You know how everything in fashion makes a comeback a generation later? Integration technology is sort of like that, but the turnaround is much, much faster.
For instance, it’s been 30 years and we’re just now seeing skinny ties and skinny jeans make a comeback from the 80s.
But it’s only been about 10-15 years since EAI was the big “it” approach for handling integration. And guess what: It’s back in vogue, thanks to the integration challenges of cloud and mobile applications, as well as the focus on integrating business processes between different applications.
But it’s also important to learn from the past. Skinny jeans and ties are OK, but who would willingly resurrect the mullet or David Lee Roth’s bottom-less chaps? (Yes, kids, these horrors really happened. Oh, the humanity!)
He looks at the five most common reasons EAI projects fail — aka, the anti-patterns for EAI. He goes into detail about each, but, in brief, the five mistakes are:
1. Trying to integrate different information systems instead of creating a unified integration enterprise system. I would say that’s the biggest mistake with integration in general: thinking about it as connecting this to that rather than building a more general solution.
2. Believing EAI is a tool. It’s not something you buy. It’s a “system that plans enterprise architecture, business processes and organizational issues.” I take this to mean it should evolve from your enterprise architecture.
3. Trying to build or replicate what you do not need.
4. Dividing applications into parts and then integrating.
5. Using EAI for enterprise storage and warehouses. In other words, EAI is not the same thing as data integration, and you shouldn’t confuse the two. (Although, there is some convergence in the data integration and EAI capabilities.)
Kropyvnytskyy isn’t the only one to make a list of EAI anti-patterns, of course. EAIIC European Chairman Steve Craggs outlined a list of seven common mistakes with EAI back in 2003. Craggs called these “bear traps,” because they trip you up.
You could extend that metaphor as a warning against more dire consequences, since bear traps will also clamp down on you, making it impossible for you to escape without maiming yourself. But maybe I’m being dramatic.
Craggs’ list also includes “Thinking of EAI as a tool as opposed to a system” as a common mistake. So you’ve been warned. Twice.
Another nice point about Kropyvnytskyy’s piece is he explains the different integration options before talking about EAI. So if you have staff members who are unclear on the options for whatever reason, this would be a good piece to share.