On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-Managing Networks [Elektronische Ressource] / Ranganai Chaparadza. Betreuer: Ina Schieferdecker

De
On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-Managing Networks vorgelegt von Diplom-Ingenieur Ranganai Chaparadza aus Simbabwe, (derzeit aus Berlin) Von der Fakulta¨t IV - Elektrotechnik und Informatik der Technischen Universita¨t Berlin zur Erlangung des akademischen Grades Doktor der Ingenieurwissenschaften - Dr.-Ing. - genehmigte Dissertation Promotionsausschuss: Vorsitzender: Prof. Dr. Küpper Berichter: Prof. Dr.-Ing. Ina Schieferdecker Berichter: Prof. Dr.-Eng. Symeon Papavassiliou Berichter: Prof. Shiduan Cheng Tag der wissenschaftlichen Aussprache: 27.07.2011 Berlin 2012 D 83 ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011 Page i On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-Managing Networks PhD-Thesis, TU Berlin, 27.07.2011 Copyright By Ranganai Chaparadza 2011 ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011 Page ii The Dissertation Committee for Ranganai Chaparadza Certifies that this is the approved version of the following dissertation: On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-Managing Networks Committee: Chairman: Prof. Dr. Küpper (TUB) Prof. Dr.-Ing.
Publié le : dimanche 1 janvier 2012
Lecture(s) : 23
Tags :
Source : D-NB.INFO/1019398574/34
Nombre de pages : 243
Voir plus Voir moins

On-Demand Monitoring: a Monitoring Paradigm for Traffic
Flows in Multi-Service Self-Managing Networks




vorgelegt von
Diplom-Ingenieur
Ranganai
Chaparadza
aus Simbabwe,
(derzeit aus Berlin)



Von der Fakulta¨t IV - Elektrotechnik und Informatik der
Technischen Universita¨t Berlin
zur Erlangung des akademischen Grades
Doktor der
Ingenieurwissenschaften
- Dr.-Ing. -
genehmigte Dissertation


Promotionsausschuss:

Vorsitzender: Prof. Dr. Küpper
Berichter: Prof. Dr.-Ing. Ina Schieferdecker
Berichter: Prof. Dr.-Eng. Symeon Papavassiliou
Berichter: Prof. Shiduan Cheng



Tag der wissenschaftlichen Aussprache: 27.07.2011





Berlin 2012

D 83

ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page i


On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows
in Multi-Service Self-Managing Networks



PhD-Thesis, TU Berlin, 27.07.2011



Copyright
By
Ranganai Chaparadza
2011



ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page ii

The Dissertation Committee for Ranganai Chaparadza Certifies that this is the approved
version of the following dissertation:


On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-
Managing Networks





Committee:
Chairman: Prof. Dr. Küpper (TUB)
Prof. Dr.-Ing. Ina Schieferdecker,
Supervisor (TUB, Germany)
Prof. Dr.-Eng. Symeon Papavassiliou
Second Advisor/Reviewer (NTUA,
Greece)
Prof. Shiduan Cheng
Third Advisor/Reviewer (BUPT,
China)




ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page iii
On-Demand Monitoring: a Monitoring Paradigm for Traffic Flows in Multi-Service Self-
Managing Networks

PhD-Thesis, TU Berlin, 27.07.2011

By
Ranganai Chaparadza



Dissertation
Presented to the Faculty of Electrical Engineering and Computer Science (German: Fakultät
Elektrotechnik und Informatik der TU Berlin, FR5-1: Fakultät IV) of
The Technical University of Berlin (TUB), Germany
in Partial Fulfilment
of the Requirements
for the Degree of

Doctor of Engineering (Dr.-Ing)


Technical University of Berlin (TUB), Berlin, Germany
2011

ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page iv




Dedication

To my father, the late Ezekiel, and my mother Tabeth, my brothers and sisters, my whole
family and friends.

My former school: Warsaw University of Technology (Politechnika Warszawska), Poland:
Faculty of Electronics and Information Technology: Institute of Telecommunications:
Department of Switching Systems, Routing Systems, Network Management, Formal Methods for
Specification and Validation of Communication Systems and Protocols. I obtained my
M.Eng/MSc degree in Telecommunications Engineering: Thesis Topic: Specification and
Validation of Multicast Mechanisms of the XTP Transport Protocol for High Speed Networks.

