A Unified Flow Control Approach for QoS Balance in Differentiated Services Jiong Jin ∗ , Yee Wei Law ∗ , Marimuthu Palaniswami ∗ and Zhihong Man † ∗ Department of Electrical and Electronic Engineering, The University of Melbourne, Parkville, VIC 3010, Australia Email: {j.jin, y.law, swami}@ee.unimelb.edu.au † Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, VIC 3122, Australia Email: zman@swin.edu.au Abstract—Proportional, TCP friendly (minimum potential de- lay) and max-min fairness are three most commonly used fairness criteria for resource allocation in communication networks. In this paper, we generalize the above fairness criteria in terms of utility and study the resource allocation problem for hetero- geneous networks where contending users may have different Quality of Services (QoS) requirements and the utility functions may not necessarily satisfy the strict concavity condition, such as real-time applications. We propose a QoS based flow control algorithm and with different link price feedback mechanisms, utility weighted proportional, TCP friendly and max-min fairness is achieved in this unified approach. In addition, the new algorithm is not only suitable for elastic data traffic, but also capable of handling real-time applications, and therefore it can be treated as an efficient flow control mechanism to provide congestion control and QoS balance for Differentiated Services in the future Internet. Keywords — Network flow control, Quality of Services, real-time applications, utility fairness, Differentiated Services. I. INTRODUCTION Powered by error detection coding, acknowledgment (ACK) feedback and lost/corrupt packet retransmission, nowadays Internet has made a great success in providing fast, long- distance, low-cost and error free data transmissions for various customers all over the world. However, due to the significant delay and a lack of bandwidth reservation policy, currently “Best Effort” Internet is not able to support real-time ser- vices such as audio/video streaming efficiently. To meet this challenge, IETF adopted an architecture named “Differentiated Services” (Diff-Serv) [1] to support real-time traffic without disturbing the current IP structure. Most significantly, Diff- Serv replaces the first 6 bits (for all 8 bits potentially) in the IPv4 ToS (Type of Service) octet or the IPv6 Traffic Class octet with a Differentiated Services Code Point (DSCP). One suggestion on DSCP is for priority assignment, in which applications with strict Quality of Services (QoS) requirements are intuitively assigned with a higher priority in DSCP and hence receive a better and faster service. Another possible suggestion, which this paper fits into, is to treat DSCP as an explicit congestion feedback mechanism in order to provide a better solution for congestion control and resource allocation in the future Internet. Based on the explicit congestion feedback, many efforts have been devoted to develop a framework for designing Internet flow/congetion control protocols. With more detailed congestion information, users may deploy an accurate control scheme on their transmission rates, not only handling the congestion, but also addressing their own QoS requirements regarding the bandwidth. Meanwhile, advanced control tech- niques may be further enforced at each link to manage the buffer backlog. Thus, the network not only is able to offer low-loss, low-delay, QoS balanced services, but also have potential advantages to even support real-time applications without complicated admission control, resource reservation or packet scheduling mechanisms. In general, the aim of flow control is to allocate bandwidth resource optimally and fairly among competing users without incurring network congestion. There are three types of well- known fairness criteria in the networking literature, i.e., max- min fairness [2], proportional fairness [3] and TCP friendly fairness [4] which also has an alternative notion of “potential minimum delay fairness” [5]. However, fairness in bandwidth allocation is not sufficient to provide a good QoS balance for various network applications with different bandwidth requirements. Therefore, we will study the QoS based flow control problem and propose a new algorithm for congestion control and resource allocation in Differentiated Services. With the deployment of different congestion feedback mechanisms, proportional, TCP friendly and max-min fairness can be achieved respectively for QoS balance in a unified approach. This paper is organized as follows. Section II provides the preliminaries of Quality of Services and various bandwidth sharing policies. In section III, we give the motivation of this work and formulate the problem. In Section IV, we investigate the resource allocation policy and propose a unified algorithm to achieve QoS max-min fairness, QoS proportional fairness and TCP friendly fairness, respectively. Finally, we present the numerical result to evaluate the performance of our algorithm in Section V and draw conclusion in Section VI. II. PRELIMINARIES A. Quality of Services For a practical network application, the user may concern about the bandwidth allocation, but a more important and direct metric is actually its QoS performance. Quality of Services (QoS) provide a measure of the application’s perfor- mance given certain network conditions such as bandwidth, transmission delay and loss ratio. For the interest of network 978-1-4244-6404-3/10/$26.00 ©2010 IEEE This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2010 proceedings Authorized licensed use limited to: UNIVERSITY OF MELBOURNE. Downloaded on July 24,2010 at 10:48:47 UTC from IEEE Xplore. Restrictions apply.