| Jun 2006 - Aug 2006 | Consultant | Mozilla Corporation |
| Jan 2002 - Apr 2006 | Chief Executive and Director | Cluster File Systems, Inc. |
| Sep 2001 - Dec 2001 | Senior Architect (Contractor) | Tacit Networks |
| Jul 2000 - Aug 2001 | Strategic Software Projects | Zero-Knowledge Systems, Inc. |
| Sep 1999 - Jun 2000 | Senior Linux Consultant | The Puffin Group / Linuxcare Inc. |
| May 1999 - Aug 1999 | Senior Systems Software Engineer (Contractor) | Stelias Computing |
| Oct 2006 | FSOSS 2006: Curious George and the $100M Supercomputer |
| Jul 2005 | Commodity Cluster Symposium: Lustre |
| Feb 2005 | Linux World: Lustre |
| Jul 2004 | Usenix 2004: Lustre |
| Jul 2003 | OLS 2003: Lustre |
| Sep 2002 | IEEE Cluster 2002: Portals and Networking for Lustre |
| Jul 2002 | OLS 2002: Lustre |
| Aug 2001 | The RIM BlackBerry Protocol |
| Aug 1999 | O'Reilly Perl Conference: The InterMezzo File System |
| Jul 1999 | OLS 1999: hFTPd |
| Jul 1999 | OLS 1999: The InterMezzo File System |
| Jun 1999 | 6th Annual Linux Kongress: The InterMezzo File System |
| Jun 1999 | Linux Storage Management Workshop: The InterMezzo File System |
| May 1999 | YAPC 1999: The InterMezzo File System |
| SC2005 | HPCwire 2005 Editors Choice Award for Lustre (with HP) |
| SC2004 | Most Innovative Use of Storcloud for Lustre Active Storage (with PNNL and HP) |
| SC2003 | Bandwidth Challenge Award for Lustre over a Wide-Area Network (with LLNL and NCSA) |
| OPC 3.0 | Best Paper, System Category for InterMezzo |
| The success and failure of business models that incorporate open source and open development |
| Software licensing; the pros and cons of various open and proprietary licenses for different business goals |
| Advanced software engineering processes, specifically the CMU Software Engineering Institute's PSP and TSP CMM level 5 processes |
| Distributed and internationally diverse work environments |
| Network and disk file systems; the storage needs of high-performance computing; disconnected network file systems |
| Virtual environments, as used in entertainment, productivity, and/or education |
| Consumer applications delivered via the web |
| General aviation in small, single-engine airplanes |
The Firefox web browser is used by 20% of the world's internet users, second only to Internet Exploror. It is the poster child for open source and open development projects, and provides a convincing case study of a successful business based on open software.
At the Mozilla Corporation, I was responsible for an overhaul and management of the localization process for Firefox 2.0. For the first time, the team was able to ship Firefox simultaneously in 23 languages on 3 platforms. For non-English users, Firefox 2.0 represented a substantial improvement in the overall user experience.
Lustre is an open-source cluster file system, primarily for Linux. It is designed to scale to tens of thousands of client nodes, with access to petabytes of storage at transfer rates of tens of gigabytes per second. It is presently used as primary storage on most of the world's largest Linux supercomputers, including the largest clusters in the United States, Europe, and Asia.
I was half of the original design team, and implemented large portions of Lustre 1.0, across every subsystem. As the company and software grew in complexity, I shifted my focus to deployment, then sales and management.
I was responsible for Lustre's deployment and customer integration on many of the top-50 supercomputers of the day, including the world's largest, BlueGene/L.
I developed a business and sales model, intellectual property strategy, and customer support program for Lustre and CFS, before resigning in 2006. I remain a shareholder and supporter of CFS.
With Zach Brown, Chainsaw was written to be a very small, simple, and understandable successor to Zero-Knowledge Systems' Freedom network server.
Like other onion-routing networks, each packet is encrypted multiple times and routed through multiple hops, each of which would remove a layer of encryption and forward the packet. By selecting hops in multiple, hopefully uncooperative, jurisdictions, one can be assured of both security and privacy as long as one hope remains uncompromised.
Cayman implemented Stefan Brands' e-cash patents to enable anonymous electronic cash transactions, using the Blackberry wireless network as a demonstration platform.
Mike Shaver and I developed this prototype to be unveiled at Financial Cryptography 2001, which generated interest in the Brands technology from both online and traditional banks.
Our goal with InterMezzo was to build a distributed file system with a local disk cache and full disconnected operation. It was similar to Coda, but with an order of magnitude less code, and thus hopefully more stable and easier to maintain.
Two successful prototypes were developed: one in Perl, which was integrated into the Linux kernel while it was being maintained, and a second in C that had an HTTP-based protocol.
Although the InterMezzo architecture was clearly technically successful, and there was significant community interest, it did not have sufficient commercial support to continue as a product.
With Joe Shaw and Mike Kestner, Achtung became part of a growing suite of GNOME business tools in 1999-2000. Although it advanced sufficiently far that it was used for its intended purpose, it was eventually overtaken by projects like OpenOffice.