Shelby Thomas

PhD Candidate
Computer Science
University of California, San Diego

Shelby Thomas

shelbyt@ucsd.edu CSE 3142 Professional

I am a third year Computer Science PhD student at the University of California, San Diego (UCSD). I work with some awesome professors at SYSNET and my primary advisors are George Porter and Geoffrey M. Voelker. My current work focuses on Terabit/s networks and the communication channels between the hardware and the OS to make this happen.

I developed an open source benchmark suite for machine learning and hardware researchers to leverage. To date it has been in use by over 100 universities and companies. In my spare time I work on improving undergradute engagement and diversity in computer science research.

Recent Publications


Dark Packets and the End of Network Scaling
Shelby Thomas, Rob McGuinness, Geoffrey M. Voelker, George Porter
ANCS, 2018.

CacheCloud: Towards Speed-of-Light Datacenter Communication
Shelby Thomas, Geoffrey M. Voelker, George Porter
HOTCLOUD, 2018.

Recent Posts


Tina Seelig: Creativity (Morning Mentors #2)
14 Jul 2018
talks research business

Qualifications: Management Science and Engineering Professor, Directory of Stanford Technology Ventures Program

Seelig’s talk on creativity has applications in entrepreneurship and research. It’s fascinating how something simple as reframing problems by asking what would change if I solved this problem can stimulate new ideas and jump-start creativity.

Knowledge
  • You gain knowledge by paying attention to the world and being more observant than ANYONE else
  • Knowledge is a toolbox for imagination. The more you know, the more you can work with. You need to know things deep enough to create metaphors for other areas
More...
Richard Hamming: You and Your Research (Morning Mentors #1)
06 Jul 2018
talks research

Qualifications: Turing Award winner, Bells Labs Alumni, Hamming code inventor, Professor

One of the most important talks that I have ever watched about the nature of research and success by a giant of computing, telecommunications, and mathematics. Very rewatchable and tremendously inspiring.

Research
  • You MUST Work on important problem in your field
  • Consider the implications of your work on the future, is it worth doing?
  • Turn defects into assets (digital vs. analog hamming differentiation)
  • Change the nature of the problem to find the underlying question and nature
  • Spend time studying related research domains (Jon Tucci studied all layers at Bell Labs)
  • Work on the right problem, at the right time. “…million races being run, just get in one and win”
More...
RDMA Explained: Part 1
23 Apr 2018
RDMA networks hardware

This is the first part of a multi-part post going over RDMA, current research and RDMA’s role in the future of networking.

“Bandwidth problems can be cured with money. Latency problems are harder because speed of light is fixed—you can’t bribe God” - Anonymous

More...

Shelby Thomas

PhD Candidate
Computer Science
University of California, San Diego

Shelby Thomas