Improving TCP performance in heterogeneous mobile environments by exploiting the explicit cooperation between server and mobile host

Abstract

In recent years, many different kinds of wireless access networks have been deployed and become inseparable parts of the Internet. But TCP, the most widely used transport protocol of the Internet, was designed for stationery hosts. It faces severe challenges when user moves around in these networks and handoff occurs frequently. In this paper, we investigate the potential benefits of bringing explicit cooperation between TCP server and mobile host.

For this purpose, TCP HandOff (TCP-HO), a practical end-to-end mechanism, is designed for improving TCP performance in heterogeneous mobile environments. TCP-HO assumes that a mobile host is able to detect the completion of handoff immediately and has a coarse estimation of new wireless link’s bandwidth. When a mobile host detects handoff completion, it will immediately notify the server through two duplicate ACKs, whose TCP option also carries the bandwidth of new wireless link. After receiving this notification, the server begins to transmit immediately and keeps updating ssthresh according to the bandwidth from mobile host and its new RTT samples. This update will end after four RTT samples or after congestion is detected.

TCP-HO has been implemented in FreeBSD 5.4. Experimental results indicate that in heterogeneous mobile environments, TCP-HO can improve TCP performance a lot without adversely affecting cross traffic even when mobile host only has a coarse estimation of new wireless link’s bandwidth. Considering that more and more users are accessing the Internet through heterogeneous wireless networks and mobile host could have a coarse estimation of wireless link’s bandwidth, it should be worthwhile to change both server and mobile host for improving TCP performance.

Introduction

In recent years, many kinds of wireless networks, such as cellular network (WCDMA [1], GPRS [3], etc.), Wireless LAN (Wi-Fi [4], [5], [6], [7], HiperLAN [8], etc.), and Wireless MAN (WiMax [9]), have been deployed and have become integral parts of the Internet. These wireless networks complement each other in terms of coverage, bandwidth, latency, etc. and form a heterogeneous mobile environment. These networks along with portable and affordable computing devices, such as laptops, PDAs, and smart phones, enable the wide spread and affordable mobile Internet access. But the lossy wireless links and mobile hosts violate the assumptions of TCP [15], the most widely used transport protocol of the Internet. And TCP performs very bad when users move around in these wireless networks.

TCP was designed for reliable links. When packets are transmitted on lossy wireless link, they are corrupted frequently. These corrupted packets are wrongly regarded as congestion signals and TCP sender reduces its sending rate unnecessarily. Hence, TCP can not fully utilize the precious bandwidth of wireless link and provide very bad performance to applications. Many mechanisms, such as I-TCP [28], Snoop [17], TCP Veno [18], TCP Hack [19], and TCP ELN [20], [21], [22], had been proposed for enhancing TCP performance over lossy wireless links. Based on the above investigations, new wireless networks normally adopt FEC (forward error correction) and/or ARQ (automatic retransmission request) in link layer with the aim of hiding lossy characteristic of wireless link. For example, RLC [2], the link layer protocol of GPRS/UMTS, adopts link layer ARQ. Link layer ARQ is also used by the MAC layer of IEEE 802.11 [4]. Hence, the current wireless networks have already been reliable wireless networks.

In reliable wireless networks, link layer ARQ may bring large bandwidth and delay variation. This problem has been solved by regulating ACK and TCP window field at base station [23], [24].

The challenges brought by lossy wireless link had been well studied. But TCP was also designed for stationery hosts. When a user moves around in these heterogeneous wireless networks, TCP performance is very poor due to the challenges brought by handoff. In this paper, we will investigate how to enhance TCP performance when users move around in the current heterogeneous mobile environments.

Within the current heterogeneous mobile environments, many kinds of handoff may occur and they can be classified in many different ways. For example, in [25], handoff is classified into vertical handoff and horizontal handoff according to the techniques used by the wireless networks. In this paper, we classify handoff from TCP point of view. More specifically, since the BDP (bandwidth-delay product) of a network path affects TCP performance significantly, we classify handoff according to BDP of old network path (before handoff) and that of new network path (after handoff).

Handoff is first classified into HH (horizontal handoff) and VH (vertical handoff) based on whether BDP changes significantly during handoff. Within HH, BDP of new and old paths do not vary appreciably. For VH, the difference in BDP is large. According to the value of BDP, HH is further divided into L-HH and H-HH. Within L-HH, BDPs of both paths are low. Within H-HH, BDPs of both paths are high. According to the direction of BDP change, VH is further divided into D-VH (downward-VH) and U-VH (upward-VH) [26]. During all kinds of handoff, there is normally a long disconnection time, which means a period of zero BDP. Fig. 1 shows BDP fingerprint of our four kinds of handoff.

Although TCP reacts to changes of network capacity through congestion control, abrupt BDP changes during handoff can still bring severe challenges. In this paper, we study the potential benefits of bringing explicit cooperation between server and mobile host. TCP-HO, a practical end-to-end mechanism based on explicit cooperation, is proposed for solving the challenges brought by all kinds of handoff with the aim of improving TCP performance in heterogeneous mobile environments.

The rest of this paper is organized as follows. In Section 2, the challenges faced by TCP during each kind of handoff are analyzed. Section 3 presents related work, and the details of our proposal, TCP handoff, are presented in Section 4. The testbed and experimental results are described in Section 5, and the issues related with wireless link bandwidth estimation are discussed in Section 6. This paper is concluded in Section 7.