Efficient Protocol Decomposition and Performance Analysis of the SSCOP Layer

This paper presents a decomposition of the SSCOP [1] protocol, which is based on the detailed analysis of the performance of the layer. The protocol decomposition aims at mapping the various functions of the SSCOP layer onto modules with optimized running speed when these are applied on modern network processors ([2], [3], [4]). The SSCOP layer is a mandatory part of the ATM Signalling Adaptation Layer (SAAL [5]). The running time performance of the SSCOP has a significant impact on the execution speed of the entire layer. A mapping of the SSCOP most complex functions onto dedicated hardware units leads to accommodating the requirement of speeding up the SSCOP execution. This work analyzes the protocol into primitive functions and presents the performance of the functions that are executed frequently during ATM call setup and tear-down. The measurements from the protocol’s profiling lead to a partitioning of the functions, in order to implement a subset of these in hardware and accomplish their execution in wire speed. The results of this performance analysis have been utilized to improve on the performance of the PRO31. ATM signalling applications can benefit from the proposed hardware software partitioning because of the significant acceleration of the ATM signalling stack that can be achieved in core network subsystems. Key-Words: SSCOP, Performance analysis, Network Processor, Protocol Profiling, Hardware and Software Decomposition.