My former school: Gdansk Technical University (Politechnika Gdanska), Poland, where I
completed my early semesters in Telecommunications Engineering Studies, and then transferred
to Warsaw University of Technology.

My former school: University of Łódź, Poland, where I learnt polish language for foreigners.

My former school: Bernard Mizeki College (BMC), Zimbabwe, where I worked temporarily,
teaching form-5 science students.

My former high school: Murewa High School, Murewa, Zimbabwe, where I obtained my GCE
O-Level and A-Level Certificates by University of Cambridge (United Kingdom) in
Collaboration with the Ministry of Education of Zimbabwe.

My Primary School: Musvaire Primary School, Mutoko, Zimbabwe.



ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page v




Acknowledgements

I completed this research work during the time I had many other responsibilities to shoulder. It
required a lot of energy and strong spirit to stay committed to achieve a number of objectives
simultaneously. I therefore thank God for the energy and strength He gave me to complete the
work under tight schedules and multi-tasking. I thank my advisor Prof. Dr.-Ing. Ina
Schieferdecker from Technical University of Berlin, whom I worked with for a number of years
in the Group she leads in Fraunhofer FOKUS Institute of Open Communication Systems, and we
achieved a lot together in both academic-oriented research as well as applied-research for
industrial solutions in the fields of software engineering, model-driven systems design and
testing methods for complex communication systems and applications as well as networking
research. Her advice was very much appreciated. I also want to thank my second
advisor/reviewer Prof. Dr.-Eng. Symeon Papavassiliou from the National Technical University
of Athens, Greece, whom I worked with in the European Commission funded FP7 EFIPSANS IP
Project, and he exchanged some ideas with me in the area of Autonomic Networking and Self-
Management. The technical discussions with him in the area were very good and inspiring. His
review comments to this dissertation were very much appreciated. I also want to thank my third
advisor/reviewer Prof. Shiduan Cheng from The State Key Lab of Networking & Switching
Technology of Beijing University of Posts and Telecommunications (BUPT), China, whom I
have worked with in the European Commission funded FP7 EFIPSANS IP Project. Her advice
and discussions we had in the field of Autonomic Networking and Self-Management were very
much appreciated. I co-organized a workshop with her in Beijing at the IPv6 Forum Summit in
2009, where she also organized a very good panel of experts from the academia and the industry,
and we discussed a lot of interesting views on the Self-Managing Future Internet powered by
IPv6 and Extensions to IPv6 towards IPv6++. I also want to thank my former student in Security
Testing in IPv6 Networks & friend Mikhail Peter who got very enthusiastic to join me in the
prototyping of the ODM-Probe presented in this work. I also want to thank people I made friends
with in this field over the years, and discussions with them were good and inspiring. One of
which is Prof. Vassilis Magraris from NTUA, Greece. Other names I always acknowledge for
their support are my mother Tabeth and father Ezekiel (late), my whole family (brothers and
sisters & their families in Zimbabwe, New Zealand, and Australia), Pastor P. Mugadza, Pastor
Kingsley Nimo, Tyson Nkani, and Małgorzata (Gosia) Kowalska— a dear friend who has always
been supportive to me. To other names I might have missed, I appreciate all your support you
gave me.


ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page 1


On-Demand Monitoring Paradigm: a Monitoring Paradigm for Traffic Flows in Multi-
Service Self-Managing Networks

Publication No._____________


Ranganai Chaparadza, degree obtained: Dr.-Ing. (Doctor of Engineering).
Technical University of Berlin, Germany, 2011

Advisor: Prof. Dr.-Ing. Ina Schieferdecker (Technical University of Berlin, Germany)
Second Advisor/Reviewer:
Prof. Dr.-Eng. Symeon Papavassiliou (National Technical University of Athens, Greece)
Third Advisor/Reviewer:
Prof. Shiduan Cheng (The State Key Lab of Networking & Switching Technology of Beijing
University of Posts and Telecommunications (BUPT), China)

