Prof. Ramesh K. Sitaraman



Department of Computer Science

University of Massachusetts

Amherst, MA 01003-4610, USA

Phone:  (413) 545 3279

FAX: (413) 545 1249

ramesh at cs dot umass dot edu

Assistant: Karren Sacco

Office: CS Building, Room 300

Phone: (413) 577 0292

ksacco at cs dot umass dot edu

research interests

  1.   Internet-scale Distributed Systems

  2. Content Delivery Networks

  3. Cloud Computing

  4. Algorithms and Theoretical Computer Science



Prof. Sitaraman is currently in the School of Computer Science at the University of Massachusetts at Amherst. His research focuses on all aspects of Internet-scale distributed systems, including algorithms, architectures, performance, energy efficiency, and economics. Previously, as a principal architect, he helped create the Akamai network and is an Akamai Fellow. He is best known for helping pioneer content delivery networks and other Internet-scale distributed networks that currently deliver much of the world’s web content, streaming videos, and online applications.   

Prof. Sitaraman is a recipient of an NSF CAREER Award, the College Outstanding Teacher Award, and a Lilly Fellowship. He has served on numerous program committees and editorial boards of major conferences and journals. He received a B. Tech. in electrical engineering from the Indian Institute of Technology, Madras. and a Ph.D. in computer science from Princeton University.


Current Favorites 

Understanding the Effectiveness of Video Ads: A Measurement Study, SS. Krishnan and R.K. Sitaraman, ACM Internet Measurement Conference (IMC), October 2013.
    Download Paper   Download Presentation
One of the largest scientific studies of video ads involving 367 million video views, 257 million ad views, and 65 million unique viewers around the globe. Shows what factors lead viewers to complete watching video ads and what could cause ads to be abandoned in the middle. Uses novel analytic techniques to extract causal information from ``big data’’. 

2) Video Stream Quality Impacts Viewer Behavior: Inferring Causality using Quasi-Experimental Designs.   
   Watch Presentation Download Paper   Download Presentation
Shows that poor video performance at a site can cause viewers to abandon videos more often, view videos for a lesser time, and return to the same site less frequently. Also, develops novel tools based on QEDs for inferring causality in this domain.

Media Coverage:
 NPR Morning Edition: In Video-Streaming Rat Race, Fast Is Never Fast Enough
UMass Press Release:  UMass Amherst Computer Science Research Quantifies How Online Video Stream Quality Affects Viewer Behavior 
Boston Globe: Instant gratification is making us perpetually impatient
CNN: Online viewers ditch slow-loading video after 2 seconds
Science Daily: How Online Video Stream Quality Affects Viewer Behavior
GigaOm: Online viewers start leaving if video doesn’t play in 2 seconds, says study
Boston Globe: MIT grad led team that built faster YouTube player
News India Times: Online Video Speed Affects Viewer Behavior, Says Computer Scientist.
Broadcast Engineering: Is streaming TV worth the wait?
India Abroad: Despair when a video buffers? Ramesh Sitaraman has some answers
More coverage:
    Akamai Blog,  Yahoo News, PC Magazine,  Hindustan Times, 
     The Verge,  Media Post, Xinhua, SF Chronicle  CNET,
     Broadcast Engineering, Slashdot,

3) Using Batteries to Reduce the Power Costs of Internet-Scale Distributed Systems. 
    Download Paper  Download Presentation
Shows how batteries can be used to reduce the power costs of a large distributed network. Helps establish batteries as a key part of the network architecture of CDNs and other large distributed networks. 

 Media Coverage:
    Giga OM: Batteries should be part of the Internet Infrastructure
    Akamai Blog: Batteries Included: A Leaner and Greener Internet using  Smart Batteries

4) The Akamai Network: A Platform for High-Performance Internet Applications. 
    Download Paper  
The first comprehensive description of the system architecture of one of the world’s largest distributed networks that consists of more than a hundred thousand servers deployed in more than a thousand locations around the world serving roughly 25% of the web today.
Media Coverage:
    GigaOM, High Scalabilty, ReadWrite  

Leading the college commencement parade as an Outstanding Teacher Awardee