An Approach to Automating Transaction Management in a Data Services Platform

In today's enterprises data lies across multiple sources in custom formats, often there exists no means to access (query and update) these data sources in a common enterprise wide uniform format. The requirement has led to emergence of data services platforms that allow real time access of data from multiple data sources. These platforms provide service oriented data access interfaces to both XA compliant and non XA compliant resources towards creation of data service interfaces. Specifically since these platforms provide SOA based shared access to data, these are termed as shared data services (SDS) platforms. One of the primary challenges in such a platform is managing access and transaction context across both XA compliant and non XA compliant resources. In this paper we discuss how these issues are presently addressed and propose enhancements over existing techniques to automate distributed transactions in SDS platforms. The proposed approach is showcased using a pilot application.