Project Proposals for William Flynn Scholarship: IIndex page

Project Number: 7
Project Title: A Framework for Optimal Dynamic Configuration of Protocol Stacks for Streaming Media to Mobile Devices in conjunction with Motion Prediction Algorithms
Project Supervisor: Mr Kevin Curran, Professor Paul Mc Kevitt

Research to be conducted
Seamless connectivity to multiple wireless networks is becoming increasingly important with the proliferation of mobile multimedia devices such as PDA's and laptops. Wireless networks will differ in bandwidth, size and location of the coverage area, access costs and network characteristics as indeed the processing power of the mobile devices. However all these devices will require the optimal delivery of networked multimedia data. One method common on wired networks is for sources to limit their transmission rates to accommodate the lowest bandwidth links, even though high-bandwidth connectivity might be available to many of the participants. However, this method does not provide optimum throughput to heterogeneous clients due to its quest for a common denominator bandwidth.

One possible solution for streaming multimedia over IP and Mobile IP Networks is to cater for rate adaptive applications by distinguishing between the various media components (e.g. video, audio) providing separately composed protocol stacks for each of the media which deliver the data to separate quality (high, medium and low) multicast groups allowing heterogeneous clients to connect according to their resources. Multicast groups overcome the scaling problem and clients subscribe to each group in accordance with resource availability and move between groups according to bandwidth availability. In addition by adhering to the principles of Reflection, one can provide a reflective framework with hooks for interchanging Quality of Service (QoS) modules at run-time. With reflection, the QoS modules may be developed from the outset as meta-objects thus allowing the systems functional requirements to be separated from the non-functional (i.e. QoS) requirements permitting future QoS modules to be incorporated at a later date with little or no disruption to the system. Active proxies could be investigated in to see whether they can perform bandwidth adaptation through transcoding and rate-control.

In addition it is hoped that the research will cover the problem of the delay imposed upon mobile receivers when switching between wireless cells in the form of an extension of Mobile IP's hand-off algorithm. A possible solution might involve the exploitation of mobility prediction to predict a mobile terminal's future location based on its previous history (i.e., the last cell that it has been in) and for the media stream to be already present and cached by next cells base station ready for receiving by the mobile device. The researcher could evaluate the dynamic reconfigurability of the middleware framework and present the results from a series of streaming applications utilising dynamic adaptation at runtime. It is hoped that these results would show that run-time adaptivity introduces only a small overhead and that throughput increased as a result of having best-fit protocol stacks in place for various environments.

Work Plan
The following schedule shows the proposed work plan.

Stage 1 : (1-12 Months) : initial investigation
¨ Literature Review
¨ New installation procedures and support for all platforms (PC, and Unix).
¨ Collaborator enhancements and HQ media streams (e.g., H.263 and MP3).
¨ Initial early background chapters
¨ Bug fixes and integration of other code.

Milestone 1: (re)Specification of software requirements / early chapters (draft)
Milestone 2: Publication documenting state of the art in multimedia frameworks and future research

Stage 2 : (13-24 Months) : Wireless CODEC/Framework Development
¨ Layered wireless CODEC and MPEG support.
¨ Real Networks gateway - reflective API hooks
¨ Broadcast management tools (i.e., broadcast manager and director's console).
¨ Advanced JMF development - Predictive movement algorithm
¨ Bug fixes and integration of other code.

Milestone 4 : Present paper at conference of work to date on framework
Milestone 5: Journal paper documenting latest findings and describing framework

(25-36 Months) : Testing/Evaluation and Thesis Write-up
¨ Command stream support. - NetShow gateway - GSM/GPRS protocols
¨ Initial media effect processing service
¨ Test mobile movements - test prediction algorithm - test video/audio
¨ Evaluation of framework
¨ Produce Final Version of Thesis

Milestone 6: Release software to general research community
Milestone 7: Journal papers detailing new software and contributions to the field

Supervision
Professor McKevitt and Kevin Curran will be responsible for directing the activities of the research and provide guidance to the student throughout the project. Professor McKevitt and Kevin Curran will manage the project and supervise the student; be responsible for policy, budget, and scheduling and provide a mechanism for regularly monitoring and evaluating the progress of the student as well as creating a "team-based" environment in which to work. Regular meetings will review progress and, if appropriate, revise plans and resource allocations. It is essential that the activities of the student meet the requirements of the project specifications. As appropriate, workshops, training and other activities will be scheduled in conjunction with conferences and meetings attended by the student.

Supervisors Background
Kevin's research with the Distributed Systems Group within the University of Ulster has been in a diverse range of fields such as Java, distributed objects, networking, multimedia, object orientated frameworks, middleware, CORBA, distributed systems algorithms, security, WAP, mobile networking, and other areas involved with the real-time nature of multimedia on a wide-area network. To date - I have over 30 publications in journals and international conference proceedings. I continue to work on research into reflective middleware and keeps abreast of the latest networking technology.

If you are interested in being considered for a studentship please contact
the Group Director, Professor T.M. McGinnity by email:
tm.mcginnity@ulst.ac.uk

or telephone: +44-(0)28-71375417.

See the current research section of this website for details on research projects pursued by existing PhD students