Abstract
This chapter describes the design and implementation of a novel query engine that treats data and query streams analogously, and performs multiquery evaluation by joining them. This allows PSoup to support queries that require access to both data that arrived before the query specification, and data that appears after. PSoup also separates the computation of the results from their delivery by materializing the results; this allows PSoup to support disconnected operation. These two features enable data recharging and monitoring applications that intermittently connect to a server to retrieve the results of a query. This chapter describes techniques for sharing both computation and storage across different queries. Psoup is currently implemented as a main memory system. One would like to be able to archive data streams to disk and support queries over them. Disk based stores raise the possibility of swapping not only data, but also queries between disk and main memory. Swapping queries out of main memory would effectively deschedule them, and can be used as a scheduling mechanism if some queries are invoked much more frequently than others.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
