Tu Ouyang Olin Building #505 Case Western Reserve University 10900 Euclid Ave. Cleveland, OH 44106 Email: tu.ouyang AT case DOT edu Email: xfe86d5x@swdr.case.edu

I am a Ph.D. candidate in EECS department of Case Western Reserve University.
I received M.S and B.S from Computer Science Department, Peking University, China, respectively in 2006 and 2003.
My research advisor is Prof. Michael Rabinovich.
I have been also working with Prof. Shudong Jin, Prof.Soumya Ray and Mark Allman.

Research
My research in Case Western Reserve University is to uplift network quality, covering network measurement, performance and security analytics.
Besides a piece of work to improve TCP performance in Mobile Ad-hoc network via strategically placing TCP proxies, I have been devoting to fighting botnets and related security issues, e.g., email spam and URL spam, by building machine learning algorithm based classifiers on top of network features.

• SpamWeeder: a precise approach to prevent spam from the root
  A running system is deployed in http://spamweeder.case.edu,
  to facilitate spam prevention for students in Case Western Reserve University.
  SpamWeeder can help to track the email address trafficking, expose the parties at the source of these channels,
  and precisely block all email from all and only parties belonging to a given distribution channel.
  

• A large-scale analysis of email spam detection through network characteristics
  A research project that tried to make use of network features of SMTP session,
  such as sources IP address, 3-way handshake time, round-trip time variance,
  to classify the email message delivered by this session is spam or ham.
  Through thorough analysis using machine learning methods,
  We demonstrated the network features are capable of classifying spam/ham;
  in addition feature characteristics were stable over period of times. 
  Classification through network features has much less computational cost as compared to content-based approach,
  and the flow-feature-based filter we have built up has an acceptable accuracy
  as a pre-filter to filter out obvious spam and ham before passed to content-based filters.

• Dynamic TCP Proxy in MANETs
  Standard TCP suffer from low-quality, long routing path in Mobile and ad-hoc network (MANETs).
  We propose dynamically installing proxies in some carefully selected nodes along the routing path.
  We demonstrate the potential benefits of this dynamically proxy mechanism, by implementing our protocol in NS2.
  [The protocol implemented in NS2]

Publications

• Tu Ouyang, Soumya Ray, Michael Rabinovich, Mark Allman. "Can Network Characteristics Detect Spam Effectively in a Stand-Alone Enterprise". The Passive and Active Measurement Conference (PAM 2011).
• Tu Ouyang, Soumya Ray, Mark Allman, Michael Rabinovich. "A Large-Scale Empirical Analysis of Email Spam Detection through Transport-Level Characteristics". Technical Report 10-001, International Computer Science Institute, January 2010.
• Tu Ouyang, Shudong Jin, Michale Rabinovich. "Dynamic TCP Proxies: Coping with Disadvantaged Hosts in MANETs". ICDCSW 2009.
  
• Tu Ouyang, Michael Rabinovich. "Weeding Spammers at the Root: A Precise Approach to Spam Reduction". Global Internet Symposium' 08 (In conjunction with INFOCOM' 08).

-------------------------------------------
Last Updated: 08/10/2011