Simulation-Based Comparison of TCP and TCP-Friendly Protocols

Simulation-Based Comparison of TCP and TCP-Friendly Protocols

Gábor Hosszú (Budapest University of Technology and Economics, Hungary)
DOI: 10.4018/978-1-60566-014-1.ch176
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Internet streaming media changed the Web from a static medium into a multimedia platform, which supports audio and video content delivery. In our days streaming media turns into the standard way of global media broadcasting and distribution. The low costs, worldwide accessibility, and technical simplicity of this telecommunication way make media streams very attractive for content providers. Streaming works by cutting the compressed media content into packets, which are sent to the receiver. Packets are reassembled and decompressed on the receiver side into a format that can be played by the user. To achieve smooth playback, packets are buffered on the receiver side. However, in case of a network congestion, the stream of packets slows down, and the player application runs out of data, which results in poor playback quality. This article presents the comparison of different transport level congestion control schemes, including variants of the TCP. The protocol mechanisms, implemented in various protocols, are hard to investigate in a uniform manner; therefore, the simulator SimCast (Simulator for multiCast) is developed for traffic analysis of the unicast and multicast streams. In this article the TCP and other transport protocol mechanisms will be compared using the SimCast simulator (Orosz & Tegze, 2001). The simulated results are presented through examples. Due to spreading of traffic lacking end-to-end congestion control, congestion collapse may arise in the Internet (Floyd & Fall, 1999). This form of congestion collapse is caused by congested links that are sending packets to be dropped only later in the network. The essential factor behind this form of congestion collapse is the absence of end-to-end feedback. On the one hand an unresponsive flow fails to reduce its offered load at a router in response to an increased packet drop rate, and on the other hand a disproportionate-bandwidth flow uses considerably more bandwidth than other flows in time of congestion. In order to achieve accurate multicast traffic simulation—being not so TCP-friendly yet—the effects of the flow control of the TCP protocol should be determined (Postel, 1981). However, there are many different kinds of TCP and other unicast transport protocol implementations with various flow control mechanisms, which make this investigation rather difficult (He, Vicat-Blanc Primet, & Welzl, 2005).
Chapter Preview
Top

Introduction

Internet streaming media changed the Web from a static medium into a multimedia platform, which supports audio and video content delivery. In our days streaming media turns into the standard way of global media broadcasting and distribution. The low costs, worldwide accessibility, and technical simplicity of this telecommunication way make media streams very attractive for content providers.

Streaming works by cutting the compressed media content into packets, which are sent to the receiver. Packets are reassembled and decompressed on the receiver side into a format that can be played by the user. To achieve smooth playback, packets are buffered on the receiver side. However, in case of a network congestion, the stream of packets slows down, and the player application runs out of data, which results in poor playback quality.

This article presents the comparison of different transport level congestion control schemes, including variants of the TCP. The protocol mechanisms, implemented in various protocols, are hard to investigate in a uniform manner; therefore, the simulator SimCast (Simulator for multiCast) is developed for traffic analysis of the unicast and multicast streams. In this article the TCP and other transport protocol mechanisms will be compared using the SimCast simulator (Orosz & Tegze, 2001). The simulated results are presented through examples.

Due to spreading of traffic lacking end-to-end congestion control, congestion collapse may arise in the Internet (Floyd & Fall, 1999). This form of congestion collapse is caused by congested links that are sending packets to be dropped only later in the network. The essential factor behind this form of congestion collapse is the absence of end-to-end feedback. On the one hand an unresponsive flow fails to reduce its offered load at a router in response to an increased packet drop rate, and on the other hand a disproportionate-bandwidth flow uses considerably more bandwidth than other flows in time of congestion. In order to achieve accurate multicast traffic simulation—being not so TCP-friendly yet—the effects of the flow control of the TCP protocol should be determined (Postel, 1981). However, there are many different kinds of TCP and other unicast transport protocol implementations with various flow control mechanisms, which make this investigation rather difficult (He, Vicat-Blanc Primet, & Welzl, 2005).

Up to now a lot of comparisons have been done. For example, Wang et al. (2001) reviewed the TCP-friendly congestion control schemes in the Internet. They differentiated two groups of the TCP-friendly congestion control algorithms as follows: (1) end-to-end and (2) hop-by-hop congestion control mechanisms. The end-to-end mechanisms are grouped into (1) AIMD-based schemes (AIMD: additive increase multiplicative decrease) with the window- and rate-adaptation schemes, (2) modeling-based schemes, including equation-based congestion control schemes and the so-called model-based congestion schemes, and (3) a combination of AIMD-based and modeling-based mechanism. Wang’s classification is mostly used in our discussion, too.

In this article various TCP congestion control mechanisms as well as congestion control mechanisms for media streams are reviewed. Then a novel simulator for transport protocols is described and the various simulation results summarized. Lastly, conclusions are drawn and work to be done identified.

Key Terms in this Chapter

Virtual private network (VPN): Virtual private communications network established over a public infrastructure through an authentication mechanism, for the secure exchange of information between two entities.

Next Generation Network (NGN): Telecommunications packet-based network that handles multiple types of traffic and decouples services from transport details.

Fixed Mobile Convergence (FMC): The integration of mobile and fixed technologies to enable seamless distribution of services over fixed and mobile broadband networks.

Public Switched Telephone Network (PSTN): Voice oriented international network for circuit switched telephony.

Internet Network (IN): Architecture for fixed and mobile networks that allows provision of value added services in addition to standard telecommunication services, without having to redesign switching equipment.

Generalized mobility: The ability of a mobile user to communicate and access services, irrespective of changes in the location or technical environment, with or without service continuity.

IP Multimedia Subsystem (IMS): Open-systems architecture that supports a range of IP-based services over the packet switched domain, employing both wireless and fixed access technologies.

Internet Protocol (IP): Network layer protocol related to the communication of packet switched data.

Complete Chapter List

Search this Book:
Reset