Abstract:Abstract-Networks need to accommodate diverse applications with different Quality-of-Service (QoS) requirements. New ideas at the physical layer are being developed for this purpose, such as diversity embedded coding, which is a technique that combines high rates with high reliability. We address the problem of how to fully utilize different rate-reliability characteristics at the physical layer to support different types of traffic over a network and to jointly maximize their utilities.We set up a new framewo… Show more
“…Consider the requirement of the average queuing delay on a link to be less than a specific amount, i.e., E(T ) ≤ d , where d is the local average delay allowed on link and T is the queuing delay on link . Let us assume a queuing delay model based on the assumption of general packet length distribution (with mean 1/μ and variance σ 2 on every link) and each link modeled as an M/G/1 queue [7,12,22,23]. Therefore the average queuing delay on each link is equal to (see [7] or [22])…”
Section: System Model and Optimization Problemmentioning
confidence: 99%
“…In [10], a fixed symbol rate is assumed and different bit rates are achieved by choosing the transmitted symbols from the appropriate signal constellation (adaptive modulation); and in [11], several extensions of the NUM problem including queuing delay are outlined. In [12], the authors incorporate the delay in addition to rate and reliability in the NUM problem. However, they assume fixed capacity links that consist of sub-links with different ratereliability characteristics.…”
Allocating limited resources such as bandwidth and power in a multi-hop wireless network can be formulated as a Network Utility Maximization (NUM) problem. In this approach, both transmitting source nodes and relaying link nodes exchange information allowing for the NUM problem to be solved in an iterative distributed manner. Some previous NUM formulations of wireless network problems have considered the parameters of data rate, reliability, and transmitter powers either in the source utility function which measures an application's performance or as constraints. However, delay is also an important factor in the performance of many applications. In this paper, we consider an additional constraint based on the average queueing delay requirements of the sources. In particular, we examine an augmented NUM formulation in which rate and power control in a wireless network are balanced to achieve bounded average queueing delays for sources. With the additional delay constraints, the augmented NUM problem is non-convex. Therefore, we present a change of variable to transform the problem to a convex problem and we develop a solution which results in a distributed rate and power control algorithm tailored to achieving bounded average queueing delays. Simulation results demonstrate the efficacy of the distributed algorithm.
“…Consider the requirement of the average queuing delay on a link to be less than a specific amount, i.e., E(T ) ≤ d , where d is the local average delay allowed on link and T is the queuing delay on link . Let us assume a queuing delay model based on the assumption of general packet length distribution (with mean 1/μ and variance σ 2 on every link) and each link modeled as an M/G/1 queue [7,12,22,23]. Therefore the average queuing delay on each link is equal to (see [7] or [22])…”
Section: System Model and Optimization Problemmentioning
confidence: 99%
“…In [10], a fixed symbol rate is assumed and different bit rates are achieved by choosing the transmitted symbols from the appropriate signal constellation (adaptive modulation); and in [11], several extensions of the NUM problem including queuing delay are outlined. In [12], the authors incorporate the delay in addition to rate and reliability in the NUM problem. However, they assume fixed capacity links that consist of sub-links with different ratereliability characteristics.…”
Allocating limited resources such as bandwidth and power in a multi-hop wireless network can be formulated as a Network Utility Maximization (NUM) problem. In this approach, both transmitting source nodes and relaying link nodes exchange information allowing for the NUM problem to be solved in an iterative distributed manner. Some previous NUM formulations of wireless network problems have considered the parameters of data rate, reliability, and transmitter powers either in the source utility function which measures an application's performance or as constraints. However, delay is also an important factor in the performance of many applications. In this paper, we consider an additional constraint based on the average queueing delay requirements of the sources. In particular, we examine an augmented NUM formulation in which rate and power control in a wireless network are balanced to achieve bounded average queueing delays for sources. With the additional delay constraints, the augmented NUM problem is non-convex. Therefore, we present a change of variable to transform the problem to a convex problem and we develop a solution which results in a distributed rate and power control algorithm tailored to achieving bounded average queueing delays. Simulation results demonstrate the efficacy of the distributed algorithm.
“…There are many applications of NUM in network engineering, such as optimal network resource allocation, Internet congestion control and protocol stack design [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. For example, in [2], the authors show that even with simple solutions that can only provide suboptimal utility maximization, one can enhance network throughput and reduce link saturation.…”
Section: Introductionmentioning
confidence: 99%
“…Congestion control in a network with delaysensitive traffic is studied in [4], which is modeled by adding explicit delay terms to the utility function measuring user's satisfaction of the Quality of Service (QoS). A new framework based on utility maximization for networks with composite links is set up in [5]. An optimal scheduling algorithm to maximize network utility for mesh network is developed in [6], and joint end-to-end congestion and contention control method based on network utility maximization for ad hoc networks is proposed in [7].…”
Section: Introductionmentioning
confidence: 99%
“…In this paper, we study the network resource allocation problem by extending the basic network utility maximization (NUM) framework developed in [5]. The basic NUM problem can be formulated as in (1).…”
Ideally, networks should be designed to accommodate a variety of different traffic types, while at the same time maximizing its efficiency and utility. Network utility maximization (NUM) serves as an effective approach for solving the problem of network resource allocation (NRA) in network analysis and design. In existing literature, the NUM model has been used to achieve optimal network resource allocation such that the network utility is maximized. This is important, since network resources are at premium with the exponential increase in Internet traffic. However, most research work considering network resource allocation does not take into consideration key issues, such as routing and delay. A good routing policy is the key to efficient network utility, and without considering the delay requirements of the different traffic, the network will fail to meet the user's quality of service (QoS) constraints. In this paper, we propose a new NUM framework that achieves improved network utility while taking into routing and delay requirements of the traffic. Then, we propose a decomposition technique-based algorithm, D-NUM, for solving this model efficiently. We compare our approach with existing approaches via simulations and show that our approach performs well.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.