Investigated Hadoop cluster computing with HP Labs for a single threaded, long running, multi-pass computation, where individual parallel computations could depend on the results of subsets of other parallel computations (non-disclosure agreement prohibits more detail). I specified the necessary multiple phase Mapping and Reduction operations, and the definitions of the global objects to provide for the required global state. These global objects were automatically created by the reduction of multiple parallel object fragments. 2008

At RAHD Oncology Products, the software developers and support staff were spread around the country (St. Louis, Salt Lake City, St. George UT, New York, Boston), and sometimes the world. In 1999, I redesigned and reimplemented the distributed software development and support environment, which reduced our development cycle times by orders of magnitude, and provided alpha and beta test capabilities never before possible. In 2000, I designed (in conjunction with the Customer Support Manger), a comprehensive distributed project management system for our geographically distributed software and hardware development groups and customer support group. This provided a web interface to a relational database backend, usable by both managers and staff, with additional support through integration with other tools, such as a version control system (CVS), and the raw file system. It was implemented by another developer and myself. 1999-2006

One of a three person team who designed a distributed information system testbed architecture for the CIA. We produced detailed documentation and numerous presentations, which led to funding for the next 6 years. The project grew to 12 members, and working systems were in use at CIA headquarters. In addition to the architecture, I also designed the initial application-level communication protocol, and later designed and implemented a robust network-transparent communication system to support it. This research was carried out on a distributed, networked, workstation environment (Apollo Computers). 1983-1990

As system adminstrator for the Apollo workstation environment, I greatly reduced downtime and improved performance through aggressive reengineering, based on knowledge gained from our distributed system research. 1986-1988