Dr. Shweta Jain

Abstract: Since the commercialization of the Internet, we have seen rapid transformation both in the way people connect to the Internet and the applications they use over the network. End-hosts are primarily wireless and mobile devices rather than wired clients and applications have shifted from primarily conversation based to content centric. In order to support this transformation, there is a significant amount of research focus on network architecture that is designed for user mobility and content centric applications. Researchers believe that the solution lies in caching content closer to its consumer so that it may be served from a wireless access point, optical network terminator, cellular base station or a storage router in the network. Fortunately the cost of storage continues to fall and hence storage on routers and other point of connection devices is becoming feasible. This talk is about a Storage Aware Routing (STAR) protocol that uses storage and historical information about path costs to make routing decisions. When downstream paths are expensive, STAR freezes content transmission and caches the data at intermediate routers. It then waits for better/cheaper paths to be discovered; perhaps the existing path may improve when the mobile user switches to a faster network, or network conditions may improve. As seen in simulation using ns2 and emulation on the ORBIT testbed, STAR works well with a hop by hop transport protocol, and performs better in terms of throughput and delay compared to a link state routing protocol. STAR is an address based routing protocol but can be extended to hybrid address and content routing. As a content routing protocol, STAR uses the same principles of history and storage to find multiple cached locations for the requested content. It then selects the best location and resumes its address based routing functionality to fetch the requested data. As STAR automatically stores content when routes become expensive, the most busy/expensive routes are likely to have more copies of the most popular content. Intuitively, this approach may result in an overall reduction in content delivery latency.

Biography: Shweta Jain is Assistant Professor in the Math and Computer Science Department at York College CUNY since Fall 2010. Prior to York College, she was a Research Associate at WINLAB Rutgers University from 2008-2010 and Senior Engineer at Staccato Communications (now Veebeam Ltd.) from 2007-2008. She completed her PhD from Stony Brook University in 2007. Her research areas include medium access and routing for wireless ad-hoc, mesh and RFID networks. She is an educator and researcher. She is currently teaching Introduction to Computing (C++) and Data Communications at York College. She is also working on protocol design for the next generation networks, measuring performance of competing/cooperating wireless technologies in the ISM band from an application layer perspective. She is an area editor for the ICST Journal on Emerging Internet Infrastructures and an active reviewer for several IEEE and ACM journals. She has served in the organizing committees of several conferences including the IEEE Sarnoff Symposium, IEEE International Conference on Internet Multimedia Systems Architecture and Application and ACM International Symposium on Mobile Ad Hoc Networking and Computing.

Email:

Address:Assistant Professor, Math and Computer Science, York College CUNY, NY, New York, United States

Dr. Shweta Jain

Biography:

Email:

Address:NY, New York, United States

Dr. Shweta Jain

Biography:

Email:

Address:NY, New York, United States