Docstoc

Leadership

Document Sample
Leadership Powered By Docstoc
					    Engineering Doctorate Annual Report
                    Radhakrishnan Kadengal




Optimal Infrastructure for a Universal Electronic Enterprise

               Reporting Year: 2000 – 2001

                    Batch: 1999- 2003




                        Supervisors:

Prof.   Jon Crowcroft       CS         UCL          Acad
 Dr.    Lionel Sacks        EE         UCL          Acad



                    Nortel Networks
                   Harlow Laboratories
                      London Road
                  Essex CM17 9NA, UK
                             Executive Summary



       The topic of research is the design, development and implementation of
system architectures to efficiently operate, control and manage a Universal
Electronic Enterprise based on Computer Communications Control.
       The project envisages to make original contributions in inter-disciplinary
systems science, distributed engineering technology and real-time protocols. The
main thrust is to find a confluence of Communications, Control, Signal processing
and economics/ game theory/ management in the emerging internet.
       Value based service provisioning and resource optimisation for computing
and multi-media application services and network resources is a high touch
functionality. Such a system operating with appropriate pricing models will
provide optimum quality of service and balanced network growth.
       Work conducted last year has progressed to the realisation of a network
management and control middleware for a client IP network running over a
photonic network with differentiated provisioning and fast restoration. An
isarithmic model for QoS representation and portfolio charging is being studied.
The work is progressing to the development of new protocols in fast IP packet
forwarding over automatically switched telecommunication networks with
wavelength multiplexing and routing and proper accounting for this always-on
networks. This has resulted in a set of invention submissions.
                                    Acknowledgements

       Author wishes to thank Nortel Networks and UCL for the opportunity given
for conducting the research. He thanks his academic supervisors and industrial
supervisor for the encouragement and support during the turbulent times caused by
the slow down in the industry.
       Author wishes to thank the research community in this vast field. At this point
in time, the references are mainly given in the papers cited in section H of this
document.
      Engineering Doctorate Research programme - Annual Report for 2000 - 2001


Title: Optimal Infrastructure for a Universal Electronic Enterprise

Objectives: Design, development and implementation of a system architecture to operate,
control and manage a Universal Electronic Enterprise based on Computer Communications
Control

Subjects: Envisages to make original contributions in inter-disciplinary systems science,
sentient computing, distributed engineering technology, real-time protocols and transmission
systems. I would like to find a confluence of Communications, Control, Signal processing and
economics/ game theory/ management in the emerging internet.

Significance: As the internet market is readying itself for high touch services (versatile, value
added and interactive services) in about an years time, the timing to undertake the research
programme in the theoretical and implementation aspects of intra-domain and inter-domain
internet and IP traffic control and management to meet the quality requirements of multi-
service networks seems to be apt.
Resource management is envisaged as a high touch provisioning functionality, rather than
control of Bandwidth/ Buffer/ CPU alone. Appropriate pricing models and control techniques
are to be devised. The functionalities may be enabled using distributed agents, multi-level co-
ordination, System identification, State-space techniques, event-driven supervisory control
system, Chaotic signal processing, Wavelets, AI for decision support, technologies like Gini,
active networking etc, and electronic/ photonic hardware.

Progress of the Project:

A. Specific work during the reporting year:

The developments in the communication industry in the current academic year provided
opportunity for more commercial involvement so as to help improve the cash-flow.
Involvement with end customers, retailers and wholesalers in the requirements
capture, assessments, bid support and devising strategies- both commercial as well as
technical- provided better understanding of the industry. This year also provided the
opportunity to act as a consultant across a broad spectrum of technology- from
multimedia to core network issues with IP and Optical technologies.

Consequent to alignment process in the industry the laboratory environment is
changed from Linux based software routers to a massive captive office having almost
the full range of Nortel product portfolio (both enterprise and network) for very large scale
experiments and customer demonstrations of a real world electronic enterprise.

