TCP friendliness of advanced TCP in high speed network simulated by ns-2:

Presentation slides, demo slides, and final report (PDF files).

With usage of high speed network in the world, current TCP congestion control algorithm is facing some difficulties, such as low link utilization, instability and oscillation. In order to solve these problems, some advanced TCP protocols like Fast TCP, HTCP, BIC TCP, CUBIC TCP are designed. Since the TCP is now widely used in Internet, it is important to evaluate the TCP friendliness of these new designs so as to be fair to end-users with different protocols. In this project, I plan to choose one or two of above mentioned protocols, download the relative protocol code from internet and merge them into ns-2, use ns-2 to simulate a network with TCP users and users with one of advanced protocols, and analyze TCP friendliness performance of the protocols.
Simulation tool: ns-2

References:
http://www.hamilton.ie/net/htcp2005.pdf
http://www.hamilton.ie/net/htcp3.pdf
http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/cubic-paper.pdf
http://www.csc.ncsu.edu/faculty/rhee/export/bitcp.pdf
http://www.cs.usc.edu/Research/TechReports/05-854.pdf

Comparative analysis between client-server and P2P networks using ns-2:

Presentation slides, demo slides, and final report (PDF files). P2P networks are emerging systems destined to replace the current client-server network. It offers several critical advantages over traditional networks such as fault-tolerence, scalability and improved congestion control.

In this project, we would like to analyze and compare the traditional client-server network with a P2P network using NS-2. We would like to see the improvements and advantages, if any, that a P2P network has over a traditional networks. There are three main areas of interests:

References:
[1] E. Cortese, F. Ouarta, and G. Vitaglione, "Scalability and Performance of the JADE Message Transport System," Proc. Of the AAMAS Workshop on AgentCities, Bologna, Italy, July 2002
[2] E. Chen, Jade and JXTA Extensions for Implementing a Better Distributed Systen, Master's Thesis, School of Engineering Science, Simon Fraser University, Canada, 2005
[3] AODV, http://moment.cs.ucsb.edu/AODV/aodv.html, accessed March, 2006
[4] DSDV, http://www.cs.virginia.edu/~c17v/cs851-papers/dsdv-sigcomm94.pdf, accessed March, 2006
[5] DSR, http://www.cs.cmu.edu/~dmaltz/internet-drafts/draft-ietf-manet-dsr-09.txt, accessed March, 2006

Evaluation of Interactive Gaming Traffic over 802.11 Network:

Presentation slides, demo slides, and final report (PDF files).

IEEE 802.11 Standard has become the major wireless local area network (WLAN) protocol in the institute or home environment. With rapid increasing number of mobile devices in the market, the WLAN will be carrying vast varieties of traffic types such as voice and video in addition to pure data. However, the type of traffic that often neglected by people is the interactive gaming traffic. The interactive gaming traffic is subject to an even more strict quality of service than video or voice. With increasing popularity of laptops, interactive gaming traffic is expected to grow. It would be interesting to see the how well can 802.11 handle this type of traffic, and what aspect of quality of service does the 802.11 protocol need to improve. This project will focus on evaluating the 802.11 network performance with the interactive gaming traffic load.

References:
[1] C.Heyaime, V.K.Prabhu,=20 "Traffic Models for Video Game Applications", Broadband Communications for the Internet Era Symposium digest, 2001 IEEE Emerging Technologies Symposium on10-11 Sept. 2001 Page(s):28 - 32
[2] K.T. Chen, P. Huang, C.Y. Huang, C.L. Lei, "Game Traffic Analysis: An MMORPG Perspective", Department of Electrical Engineering National Taiwan University, Institute of Information Science Academia Sinica, Department of Electrical Engineering and Graduate Institute of Networking and Multimedia National Taiwan University
[3] S. Zander, G. Armitage, "Network gaming: A traffic model for the Xbox game Halo 2", Proceedings of the international workshop on Network and operating systems support for digital audio and video, 2005, Page(s):13 - 18
[4] R.A. Bangun, E. Dutkiewicz, "Modelling multi-player games traffic", Information Technology: Coding and Computing, 2000, Page(s):228 - 233
[5] J. Farber, "Network game traffic modelling", Proceedings of the 1st workshop on Network and system support for game, 2002, Page(s): 53 - 57

Scalability and Robustness of the Gnutella Protocol:

Presentation slides and final report (PDF files).

Gnutella is a decentralized p2p network protocol for file sharing. There are two major advantages of decentralized protocols. The first one is scalability, and the second one is robustness to failure.

In this project, I will simulate the Gnutella protocol. I will test its scalability by adding a large number of nodes. To test the robustness to failure, I will arbitrary fail some nodes to see the effect on the network. Statistics such as throughput and delay will be compared.

