Thomas Karagiannis                                       Curriculum Vitae

 

Microsoft Research

7 JJ Thomson Ave

Cambridge, CB30FB

UK

                   

 

  E-mail: thomkar@microsoft.com

URL: http://research.microsoft.com/~thomkar/

 

 

 

 

 

 

 

 

 

Research Interests

 

  • Opportunistic communications and social networks.
  • Internet measurements and monitoring.
  • Analysis and modeling of Internet traffic dynamics.
  • Traffic analysis, identification and classification of Internet applications (e.g., web, peer-to-peer, streaming, etc.)
  • Peer-to-peer networks: Measurements, characterization and identification of peer-to-peer Internet traffic; economics of peer-to-peer networks.
  • Security: Intrusion detection and prevention methodologies for large enterprise networks.

 

 

Education

2000  - 2006

Ph.D. in Computer Science at the University of California, Riverside.

Supervised by Associate Professor Michalis Faloutsos.

Dissertation Title: Novel Techniques and Models for Network Traffic Profiling: Characterizing the Unknown.

 

1996 - 2000

B.S. at the University of Macedonia, Greece.

Department of Applied Informatics with GPA 8.8/10 (Excellent).

 

 

Conference Publications

  • Power law and exponential decay of inter contact times between mobile devices.

Thomas Karagiannis, J-Y Le Boudec, Milan Vojnovic

ACM MOBICOM, Montreal, Canada, 2007.

 

  • Profiling the End Host.

Thomas Karagiannis, Dina Papagiannaki, Nina Taft, Michalis Faloutsos

Passive and Active Measurement Conference (PAM), Louvain-la-neuve, Belgium, 2007.

 

  • Multi-level application-based traffic characterization in a large-scale wireless network.      

Manolis Ploumidis, Maria Papadopouli, Thomas Karagiannis

IEEE WoWMoM, Helsinki, Finland, 2007.

 

  • Planet Scale Software Updates.

Christos Gkantsidis, Thomas Karagiannis, Pablo Rodriguez, Milan Vojnovic
ACM SIGCOMM, Pisa, Italy, September 2006.

 

  • Should Internet Service Providers Fear Peer-Assisted Content Distribution?

Thomas Karagiannis, Pablo Rodriguez, Dina Papagiannaki
ACM/USENIX Internet Measurement Conference (IMC), Berkeley, CA, October, 2005
.

 

  • BLINC: Multilevel Traffic Classification in the Dark.

Thomas Karagiannis, Dina Papagiannaki, Michalis Faloutsos

ACM SIGCOMM, Philadelphia, PA, USA, August 2005.

 

  • Transport layer identification of p2p traffic.

      Thomas Karagiannis, Andre Broido,  Michalis Faloutsos, kc claffy

ACM/USENIX Internet Measurement Conference (IMC), Taormina, Italy, October, 2004.

 

  • Is P2P dying or just hiding?

      Thomas Karagiannis, Andre Broido, Nevil Brownlee, kc claffy, Michalis Faloutsos

      IEEE GLOBECOM, Global Internet/Next Generation Networks, Dallas, TX, Nov. 2004.

 

  • A Nonstationary Poisson View of Internet Traffic.
    Thomas Karagiannis, Mart Molle, Michalis Faloutsos, Andre Broido
    IEEE INFOCOM,
    Hong Kong, March 2004.

 

·         Long-Range Dependence: Now you see it, now you don’t!

Thomas Karagiannis, Michalis Faloutsos, Rudolf H. Riedi

IEEE GLOBECOM, Global Internet Symposium, Taipei, Taiwan, November 17-21, 2002.

 

Journal Articles & Editorials

  • Comparing Traffic Classifiers.

Luca Salgarelli, Francesco Gringoli, Thomas Karagiannis

ACM SIGCOMM Computer Communication Review, July 2007.

 

  • Long-range dependence: Ten years of Internet traffic modeling.

Thomas Karagiannis, Mart Molle, Michalis Faloutsos

 IEEE Internet Computing, Special Issue -- Measuring the Internet, September, 2004.

 

  • A User-Friendly Self-Similarity Analysis Tool.

