Service Oriented Architecture (SOA) is not just an enabler of business process and data integration. It allows your business intelligence engine to function as a single entity rather than multiple components.
Historically, you’ve always had a reporting/query tool, an OLAP tool, and a dashboard/scorecard tool. In some cases, each of these may have had multiple components themselves. Add the fact that you might have different products functioning in your operating units, and the number of BI vendors and tools increase.
SOA really allows a single tool to drill down to the multiple data layers through the SOA middleware which acts as an integration point. From the UI level (assuming you have a single portal) the integration and interaction is seamless. Driving through the middleware where the directory access, security, data permissions, etc are controlled, a user with access to reporting may have started off at the same portlet as a user accessing data cubes, but they wind up vastly different tools with vastly different presentations returned to them.
The value of scorecards and dashboards is that they should be presented at the user level with an enterprise wide comparator in mind. Your beautiful scorecards are meaningless if the baseline for measurement to strategy changes based on the business unit. Sure, some scorecards are meaningful at a business unit or cost center level only, but if you are unable to get to enterprise wide comparators because of a lack of integration, you’re in trouble. SOA solves this.
One response to “Changing the Game: SOA and Business Intelligence”
Great series of posts. SOA’s impact on BI is an interesting topic. I presented at knowledge management conference a few weeks ago. One of the speakers repeated the frequently cited assertion that upwards of 80 percent of enterprise knowledge is in unstructured formats such as e-mail, word docs, and blogs. The idea of getting all throwing all this stuff in data warehouses and mining it as a source of who knows what as well as problems, trends, and opportunities was briefly discussed. I guess the power of data mining is its power to put together and discover relationships that aren’t obvious in a top-down way — but the suggestion struck me as a little off base — particularly when many enterprises have structured sources of data that they just aren’t connecting. Learning management systems, employee performance management systems, project management systems all are potential sources of data about who knows what and what gaps and challenges exist. Agile integration techniques such as web services and Ajax have the potential to link structured, verified information with the unstructured (such as blogs). As you envision, the impact of these agile integration approaches could be to give executives and managers new UIs with which to navigate through their organizations to better understand things such as knows what and problems, trends, and opportunities facing the enterprise.