Solving GMPLS Multi-Region Path Constraints

In this letter, we propose a novel binary integer program (BIP) to obtain the optimal and correct combination of switch type per link, and adaptation capabilities per node, along a multi-region traffic engineered label switched path (TE-LSP) in a generalized multiprotocol label switching (GMPLS) network. The proposed BIP considers constraints such as the nesting/un-nesting order of LSPs, the adaptation capacity per node, the bandwidth capacity per link, as well as efficiency. By efficiency constraint we refer to opting primarily for nesting/un-nesting as opposed to conversion which is a resource demanding process. Numerical results show that applying the proposed BIP in a constraint shortest path first (CSPF) algorithm achieves lower blockage rates and outperforms an existing graph transformation method which is the closest related work to our proposal.