Elsevier

Computer Communications

Volume 27, Issue 10, 20 June 2004, Pages 1012-1024
Computer Communications

LS-SCTP: a bandwidth aggregation technique for stream control transmission protocol

https://doi.org/10.1016/j.comcom.2004.01.021Get rights and content

Abstract

Stream Control Transmission Protocol (SCTP) specifications utilize the multiple paths capabilities between the sender and receiver for retransmission of lost data chunks and as a backup in case of primary path failure. Under normal conditions, all data chunks are sent on the primary path chosen by the SCTP user during the transport connection initiation. In this paper, we address in detail various aspects related to extending and engineering SCTP in order to utilize the available paths for simultaneous transmission of data chunks, while maintaining the SCTP congestion control on each path so as to ensure fair integration with other traffic in the network. The extended SCTP, referred to as Load-Sharing SCTP (LS-SCTP), is able to aggregate the bandwidth of all the active transmission paths between the communicating endpoints. LS-SCTP monitors the paths, and accordingly it chooses the paths that are suitable for load sharing. LS-SCTP retransmission mechanism accelerates the delivery of missing data to the receiver in order to prevent stalling the transport connection while waiting for missing data chunks. Simulation results show that LS-SCTP is extremely beneficial for networks with limited bandwidth, high loss rate and failure prone links.

Introduction

For many years SS7 has been the dominant bearer of signaling traffic for telecommunication networks, but recently many proprietary solutions for transporting signaling traffic over IP have appeared. This approach promises tighter integration with Voice over IP (VoIP) solutions and ultimately the possibility of a common core network capable to transport signaling and media traffic. Driven by industry interest and general agreement on the unsuitability of TCP and UDP for signaling transport, the IETF Signaling Transport (SIGTRAN) group was formed in 1999 to standardize a suitable transport protocol for signaling traffic over IP. Stream control transmission protocol (SCTP) was the result of this work, and it was recently published as RFC 2960 [1] by the Internet society.

SCTP is a reliable, message-oriented data transport protocol that supports multiple streams within a single transport layer connection, an ‘association’ in SCTP terminology, and hosts with multiple network interfaces (multi-homed hosts). These properties make SCTP more suitable for signaling transport, as well as for providing transport benefits to other applications requiring additional performance and reliability. SCTP features will be described in detail in Section 3.

SCTP support for multi-homed hosts is intended to provide communication reliability for the hosts engaged in the association. Initially, two interfaces, one at each host, are chosen to form the primary path that is used for transmission of the data units, ‘data chunks’ in SCTP terminology. The other interfaces, which form the secondary paths, are only used for retransmission of lost data chunks or as a backup for the primary path. This means that although these paths exist, they are only utilized for retransmission or for failure recovery. In this paper, we propose extending SCTP to utilize the available paths for simultaneous transmission of data chunks, i.e. load sharing, while maintaining the SCTP congestion control on each path, in order to ensure fair integration with other traffic in the network. We believe that this form of bandwidth aggregation is extremely beneficial for networks with limited bandwidth and high loss rates. In order to aggregate the available bandwidth, taking in account the differences in the characteristics of the paths, in terms of bandwidth, latency and loss rates, we propose a separation between the association congestion control and flow control. In our SCTP extension, which we refer to as Load Sharing-SCTP (LS-SCTP), the congestion control is performed on a path basis, while the flow control is on association basis. Standard SCTP does not separate between the flow and the congestion control, as both mechanisms work together on a single path. As the failure of a transmission path or the increase in the path loss rate can affect the throughput of the whole association, LS-SCTP monitors the paths, and accordingly it dynamically determines the paths that are suitable for load sharing. In addition, LS-SCTP retransmission mechanism accelerates the delivery of missing data to the receiver in order to prevent stalling the association, while the receiver is waiting for a missing data chunk. As will be shown in our performance study, in Section 5, that these features make LS-SCTP robust to the variations in the characteristics of the transmission paths.

During the association initialization, LS-SCTP allows the user to enable/disable the load sharing capability, as well as controlling the number of interfaces that can be used simultaneously for load sharing. This feature is important for battery-powered hosts, as it enables the LS-SCTP user to conserve the battery power by controlling the simultaneous use of the interfaces.