Thomas Karagiannis, Michalis Faloutsos, Mart Molle

ACM SIGCOMM Computer Communication Review, Special issue -- Tools and technologies for networking research and education, 2003.

 

 

Technical Reports

  • Sampling Strategies for Epidemic-Style Information Dissemination

Milan Vojnovic, Varun Gupta, Thomas Karagiannis, Christos Gkantsidis

MSR-TR-2007-82, July 2007.

  • Power law and exponential decay of inter contact times between mobile devices.

Thomas Karagiannis, J-Y Le Boudec, Milan Vojnovic

MSR-TR-2007-24, March 2007.

  • Understanding the Limitations of Estimation Methods for Long-Range Dependence.

Thomas Karagiannis, Mart Molle, Michalis Faloutsos

UCR-CS-2006-10245, November, 2006.

  • BLINC: Multilevel Traffic Classification in the Dark.

Thomas Karagiannis, Dina Papagiannaki, Michalis Faloutsos

January, 2005.

  • File-sharing in the Internet: A characterization of P2P traffic in the backbone.

Thomas Karagiannis, Andre Broido, Nevil Brownlee, Kc Claffy, Michalis Faloutsos

November. 2003.

 

Honors and Awards

  • Chancellor’s Distinguished Fellowship offered by UC, Riverside (2000).
  • Scholarship from the Greek State Scholarships Foundation (1996).

(Top score in the Greek National Examinations out of approx. 150,000 participants.)

  • Scholarship from the Greek Telecommunications Organization (1999, 2000).

(Awarded to the top three students of the dept. of Applied Informatics,

 Univ. of Macedonia, Greece.)

  • Annual awards from the Greek State Scholarships Foundation (1996 - 2000).

(Awarded to the top three students of the dept. of Applied Informatics,

 Univ. of Macedonia, Greece.)

 

 

Invited talks and acclaims

  • Press interviews: Wired magazine and MIT Technology Review on trends of peer-to-peer file-sharing networks.
  • News and articles: The peer-to-peer measurement work appeared in numerous articles in newspapers,

professional magazines and Internet news sites such as USA Today – “The rise and fall (?) of P2P music downloading”,

Advisory Committee Congressional Internet Caucus, Electronic Frontier Foundation, ACM tech news, Slashdot.org and others.

  • Courses syllabus: The papers on peer-to-peer traffic measurements and traffic analysis have been included

in the reading material of computer science and law courses (e.g., Infosys 296A section 2,

School of Information Management and Systems, UC Berkeley, CPSC 641: Performance issues in High Speed Networks, Univ. of Calgary,

Advanced Computer Networks, Univ. of Delaware, etc.).

  • NANOG Fall 2005: BLINC: Multilevel Traffic Classification in the Dark.
  • NANOG Fall 2005: Should Internet Service Providers Fear Peer-Assisted Content?
  • ISMA 2004 Workshop on Internet Signal Processing (WISP)
  • SAMSI Workshop on Congestion Control and Heavy Traffic Modeling (2003)

 

 

 

 

Work and Research Experience

September 2006 –present.

Microsoft Research. Associate Researcher.

Opportunistic communications, social networks, traffic modeling and characterization, P2P networks, end-host profiling, security and anomaly detection.

 

July 2006-August 2006

Foundation for Research and Technology, Hellas. (Visiting Researcher)

Application classification of WLAN traffic, characterization and profiling of mobile users.

October 2005 –

April 2006

 

Intel Research at Cambridge.  (internship)

Anomaly detection and security for enterprise networks through end-host profiling. Proposed a novel methodology to profile individual host behavior through a graph-based footprint that generates a compact, robust and intuitive description of the behavior with the goal of anomaly detection.

June 2005 –

September 2005

 

Microsoft Research at Cambridge. (internship)

Analysis of the properties of Microsoft's patching service with a goal to design a system for rapid patch propagation. Analyzed patch-request patterns and user profiles, and studied the benefits and impact of alternative architectures for efficient patch distribution such as peer-to-peer and caching.

