PRISM: Improving the Performance of Inverse-Multiplexed TCP in Wireless Networks

  • Authors:
  • Kyu-Han Kim;Kang G. Shin

  • Affiliations:
  • -;-

  • Venue:
  • IEEE Transactions on Mobile Computing
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Multi-homed mobile hosts in physical proximity may spontaneously team up to form a community and run high-bandwidth applications by pooling their low wireless widearea network (WWAN) bandwidths together for communication with a remote application server. Utilizing their high-bandwidth wireless local-area network (WLAN), the thus-teamed mobile hosts can aggregate and distribute the application content among themselves. This paper first justifies the need for such a mobile collaborative community (MC2), or a community, to improve user-perceived network bandwidth and utilization. Then, existing one-to-one communication protocols like TCP are shown to suffer significant performance degradation due to frequent out-of-order packet deliveries. To address this TCP problem, we propose a proxy-based inverse multiplexer, called PRISM, that enables TCP to efficiently utilize the community members' WWAN connections while avoiding the performance degradation. PRISM runs at the proxy's network layer as a routing component and stripes each TCP flow over multiple WWAN links by exploiting the transport-layer feedback information. Moreover, it masks a variety of adverse effects specific to each WWAN link via an intelligent ACK-control mechanism. Finally, PRISM enables TCP to respond correctly to dynamically-changing network states through a sender-side enhancement of congestion control. PRISM has been evaluated with experimentation on a testbed as well ns-2-based simulation. Our experimental evaluation has shown PRISM to improve TCP's performance by up to 310% even with two collaborative mobile hosts. Our in-depth simulation study has also shown that PRISM delivers a near-optimal aggregated bandwidth in the community, and improves network utilization significantly