Overview:

The broad scope of the project is to design a reliable transport protocol over multi-hop wireless networks. The various logical phases leading to the development of such a protocol have been identified as:

1. Performance analysis of the de-facto transport protocol (TCP).
2. Enhance the performance of TCP
   1. Investigate factors in the MAC, routing and transport layers that could cause throughput degradation. The various metrics/parameters used for this investigation are random loss probability, path length, route/link failures, route computation time, link failure detection time and reaction of transport layer to failures which includes packet losses and changes to traffic parameters of transport layer protocol.
   2. Propose a suite of mechanisms at the MAC and routing layers called the ATRA framework to enhance TCP performance, based on the above investigation.
3. Design a reliable transport protocol called ATP (Ad-hoc Transport Protocol) using the insights gained from the first two phases. While phase II aims to improve the transport protocol without actually making any changes to it (the mechanisms proposed are only at MAC and routing layers), this phase is concerned with the design of a new transport protocol itself. Though phases II and III have been proposed as independant solutions they can work in tandem.

ATRA:

Results

• Throughput results for 1 connection
  

  Figure 1: Throughput for 1 Connection

  Figure 1 presents throughput results for 1 connection. It can be observed that the Atra framework and its several variations outperform the default protocol stack and ELFN by a large margin. Specifically, Atra exhibits throughput improvement of around 100% over the default case is seen at a mobility rate of 20 m/s. The improvement is around 70% when compared to the ELFN approach.
• Throughput results for 25 connections
  

  Figure 2: Throughput for 25 Connections

  Figure 2 presents results for 25 connections. It compares the Atra framework with the other schemes when the load in the network is increased to 25 connections. The improvement in throughput performance over the default scenario is around 50%. The reduction in performance improvement can be attributed to the fact that other flows can occupy the bandwidth given up by a suffering flow.
• TCP loss percentage
  

  Figure 3: Loss for 1 Connection

  Figure 4: Loss for 25 Connections

  Figures 3 & 4 present the loss percentage for 1 and 25 connections for all the schemes. The loss percentage for Atra is lowered by a factor of 75% when compared to ELFN in the 25 connections case.

ATP:

Overview

In this project we address the problem of reliable transport in mulit-hop wireless networks. Existing works have approached the problem of reliable transport in ad-hoc networks by proposing mechanisms to improve TCP's performance over such networks. We take the stand point that several of the design elements in TCP are fundamentally inappropriate for the unique characteristics of ad-hoc networks. Given that ad-hoc networks are typically stand-alone, we approach the problem of reliable transport from the perspective that it is justifiable to develop an entirely new transport protocol that is not a variant of TCP. Toward this end, we propose a new reliable transport layer protocol for ad-hoc networks called ATP (ad-hoc transport protocol). The design elements of ATP are such that, they address each of the issues identified with TCP in the target environment.

Factors that highlight the inappropriateness of TCP:

• Window based transmissions
  
• Slow start
  
• Loss based congestion indication
  
• Linear increase multiplicative decrease
  
• Dependece on ACKs
  

The design elements in ATP are:

• Layer co-ordination
  
• Rate based transmissions
  
• Decoupling of congestion control and reliability
  
• Assisted congestion control
  
• Fairness
  

Results

Figure 5(a): Throughput for 1 Connection

Figure 5(b): Throughput for 5 Connections

Figure 5(c): Throughput for 25 Connections

Figure 6(a): Fairness for 5 Connections

Figure 6(b): Fairness for 25 Connections

Figure 5 presents the throughput results for 1, 5 and 25 connections respectively. It can be clearly seen that ATP outperforms both default TCP and TCP-ELFN under different load conditions. Figure 6 presents the fairness results for the scenarios of 5 and 25 connections. It can be inferred that ATP is able to provide a better degree of gobal fairness than default TCP and TCP-ELFN.

Publications & Presentations:

• K. Sundaresan, V. Anantharaman, H.-Y. Hsieh, and R. Sivakumar,
  "Sensitivity Analysis of the ATP Protocol Parameters."
  GNAN Technical Report, School of ECE, Georgia Tech, Mar 2004.
• K. Sundaresan, V. Anantharaman, H.-Y. Hsieh, and R. Sivakumar,
  "ATP: A Reliable Transport Protocol for Ad-hoc Networks."
  ACM International Symposium on Mobile Ad Hoc Networking and Computing (MOBIHOC), Annapolis, MD, USA, June 2003.
• V. Anantharaman, S.-J. Park, K. Sundaresan, and R. Sivakumar,
  "TCP Performance over Mobile Ad-hoc Networks: A Quantitative Study."
  Wireless Communications and Mobile Computing Journal (WCMC), Special Issue on Performance Evaluation of Wireless Networks, 2003.
• V. Anantharaman and R. Sivakumar,
  ``A Microscopic Analysis of TCP Performance over Wireless Ad-hoc Networks.''
  ACM International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS) - Poster Paper, Marina Del Rey, CA USA, June 2002.

Software Downloads:

People:

• Karthikeyan Sundaresan (Student)
• Seung-Jong Park (Alumnus)
• Vaidyanath Anantharaman (Alumnus)
• Raghupathy Sivakumar (Professor)

References & Related Work:

• G. Holland and N. Vaidya, "Analysis of TCP Performance over Mobile Ad Hoc Networks," Proceedings of ACM MOBICOM, Seattle, WA USA, August 1999.
• T. Goff, N. Abu-Ghazaleh, D. Phatak and R. Kahvecioglu, "Preemptive Routing in Ad Hoc Networks," Proceedings of ACM MOBICOM, Rome, Italy, July 2001.
• M. Grossglauser and D. Tse, "Mobility Increases the Capacity of Ad-Hoc Wireless Networks," Proceedings of IEEE INFOCOM, Anchorage, AK USA, April 2001.
• K. Chandran, S. Ragbunathan, S. Venkatesan and R. Prakash, "A FeedBack Based Scheme for Improving TCP Performance in Ad-Hoc Wireless Networks," Proceedings of IEEE ICDCS, Amsterdam, The Netherlands, May 1998.
• B. Bakshi, P. Krishna, N. Vaidya and D. Pradhan, "Improving Performance of TCP over Wireless Networks," Proceedings of IEEE ICDCS, Baltimore, MD USA, May 1997.
• J. Monks, P. Sinha and V. Bharghavan, "Limitations of TCP-ELFN for Ad Hoc Networks," Proceedings of MoMuC, Tokyo, Japan, October 2000.
• G. Holland and N. Vaidya, "Impact of Routing and Link Layers on TCP Performance in Mobile Ad Hoc Networks," Proceedings of IEEE WCNC, New Orleans, LA USA, September 1999.
• M. Gerla, K. Tang and R. Bagrodia, "TCP Performance in Wireless Multi-Hop Networks," Proceedings of IEEE WMCSA, New Orleans, LA USA, February 1999.
• M. Ostergren, "TCP Performance in Ad-Hoc Networks," M.S. Thesis, Swedish Instutie of Computer Science, Sweden, November, 2000.