Efficient routing in MANETs using ordered walks

Abstract

A new approach for routing protocols operating in MANETs is presented in which flooding is not required to establish paths from sources to destinations on demand in MANETs of moderate size. The concept of ordered walk is introduced as a depth-first search (DFS) that does not rely on geographical or virtual coordinate information and is much more efficient than mere random walks. The benefits of using DFS as the building block of the signaling in MANET routing protocols are exemplified by the introduction of the Ordered Walk Search Algorithm (OSA), which is used as part of the proposed Ordered Walk with Learning (OWL) protocol. OWL integrates OSA with the learning of paths from prior successful and failed attempts, and performs one or multiple concurrent ordered walks to search for destinations. Simulation experiments are used to compare the performance of OWL against that of well-known MANET routing protocols based on BFS (e.g., OLSR and AODV). The results show that OWL can achieve a performance comparable to traditional protocols that rely on some form of flooding of link states or network-wide dissemination of distance information in terms of packet delivery ratios and average end-to-end delays, while incurring up to ten times less overhead than AODV.