Abstract:
Traffic monitoring in information and telecommunication networks continues to experience a lot
of research as it increasingly plays a very significant role in implementing emerging and future
networks that self-adapt to changes in their usage context as well as to challenging adverse
conditions experienced by the network elements. Self-managing networks call for research on
the key design and operational principles for traffic monitoring components and platforms that
are suitable for multi-service self-managing networks.
A self-managing network is one whose network elements and network management systems
work co-operatively to perform so-called self-* operations, such as self-configuration, self-
diagnosing, self-repairing, self-optimization, all aimed at eliminating or drastically reducing
human intervention in the complex aspects of manual and error prone, or daunting tasks of
network management. In contrast to today’s non-self-managing networks, self-managing
networks require more computing resources for the computation of decisions taken by individual
network elements and by the entire network under adverse conditions or context changes, based
on large and diverse information sets collected by monitoring components. In self-managing
networks, depending on the network dynamics, a lot more information exchange than in non-
self-managing networks may be required to flow between the network elements e.g. routers,
hosts, or switches. Information has to flow between the elements and the manager components
that aggregate global network state information, compute and perform more sophisticated
decisions for the entire network and, communicate control information to the network elements
for global self-adaptation behaviour of the whole network. More storage resources are required
than in non-self-managing networks, for the storage of the network state information including
historical network state data. Therefore, they require intelligent and opportunistic use and sharing

ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page 2


of resources available at individual network elements as well as in the entire network. Flexibility
to support context-driven (re-)configuration and self-adaptation of the network tasks, is also
required. To guarantee intelligent and opportunistic use and sharing of resources of the network,
as well as context-driven (re-)configuration and self-adaptation of network tasks, we propose to
design functions and components of the individual network elements and the overall network
such that their operational principles allow functions, including monitoring functions, to be
invoked on-demand by automated tasks that drive a self-managing network.
The dissertation presents the on-demand monitoring paradigm (ODM), a monitoring paradigm
we researched and demonstrated its suitability for collecting information in self-managing
networks dynamically as needed. The ODM-Paradigm takes into account the need to use
network resources intelligently and opportunistically throughout the network, as well as the need
for monitoring components to self-describe their monitoring capabilities to the network. Self-
description of capabilities enables automated tasks to locate and select a monitoring component
of desired capabilities (on-demand), trigger monitoring functions and manage their execution,
and free resources on the targeted component whenever monitoring is temporarily not required
or no longer required by an automated task(s).
Artefacts developed are: a mathematical foundation upon which ODM is based; design and
operational principles that enable a traffic monitoring component to realize ODM; principles that
enable the component to self-describe its capabilities and point of attachment to the network, to
enable network tasks to request monitoring on desired components; a language for describing
capabilities; a composition language for specifying requested monitoring-behaviours; the
concept of on-demand SNMPMIBs for dynamic monitoring data models; admission-control on
monitoring-requests; a mapping of ODM within the GANA (Generic Autonomic Network
Architecture) reference model. To demonstrate the feasibility of the approach, an ODM-capable
probe that serves as proof of concept was developed. Evaluations of metrics associated with
ODM were conducted. A methodology was developed on how to measure and use estimations of
dynamics of resource demands on a system, in queuing and processing monitoring-requests, and
tuning the admission control behavior of an ODM-capable component. Example network tasks
that demonstrate applicability of ODM were provided. Arguments are presented on how
intelligent and opportunistic use of resources is achievable when traffic monitoring components
are ODM-capable. The evaluations and case study with on-demand SNMP MIBs show that
ODM can be feasibly implemented and applied in practice.



ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page 3


Table of Contents
1 Structure and summary of contribution of this thesis ...................................... 10
1.1 Overview ....................................................................................................... 10
1.2 Structure of this thesis ................................................................................... 10
1.3 Summary of the contribution of this thesis..................................................... 14
1.3.1 Overview .................................................................................................. 14
1.3.2 The big-picture of the ODM-Paradigm (a nutshell view) .......................... 17
2 Introduction .......................................................................................................... 20
2.1 Motivation and Overview ............................................................................... 20
2.2 Monitoring in general—a brief overview ........................................................ 22
2.3 Problem Formulation ..................................................................................... 23
2.4 The limitations of today’s monitoring paradigms ........................................... 25
2.4.1 Limited monitoring data-sets .................................................................... 25
2.4.2 Limited flexibilities in monitoring components or subsystems .................. 26
2.4.3 Limited customizability of monitoring-behaviours ..................................... 27
2.4.4 Limited support for programmability of monitoring services at run-time ... 28
2.4.5 Limited possibilities in deriving data-sets from raw monitoring data ........ 29
2.4.6 Some remarks ......................................................................................... 30
3 Design principles for self-managing networks .................................................. 32
Overview of design principles and introducing the GANA Reference Model ............. 32
3.1 Requirement for traffic flow monitoring .......................................................... 47
3.2 Requirements for a traffic monitoring platform .............................................. 54
4 The On-Demand Monitoring (ODM) Paradigm ................................................... 61
4.1.1 Overview and Fundamental Definitions ................................................... 61
4.1.2 Principle-P1: Support for Customizable Monitoring ................................. 66
4.1.3 Principle-P2: Support for Programmable Monitoring ................................ 66
4.1.4 Principle-P3: Support for on-demand creation and destruction of gathered
monitoring-data and in-memory data models ......................................................... 67
4.1.5 Principle-P4: Support for Triggerable, Configurable and Re-Configurable
Monitoring .............................................................................................................. 67
4.1.6 Principle-P5: Support for Adaptive Monitoring ......................................... 68

ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page 4


4.1.7 Principle-P6: Support for Intelligent and Opportunistic use and allocation
of resources ........................................................................................................... 68
4.1.8 Principle-P7: Support for Self-description and Self-advertisement of
Capability Models ................................................................................................... 69
5 Technical Solutions for the ODM Concepts and Principles ............................. 70
5.1 Overview ....................................................................................................... 70
5.2 Customizable Monitoring and Programmable Monitoring .............................. 70
5.2.1 The EDBSLang Language and its XML-Schema ..................................... 73
5.2.2 Some remarks ......................................................................................... 80
5.3 On-demand creation and destruction of gathered monitoring-data and in-
memory data models ................................................................................................. 80
5.3.1 On-Demand SNMP MIBs ......................................................................... 82
5.4 Triggerable, Configurable and Re-Configurable Monitoring .......................... 90
5.5 Intelligent and Opportunistic use and allocation of resources ....................... 92
5.6 Self-description and Self-advertisement of Capability Models....................... 93
5.7 Programmable Traffic Flow Monitoring in Multi-Service Self-Managing
Networks .................................................................................................................... 94
5.7.1 An example scenario on the use of the EDBSLang Language ................ 96
5.8 The Conceptual Architecture of an ODM-Capable Monitoring Component, and
a reflection on the GANA Model ................................................................................ 97
5.9 Contrasting ODM to Related Work ................................................................ 99
6 Example Application Areas and Real world Scenarios .................................. 105
6.1 Overview ..................................................................................................... 105
6.2 Applying the ODM-Paradigm ....................................................................... 105
6.3 Example Application Areas and Scenarios.................................................. 111
6.3.1 Application of On-Demand SNMP MIBs to Traffic Engineering ............. 111
6.3.2 On-demand monitoring in dynamic automated Network Configuration
Management and Troubleshooting tasks ............................................................. 117
6.3.3 An example Scenario on Network Configuration Management .............. 118
6.3.4 Automated Troubleshooting Scenario .................................................... 119
7 Specification and Implementation of an ODM-Capable Prototypical System:
The ODM-Probe ......................................................................................................... 120
7.1 Overview ..................................................................................................... 120
7.2 The ODM-Probe: Specification, Architecture, and Implementation ............. 121
7.2.1 The ODM Request Handler & Admission Control Component ............... 123
7.2.2 The Repository for storing Monitoring-Behaviour-Specifications ........... 131

ODM-Paradigm for Traffic-Flow-Monitoring in Multi-Service Self-Managing Networks: Copyright by Ranganai Chaparadza: 2011

Page 5


Soyez le premier à déposer un commentaire !

17/1000 caractères maximum.