Cross-Layer Quality of Service Optimization in Satellite Networks

Franco Davoli
Department of Communications, Computer and Systems Science (DIST), University of Genoa, Italy

Abstract

Resource allocation is one of the main tasks of a telecommunication network, where different users and services must share a pool of common resources. In wireless networks, where bandwidth may be relatively scarce with respect to cabled networks and environmental conditions may affect channel quality, the dynamic control of allocated resources becomes a challenge. In particular, satellite systems not only have to face variable load multimedia traffic, but also variable channel conditions and large propagation delays. It is therefore crucial to make use of adaptive network management and control algorithms to maintain the Quality of Service (QoS) of the transmitted data. The combined action among various layers of the network (from the physical layer up to the application layer) might improve the performance of the overall system, through a coordination of control actions, QoS mapping and cross-layer information exchange for control purposes.

In order to obtain optimized policies for satellite bandwidth allocation, we coordinate the actions taken in a satellite network at the physical layer (where the fade countermeasure technique is applied) with the work that is done at the data link layer (where the satellite bandwidth is allocated), thus obtaining a cross-layer optimization. The operating scenarios that we consider may be divided into two categories: one assumes the presence of both guaranteed performance (inelastic, or Continuous Bit Rate, CBR) and best-effort (BE) traffic, without explicitly characterizing the elastic behavior of the latter (as induced by TCP congestion control mechanisms); the other considers only best-effort services, by explicitly modeling the effects of TCP congestion control.

More specifically, after introducing the fade countermeasure and the traffic models adopted, we first treat the case of mixed inelastic and BE traffic. The second part of the talk is devoted to the analysis of long-lived TCP connections, experiencing diverse fading conditions, and to their fair bandwidth allocation, performed by taking into account the tradeoff between Bit Error Rate (BER) and Information Bit Rate (IBR) upon TCP goodput. We present and discuss some numerical results, based both on analytical modeling and discrete event simulation.

Back to Conference main

Abstract

Franco Davoli CV