Simple Models for the Performance Evaluation of a Class of Two-Hop Relay Protocols

Abstract

Abstract. We evaluate the performance of a class of two-hop relay protocols for mobile ad hoc networks. The interest is on the multicopy two-hop relay (MTR) protocol, where the source may generate multiple copies of a packet and use relay nodes to deliver the packet (or a copy) to its destination, and on the twohop relay protocol with erasure coding. Performance metrics of interest are the time to deliver a single packet to its destination, the number of copies of the packet at delivery instant, and the total number of copies that the source generates. The packet copies at relay nodes have limited lifetime (time-to-live TTL). Via a Markovian analysis, the three performance metrics of the MTR protocol are obtained in closed-from in the case where the number of the copies in the network is limited. Also, we develop an approximation analysis in the case where the inter-meeting times between nodes are arbitrarily distributed and the TTLs of the copies are constant and all equal. In particular, we show that exponential intermeeting times yield stochastically smaller delivery delays than hyper-exponential inter-meeting times, and that exponential TTLs yield stochastically larger delivery delays than constant TTLs. Finally, we characterize the delivery delay and the number of transmissions in the two-hop relay protocol with erasure coding and compare this scheme with the multicopy scheme.

A Framework for Routing Performance Analysis in Delay Tolerant Networks with Application to Noncooperative Networks

“…This is in sharp contrast with existing work, which either considers only expected value of the metrics of interest ( [2], [3], [9], [15], [16]), or provides an asymptotic approximation of their probability distribution [19].…”

“…The research community has devoted substantial efforts to accurately estimating expected performance metrics in DTNs (see, e.g., [2], [3], [9], [15], [16], [19]). In some cases, though, knowing the expected value of a performance metric is not sufficient to fully understand DTN behavior in different conditions, and a characterization of the distribution of the metric of interest is needed.…”

“…However, most of these papers [3], [9], [15], [16] considered only derivation of the expected value of the metric of interest (typically, packet delivery delay) in their analysis. To the best of our knowledge, only a few papers [2], [19] attempted at characterizing a performance metric's distribution. However, the authors of [2] consider a network model in which the T T L is associated with each single packet replica, and thus their results cannot be used to provide bounds on the delivery time since initial packet generation.…”

Impact of source counter on DTN routing control under resource constraints

“…As to the performance modeling of DTN routing schemes, the majority of work (e.g., [5,20,8]) assume there is no bandwidth constraint, i.e., when two nodes meet, they can transfer unlimited number of packets. To the best of our knowledge, [10] is the only modeling work taking into account bandwidth constraint.…”

“…Furthermore, [3,11] show that under a large class of mobility scenarios, the inter-meeting time follows a power-law up to a point, and then exhibits an exponential decay. Based on these previous studies, we assume in this paper that the pair-wise inter-meeting time follows exponential distribution with rate β, an assumption that has been made by most previous modeling work in DTNs (e.g., [5,20,8]). We leave the study of more general inter-meeting time distribution as future work.…”