1998 in the NOW lab at UC Berkeley
I’ve had a lot of odd jobs since I was very young, but this list covers my career in the tech industry:
UC Berkeley Computer Science Department (1996-1998)
Research Assistant, NOW / Millennium Research Group
Microsoft Research (1998)
Intern, Bay Area Research Center (BARC)
Software Engineer, Traffic Server
Scale Eight (1999-2003)
Founder, CTO, Director
Shut down in 2003.
Internet Archive (non-profit) (2003-2004)
VP/Director of Engineering Operations
Berkeley Data Systems (Mozy) (2005-2007)
Sold to EMC for $76 million in 2007.
Brigham Young University (Computer Science Department (2008-2010)
Adjunct Instructor (CS405 Intro to Venture Startups)
Instructure, Inc. (2010-2020)
CEO, Executive Chairman
IPO on NYSE in 2015
Sold to Thoma Bravo for $1.9 Billion in 2020
University of California, Berkeley
BA Computer Science (1998)
I've only published a single peer reviewed paper because I ended up not pursuing a career in academia. This paper was the result of a project where we broke the world record in data sorting. We sorted 1 million, 100 byte records in 1.18 seconds, disk to disk on a cluster of Pentium machines. It was pretty neat at the time.
Millennium Sort: A Cluster-Based Application for Windows NT Using DCOM, River Primitives, and the Virtual Interface Architecture. Philip Buonadonna, Joshua Coates, Spencer Low, and David E. Culler, University of California, Berkeley, In Proceedings of the 3rd USENIX Windows NT Symposium, July, 1999
This paper was an internal Microsoft document, so it wasn’t peer reviewed.
Performance/Price Sort and PennySort. Gray, J., Coates, J., and Nyberg, C. Technical Report MSTR -98-45, Microsoft Research, August 1998 (word document)
This isn’t a paper at all, but it’s an article I wrote shortly after I moved to Utah. I was irritated at the state of funding in the local venture startup community so I “wrote a letter to the editor”:
Poison in the Well:Could Utah's Unorthodox Debt-Funding Term Sheet Tactics Be Hurting Economic Growth? Josh Coates, Connect Magazine, October 2005 (also see the "rebuttal" by BYU/Utah Angels.)
I don't think software patents are very useful, but when I was young and didn't know any better and I thought they were important. When Scale8 shut down we sold the pending patents to Intel for a couple of hundred thousand dollars and eventually they ended up getting executed. They are pretty useless, but for the sake of posterity, here they are:
Dynamic Replication of Files in a Network Storage System; Patent Number 7203731; Filing Date 12/29/00; Issued Date 4/10/07
Methods and Apparatus for Accessing Remote Storage in a Distributed Storage Cluster Architecture; Patent Number 6952737; Filing Date 12/29/00; Issued Date 10/04/05