IT channel takeaway: One consultant warns not to assume enterprise information integration (EII) is a panacea; there are various gotchas to take into account when tackling an EII project.
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
With Michael Schiff, principal consultant for MAS Strategies.
Question: You recently wrote an article that was somewhat critical of enterprise information integration. What are some of the weaknesses of EII?
Schiff: The concept makes sense, but don't assume it's more than it is. Say you have to figure out how to pull historical data from multiple systems together. That's usually in a data warehouse, and it's usually cleansed. You have to put that data into some sort of consistent format. That's one of the things that a DW does that people sometimes overlook. To think that you're going to have the same definitions in multiple systems, like sex codes M and F or 0 and 1 — it's not going to happen. And it gets more complicated if you go to purchasing systems. You're not going to get a vendor to modify codes.
When you're going with an EII system, in theory, you're going to the operational systems as they exist today. But the units of value are going to be implicit, and they might be dollars, or pounds, or Euros or marks. When you start pulling these discrete values together in an EII system, you have to know that they're different units of measure and you have to do the conversion. This can be misleading.
Another thing that's nice about a data warehouse is that it gives you a point-in-time figure. The values from an EII system can be transient values that can change.
Question: Let me ask you the opposite question. What is the place of EII in the enterprise?
Schiff: Everything I said applies in a different way. EII gives you the "then-current" value. You might have the sales by month in your data warehouse. But let's assume one of these systems is storing month-to-date balances as it goes along. You can tell by looking at the operational systems where things stand at this point in time. And you could potentially compare that to what was in the data warehouse at the end of the month a year ago. So you could say, it's the middle of the month now, we should be halfway there. So I'm not saying EII is bad. It's just part of an overall information architecture.
Question: What about the concept of the virtual data warehouse, which seems to give the best of both worlds?
Schiff: In the old days — say, 10 years ago — people used to say, there's no need for a data warehouse with all the extra storage. We have products that will help you build a virtual data warehouse so you can look into your operational systems and pull that data together. The problem is, it's EII revisited. You don't have historical values. It's a dangerous concept, because it gives you the impression that you don't need a data warehouse.
This 3 Questions originally appeared in a weekly report from IT Business Edge.