A mobility-transparent deterministic broadcast mechanism for ad hoc networks

Abstract

Broadcast (distributing a message from a source node to all other nodes) is a fundamental problem in distributed computing. Several solutions for solving this problem in mobile wireless networks are available, in which mobility is dealt with either by the use of randomized retransmissions or, in the case of deterministic delivery protocols, by using conflict-free transmission schedules. Randomized solutions can be used only when unbounded delays can be tolerated. Deterministic conflictfree solutions require schedule recomputation when topology changes, thus becoming unstable when the topology rate of change exceeds the schedule recomputation rate. The deterministic broadcast protocols we introduce in this paper overcome the above limitations by using a novel mobility-transparent schedule, thus providing a delivery (time) guarantee without the need to recompute the schedules when topology changes. We show that the proposed protocol is simple and easy to implement, and that it is optimal in networks in which assumptions on the maximum number of the neighbors of a node can be made.Index Terms-Ad hoc networks, broadcast protocols, distributed algorithms, mobile computing.

Reliable Broadcast in Wireless Mobile Ad Hoc Networks

“…Additionally, each round has R+3 slots out of which the first control slot has a duration of O(log n) and the others O (1). Hence, the overall duration of a round is in the order of O(D(log n + R + 2)) and finally O(D log n) for a constant R. The synchronous network model we use in this paper has been widely considered to analyze the complexity of the broadcasting problem [16], [17], [5], [18], [19].…”

“…They prove a lower bound of Ω(D log n), for D ≤ n/2. A broadcast protocol for this model was given by Basagni et al [17]. The total time required by this broadcast is O(DΔ log log Δ n) slots, where Δ is the maximum degree of the network.…”

“…Although these solutions claim to operate correctly even if topology changes due to mobility, this is true in a limited sense. For example, the solution described in [17] works correctly if mobility induced topology changes occur only at inter-frame boundaries, where a frame consists of a number of time-slots. Similarly, the solutions by Chlebus et al [18] do not explicitly take mobility into consideration.…”

“…Besides [17], the above algorithms consider networks where all the nodes are static. Our solution, on the other hand, allows a more flexible mobility pattern than [17]: mobility is no longer restricted to inter-round boundaries but permitted during the rounds themselves.…”

“…Besides [17], the above algorithms consider networks where all the nodes are static. Our solution, on the other hand, allows a more flexible mobility pattern than [17]: mobility is no longer restricted to inter-round boundaries but permitted during the rounds themselves. Our algorithm requires that a maximum of only 2 transmissions per node be executed at full transmission range R. The remaining O(log n) transmissions are executed at a very small range (r 1), leading to an overall energy-efficient scheme.…”