A Run-Time Based Technique to Optimize Queries in Distributed Internet Databases

Abstract

An adaptive probe-based optimization technique is developed and demonstrated in the context of an Internet-based distributed database environment. More and more common are database systems, which are distributed across servers communicating via the Internet where a query at a given site might require data from remote sites. Optimizing the response time of such queries is a challenging task due to the unpredictability of server 701 E. Chocolate Avenue, Hershey PA 17033-1240, USA Tel: 717/533-8845; Fax 717/533-8661; URL-http://www.idea-group.com IDEA GROUP PUBLISHING This chapter appears in the book, Advanced Topics in Database Research, edited by Keng Siau. Copyright © 2003, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Queries in Distributed Internet Databases 129 Copyright © 2003, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. performance and network traffic at the time of data shipment; this may result in the selection of an expensive query plan using a static query optimizer. We constructed an experimental setup consisting of two servers running the same DBMS connected via the Internet. Concentrating on join queries, we demonstrate how a static query optimizer might choose an expensive plan by mistake. This is due to the lack of a priori knowledge of the run-time environment, inaccurate statistical assumptions in size estimation, and neglecting the cost of remote method invocation. These shortcomings are addressed collectively by proposing a probing mechanism. Furthermore, we extend our mechanism with an adaptive technique that detects sub-optimality of a plan during query execution and attempts to switch to the cheapest plan while avoiding redundant work and imposing little overhead. We demonstrate that this probe technique can be extended in a client-server environment as a basis for choosing the right place for the execution of user defined functions (UDFs). An implementation of our run-time optimization technique for queries was constructed in the Java language and incorporated into an experimental setup. The results demonstrate the superiority of our probebased optimization over a static optimization. INTRODUCTION A distributed database is a collection of partially independent databases that share a common schema, and coordinates processing of non-local transactions. Processors communicate with one another through a communication network (Silberschatz, Korth, & Sudarshan, 1997; Yu & Meng, 1998). We focus on distributed database systems with sites running homogeneous software (i.e., database management system, DBMS) on heterogeneous hardware (e.g., PC and Unix workstations) connected via the Internet. The Internet databases are appropriate for organizations consisting of a number of almost independent suborganizations, such as a university with many departments or a bank with many branches. The idea is to partition data across multiple geographically or administratively distributed sites where each site runs an almost autonomous database system. In a distributed database system, some queries require the participation of multiple sites, each processing part of the query as well as transferring data back and forth among themselves. Since usually there is more than one plan to execute such a query, it is crucial to obtain the cost of each plan, which highly depends on the amount of participation by each site as well as the amount of data shipment between the sites. Assuming a private/dedicated network and servers, this cost can 32 more pages are available in the full version of this document, which may be purchased using the "Add to Cart" button on the publisher's webpage: www.igi-global.com/chapter/run-time-based-techniqueoptimize/4344