A True Cost Benefit Approach to IT Architecture
The first step to a more business-focused outcome on the mainframe is to identify the costs and benefits of solutions in place.
12/6/2018 12:19:39 PM |
By Reg Harbeck
Remember copy-proof checks? They’re not so common anymore, with every kind of electronic funds transfer, not to mention picture-quality scanning and photocopying—but I believe you can still find them.
The reason you couldn’t photocopy them is because their background was made of small dots close together and larger dots further apart, usually in an innocuous color such as green or blue, creating the impression of an even color and shade to the human eye. But when you passed them through a photocopier, the larger dots were sufficiently visible to be reproduced, while the smaller ones were missed. And taken together, the larger ones created a word such as “VOID” across the check.
Copy-Proof Checks and the Mainframe
Similarly, you can’t photocopy the mainframe. Think of petty cash sized electronics such as personal computers, and the individual pieces of software and peripherals you’d buy for them, as small dots. In fact, any computing related solution that fits on or under an employee’s desk generally fits under the petty cash ceiling, or at least requires the lowest level of signing authority. Likewise, the effort it takes each individual user to install, configure, troubleshoot and learn these small systems all fits “under the radar.”
We all know why we went out and bought all this stuff for our offices without going through IT in the first place, even if we’re not comfortable clearly explaining it. It was fun, it felt like having control, it felt like status and it felt like getting results that didn’t involve running a gauntlet of nerdy bureaucracy. And even when we grudgingly handed off the support (and maybe some budget) for it to IT, we still kept back a fair amount of local ownership—particularly the fun parts, or as much as we could get away with.
So, all that stuff—those are the small dots that avoided the checks and balances of organizational scrupulousness. The mainframe, of course, is a big dot. So is every expense associated with it, not to mention the detailed and exacting chargeback that comes with using it. None of this explains how the corporate jet ended up hidden in the mainframe budget, but hey, once you’ve got a big “VOID” written across the easy way out, you just stay in your original status, even if that means sometimes getting lumped together with other things that are in a similar position.
Creating a Business-Focused Outcome
So, on the one hand, we have an object too big to be hidden in petty cash, but which gets other big things bolted to it because they can hide in its shadow. On the other hand, we have this pervasion of little things whose cost is like the grains of sand on a beach, and can be readily hidden by the tides of interest and petty cash politics.
And if you’re wondering why no one has created tools to deal with these issues, my answer is that they have; we just resist using them properly.
Yes, we have a single, massive number that represents all the costs of the mainframe, and maybe even a few extras piggy-backed on, which we can then divide up using SMF-based chargeback plus various fudge factors. Yes, we have every digital device known to the business and personal computing world hanging off the edges of the corporate network and VPNing into it.
And we have the great mainframe SMF-processing solutions, and we have similar reporting solutions in the distributed world, and we have spreadsheets.
All of these can be driven forward to achieve a more business-focused outcome, and sometime between now and when computing becomes a true profession (which I’m betting will happen around December 31, 2099), we’ll have to take all of this bull by the horns and bear with the consequences of finding out where our true expenses and value are happening.
The time has come to stop idling and start driving these things forward while requiring the same level of accountability and justifiability for all our IT initiatives as we continue to plan, architect and move to the future.
Taking the Next Steps
The first step—one which has been taken a number of times by mainframers and soundly ignored by other parts of their organizations—is to identify the costs and benefits of the solutions in place and any new initiatives, from the hardware, systems and related software, applications, support staff, and user training and activity.
You have the spreadsheets and other cost/benefit comparison reports, so it’s time to dust them off and start developing them further so you can clearly characterize any solution using the same criteria. And even if you don’t have the ones you created, you have mainframe colleagues who have them—so start bringing them together to create a consistent picture of what matters and what it costs.
So here’s my challenge to you: It’s time to further refine and enhance those spreadsheets so they can be used to measure non-legacy alternatives with the same level of objectivity and comprehensiveness. Don’t wait for the next project when people can claim it’s biased for a specific result.
Everything Is a Factor
Meet with your mainframe and non-mainframe colleagues and start putting together a list of real resources, their real costs and their required benefits. Don’t skip anything that’s a factor just because it seems too small.
Half a day of installation, configuration and troubleshooting per user times twenty-thousand users is a heck of a lot of difference from a few hours of systems programmer time to do it once on the mainframe.
Supporting a different unique configuration on each user’s computer, using virtual sessions and having to house them across numerous similar but not identical servers or having a single environment that houses an entire application all have very different resource requirements. And how about restoring it all from backup? Can you even do that?
Understanding the true costs benefits us all, and with the growing demands on IT which are increasingly outstripping budgets, hardware capacity, and personnel availability and experience, the necessity of this is getting more and more urgent.
Of course there are some things that belong off of the mainframe—and some things that are vastly more affordable on the mainframe. It’s time to face the data and find out which is which.