MeTHODICAL: Towards the next generation of multihomed applications

We introduce MeTHODICAL, a multihoming-aware decision-making mechanism which enables applications to capitalize on multihoming availability. MeTHODICAL employs an algorithm that establishes the foundation for assigning weights to multiple criteria, and a path optimization algorithm that performs path selection restricted by multiple multihoming and traffic performance criteria, a well-known NP-hard problem. The feasibility of the MeTHODICAL weighting criteria algorithm is demonstrated in this paper for different classes of service as specified in ITU-Y.1541. The algorithm is shown to have a time complexity of O(mcnv)Omcnv. The path optimization algorithm, with a time complexity of O(n·m)O(n·m), is evaluated for the 1:1 and 1 + 1 protection models and is compared with previous proposals in this area. Results from our testbed evaluation demonstrate that the MeTHODICAL path optimization algorithm does not suffer from ranking abnormalities, nor does it require high-volume data to be efficient. The results show that heuristics can enhance the performance of MeTHODICAL and eliminate handover side-effects.