address book, a replacement for an existing application. The agency describes it as “half in” but not yet well integrated or documented. They will need an ESB and a Registry for that.
In addition to developing services and emphasizing reuse and governance, the agency has been creating integration points for legacy applications. They’d like to finish that part of their effort and have consistent entry points for future applications in the SOA style.
The agency emphasizes that so far SOA has been a migration not a Big Bang project. They believe that they will need such a large scale, across-the organization project, to justify investments in more governance and an ESB. They note that paying for the ESB itself (software) is not the issue, but rather the expense of negotiating and writing the policies which will make the ESB a valuable piece of the SOA environment, providing connectivity across assets and services.
A candidate for a large-scale project that could justify the ESB is a system that will provide an interface to their agency services to all of the businesses which are their customers. Here they are in preliminary design mode in an SOA style. If the ESB is approved this year and implemented in 2007, it will be part of the implementation. The project, intended to replace a legacy system, is scheduled to go into its coding phase in late 2007.
The architect emphasized the need to draw a visible line of sight between IT investments and business needs. SOA does that but there are costs to get started, including the investment in an ESB. They are considering these issues:
What funding model should they use? Are shared SOA assets an organizational investment or must they be sponsored by a business unit? If the first users pay the price for those who get the future use of the asset, where’s his incentive?
Today, business applications live in business unit stovepipes and because each business unit has gotten good service from IT they are happy continuing with that strategy. How do you get to an enterprise-wide view?
How to get LOB (Line of Business) managers to change from wanting to invest in their own silos to investing in across-the company SOA investments. The agency has sought the endorsement of its vice president – and received it – but getting the LOB heads to hear that message isn’t easy, especially when they’re satisfied with the current state of things. Our architect is looking for a project with cross-department appeal.
Promoting Reuse in a multi-location insurance business
A large international insurance company provides sales and customer service activities through more than 150 local offices, geographically dispersed worldwide. They have decided to use SOA to manage their customer application, rate quotation, and contract processing workflow. The project will start with a pilot project in three North American offices located near their U.S. headquarters.
The headquarters business analysts, IT, and local office management are in agreement that the customer application process requires substantial attention. It was identified as a key opportunity in an SOA evaluation using the Component Business Model (CBM). A new process has been designed, using SOA concepts, as a series of web services, and implementation in the pilot sites is under way.
In just a few weeks managers in other locations were inquiring as to when they might have the updated process, too. IT is concerned that some ambitious managers may try to start their own projects, using local resources. That would be unfortunate, because the whole idea is to use the experience of the first pilots to fine tune the new applications before an orderly rollout through the
October 18, 2006