Delay Constraint Multicast Routing for Wireless Ad Hoc Networks
Klangjai Sithitavorn, Bin Qiu · 2006
In this paper, we discuss the issue of delay- throughput tradeoff from providing end-to-end delay guarantee multicast service in wireless ad hoc network. Efforts are made to develop a Delay Constraint Multicast (DCM) routing protocol, focused on constructing a delay based multicast tree, which is able to provide soft end-to-end delay guarantees for all multicast receivers. Forwarding decisions are made based on local access delay and QoS information from a source node, resulting in a high ratio of timely delivered packets. Based on simulation results, DCM is able to deliver data packets with average latency right below the delay constraint while keeping communication overhead at minimum. I. INTRODUCTION Multicast in Mobile Ad-hoc NETworks (MANETs) has recently emerged as a premier research topic due to several attractive features of MANETs and the increase in avail- ability of many multicast-based mobile applications such as video/audio streaming. However, designing a QoS multicast routing protocol is not an easy task due to the path-sharing nature of multicast and additional challenges of MANETs. Contrasting to unicast routing where searching and reserving a QoS path can be done per receiver, constructing a multicast tree with QoS paths that can deliver guaranteed QoS from a source node to all receivers is somewhat more complicated. In multicast routing, packets are multiplexed when it is necessary to reach two or more receivers on disjoint paths. Forwarding decisions, therefore, are made in a distributed manner. Since multicast packets are meant for all multicast receivers not per destination, forwarding nodes are mostly used to apply a QoS path assignment. QoS support for multimedia application is closely related to topologies of ad hoc networks (1). In particular to a delay requirement, QoS guarantee is meaningless if nodes are too mobile or located too far from the multicast source. Without an option of adjusting transmitting power (adjustable transmission ranges) or antenna directions, topology control to enable QoS guarantees is almost impossible. Since a QoS multicast tree construction is not only based on connectivity but the delay requirement, it also has a negative impact on multicast throughput (2)(3). In this paper, we address the issue of whether multicast routing protocol can obtain a soft end-to-end delay guarantee without trading off much throughput. Efforts are made to develop a Delay Constraint Multicast (DCM) routing protocol, focused on constructing a delay based multicast tree which is able to provide soft end-to-end delay guarantees for all multicast receivers. In DCM, forwarding decisions are made based on an estimation of MAC layer access delay and QoS information from a source node. Given the throughput tradeoff, we introduce a topology-transparency approach providing the option of receiving delay-guarantee with the cost of changing node positions. We however stress that such change in node position is based on an assumption that the multicast receivers/users are willing to cooperate and move to suggested positions to be able to obtain the required end-to-end delay guarantee. We also emphasize that DCM only supports soft QoS instead of hard guarantees and it is focused on the case in which nodes are stationary, such as a conference venue or a class room. Details on assumptions used are discussed in section II. This paper is organized as follows. The motivation and background on the access method, simulation assumptions as well as related work are provided in section II. The delay constraint multicast routing protocol is introduced in section III which is followed by its evaluation and result analysis in section IV and the conclusion is given in Section V.