Article Preview
Top1. Introduction
Nowadays, the Internet must support various multimedia applications with different quality of service (QoS) requirements. Widely differing types of services such as real-time communications, interactive communications, and streaming media pose their own QoS requirements including bounded end-to-end delay and jitter, and high bandwidth (Kanellopoulos et al., 2009). For example, video streaming and voice over Internet protocol (VoIP) service require bounded end-to-end delay and jitter. On the other hand, reliability is a vital requirement for data transfer applications. The current Internet architecture is designed around a best-effort service paradigm, which does not support guaranteed delivery inherently. Meanwhile, QoS requirements can be addressed in different OSI-RM layers ranging from the Application to Physical layers in order to improve the perceived media quality by end-users over IP networks (Kanellopoulos et al., 2009). Therefore, delivering multimedia over IP networks, such as the Internet, is a formidable task.
In this paper, we focus on the QoS provided by the transport layer protocols over wired IP networks (Iren et al. 1999). The majority of the Internet applications uses two well-known transport layer protocols: (1) the Transmission Control Protocol (TCP) (Postel, 1981); and (2) the User Datagram Protocol (UDP) (Postel, 1980). TCP provides a reliable in-order delivery of the data, while UDP provides faster delivery of packets as compared to TCP. However, UDP does not provide any congestion control mechanism (Kumar & Rai, 2012). Real-time multimedia applications need an enhanced transport protocol that should reliably handle packet loss, minimize errors, manage network congestion and transmit efficiently. Recently, new enhanced transport protocols have been designed such as Data Congestion Control Protocol (DCCP) (Lai, 2008), Stream Control Transmission Protocol (SCTP) (Stewart, 2007), and TCP-Friendly Rate Control (TFRC) (Handley et al., 2008). These protocols are still under research whether they can be used for real-time applications practically. Actually, the choice of what transport protocol will be used depends on the QoS requirements imposed by the application on hand. Additionally, the advent of heterogeneous networks magnifies the volatility of network conditions and imposes greater challenges for multimedia delivery. However, future IP networks are expected to consist of several, potentially incompatible, access technologies that would be offered by several competing service providers. And the diversity of the transport protocols may drastically affect the QoS provided for multimedia services.
The aim of this paper is to evaluate the performance of four transport layer protocols (i.e., UDP, DCCP, SCTP, and TFRC) in a wired network environment under three different scenarios/traffic flows: (1) data traffic; (2) video streaming; and (3) VoIP traffic. We selected these traffic flows because video streaming applications are very popular in IP networks, while VoIP has been a prevalent multimedia service for the delivery of voice communications and multimedia sessions over IP networks. Through simulations, we analyze the performance of UDP, DCCP, SCTP, and TFRC protocols in order to answer in which scenario each of these transport protocols functions better than the others.
The rest of this paper is organized as follows. In Section 2, we present in brief the transport layer protocols UDP, DCCP, SCTP and TFRC. We also outline related studies that focus on the performance of these transport layer protocols. In Section 3, we discuss critical issues on video streaming and VoIP service. In Section 4, we present our simulation experiments and the simulation environment, while in Section 5 we present and analyze the simulation results. In Section 6, we conclude the paper and give directions for further work.