DEVELOPMENT OF A PROTOTYPE OF EVENT STREAM PROCESSING ACTIVITY LOGGER FOR MOBILE APPLICATION.

Abstract

One of the critical requirements of Yellow Pages Mobile Applications is the ability to process and to analyse big data acquired from massive activity logs generated by Yellow Pages Mobile applications. Conventional computing and storage model is no longer compatible to handle big data processing . This research is aimed to develop a proptoype of event stream processing activity logger for mobile applications that can be implemented and aligned with current technological developments. The prototype development approach is based on Software Development Life Cycle that provides sistematic incremental development stages to deliver a final prototype of event stream processing. In the implementation stage Apache Kafka, Apache Storm and Apache Cassandra are used to support messaging system, distributed computing system, and media storage system respectively. The developed prototype was tested using two-month data logs of 50-gigabyte log activity from 48.880 transaction activity records. The transactions come from 89 registered members and 202 anonymous users. The result of this research shows that there are 1.998 times of performance improvement compared to the current existing system. This significant better performance promotes the utilization of the developed prototype of event stream processing logger technology as a potential substitute for the current existing processing system,