Latency-Aware Information Access with User-Directed Fetch Behaviour for Weakly-Connected Mobile Wireless Clients

Mobile wireless clients have highly variable network connectivity and available bandwidth. There are times when they may be completely disconnected from the larger Internet. This dynamism of connectivity poses unique constraints for the problem of information access in general, and Web access in particular. These and other factors (such as loaded servers and congested networks) contribute to unpredictably high response times in content retrieval. We examine the problem of improving the utility of information access applications for these imperfectly connected devices. In particular, we are concerned with three related problems: 1. Providing information to the user on the estimated response time to retrieve content, the freshness of cached content, and the status on the strength of connection to the network. This information allows the user to make informed decisions about which content to retrieve, and how to retrieve it. We propose that client applications should be latency-aware with location translucency. 2. Giving the user control over how the client application behaves when desired content is either not fresh or absent from the cache, for example, whether the user waits for the content, uses an out-of-date cached version, or both. We propose that clients should support user-directed fetch behaviour. 3. Keeping the content cache as current as practical. In addition to conventional techniques such as demand prefetching and push preloading, we propose to build a profile of user activity. During periods of strong connectivity, the client application should hoard to keep the cache fresh, so that content will be locally available when disconnected from the network. We describe the problem, through a motivating, application, and present our architecture, design, and prototype implementation, followed by a discussion of research issues to be explored. Our work draws on the experiences of designing mobile clients in other scenarios such as distributed file systems (e.g. Coda), as well the large body of work on Web caching and anticipation.