The (Un)Economics of Infrastructure Support


It's a fact of life in IT that things break. As reliable as hardware is, over time there is a "natural" failure rate for components, sub-systems and the systems that are assembled from them.

Nothing lasts forever, and we have to be prepared for the inevitable, hopefully infrequent, failures. High availability designs will ensure that no one failure can take us down hard, but we'll still need to eventually replace failed parts. That's why we have maintenance and support contracts.

Enterprise-class hardware increasingly comes with continuous monitoring and predictive failure capabilities. The boxes can be set to warn you when things are looking suspect--or even phone home and request an engineer to swap out a failing part before it's an emergency. Manufacturers and third-party support services want to avoid unscheduled truck rolls if at all possible, and they're getting pretty good at the preventive aspects of support. If you are comfortable with the necessary connectivity and can integrate the vendors' systems with your own service desk processes, things on the hardware side can run pretty smoothly.

And then there's software.

One of the most frustrating costs of ownership to manage for enterprise IT infrastructure is the annual fees we pay to software vendors to fix bugs, add required features (usually the reasons we selected their product in the first place) improve performance, and so on. It takes a lot of software to run a modern large business, and the aggregate cost of support and maintenance is material--maybe too much so.

Add in the wide range or seemingly arbitrary pricing models (CPU-based, core-based, power-based, user-based, etc.) and you can't really know if you're getting a good deal (unlikely) or getting screwed (probably).

An account rep for one of the major software vendors told me last year that their CEO had issued an edict that they could negotiate on license fees to win business, but could not touch the support contract pricing. Too much reliable profit was at risk from the stream of support revenue flows.

I have come to the conclusion that this model is close to unsupportable. At the "average" support rate of 22 percent of list price (OK, so sometimes you can get it down closed to 22 percent of license cost) I pay for the software again every four years or so. If I keep using it for 20 years--not unusual in the enterprise--I will have paid for it six times over. Sure, I'll have new features and better performance and (hopefully) fewer bugs, but I'm really "renting"--not buying--at this rate.

As hard as I have tried to get that 22 percent number down, I've generally failed. If anything, the rate is increasing. And I don't see how we as users of these products can afford it any more.

To compensate, I think we are all going to have to concentrate on fewer vendors (to get more scale and some additional pricing leverage) and use more open-source-like support strategies. And we really ought to stop selecting products on the basis if features that don't have yet--which just encourages the vendors to persist with a broken model.

And finally, we ought to start insisting that the software we buy actually works reliably. Perfection may be too much to expect, but "fit for use" certainly isn't.

Maybe I should start charging the vendor for every undeclared bug I discover.

John Parkinson, the former CTO of TransUnion LLC, has been a technology executive and consultant for over 30 years, advising many of the world's leading companies on the issues associated with the effective use of IT. Click here to read his columns in CIO Insight's print edition.


0 Comments for "The (Un)Economics of Infrastructure Support"

No one has commented yet. Be the first one to comment.

Leave a Comment