Rate-Based Query Optimization for Streaming Information Sources

Query optimizers typically attempt to minimize response time. While this approach has been and continues to be very successful in traditional environments, in the presence of information sources of differing rates and non-blocking query execution something else is needed. In this paper, the authors propose a framework for rate-based optimization, in which the goal is to choose a plan that maximizes the rate at which answers are produced over a specified time interval. They demonstrate that different plans can have substantially different behavior with respect to the rate at which they produce tuples as a function of time.