Section 12.3. Four Pillars for Success

12.3. Four Pillars for Success

Although a wide variety of factors determines the success of an enterprise's SOA strategy, four main pillars for success can be highlighted: securing a budget, choosing a suitable project to start with, setting up a strong SOA team, and finding backers and buddies (see Figure 12-4).

Figure 12-4. Four pillars of an SOA's success: budget, project, team, and buddies.

12.3.1. BUDGET

Obviously, securing a budget is a sine qua non for any successful introduction of new technology within an enterprise. For one thing, this budget will be needed to finance one or more initial projects acting as pilot applications of the SOA. Because the success of these projects will have a considerable impact on the (perceived) success of the overall SOA introduction, they should be chosen with great care, as we will explain in detail later. It is also crucial that they are equipped with sufficient budget to meet the challenges inherent in the use of new technologies, methodologies, and processes.

In addition, a budget is needed to compensate for initial overheads caused by applying an SOA. Such overheads are caused by different factors. For one thing, employees have to familiarize themselves with new standards and processes. More important, however, is the initial overhead caused by efforts required to increase reusability. Instead of merely focusing on immediate requirements, potential future applications must be taken into account to ensure that the implemented service is reusable.

Even if a business department is supportive of an SOA in principle, it might have problems providing the resources needed to account for this overhead. In such a case, it is good to have a budget available to cover the necessary resources.

12.3.2. INITIAL PROJECT

The second pillar is the choice of a suitable project piloting the implementation of the enterprise SOA.^[1] You must take into account several criteria to determine good candidates. First, the chosen project should be as visible as possible. This does not necessarily mean that it has to be particularly prestigious. On the contrary, choosing a less prestigious project might be advantageous because it diminishes the immediate pressure during the project's realization. However, the functionality provided by the implemented services should be widely used in the enterprise. On one hand, this ensures that the results achieved in the project will be highly visible. On the other hand, it will guarantee a significant reuse potential of the implemented services, which in turn will contribute to the validation of the benefits of the SOA and will help to sell it.

^[1] Obviously, more than one project might be chosen to pilot the SOA. For ease of presentation, our discussion will focus on a single pilot project.

Ideally, the project should run no longer than two or three years, with a first delivery after six months. There should be a clear technological scope based on equally clear business requirements. In fact, it is crucial that the project have a business focus with measurable benefits instead of just being technology-driven. This not only concerns the project as a whole but also holds true for the individual services developed in the project. The more obvious the benefit of these services, the easier it will be to prove the SOA's ROI.

It is also a good idea to carefully evaluate the business department that will be responsible for the realization of the pilot project. Ideally, it should be enthusiastic toward the SOA idea, but at the very least, it should be open and positively biased. Otherwise, you risk too much friction during the delivery of the pilot scheme, which might subsequently jeopardize the entire SOA endeavor.

12.3.3. SOA TEAM

The third pillar is setting up a special SOA team. Such a team should focus exclusively on how to best support and establish the SOA in the enterprise. Naturally, this includes first and foremost the design and specification of overall architecture principles, standards, and processes as well as the careful monitoring of the actual application of the SOA.

The SOA team will therefore have to contain evangelists whose task it is to explain the benefits of the SOA to the different departments of the enterprise. This includes education on the fundamental standards and processes making up the SOA, as well as support in the actual implementation of these principles. The SOA team will also play an important role in setting up organizational structures within the enterprise to monitor the "correct" application of the SOA.

12.3.4. BACKERS

Finally, the fourth pillar consists of having backers and buddies. This is important at all levels, beginning with active support by top management. It is crucial to acquire budget approval and sufficient enterprise-wide awareness to get the SOA off the ground. This support will also come in handy should there be substantial opposition to the SOA as a whole or to some of the various processes and standards necessary to establish it. It is therefore important for top management to explicitly include the SOA in the strategy planning, covering a three- to five-year perspective.

Equally important is the backing of the key business departments. After all, business-driven projects will implement and use service functionality. If the departments providing and consuming the business functionality are not supportive of the SOA, it will be next to impossible to successfully establish it. Furthermore, while an SOA is not primarily a technical issue, it is infeasible to implement it without assistance from key IT management and staff.

You should always keep in mind that backers have their own interests and that they have to sell the SOA to others. In order to successfully introduce an SOA, it is therefore crucial to have arguments for various target groups, for example in the form of fact sheets or elevator pitches (see Chapter 11, "Motivation and Benefits"). On one hand, these pitches should be used to convince relevant actors in the enterprise to become supporters. On the other hand, they should be made available to supporters to facilitate their own selling of the SOA to others. The easier it is for backers to sell the SOA, the more likely it is that they will support it.

Thus, before approaching potential supporters, the following questions should be carefully answered:

• Who are the target groups and key actors?

• What are the three main arguments for and against an SOA (per target group/key actor)?

• Whose support will be likely?

• Who is to be convinced or overruled?