Layered Multiplexing Considered Harmful

Traditionally, computer communication networks have been optimized with respect to throughput, robustness and absolute delay, with little or no concern for the variation in delay (jitter) induced by the network. It is now desirable that high speed networks support a greater range of telecommunication services by providing a multi-service environment. If the transmission of jitter-sensitive tra c is not to be arbitrarily precluded, then new protocol architectures must take account of, and provide support for, the constraint of jitter. The ATM approach to broadband networking is presently being pursued within the CCITT (and elsewhere) as the unifying mechanism for the support of service integration, rate adaption, and jitter control within the lower layers of the network architecture. This position paper is speci cally concerned with the jitter arising from the design of the middle and upper layers that operate within the end systems and relays of multi-service networks (MSNs). 1 What is Layered Multiplexing? Multiplexing can be viewed as a layer function that maps some number of upper layer tributary associations onto a single lower layer aggregate association. The principal e ect of multiplexing is the sharing of lower layer resources by upper layer clients. Clearly some degree of inter-system multiplexing is justi ed when a common transmission medium is shared by a number of independent systems. This style of multiplexing is normally accommodated within the media-dependent components (MAC) of the Data Link layer. In most environments, there is a further requirement for intra-system multiplexing to support the shared use of a system's points of attachment to its transmission media. In traditional layered architectures, intra-system multiplexing is not restricted to a single layer. Instead, the functionality is supported in a layered fashion at many points within the upper and middle layers of the protocol stack. Title suggested by Dr David Clark