PERFORMANCE EVALUATION OF DISRUPTION TOLERANT NETWORKS WITH IMMUNITY MECHANISM AND CODING TECHNIQUE

Abstract

We examine the performance of a Disruption Tolerant Networks (DTNs) with an epidemic routing (ER) scheme with the coding technique and/or immunity mechanism under the various network environments. We are interested in the scenarios of opportunistic dissemination of large files. First, we study how the different implementations of the ER scheme perform in diverse network settings. We compare the performance of ER with its summary vector implemented as both a list and as a Bloom filter. Second, we examine how network coding affects the performance of the ER scheme. To this end, we investigate the performance of encoding-based routing (EBR), a variant of the ER scheme which uses random linear coding at source nodes. EBR is expected to mitigate what is commonly known as the coupon collector’s problem, which arises when a large file is chopped into small fragments and then the fragments are disseminated throughout the network. We

compare this to the case where intermediate non-source nodes are allowed to create new linear combinations from the ones it already holds. Lastly, we evaluate the benefits of two different types of immunity mechanisms – one based on file ID and the other based on bundle ID – with not only the ER scheme but also two different EBR schemes in various network scenarios and settings. We also investigate the performance gain from compressing the immunity list.

By presenting and analyzing extensive simulation results, we provide information that could provide a guideline for employing each of the aforementioned techniques in routing schemes of interest in various network settings.