July 2004 –

May 2005

 

Intel Research at Cambridge.  (internship)

Characterization, analysis, traffic profiling and classification of Internet applications. Proposed and implemented a novel methodology for traffic flow classification according to the applications that generate them based on exploiting host-interactions and traffic patterns at the end-hosts.

Peer-to-peer networks: Analysis of the impact of peer-assisted content distribution on Internet service providers, content providers and end-users based on real Internet traces.

July 2003 –

September 2003

Cooperative Association for Internet Data Analysis (CAIDA).(internship)

Identification, measurement and characterization of peer-to-peer traffic in the Internet backbone. Proposed and implemented payload-based and transport-layer based methodologies for the identification of peer-to-peer file-sharing traffic at the Internet backbone.

 

March 2001 –

July 2004

University of California, Riverside.

Research Assistant at the Networking & Communications Laboratory.

Analysis and modeling of Internet traffic dynamics: Studied the self-similar and long-memory properties of Internet traffic at the backbone and proposed the modeling of backbone traffic as a nonstationary Poisson process.

 

1998

University of Macedonia, Greece.

Career and Liaison office, Full-time. Technical support and supervision of student internships. Design and implementation of intern database.

 

July 1997 –

October 1997

National Bank of Greece.

Department of Commercial Loans. Risk analysis and assessment.

 

 

 

Teaching Experience

2000-2001

University of California, Riverside. Teaching Assistant.

Principles of Programming Languages, Intermediate Data Structures & Algorithms.

 

1999-2000

University of Macedonia, Greece. Teaching Assistant.

Dept. of International & European Economic & Political Studies.

MS Windows, MS Office.

 

 

Software Development

  • The SELFIS tool: Designed and implemented the SELF-similarity analysIS software tool. SELFIS is a java-based tool

for the analysis of the self-similar and long-range dependence properties of time-series with focus on networking.

It has been downloaded over 600 times and is in use by researchers spanning various disciplines and institutions.

 (URL: http://www.cs.ucr.edu/~tkarag/Selfis/Selfis.html.)

 

  • BLINC: Designed and implemented the BLINd Classification tool, a multilevel transport-layer methodology to classify

Internet traffic flows according to the applications that generate them without the knowledge of port numbers

or the use of payload signatures. BLINC was implemented in C++.

 

  • PTP: Design and implementation of the P2P Traffic Profiling tool.  The PTP tool is a methodology for the identification

of P2P file-sharing traffic at the transport-layer without the use of payload signatures and is implemented in C++ and Python.

 

  • Peer-to-peer file-sharing and application-specific protocol dissectors: Reverse-engineering, design and implementation

of payload signature-based traffic identification of the nine most popular file-sharing peer-to-peer networks

(Kazaa, Gnutella, BitTorrent, eDonkey, Ares, WinMx, OpenNap, Soulseek, MP2P). Identification of the traffic of the majority

 of Internet applications (web, streaming, ftp, chat, mail, news, gaming, etc). Implemented in C/C++ on top of the

 CoralReef Internet Traffic monitoring software suite.

 

Professional service

  • TPC member of IEEE Infocom 2008, QoSim08.
  • Member of ACM Sigcomm shadow PC (2005).
  • Journal referee for Transactions on Networking, JSAC special issue on Delay Tolerant Networks (2007), ACM CCR (2007),

Journal of Systems and Software (2007), Computer Communications (2007), JSAC special issue on “High-speed Network Security – Architecture,

Algorithms and Implementation” (2005) , Performance Evaluation special issue on Long-Range Dependence and Heavy Tail Distributions (2004),

IEEE Transactions on Signal Processing (2003), Machine Learning (2004).

  • Referee for IEEE Infocom (2004, 2005, 2006, 2007), PAM (2007), Globecom-Global Internet and Next Generation Networks (2004),

1st EuroNGI Conference on Next Generation Internet Networks - Traffic Engineering (2005), COMSWARE (2006).

 

Personal Information

Nationality

Greek

 

Languages

 

  • English (Cambridge First Certificate in English & Cambridge Certificate of Proficiency in English)
  • German (Grundstufe & Mittelstufe)
  • Greek

 

References

  • Available on request