New QoS Measures for Routing and Wavelength Assignment in WDM Networks Shi Zhong Xu and Kwan L. Yeung Dep. Of EE, The Univ. of Hong Kong ICC’2002
Example 1: Mesh-Torus Network
The Disadvantages of OABP Measurement The overall average blocking probability (OABP) is not a good performance measure. It hides the information of service deviation among node pairs. It does not consider the revenue generated by different types of calls.
Example 1: Mesh-Torus Network (continued)
Example 2: ShuffleNet
Example 2: ShuffleNet continued)
Requirements for A New QoS Measure Unbiased. Make reference to the individual requirements of SD pairs. One-sided. Take the potential revenue into consideration.
New Performance Measure B * ij : QoS requirement of a SD pair ij. λ ij : arrival rate for SD pair ij. Λ = Σ ij λ ij. The overall average blocking probability (OABP) is biven by
New Performance Measure (continued) A new class of composite performance measures is designed as follows, I is the set of node pairs with blocking probabilities exceeding their QoS requirements. L ij is a punishing factor proportional to the shortest distance between nodes i and j.
New Performance Measure (continued) EB(1) can be considered as the weighted average excessive blocking. EB(2) can be considered as the mean square weighted excessive blocking.
Case Studies Least-loaded routing (LLR). Fixed path least congestion routing (FPLC). New fixed path least congestion routing (NFPLC). Fixed routing with random wavelength assignment (FR_RAND). SPREAD routing based on layered graph (LG).
Case Studies (continued)
Conclusion A new class of composite measure EB(p) is proposed. Compared with OABP, EB(p) is unbiased, makes reference to the QoS requirements, one-sided, and takes the potential revenue loss into consideration.