This paper is organized as follows. Section 2 provides a review for similar works in bandwidth aggregation. Section 3 presents a quick overview on SCTP features and the differences between SCTP and TCP. Section 4 describes in detail LS-SCTP design. Section 5 provides a performance study for LS-SCTP. Finally, Section 6 concludes the paper.

Section snippets

Related work

The idea of resource aggregation in order to obtain higher performance has been used in different areas in the computer and communication fields. Ref. [2] introduces a striping technique for the disk subsystem, which is now a key aspect in Redundant arrays of inexpensive disks (RAID) architectures. In Ref. [3], authors provide an overview on the use of resource aggregation in the network subsystem, and introduce an evaluation criteria to judge the benefits provided from the resource aggregation

SCTP overview

SCTP is the fundamental member of a family of protocols designed by the SIGTRAN group to allow SS7 messages to be transported over an unreliable IP infrastructure [12], [13]. In SCTP, data transfer between two hosts takes place in the context of an association, as shown in Fig. 1.

All data transferred between the hosts is encapsulated in SCTP packets. SCTP packet contains a common header and a sequence of structures called ‘chunks’. The common header has source and destination port numbers to

LS-SCTP design

In this section we present our extension for SCTP (LS-SCTP) to support aggregating the available bandwidth of the transmission paths between the transport endpoints. First, we will discuss the reasons that make SCTP, in its current form, not suitable for load sharing. Then we present solutions to make SCTP load sharing capable.

Our motivations to extend SCTP for load sharing are:

  • (i)

    The increase in the number of multi-homed devices. For instance, laptops commonly come with a built in infrared and

Performance study

In order to examine the performance of LS-SCTP, we extended our SCTP implementation, in OPNET network simulation software [25]. In our simulation, we used the network topology shown in Fig. 11. We assume that an SCTP association is already initiated between the two hosts, and the association is unidirectional, which means that data chunks will only be sent from the host A (sender) to host B (receiver). In addition, we use only one stream within the association.

Nodes 1-N are used to configure

Conclusions and future research

In this paper, we propose an extension for SCTP, LS-SCTP, which aggregates the available bandwidth on the paths between the sender and receiver. We first showed that SCTP in its current form is not suitable for load sharing. Then, we suggested remedies for SCTP to make it load sharing capable. We proposed separating the association flow control and congestion control. The congestion control is performed per path while the flow control is performed per association. Also we proposed modifying the

Disclaimer

The views and conclusions in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research laboratory or the US Government.

References (25)

  • R. Stewart et al.

    Stream Control Transmission Protocol

    IETF

    (2000)
  • R. Katz et al.

    Disk system architecture for high performance computing

    Proceedings of the IEEE

    (1989)
  • C. Brenden et al.

    Striping Within the Network Subsystem

    IEEE Network

    (1995)
  • Computer Staff

    Gigabit network testbeds

    IEEE Computer

    (1990)
  • J. Duncanson

    Inverse Multiplexing

    IEEE Communications Magazine

    (1994)
  • A. Snoeren

    Adaptive inverse multiplexing for wide-area wireless networks

    Proceedings of the IEEE GLOBECOM’99

    (1995)
  • L. Magalhaes et al.

    Transport level mechanisms for bandwidth aggregation on mobile hosts

    Proceedings of the IEEE ICNP’01

    (2001)
  • J. Lee et al.

    Applied techniques for high bandwidth data transfers across wide area networks

    Proceedings of the Computers in High Energy Physics CHEP’01

    (2001)
  • M. Allman et al.

    An application-level solution to TCP's satellite inefficiencies

    Proceedings of the Workshop on Satellite-Based Information Services WOSBIS’96

    (1996)
  • H. Sivakumar et al.

    PSockets: the case for application-level network striping for data intensive applications using high speed wide area networks

    Proceedings of the IEEE Supercomputing SC’00

    (2000)
  • T. Hacker et al.

    The end-to-end performance effects of parallel tcp sockets on a lossy wide-area network

    Proceedings of the IEEE IPDPS’02

    (2002)
  • L. Ong et al.

    Framework Architecture for Signaling Transport

    IETF

    (1999)
  • Cited by (0)

    Prepared through collaborative participation in the Communications and Networks Consortium sponsored by US Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD19-01-2-0011.

    View full text