Enterprise data: A key factor for a successful SOA implementation

I find it very interesting to know more about service oriented architecture, since it is an architectural style the creates business processes as packaged service throughout their life cycle. One beautiful feature of SOA is, It is loosely coupled with operating system and programming languages underlying it.

OOP VS SOA

OOP is a fine way to create applications that perform specific tasks. However, it is a terrible architecture for creating robust network applications. Every OOP-based architecture for distributed applications (using CORBA, DCOM, or EJB) has failed to deliver on its promises... and made network programming vastly more difficult than it needed to be.

A service-oriented approach, on the other hand, gives the developers and architects an easy way to integrate systems. They are far less complex, faster, more robust, more scalable, more easily maintained, and they are much easier to for designers to conceptualize.

Having looked at the advantages of SOA we need to discuss how to avoid the hidden pitfalls while implementing an application using SOA.

SOA holds great promise as the enabler for a new generation of more agile and cost-effective IT solutions. Unfortunately, it has failed so far to live up to these high expectations. That’s because, in their zeal to realize the benefits promised by early SOA hype, customers and technology providers have overlooked a crucial component for the success of any SOA initiative. That is, they’ve ignored the data.

LACK OF DATA ACCURACY AND CONSISTENCY

SOA offers a way to decompose enterprise applications into reusable and sharable “services,” bringing agility to business processes by eliminating the walls between applications. But in breaking down these silos, SOA forces the data contained within to be exchanged seamlessly between disparate applications and business processes. This free-flowing exchange
requires a high degree of data accuracy and consistency, but SOA technology providers have largely ignored this need for a holistic approach to data management and integration. This was one of the main issues that came to light during an
extensive survey and in-depth interviews with IT executives on SOA implementations, conducted by IDG Research Services.

DATA CENTRIC CHALLENGES ON ROAD TO SOA

• Heterogeneous data sources distributed across the enterprise and beyond
• Inconsistent and constantly changing data structures
  
• Poor data quality that is often difficult to measure or monitor
  
• Lack of agreement or visibility (single-view) into critical information assets
  

Definitely data is the key driver in Service Oriented architecture, don't you feel so??