Service Intelligence Oriented Distributed Data Stream Integration

TL;DR

This paper presents a comprehensive approach for integrating distributed data streams in service-oriented architectures, addressing synchronization, aggregation, and performance evaluation to meet real-time SaaS requirements.

Contribution

It introduces systematic strategies for relating distributed data, a novel small window array mechanism, CDF-based window parameter setting, and queuing models for performance prediction.

Findings

High accuracy and completeness in distributed data stream processing

Effective aggregation of distributed service data streams

Acceptable performance metrics in real-time scenarios

Abstract

Software as a service (SaaS) has recently enjoyed much attention as it makes the use of software more convenient and cost-effective. At the same time, the arising of users' expectation for high quality service such as real-time information or functionality provisioning brings about new challenges: to satisfy such (near) real-time requirements, real-time monitoring and effective processing of streaming data is necessary. However, due to the composition structure and multi-instance property of services, service data streams are often distributed, hard to synchronize and aggregate. We tackles these challenges by (1) proposing systematic associate strategies for relating distributed data; (2) introducing a new small window array mechanism for aggregating distributed data; (3) setting window parameters based on the cumulative distribution function (CDF) method; and (4) modeling streaming operators with queuing models for performance evaluation and prediction. Experiments show that our approach has good accuracy, completeness and acceptable performance measurement in processing distributed service data streams.