Login
Computer Science @ Rochester
Monday, March 24, 2003
11:00 AM
CSB 209
Laura Bright
U. Maryland
Profile-Based Data Delivery in Wide Area and Mobile Environments
The growing popularity of the Internet has led to an increased diversity of caching technologies, for example middle-tier caches and web proxy caches. An important characteristic of data access in wide area environments is that applications often have different requirements for the latency and recency of data. Traditional caching and scheduling schemes do not provide interfaces to express these requirements, and treat all clients and applications alike. Further, it is difficult to accurately estimate the recency of cached objects. In this talk I present a scheme to support diverse clients and applications on both fixed and mobile networks using client profiles. In the first part of this talk I introduce latency-recency profiles, a set of parameters that allow clients to express preferences for their different applications. These profiles are used in caching on fixed networks, e.g., a proxy cache. A cache uses profiles to determine whether to deliver a cached object to the client or to download a fresh object from a remote server. I present an architecture for profiles that is both scalable and straightforward to implement at a cache. Next, I present techniques to improve estimates of the recency of cached objects by increasing the amount of information servers provide to caches. This improves the effectiveness of using latency-recency profiles by simultaneously reducing bandwidth consumption and increasing the recency of data delivered to clients. Finally, I describe how profiles can be used to support diverse applications in mobile environments. I present a framework for incorporating profile-based decision making into the cache utilization, downloading, and scheduling decisions at a mobile base station. Experiments with trace data and implementation results validate the effectiveness of using profiles.