Blocking probabilities for inverse multiplexer in public switched digital networks

New application services, such as multimedia, digital images, video on demand, and high bandwidth data services, demand bandwidth much higher than those available from existing public switched digital networks (PSDN). These new services have a wide range of traffic flow characteristics and quality of services (QoS) requirements. To meet the new requirements, network providers use a technique called inverse multiplexing which combines two or more independent switched digital channels into a single high bandwidth end-to-end connection. In this paper we focus our study on the dynamic allocation and management of bandwidth by inverse multiplexers to support variable bandwidth applications. We add a quality of service feature in the BONDING protocol for the inverse multiplexer. A real-time algorithm is developed to compute blocking probability experienced by network users during a connection establishment phase (external blocking) and during a data transfer phase (internal blocking). We show that the algorithm is suitable for real-time usage because it is both stable and efficient