENDON, Carlos Raniery P. dos Santos, Arthur Selle Jacobs,
Lisandro Zambenedetti Granville. Monitoring Virtual Nodes Using Mashups. Computer
Networks (COMNET), v. 64, pp. 55-70, May 2014. ISSN 1389-1286
• OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti
Granville. An Approach to Overcome the Complexity of Network Management
Situations by Mashments. The 28th IEEE International Conference on Advanced
Information Networking and Applications (AINA 2014), 13-16 May 2014, Victoria,
Canada

85 / 16
Publications
• OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti
Granville. A Mashup Ecosystem for Network Management Situations. The IEEE Global
Communications Conference (GLOBECOM 2013), 9-13 December 2013, Atlanta, United
States
• OSCARMAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Lisandro Zambenedetti
Granville. A Mashup-based Approach for Virtual SDN Management. The 37th IEEE
Annual International Computer Software & Applications Conference (COMPSAC 2013),
22-26 July 2013, Kyoto, Japan

86 / 16
Publications
• OSCAR MAURICIO CAICEDO RENDON, Felipe Estrada-Solano, Vinicius Guimarães,
Lisandro Zambenedetti Granville. Rich Dynamic Mashments: An Approach for Network
Management Based on Mashups and Situation Management. Computer Networks
(COMNET), 2015, submitted - *

Collaborations
• Wanderson Paim de Jesus, Ricardo L. dos Santos, OSCAR MAURICIO CAICEDO RENDON,
Lisandro Zambenedetti Granville. A Platform for Programmable Virtual Network
Management. The 31st Brazilian Symposium on Computer Networks and Distributed
Systems (SBRC 2013), 6-10 May 2013, Brasilia, Brazil
• Armando Ordóñez, Jesus David Ramirez, Paolo Falcarin, OSCAR MAURICIO CAICEDO
RENDON, Lisandro Zambenedetti Granville. An Architecture for Self-Reconfiguration of
Convergent Telecom Processes. The 7th International Workshop on Principles of
Engineering Service-Oriented and Cloud Systems (PESOS 2015), 23 May 2015, Florence,
Italy, accepted *

Collaborations
• Armando Ordóñez, Vidal Alcazar, OSCAR MAURICIO CAICEDO RENDON, Paolo Falcarin,
Juan Corrales, Lisandro Zambenedetti Granville. Towards Automated Composition of
Convergent Services: a survey. Computer Communications (COMCOM), ISSN 0140-
3664, revision submitted - *
• Ricardo L dos Santos, OSCAR MAURICIO CAICEDO RENDON, Juliano Wickboldt,
Lisandro Zambenedetti Granville. App2net: A Platform to Transfer and Configure
Applications on Programmable Virtual Networks. The 20th IEEE Symposium on
Computers and Communications, 2015, submitted - *

Popayán – Colombia La Fonda Patía - Colombia
¡Gracias!

An Effective Approach for Network Management Based on Situation Management and Mashups

More Related Content

What's hot (7)

Similar to An Effective Approach for Network Management Based on Situation Management and Mashups (20)

Recently uploaded (20)

An Effective Approach for Network Management Based on Situation Management and Mashups