Connectedness-Aware Copy-Adaptive Routing Protocol in Intermittently Connected Networks

Abstract

In the intermittently connected networks where the complete paths from a source to a destination do not always exist, multi-copy forwarding is considered to be capable of providing more reliable transmission than single-copy forwarding. However, injecting too many copies into a network can cause unnecessary network resource consumption that degrades the network performance, especially if the network nodes are low-cost and resource-constrained. Therefore, copy control is a critical efficiency mechanism for the forwarding protocols in such situations. Using a preset threshold is a common copy-control mechanism used by many multi-copy forwarding protocols. The preset threshold is the maximum allowed number of copies of a packet in a network, calculated before running these protocols. It can effectively reduce the unnecessary copies and provide better performance than the protocols without using copy control. However, preset threshold makes the protocol less flexible with respect to the variations of network characteristics such as density or connectedness caused by node failures or new nodes joining. In this paper, we propose a multi-copy routing algorithm, Connectedness-Aware Copy-Adaptive Routing (CACAR) protocol, which adopts a novel copy-control mechanism. CACAR estimates the network connectedness locally and conducts the copy control and the next-hop selection based on the measured connectedness. This proposed protocol can achieve a higher delivery rate and lower latency without sacrificing the energy efficiency compared with other benchmark multi-copy routing protocols.