Restful Web Service-oriented Searching Of Similar Data Objects

In2000, Roy T. F. accomplished his PHD thesis Architectural Style and the Designof Network-Based Software Architectures, in which he presented a new architecturalstyle called REST (Representational State Transfer). In following years, website, Webapplications and Web services based on REST have been booming. Meanwhile, with therise of NOSQL database, a great mass fervor on JSON is engaged, and more and moredevelopers have chosen JSON as a light-weighted format for data exchanging on theWeb. However, nowadays, people have more demand on the performance of webservice, including response time, rate of data transmission and etc. It brings a greatchanllenge on the Internet where incalculable requests and responses are generatedevery day. Therefore, in this thesis, we make a discussion on improving the performanceof RESTful application by reducing the response time and network traffic.Besides describing the background and the characteristics of RESTful Web serviceand analyzing the performance of JSON based RESTful Web Services, our contributionsinvolve the follows aspects:1. After analyzing related theories and structure of JSON data, we propose aneffective data structure and define a presentation model for JSON.2. According to the relevance between JSON data and JAVA data, we define aconceptional model and implementational model to transfer JSON objects toJava objects and propose the algorithms for serializing and deserializing dataobjects when programming with Java.3. We propose a Bloom Filter-based structural similarity calculating algorithm tocompute the similarity of different JSON data objects effectively andaccurately.4. After analyzing similar JSON data objects, a dataID-based compression anddecompression algorithm is presented to decrease the network traffic andreconstruct the original data in an efficient way.