A Comparative Analysis of the Resource Consumption in Message Oriented Middleware with XML and Binary Encoded Messages

Abstract

The XML coded information is notorious for its verbosity over the actual effective payload. The solution for reducing the verbosity and the performance cost of XML has been the use of the alternative encoding schemes, other than pure text. The processing of the XML using computing resources sparingly is even more important in embedded, and otherwise restricted environments, when the services of the enterprise domain are extended to the mobile wireless domain. This paper presents and studies the empirical results of computational resource consumption by benchmarking of an implementation of the Message Oriented Middleware, MOM. This MOM solution defines and implements a service and application level messaging protocol and appropriate middleware. The benchmarked MOM implementation provides XML and two alternative binary encoding presentations for the messages, a proprietary binary encoding and an XBIS encoding. The message encoding schemes are compared by considering message compression, protocol overhead and measured computing resource usage. The measured computing resource usage includes CPU processing load and memory consumption. The results show the advantage of the binary encoding over the traditional XML, in most cases. However, the study also showed how XML is preferred for the lengthy messages that have a higher share of payload data.