Events & News
Colloquium
| Category | Lecture |
| Date | Tuesday, November 6, 2012 |
| Time | 11:00 am |
| Concludes | 12:00 pm |
| Location | Gould-Simpson 906 |
| Speaker | Justin Cappos |
| Title | Assistant Professor |
| Affiliation | Polytechnic Institute, New York University |
Seattle: Harnessing Community Resources for Cloud Computing
Traditional distributed computational models, such as client-server and cloud computing, involve moving computation from geographically distributed devices with little computational power to well-provisioned centralized servers. In this work, we explore the opposite model --- moving computation from servers to clients with little computational power. This helps to solve problems that are not well served by traditional computing models (client-server, cloud, p2p, etc.). This includes situations where the location of computation is paramount, such as in network management, responsive web caches, sensor control, emergency response, and censorship detection. Using this paradigm, we have constructed a testbed called Seattle. Seattle makes it practical for arbitrary Internet users to securely participate in our testbed without compromising the security or performance of their laptop, desktop, phone, tablet or other device. Seattle has been deployed for four years and has wide spread practical use as a testbed for researchers and educators. In the past four years, the Seattle testbed has been used by 30 classes (including 7 that are in progress). The Seattle testbed has thousands of installs, thousand of developers, and thousands of VMs available for use. After providing a quick overview of the Seattle platform, this talk will focus on some of the research challenges like security, performance isolation, network heterogeneity, and portability that were addressed to enable such a system.
Biography
Justin Cappos has been an Assistant Professor at NYU's Polytechnic Institute for the past year. During his time at NYU Poly, he has received 4 NSF grants totaling over 5M USD. Justin's research interests generally fall broadly in the area of systems security. He focuses on understanding high-impact, large-scale problems by building and measuring deployed systems. Justin completed his PhD at the University of Arizona in 2008, under the supervision of John Hartman. His dissertation work was on Stork, a secure and efficient package manager for (OS VM) cloud computing environments. Improvements pioneered in Stork have been adopted by most major Linux package managers including APT, YUM, Pacman, and YaST.