An optimized prioritized load balancing approach to scalable routing (OPLBA)

Abstract

Mobile Ad hoc networks (MANETs) are self organized multi-hop networks, without any infrastructure such as base stations or access points. Due to the mobility and absence of any central administration, the resources of MANETs are limited. If there is any congestion in the network, it puts a great strain on the already scarce resources and severely affects the performance of such networks. Multi path routing is considered as advantageous over single path routing, due to the many benefits it offers. However, these benefits do not come without their associated costs. In this paper, we propose a general metric to define scalability of a routing method. We further propose and implement a new load-balancing routing protocol, which retains the benefits of multiple paths, while at the same time keeping the overheads of routing, as close to single path routing, as possible. The proposed scheme dynamically distributes traffic through different available paths, so that no single path is flooded. Priority is assigned to available paths and paths with higher priority (better routes) are used more often than those with lower priorities. To keep our method light-weight and scalable, we control the Degree of Distribution (DoD) value (number of alternate paths used), to reap maximum benefits at minimum cost. To further reduce the overheads and decrease access time, optimized insertion and path selection are provided. An index to the RouteList table has been added, which reduces the access and insertion time to O(m ? n d ) and O(m) respectively, which is within a constant difference of Single path routing methods. Simulation results demonstrate that the proposed solution shows significant improvements in network metrics such as packet loss ratio, end to end delay, throughput and packet delivery, without any increase in routing overheads. Results also verify that this model is very efficient and scalable.

An efficient load balancing method for ad hoc networks

“…This is because maintaining each additional path incurs some time and space cost, but its usage may not be equally beneficial. This result was further extended in our previous work, 20 where we define the term DoD as the total number of profitable paths, out of the many alternate paths for a same source-destination pair. A scalability metric to evaluate usefulness of a path is also defined there.…”

“…20 Efficient load balancing method is an extension of OPLBA, with some similarities, but at the same time having very distinct core logic. The multiple paths so obtained are arranged according to increasing round trip time.…”

“…20 Both QMLB and OPLBA are comparable multiple path routing methods. 20 Both QMLB and OPLBA are comparable multiple path routing methods.…”

“…Another method we find appropriate for comparison is our own implementation: Optimized Prioritized Load Balancing Approach to scalable routing (OPLBA). 20 Efficient load balancing method is an extension of OPLBA, with some similarities, but at the same time having very distinct core logic. The similarity that ELBM shares with OPLBA is that only a few efficient paths, as defined by DoD, are used in both.…”

“…We compare our protocol with 2 load balancing protocols: QMLB 38 and our earlier implementation method OPLBA. 20 Both QMLB and OPLBA are comparable multiple path routing methods. The differences between these 3 protocols is already explained in Section 2 and further highlighted through Table 2 below.…”