If time permits, I will implement another p2p protocol, such as KaZaA, and compare the results to Gnutella results.

Since I didn't learn NS2 simulation tool yet, I'm not sure yet if I will use OpNet or NS2 for this project.

References:
1. "The Annotated Gnutella Protocol Specification v0.4," http://rfc-gnutella.sourceforge.net/developer/stable/index.html, accessed March 2006.
2. L. Gong, "Peer-to-Peer Networks in Action," IEEE Internet Computing, vol. 5, no. 1, 2002, pp. 37-39. http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/4236/21094/00978367.pdf?isnumber=21094&arnumber=978367
3. Sen, S., Jia Wang, "Analyzing peer-to-peer traffic across large networks," IEEE/ACM Transactions on Networking, vol. 12, no. 2, 2004, pp. 219 - 232. http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/90/28714/01288128.pdf?tp=&arnumber=1288128&isnumber=28714
4. Manini, D., Gaeta, R., Sereno, M., "Performance Modeling of P2P File Sharing Applications," Workshop on Techniques, Methodologies and Tools for Performance Evaluation of Complex Systems, pp. 34 - 43, 2005 http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/10582/33459/01587693.pdf?tp=&arnumber=1587693&isnumber=33459
5. He Q., Ammar M., Riley G., Raj H., Fujimoto, R., "Mapping peer behavior to packet-level details: a framework for packet-level simulation of peer-to-peer systems," IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer Telecommunications Systems, pp. 71 - 78, 2003. http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/8783/27814/01240644.pdf?tp=&arnumber=1240644&isnumber=27814
6. Juan Li, Son Vuong, “Ontology-Based Clustering and Routing in Peer-to-Peer Networks,” International Conference on Parallel and Distributed Computing, Applications and Technologies, pp. 791 – 795, 2005. http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/10544/33357/01579032.pdf?tp=&arnumber=1579032&isnumber=33357
7. Tsungnan Lin, Hsinping Wang, Jianming Wang, "Search performance analysis and robust search algorithm in unstructured peer-to-peer networks," IEEE International Symposium on Cluster Computing and the Grid, pp. 346 - 354, 2004. http://ieeexplore.ieee.org.proxy.lib.sfu.ca/iel5/8717/27590/01231525.pdf?tp=&arnumber=1231525&isnumber=27590

Dynamic Right-Sizing: TCP Flow-Control Adaptation:

Presentation slides and final report (PDF files).

Grid and networking researchers continue the practice of manually optimizing TCP buffer sizes to keep the network pipe full, and thus achieve acceptable performance over the wide-area network. Not only is this process cumbersome, but the result oftuning window sizes for a particular pair of hosts is sub-par performance for connections with larger delay-bandwidth products and the misappropriation of scarce resources to connections with smaller delay-bandwidth products. Dynamic Right sizing lets the receiver estimate the sender’s congestion window size and uses that estimate to dynamically change the size of the receiver’s window advertisements. As a result, the sender will be congestion-window-limited rather than flow-control-window-limited. This project will investigate a simple OPNET implementation of Dynamic Rightsizing.

References:
[1] Mike Fisk and Wu-chun Feng, "Dynamic Right-Sizing: TCP Flow-Control Adaptation," in Proceedings of the 14th Annual ACM/IEEE SC2001 Conference, November 2001. http://public.lanl.gov/radiant/pubs/drs/sc2001-poster.pdf
[2] Van Jacobson, "Congestion Avoidance and Control, in Proceedings, SIGCOMM '88 Workshop. ACM SIGCOMM, Aug. 1988, pp. 314 - 329, ACM Press, Stanford, CA.
[3] Jeff Semke, Jamshid Mahdavi, and Matt Mathis., "Automatic TCP buffer tuning," Computer Communications Review, vol. 28, no. 4, pp. 315 - 323, Oct. 1998.
[4] Jian Lui and Jim Ferguson, "Automatic TCP socket buffer tuning," in Supercomputing 2000 Research Gems, Nov. 2000, Awarded Best Research Gem of the Conference
[5] Brian L. Tierney, TCP tuning guide for distributed application on wide area networks, vol. 26, no. 1, Feb. 2001.

BitTorrent's peer protocol OPNET model:

Download is traditionally transmit the file from sever to client use protocols such as FTP, HTTP, PUB and so on. In this case the number of the client will limit the speed of transmit. The transmission based on BitTorrent¡¯s peer protocol is opposite circumstance. I try to use OPNET to simulate the simple process of BitTorrent's peer protocol based file transmission. I conclude by the effect of BitTorrent's peer protocol to tradition peer to peer protocol.