SOA efforts frequently enjoy early success creating relatively straightforward services that happen to have few
side effects, perhaps related to processing a payment or creating a new account. At some point, they realize a need for a data service, for example one that wraps customer account information. The usual challenges are not so much SOA related, but the SOA effort hits a snag because there are multiple opinions in the business about the meaning of data elements (what is a customer, exactly?), and because there is very likely many years of ignoring or working around data quality problems that has to be addressed. In other words, the SOA project can't succeed without cleaning up the data quality and definition issues.
My best advice in such situations is to take it in steps - don't try to create the complete data service in one lump, but instead start with one specific need and expand from there. The other likely challenge is one that is common to early SOA efforts - you will need some practice at defining services well. In other words, you should plan on some rework of initial services as you get smarter after the first year or so.
Dig deeper on Web data services and enterprise mashups
Related Q&A from Larry Fulton
Data services expert Larry Fulton discusses what happens when a business has multiple sets of data representations and how to go about narrowing them...continue reading
Larry Fulton discusses how it is inevitable that access to the many data related capabilities emerging in support of SOA be equally accessible via an...continue reading
Larry Fulton discusses best and worst practices for the ever changing core data relationships within data abstraction.continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.