Benchmarking System untuk Mengevaluasi Waktu Eksekusi dan Dukungan XQuery pada Engine XQuery Berbasis Java

Abstract

XML retrieval, based on the document to be retrieved, is divided into two different types: text-centric and data-centric. Data-centric XML retrieval is marked by strong emphasis on the document structure. These characteristics lead to a question on how a data retrieval system should handle the document. XQuery has come as a solution for this type of XML retrieval. XQuery is a query language for XML to extract data, similar to the role of SQL for databases. Developed as a W3C Standard, XQuery is now emerging into de facto standard for querying XML. data. There are already many implementations of XQuery as of now but how these implementations perform the task and how they compare each other is still barely known. This research tries to bring the matter by benchmarking some of file-processor based XQuery implementations with the hope that the result can be both a recommendation to users and inputs to developers. Using XMark as the benchmark framework, we have develop a benchmarking system to test five engines, i.e., MXQuery, Qexo, Saxon, Xbird, and XQEngine; which all of them are free and open source and based on Java programming language. The result of the benchmarking system has showed that not every engine can finish the entire benchmark query within the given time limit. Despite the partial XQuery support from some engines, a number of query executions can be finished relatively fast. For some engines and on certain queries, the growth ratio of document size to query execution time is relatively low, which is favourable when bigger documents had to be dealt.