Time Constrained Push Strategies in Client-Server Databases

• Authors: Vinay Kanitkar;Alex Delis
• Venue: Distributed and Parallel Databases
• Year: 2001

The widespread availability of networked environments and the arrival of high-speed networks have rekindled interest in the area of automatic data refresh/update mechanisms. In many application areas, the updated information has a limited period of usefulness. Therefore, the development of systems and protocols that can handle such update tasks within predefined deadlines is required. In this paper, we propose and evaluate two real-time update-propagation mechanisms in a client-server environment. The fundamental difference in these two time-constrained techniques, Client–Push and Server-Push, is in the location where the push-transactions are generated. In both these techniques, and in contrast to conventional methods, we propose the transport of the scripts of updating transactions in order to make client-cached data current. This avoids unnecessary shipments of data over the network. Instead, messages are used to maintain the consistency of cached data. In addition, the propagation of update transaction scripts to client sites is neither periodic nor mandatory, but is instead based on client-specific criteria. These criteria depend on the content of the database objects being updated. We carry out a comprehensive experimental evaluation of the suggested methods as we examine the following aspects: (a) time constrained push scheduling issues, (b) effects of various workloads on real-time push-transaction completion rates (efficiency), and (c) overheads imposed by push-transactions on the regular transaction processing. Our experiments show that Client–Push outperforms Server-Push only for a small number of clients. The opposite is true once the load is increased by attaching a large number of sites per server. The efficiency of the update push protocols is, as expected, dependent on the load on the system as well as the percentage of updates to the database. Surprisingly, the percentage of successfully completed real-time push-transactions is not affected very much by the strategy used to schedule them.