Utilizing Spatial Locality to Optimize Temporal Efficiency in OLSR Route Calculations

The Optimized Link State Routing (OLSR) protocol has received considerable attention as a viable routing solution for mobile ad hoc networks. However, OLSR suffers from high routing and computational overhead, resulting in significant energy consumption and is quite detrimental to the lifetime and service duration of mobile ad hoc networks. In this work, we provide a detailed analysis of the OLSR protocol and show that its route calculation (RC) algorithm is responsible for the majority of the overhead as it redundantly re-computes the entire routing table after each routing change. We further evaluate the performance of the classic RC algorithm as a function of network size, node mobility, and network density and determine that the RC algorithm is overwhelmingly triggered by very localized topology changes. Interestingly, our analysis shows that after a local topology change, the majority of the recomputed routing tables remain unchanged from their original state. Based on this analysis we identify two distinct types of routing table re-calculations that are not necessary: local changes and vain changes. Hence, we proposed novel algorithms to optimize route table calculation in the case of these two changes. The proposed ARC algorithms work cooperatively to reduce overall runtime (and therefore energy consumption) by utilizing spatial locality to provide incremental updates to the routing tables (as opposed to re-computing the entire table upon each route change). Extensive experiments verify that the proposed method is scalable with respect to both network size and node density and reduces overall RC runtime by 80% - 90% for practical scenarios. Moreover, the proposed RC algorithms are fully-compatible with existing OLSR implementations.