Self-tuning query mesh for adaptive multi-route query processing
In real-life applications, different subsets of data may have distinct statistical properties, e.g., various websites may have diverse visitation rates, different categories of stocks may have dissimilar price fluctuation patterns. For such applications, it can be fruitful to eliminate the commonly made single execution plan assumption and instead execute a query using several plans, each optimally serving a subset of data with particular statistical properties. Furthermore, in dynamic environments, data properties may change continuously, thus calling for adaptivity. The intriguing question is: can we have an execution strategy that (1) is plan-based to leverage on all the benefits of traditional plan-based systems, (2) supports multiple plans each customized for different subset of data, and yet (3) is as adaptive as "plan-less" systems like Eddies? While the recently proposed Query Mesh (QM) approach provides a foundation for such an execution paradigm, it does not address the question of adaptivity required for highly dynamic environments. In this work, we fill this gap by proposing a Self-Tuning Query Mesh (ST-QM) --- an adaptive solution for content-based multi-plan execution engines. ST-QM addresses adaptive query processing by abstracting it as a concept drift problem --- a well-known subject in machine learning. Such abstraction allows to discard adaptivity candidates (i.e., the cases indicating a change in the environment) early in the process if they are insignificant or not "worthwhile" to adapt to, and thus minimize the adaptivity overhead. A unique feature of our aproach is that all logical transformations to the execution strategy get translated into a single inexpensive physical operation --- the classifier change. Our experimental evaluation using a continuous query engine shows the performance benefits of ST-QM approach over the alternatives, namely the non-adaptive and the Eddies-based solutions.
distinct statistical properties, fluctuation patterns, single execution plan, Query Mesh, Self-Tuning Query Mesh, concept drift problem, single inexpensive physical operation
Date of this Version