In the first half of the year, theoretical studies and practical demonstration of a data driven
approach to service provisioning (where the boundary routers use traffic measurements to
autonomously control the number of light-paths) has been delivered using software routers
and MPLS. This was later extented to provide a fast restoration for a MEMS based photonic
network.

Studies have been carried out on a new method of charging for quality of service in an
isarithmic manner. Studies are also conducted on utility functions that incorporate dynamic
perceptive behaviour.
Advanced reliabilty analysis of system of systems, hardware as well as software dependability
issues, emergency planning and disaster recovery, echo problems in VoIP, QoS mapping and
solutions for dependable networks etc were some of the other issues handled during the year.


B. Specific conceptual contributions:

1. Proved that a distributed system operating under distributed feedback send to
   participating decision agents will achieve system optimisation
2. An „isoqos‟ charging system (own terminology) can be used to bundle service portfolios
   and this has to consider the ratio of priority service to total service load for assuring levels
   of service
3. A dynamic utility model can indicate when the system moves to negative utility and this
   can help manage the network better.
4. Architecting a real data driven IP/ Photonic embedded network control system
5. An improved scheme for optical ethernet running over WDM is proposed


C. Achievements:

1. Designed and Developed a data driven IP/ Photonic network control system in Linux.
2. An extented model of the above scheme is now being developed using an embedded
   processor for a MEMS based photonic network.
3. Developed a framework for the value analysis of a telecommunication system.
   Mathematical proof and novel presentation of advantages of resource sharing was
   produced as part of network system evaluation.
4. Designed a new equation based TCP. (This is not yet tested)
5. Designed a decision support system architecture for emergncy plan and disaster recovery
   in telecommunication systems
6. Developed a new method to identify single point failures in a system when using fault
   tree analysis
7. Designed an improved medium access system for optical ethernet using WDM
8. Designed and run simulations of some of the above systems
9. Authored one technical paper (published in UKTS17, Dublin) and two technology-
   business papers, three external presentations and eight internal reports

Ideas of network measurement systems and business policy switches can now be seen to be
available as specific products. The ideas on priority user scheduling for content provisioning
is developed elsewhere. The photonic network control system is now being developed for a
MEMS based photonic switch. The IP Client system is available as an agent in routers and the
optical network server system is available as a module in the optical/ photonic switch OS.

For further details, please see monthly reports- October ‟00 – September ‟01.

D. Intellectual Property: Filed two patents in the UK and US on distributed network data
collection and dissemination, multi-layer network control system and submitted for filing two
patents on adaptive multipath load balancing and optical ethernet.
E. Evaluation of the commercial potential of research:

Internet and IP based multimedia traffic Control and Management has become one of
the most important businesses today. Value based service provisioning and network
economics are expected to be the prime movers behind future networks providing Quality of
Service and balanced network evolution. The ability to receive network services at „will‟ has
enormous commercial potential in the wireline, wireless and optical networks. Carrier
network operators, Enterprises as well as individual users can use the architecture developed.
The system automates the provisioning and optimisation functions and hence reduces the
operation costs. The work is now directed towards developing an advanced network control
and traffic management middleware. This will result in significant strategic advantage and
financial rewards for the organisation.

The system can be used for the management of various other service provisioning and
resource optimisation systems including Application Service Provisioning, process
automation and GRID computing.

F. Expected future research direction:

I wish to further the research to the development of a middleware for quality of service
monitoring, usage control and resource optimisation for near real-time applications using
sentient computing. Stability, Scalability, reliability, risk analysis etc are also required. I wish
to define interfaces for interoperability with other management systems/ products. I also wish
to link this to automated charging. I propose to develop the system using standard platforms
like Java, Corba etc with software engineering based on UML.

I would like to concentrate in the optical and photonic networks using distributed routing
fabric carrying IP. For the immediate future I would like to work on optical ethernet and input
scheduling schemes for wavelength routing.

