Modeling Quality of Service (QoS) for Distributed Virtual Environments
by Dr. Denis Gracanin
 

QoS negotiation process balances between the QoS specified by a client, the resource capabilities of the distributed system and the functional capabilities of the distributed application. Selecting type of service received from the network includes best efforts (today's Internet) or various gradations of premium service (Internet2). In case of distributed virtual environment applications the major obstacle is dealing with many mostly subjective parameters that have an impact on overall perceived QoS (sense of presence, multidimensional sensory inputs, user interface, interactions, task performed). In most cases measurement is done in a form of questionnaires given to the users after using the application. Results gathered from the analysis of users responses are removed from the network level. However, the network services are critical for the "real-time" response lack of which may cause the break in a sense of presence and directly impact perceived quality on application level. There is a need to measure network services in quantitative terms (bandwidth, loss, delay and jitter tolerances) that represent desirable/minimal/maximal values required for different "real-time" sensory and control data (video, audio, simulation, tracking, etc.). A QoS architecture for distributed virtual environment applications is needed that integrates network and application level requirements in a single framework and establishes relationship among requirements. The QoS architecture can then be used to develop a modeling mechanism for network parameters.