PhD Studentship in Optimisation Analytics for Resource Management in IP Networks
Ulster Supervisors: Prof Gerard Parr, Prof Sally McClean, Prof Bryan Scotney
BT Advisor: Dr. Steve Appleby
Expected start date: January 2009
Overview:
The following PhD project has been created under the auspices of the
India-UK Advanced Technology Centre (IU-ATC) of Excellence in Next Generation
Networks Systems and Services
(www.iu-atc.com).
During the lifetime of the project there will be opportunities to spend
a period of time as part of a related internship at BT Adastral Park in Martlesham,
UK and/or to visit one or more partner research institutes in India.
The successful candidate should therefore be willing to travel and spend some
time overseas as part of the research project. An Internship will be
supported by BT whilst any visits to India will be supported by our associated
UKIERI-DST project for the IU-ATC Virtual Graduate Research School.
Arrangements for the internship / visits will be made by the supervisory
team of the PhD project.
Project Specification:
The rising customer demand for constant application behaviour in
the presence of dynamic hardware-software faults or degradations in the
communications fabric is placing huge pressure on the operations of the
Network Management Control Plane to deliver and support Policy-based Optimisation.
Coupled to this is a desire to have near real-time responses to exception events
at the Network Element level which impact on the ability to satisfy the QoE
requirements of higher-level large- scale distributed applications. When one
considers the ability to have full IP transport from the client through the
access network and into the optical core fabric it opens up opportunities to
have seemless end-to-end resource reservation which is mapped against the context
of prevailing applications requirements and the services that are required to
deliver these requirements. In many cases these requirements are defined in a
programmable SLA which can be wavelength, bitstream, connection, service,
application or user focused.
The integration of the management and control plane operations offers opportunities to extract efficiency gains from the underlying infrastructure in support of a specific customer application demands, but it does requires highly dynamic approaches for dedicated optimisation functions to be designed and deployed to suit the services and physical landscape involved.
The main thrust of this PhD research project will be to examine cost-benefits and opportunities for efficient mapping of available resources in the fabric against the key causes of problems from the physical level (storage, compute and network elements) and translate this into metrics that can be used to generate a real-time policy framework that will support dynamic mapping of applications to services that can cope with prevailing conditions. Within the framework will be requirements to model how any such optimisation algorithms can cope with increasing scale and to also minimize unnecessary data obesity from the control plane in support of varying classes and scale of distributed applications and enterprise solutions. The research should explore current state-of-the-art network, combinatorial and stochastic optimisation algorithms and adapt them to IP networks. Due to the increasing scale of complexity, we may also explore the use of heuristic search methods for solving such class of problems before transferring the near optimal solution to globally convergent algorithms. Constraints (linear, nonlinear, probabilistic or integral functions) in the modelling stage of the optimisation problem will also be investigated.
The approach to be taken will first look at the derivation of mathematical methods to help categorise the applications and services landscape and then to examine the opportunities to automate the procedures that map accessible and addressable resources to these demands. The resources involved will be a combination of network, storage and processing capabilities. The end target is to derive a generic approach which will be vendor, service, network agnostic.
Eligibility:
Candidates should hold a first or upper second class honours degree in
Computer Science, Electrical Engineering or a cognate area. It is desirable
that candidates will have an MSc in a related subject and technical
experience/knowledge of fixed-wireless Telecommunications Network Protocols and
Data Analytical tools and optimisation techniques for Network Management in
converged IP networks
Whilst this PhD Award is open to local and international candidates, in the context of the IU-ATC we would especially welcome applications from suitably qualified candidates from India. Successful candidates will enrol on a full-time programme of research studies leading to the award of the degree of Doctor of Philosophy.
The studentship will comprise fees and an annual stipend of £12,940. It will be awarded for a period of up to three years subject to satisfactory progress and is tenable in the School of Computing and Information Engineering at the Coleraine Campus, Northern Ireland, UK.
The closing date for receipt of completed applications is 14th November 2008.
For further information please contact Professor Gerard Parr, Chair in Telecommunications Engineering, University of Ulster - gp.parr@ulster.ac.uk.
Application materials are available from:
Research Office,
University of Ulster,
Cromore Road,
Coleraine,
BT52 1SA,
Tel: 028 7032 4729,
e-mail; hj.campbell@ulster.ac.uk,
Web:
http://research.ulster.ac.uk/info/prospective/apply.html