Integration May Not Be Cool, But It Pays

Loraine Lawson

Integration isn't exactly what you'd call a cool technology.


Lately, it seems to be becoming even more uncool. Maybe it's just the summertime blues, maybe it's an editorial hangover from the Fourth of July holiday -- but I'm having a hard time finding anything substantial about integration.


This is especially true since the SOA folks decided to distance themselves from the integration discussion. I understand their point -- SOA looked to become something of a synonym for better integration and clearly, integration is only a part of SOA's value proposition.


But at the same time, it seems a bit mean, somehow -- especially when so many companies are embracing SOA precisely because they believe it can help them with integration. It also seems a bit stubbornly purist, since research shows this approach actually pays off in the real world.


I can also understand why integration isn't such a hot topic in the general trade press. Let's face it -- it's hard to do, it tends toward the situation-specific and it can be very technical. It just doesn't have the allure of, say, an iPhone or virtual servers.


To put it bluntly, integration is a bit boring.


But, man, can it pay off. And that's why it's still worth discussing.


Take, for instance, this item about a recent Aite Group study on how technology integration impacted investment advisory firms.


As it turns out, if you're an investment adviser, you want to spend a big chunk of your time with your clients. If you're anything like me, you prefer to picture your financial adviser as worrying over the performance of your -- and only your - personal portfolio, but apparently, that's not how they want to spend their time. Go figure.


Anyway, on average, advisers spend 34 percent of their time on client-facing activities, and 25 percent on portfolio management. The rest is frittered away on administration stuff, compliance and marketing.


But the picture changes if IT does its job and creates tight integration among the supporting applications used by adviser:

"Advisers who are supported by a fully integrated technology environment spend the majority of their time with clients, about 50 percent more time than the average adviser. At the same time, a fully integrated technology environment reduces advisers' administration burden by 33 percent and their compliance efforts by 30 percent."

When you see numbers like that, you realize just how big an impact integration can make.

Add Comment      Leave a comment on this blog post
Jul 15, 2008 11:37 AM Rob Eamon Rob Eamon  says:
Nice article. Integration is indeed a dirty word. It is assumed that integration is a tactical activity. It is assumed that past integration projects gone wrong are reprentative of the majority of projects. It is assumed that SOA eliminates integration.I argue that SOA embraces integration as a core value. Services from differing ownership domains interacting is, almost by defintion, integration. A service consumer, which might be an end user application, interacting with one or more service providers is integrated with those services. Loosely coupled? Hopefully. Can a provider be swapped out for another without consumer change? Theoretically.SOA isn't enterprise architecture, IMO, though that seems to be becoming the consensus view. SOA is a style of architecture that is applied to the levels we already had. An EA can be SO. And it will be other things as well (e.g. event driven, document oriented, etc). SO can be applied to any level of architecture (it started out as an AA notion, per Gartner's definition). SO principles can be applied to an integration architecture. Or a business architecture.From your interview with Schmelzer:"The irony of this is the business fully understands the SOA perspective. Theyre like, "Yeah, Ive spent 20 million dollars on this SalesForce thing, and I spent another 15 million dollars on this portal, why are you telling me its going to take me another $8 million to get my $20 million thing to talk to my $15 million thing?"I'll ask the obvious question: how does SOA solve that? Let's assume SalesForce is structured as services. Let's assume too that the portal is structured to leverage services as a consumer. Is Schmelzer suggesting that these two things can interact without any work? That's a pretty big leap, I would think.Integration is here for the long haul. And your points about the payoffs provide evidence that that's okay. Reply
Jul 15, 2008 11:49 AM Rob Eamon Rob Eamon  says:
From the interview with Schmelzer:" can give me a service and I can consume that service.""...building the applications in the first place to be proper citizens in an environment where we have lots of other services."" we can make these capabilities be broadly consumable in environments where we dont even know how theyll be consumed.""...stop building systems that are islands onto themselves so that you have to solve problems later on just to get them to communicate."In other words, make them "integrate-able" from the get go. Expose interfaces (a classic integration activity) that others can consume. And since you have no idea who those others are, the service and its interface(s) better be flexible, have broad use and not make any assumptions about those others.He seems to make a compelling case that SOA is really "integration done right." :-) Reply

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.