Spare capacity allocation using shared backup path protection for dual link failures

This paper extends the spare capacity allocation (SCA) problem from single link failure [1] to dual link failures on mesh-like IP or WDM networks. The SCA problem pre-plans traffic flows with mutually disjoint one working and two backup paths using the shared backup path protection (SBPP) scheme. The aggregated spare provision matrix (SPM) is used to capture the spare capacity sharing for dual link failures. Comparing to a previous work by He and Somani [2], this method has better scalability and flexibility. The SCA problem is formulated in a non-linear integer programming model and partitioned into two sequential linear sub-models: one finds all primary backup paths first, and the other finds all secondary backup paths next. The results on five networks show that the network redundancy using dedicated 1+1+1 is in the range of 313-400%. It drops to 96-181% in 1:1:1 without loss of dual-link resiliency, but with the trade-off of using the complicated share capacity sharing among backup paths. The hybrid 1+1:1 provides intermediate redundancy ratio at 187-310% with a moderate complexity. We also compare the passive/active approaches which consider spare capacity sharing after/during the backup path routing process. The active sharing approaches always achieve lower redundancy values than the passive ones. These reduction percentages are about 12% for 1+1:1 and 25% for 1:1:1 respectively.