Learning To Love IO
Infrastructure optimization is more than just another way for Microsoft to pitch its products.
I first heard about IO on a trip some months back to Microsoft. They weren't
talking about I/O as in input/output. No, they were talking about infrastructure
This IT model is how Microsoft now positions, sells and develops products.
I reckoned I'd better learn what this thing was all about.
Coming to grips with IO was a long process. I read several hundred pages worth
of white papers, case studies, briefs and press releases. That was a mistake,
for as I tried to absorb the whole model -- there are actually three of them
-- and all its implications and product ties, I scratched my head so much my
co-workers got me a flea collar.
Here's the short version of what I've learned. IO is an IT maturity model that
first focuses on efficiency and driving down costs. As you move up the ladder
the focus turns to technology that helps companies do new things, launching
products and responding quickly to changing markets.
The first model is Core Infrastructure, meaning your basic server and PC architectures.
Then comes Business Productivity Infrastructure and finally Application Platform
Each model works the same way. IT goes through an elaborate questionnaire that
leads to a set of goals for the shop. In the last part, Microsoft maps your
business goals to products -- yes, in most cases, Microsoft products.
When I first looked at IO I fixated on the product mapping and decided the
model was just a way to sell more Microsoft products. I was biased against IO
from the start. The first IT pros I spoke with bashed IO for the same reason
-- they just didn't trust it.
I kept going, talking to more customers, analysts, partners and a few rather
persuasive Microsoft execs. IO is a way to sell more Microsoft products, and
the company is not ashamed of pitching its products to fit IT goals established
by IO. Once IT has those goals, it's free to seek out any product or vendor
it cares to. I also found that in the first two stages of IO, the questionnaires
and goal setting are almost entirely vendor neutral.
It took a lot of reading and talking and looking at other models, but now I'm
convinced that IO is a good thing -- a very good thing. But good isn't perfect.
Let me play analyst for a moment and humbly -- and when have you ever seen me
humble? -- suggest some changes.
- An "IO for Dummies"-style introduction: A document that lays the
model out so the basics can be understood in about five minutes.
- More work with third parties: When it comes to product mapping, I'd like
to see as many third parties as possible represented.
- An upfront discussion of competitive technologies and multi-vendor shops.
- Better do-it-yourself tools for IT, such as a self-service portal so IT
can do their own IO analysis.
- Microsoft really should work with IBM, which has been an expert in infrastructure
optimization for years.
Give me your take on IO by writing to firstname.lastname@example.org.
Doug Barney is editor in chief of Redmond magazine and the VP, editorial director of Redmond Media Group.