This week in Las Vegas at Gartner’s SOA / ADI Event I hosted a Analyst / User Roundtable on the topic of Cloud computing / SaaS integration.
One attendee which has already (impressively) leveraged the services of a dozen SaaS vendors claims that four of the twelve providers have changed API’s without notifcation, impacting production operations. A few other members of the group cited some experience with what I’ll coin here (strictly for the entertainment value, of course) “API slamming”.
By any interpretation API slamming is not good. While I understand how this can happen at a technical level, in the SaaS user’s defense I don’t understand why any SaaS provider would change its runtime API’s without notifying users (er, in advance, please). In the SaaS provider’s defense, I can imagine that implementing a proactive notifcation mechanism quickly becomes non-trivial, e.g., just maintaining user contact information is problematic
When I said that a leading SaaS vendor claimed to take a different approach, since they told me it “never” changed its API’s and instead just revisions them and keeps the old — the skepticism around the table was palatable.
This group of users considers “API Slamming” to be a Cloud computing / SaaS integration “risk and concern”.
So what do you think?
Is it possible that I just happened to run across a “tough crowd” and an associated exceptional but perhaps not so representative data point on API slamming? Or is it possible that I’ve just uncovered an ugly little secret (and risk) of Cloud computing / SaaS integration?