Fuzzy non-dominance multipath link-state routing framework for network routing management with inaccurate information

An open problem for quality of service routing management is the use of inaccurate metric information. Due to the periodic link-state updates, the router always contains out-of-date information. This paper proposes a novel framework for routing management with inaccurate information. We introduce fuzzy metric representation and the concept of non-dominance or Pareto-optimal fuzzy shortest path routing into our framework. The term fuzzy shortest path routing used here is distinct from the classic sense of fuzzy rule controlled routing. We map the crisp link-state information to a new plane by fuzzification factor, and use fuzzy non-dominated multipath routing algorithm for path selection in order to achieve a better network usage. Under our framework, a network administrator can easily switch between fuzzy or non-fuzzy routing. We will discuss and investigate how this framework could be merged with the current network infrastructure to support better QoS for emerging services and applications. A case study is shown in the paper, and simulation results are also included to validate the network preference enhancement under our framework.