The use of mobile computers is gaining popularity. The number of users with laptops and notebooks is increasing and this trend is likely to continue in the future where the number of mobile clients will far exceed the number of traditional fixed clients or desktops. Applications running on mobile clients download information by periodically connecting to repositories of data. Mobile clients constitute a new and different kind of workload and exhibit different access patterns than those seen in traditional client server systems. In this paper, we argue with particular emphasis on databases, that the workload presented by mobile clients allows partitioning of a dataset on a server. The partitions are defined by constraints on specific attributes of the data called the hoard attributes. The techniques are general enough to be applied to other non-database applications that access collections of data, including filesystems such as CODA and FICUS. We present simulation results that ill...
Shirish Hemant Phatak, B. R. Badrinath