John Giacomoni
Department of Computer Science
University of Colorado at Boulder
Engineering Center, ECCR 1B54
430 UCB
Boulder, CO 80309-0430
E-Mail: John.Giacomoni@colorado.edu
|
|
John GiacomoniDepartment of Computer Science University of Colorado at Boulder Engineering Center, ECCR 1B54 430 UCB Boulder, CO 80309-0430 E-Mail: John.Giacomoni@colorado.edu |
John Giacomoni is currently a fourth year graduate student pursuing a Ph.D. at the University of Colorado at Boulder in the Department of Computer Science and is advised by Dr. Manish Vachharajani. John Giacomoni received the M.S. degree from the Department of Computer Science and a Graduate Certificate in Science and Technology Policy from the Center for Science and Technology Policy Research at the University of Colorado at Boulder in May of 2006. John Giacomoni recieved the B.S. degree from the Department of Computer Science at the University of Illinois at Urbana-Champaign in May of 1997.
Previous to his graduate education, John Giacomoni held a position as a Professional Research Assistant with Dr. Alexander L. Wolf in the Software Engineering Research Laboratory at the University of Colorado at Boulder. John Giacomoni has also been a trusted systems kernel engineer.
My current research focus is in synthesizing new general purpose systems using principles from the areas of systems, computer architecture, security, software engineering, and public policy.
Presently my efforts are focused on operating system services to efficiently support fine-grain parallelism on general purpose commodity multicore hardware platforms. This extends my work in using general purpose commodity hardware platforms for high-rate networking (software network processors) instead of special purpose network processor systems. Previously I have worked on content-based networking and explored certification issues with software defined radios.
The FastForward engine is the core communication technique used to optimize core-to-core communication in this work. FastForward is a new cache-optimized single-producer/single-consumer concurrent lock-free queue providing low overhead operations (sub main-memory latency) wrapped in a simple polymorphic interface enabling a wide range of higher level optimizations. Presently FastForward is focused on core-to-core communication over shared memory, although other communication schemes are being investigated.
Initial experimentation with fine-grain parallelism resulted in the Frame Shared Memory architecture. This architecture is capable of supporting wire-rate processing of 1,000 Mbps Ethernet frames (64B+) in user-space with zero-drops.
Additionally I am interested in the policy, both public and private, implications of technology. Previously I have explored certification methodologies for software defined radios and the potential impacts of affordable wire-rate network frame processing.