<<Over the past few years, many organizations have invested heavily in service oriented architecture (SOA) as an approach for creating a more agile, responsive and re-usable IT infrastructure that can respond effectively and efficiently to business requirements.
But as organizations continue to focus on SOA and building services that can be re-used and shared across different business processes, there's one issue that's often overlooked; the reliability of the underlying data. At one level, SOA is all about connectivity and building composite applications out of discrete services that can be leveraged across different business processes and managed both separately and together. But SOA is also about communication-about enabling those services to convey and leverage corporate information effectively across those different business areas. And effective communication requires consistent, reliable and accurate underlying data. Unfortunately, most organizations have been fighting a losing battle when it comes to being able to create and maintain consistent and reliable data across ever-more distributed business and IT systems. And while SOA-type data services can make data more accessible and even make access to that data more consistent, there's no guarantee that the underlying data itself is reliable or consistent. The opposite of reliable is unpredictable - Even if your data is "mostly reliable" the 3 records in a thousand that break the rules will screw everything else up and waste huge amounts of time - "mostly OK" is not the same as "guaranteed reliable" and the difference can cost you a lot. (you "mostly" don't crash your car, but you still buy insurance for the chance that you might - can you afford to 'crash' your systems with faulty data?) In effect, I believe that unless organizations pursing an SOA strategy give some solid thought to their underlying data, they're setting themselves up for a wide range of potential problems down the line. Think of the analogy of a company building a larger, fancier and more complete supermarket, with all types of easy entrances and checkout lanes, but with the fruits and vegetables displays all mixed up and mismatched. Or with canned goods stuffed into the freezers and milk sitting on the shelves next to the dog food. Building a better and more efficient supermarket requires not only attention to the overall structure, design, architecture and use of the store itself, but also close attention to the underlying goods and products that make up the basis for the store. Simply put, with SOA, thinking about the quality and consistency of your information is more important than ever before. In order to have an effective SOA strategy, an organization needs to make sure that the services they are creating are consuming and distributing consistent and reliable data. But doing that takes work. In order to ensure the reliability of the underlying data, and therefore of any of the SOA services that are consuming or distributing that data, you need to be able have a process or way to analyze the data and ensure consistency. You need to be able to trap data errors and exceptions before they disrupt downstream systems, and you need a way to identify and fill out missing information. Achieving better data reliability on a consistent basis takes more than manual effort or just hoping that all your information providers will automatically take care of their data. It takes great inline data quality monitoring that can detect when "unexpected" (such as missing or improperly structured data) is detected, as well as taking (or enabling someone) to take appropriate action (such as rejecting it, re-routing it for correction, or fixing it). Done right, SOA can make a big difference in a company's ability to respond to business change. However, when done without concern for the reliability of the underlying data, SOA can be like supercharging a used car to make it go a whole lot faster, but not replacing the worn out brakes or old unbalanced tires. You might be able to go from zero-to-sixty in half the time you used to, but when it comes time to slow down or turn the corner, you might not get the results you expected. Implementing an SOA strategy without putting best practices in place to ensure the reliability and consistency of the underlying data will do the same thing-it will allow an organization to build, share and re-use services and information a whole lot faster than ever before, as well as make business decisions more quickly than ever before. But those decisions may be more unbalanced, inconsistent and deeply flawed than ever before. Instead, organizations pursing an SOA strategy should make sure to give good consideration to their underlying data-from customer data to product data-and put systems in place that can help rationalize it and ensure its reliability.>> You can read this at: http://www.ebizq.net/hot_topics/soa/features/8090.html
