Innovative resource utilization model for distributed multimedia systems.

M. Saleem Mian; M. Riaz Moghal; Sajid Aeem

The powerful end-systems and high-speed networks give a boost to a new class of Distributed Multimedia Applications, such as video-conferencing and video on-demand. These applications are highly demanding on Quality of Service and dynamic in their nature. The users generally start their sessions at their machines. The random arrival of these newly accepted sessions could cause some computers to become highly loaded creating bottlenecks while others are lightly loaded or idle. Ideally, we would like to have all the machines operating continuously these sessions that will eventually lead to the minimum execution time. The objective of load sharing is to improve the performance of parallel computations by automatically equalizing the workloads of processors during program execution i.e. minimizing bottleneck resource utilization. The Quality of service of each session in this strategy is adapted to dynamic changes of available resources based on distributed workloads and users choices. In this paper; the authors present a Resource Utilization Model Prototype (RUMP). The model shows the basic concepts proposed in the Resource Utilization Model. lt also demonstrates the capability of the Resource Utilization Model to deal with session admission (new session creation) quality adaptation (session quality profile) and integrated resource allocation in an integrated way. The prototype provides the opportunity to drop the selected session(s) to change the selected session(s) quality profile and also provides the opportunity to change the total resources available. It facilitates to negotiate with the client/customer to change his resource levels so that system can be used in an efficient way. The authors performed different tests to evaluate prototype for different session requests.