lustratus LiteBytes

IT.Director.com has an…

…interesting article discussing the IT industry’s attempts to speed up SOA adoption by identifying the killer apps that will drive business execs to make the required investments.

Taking the liberty of paraphrasing, the two killer apps candidates discussed are Business Activity Monitoring (BAM) and Software as a Service (SaaS). Now, I absolutely agree that BAM can be a significant business driver, presenting as it does a greatly increased visibility into business operations, often available real-time. The combination of this visibility with management tools to set and monitor exceptions and key performance indicators can be very appealing to business execs. However, I must point out that BAM isn’t actually anything to do with SOA. Sure, an SOA architecture provides an ideal basis for BAM, exposing IT activities in business services terms, but BAM can also be achieved a number of other ways, such as through Business Intelligence (BI) based solutions like those from Information Builders and Business Objects.

On the SaaS front, the key point is that operational processes must be of a level of granularity and ease of access that it is easy to take a particular process or set of processes and effectively outsource them by using a hosted service. In this case, the suitability of SOA is even stronger, largely because not only does it present clean interfaces but also they are standards-based.

However, It is unclear to me that these two angles will drive enterprise-wide take-up of SOA. They might be a way to get a foot in the door, but in the end I believe that the investment needed to drive widespread use of SOA will need to have a mixture of drivers. Reuse may be boring, but within the IT budget it matters, and the resulting faster time to market and associated agility is also important. The visibility point is strong, but since there are other ways to achieve it I think the key will be to focus on the fact that SOA enables MANY killer app opportunities by making IT investments more flexible, responsive, productive and better aligned with business needs, and that it does so in a much easier, faster and less risky way.

The summary in my mind is that although excitement will continue to grow around SOA, all-consuming, enterprise-wide deployments through 2007 will still be rare. A lot of toes will be dipped in the water, pilots will be carried out, individual projects will be implemented – but we are still some way off mass adoption.

Steve