Does SOA add value to the Business Users at the end of the day?
In most of the cases we see that SOA initiatives are driven by enterprise architects. The reason they are funded in the first place could be because SOA promises reduction in overall IT costs! From the tools that are available today we have Service Registries, Web Services Management,BPEL, ESBs and Governance based products to name a few. No doubt that they play a key role towards enabling SOA but do the business users really benefit from these tools? Say from a “Line of Business representative’s” perspective one could possibly use the Service specific information provided by the WSM tools such as Service Faults, SLA breaches for the service, Transactions Per Second…But they are all very service specific! what happens to the data that is passing through it? The data which is all important and is instrumental in running the enterprise itself!!! Do we see a convergence here where the concepts of Business Activity Monitoring and Web Services Management could be merged to benefit the Business Users and attribute to the overall profitability?
This is an interesting question that parallels some of the things I’ve been looking at around business operations recently. I see the same line of thinking in GE’s Change Accelleration Program (CAP) and CMMI as well. I might be stretching a bit, but here goes…
The business isn’t just interested in doing things - it wants to know how well it is doing them. The common thread in CAP and CMMI is to make a plan and then monitor the execution of the plan. For operations you have dashboards and balanced scorecards.
The idea about SOA and BAM intersecting seems to match these. Each transaction is (hopefully) useful for some business purpose. The data inside the transaction is what tells you the meaning and importance. And from this meaning, you can derive if the business is on track.
The technical measures, like transaction errors and whatnot, are only important if they represent meaningful failures. But didn’t we go through this 5 and 10 years ago when people were measuring webpage “hits” as online success?
BPM can tell you about sales volume which matters. It can support JIT procurement - and alert when you get off track. It could track a healthcare company’s performance against best practice and outcomes for specific conditions, and let you know immediately if there appears to be a problem - not next quarter after 60 manual chart reviews.
One of the biggest issues with dashboards and balanced scorecards is getting the data; getting it timely, accurately, and cheaply. To the extent that web services monitoring can supplement the dashboard with meaningful metrics you have success. If you customer service is online and your web services show transaction failures with certain types of requests, this is valuable information. If it is bad enough, that data might make an operations dashboard to be tracked and improved. Isn’t that better than a backward-looking customer survey to track your problems?