Optical packet switching using in-band signalling, game theoretic approach to improve
network survivability, cross-layer multiple failures and routing, dynamic programming for
policies and bandwidth trading etc are other areas of interest.

I would like to take appropriate scientific, technical and management modules that are
available for study. I am interested in content that deals with dynamic resource control and
management for delivering value added services viz. Real-time distributed and parallel
computing, sentient computing, embedded control, OOD, Java, Corba, signalling, switching,
photonics, free-space optics, planning, performance analysis, management, Network calculus
etc. In personal and business skills elements, I am interested in content that deals with
Research management viz. Business in competitive environments, corporate finance,
innovation case studies, research process and practise, statistical modelling and data analysis,
new ventures, decision making, research communication, managing people and change,
Intellectual Property Law etc. I have applied for the course “New Venture Development” and
“Managing Growing Business” by the Centre for Scientific Enterprise and London Business
School. I believe these courses will help me attain the skills, behaviours and knowledge
required to effectively hasten technology trajectories in the scientific and technical
enterprises. This is required to lead continuous growth and maintain strategic position in the
changing technological, economic, social and political environment. I aspire to be a leader of
strategic research in the next ten years.

G. Citation index and Bibliography: Under preparation!
H. Courses/ Conferences/ Presentations/ Papers/ Invention submissions:

Oct   ‟00       Internal doc: 1) Distributed digital control system approach for dynamic
            resource management of Carrier networks: Extensions to Optical routing project
            2) Multilayer Control and Management Demonstrator Phase 1 – Architecture
            and scenarios
                Courses: 1) Straight talk: giving and receiving feedback – one day course 2)
            Influencing and Assertiveness skills – two day course with one day follow up
Nov   ‟00       Presentation in UCL research seminar: Dynamic Resource Control and
            Management Middleware for carrier Networks
                Course: Leader in each of us: understanding the Nortel Networks
            Leadership Attributes and provides a tool to diagnose current leadership
            capability and skill development priorities
Dec   ‟00       Internal doc: 1) Value analysis of Dynamic Resource Control- framework
            and methods 2) DRC/DiffServ – Value and Traffic mix analysis using
            simulation and measurement
                Course: Working together to reach greater heights
Jan   ‟01       Internal doc: Project Report on Adaptive multipath load balancer
                Filed invention submission on adaptive multipath load balancing
                Course: Emerging Leaders Programme: a prestigious Nortel course
Feb   ‟01       Internal doc: Design of IP over Photonic network - Preliminary draft
                Course: Network and Services Management by Dr. L Sacks
Mar   ‟01       Poster presentation at UCL: Adaptive Bandwidth Management in
            IP/Photonic networks
                Submitted abstracts of a paper to the UKTS17
Apr   ‟01       Course: Leadership in Technology Organisations (exec MBA module, LBS)
            and passed the exam.
                Received membership of the IEE
May   ‟01       Internal doc: Proposal for emergency plan and disaster recovery for an
            international network operator
                Paper: Advanced Router and Switch Architectures for Quality of Service in
            distributed routing fabric, submitted to UKTS17 by IEE
Jun   ‟01       Internal doc: Report on the Fault tree analysis of a succession network
                Course: Understanding Business Finance
                Presetation: Riding the Waves: Nortel Networks and John Roth, LBS
                Paper: Riding the Waves: How Nortel Networks created successive waves
            of technology, LBS
Jul   ‟01       Paper: Increasing profit returns and dependability of System of systems: An
            overview
                Received the title of Chartered Engineer from the Engineering Council
Aug   ‟01       Internal doc: Report on troubleshooting a VoIP problem
                Invention submission on Improved protocol to maximise the efficiency of
            medium access in Optical Ethernet
                Conference: IETF Lunch time talks at UCL
Sep   ‟01       Conference: London Communications Symposium

				
DOCUMENT INFO