Wikipedia defines SOA as - an architecture that relies on service-orientation as its fundamental design principle.
Service-orientation describes an architecture that uses loosely coupled services to support the requirements of business processes and users. Resources on a network in an SOA environment are made available as independent services that can be accessed without knowledge of their underlying platform implementation.
SOA - The Challenge
So essentially SOA means integration rather than new build and as such it faces the same set of technical challenges that troubled the preceding generation of EAI & EII projects.
Specifically; as the number of point-to-point integrations increase, so does development complexity and the potential for systems fragility and higher maintenance costs.
Given that SOA is a set of architectural principles, a wide range of toolsets, both new and adapted, will be positioned by vendors as the solution to these development challenges.
But designing & building new services from existing applications is fundamentally dependent upon understanding, in meaningful detail, how these established Enterprise Applications operate and store their data.
Even when these data sources are understood in technical terms, it still leaves the considerable semantic problem of bridging to a common set of business terms.
Essentially these are data modeling, or metadata management issues and as such SOA raises the importance of such skills, techniques and tools.
Enterprise Applications – adding to the
SOA Data Modeling Challenge
There is no immediate quick fix to this metadata challenge - it is an ongoing process of discovery, documentation and debate between technical & business communities in the organisation.
To add to this challenge Enterprise Applications such as SAP, Siebel, PeopleSoft etc present unique problems for data modelers because of the complexity & opaqueness of their data architectures.
Given that such packages are now probably the major delivery mechanism of corporate business processes and sources of data, it is critical that they can be integrated in any data modeling or metadata management strategy in support of SOA initiatives.
Most EA vendors will have an approach to SOA but Saphir technology remains the only toolset dedicated to delivering such ERP metadata intelligence and directly interfacing to the leading data modeling tools.
But don’t just take our word for it
Gartner advises organizations wishing to fully exploit service-oriented business applications to start by focusing on integrating the processes and underlying data models, rather than on integrating individual application components.
BEA as a leading middleware vendor says of Enterprise-Wide Data Models “A service infrastructure must address the issue of data models. There are often as many data models as there are enterprise applications in the company. For example, a billing system, inventory system and product catalog all have different ideas of what a product looks like. This is clearly a barrier to the free flow of information between these systems.”
|
