ViewYonder

Put your hand down if you’re a vendor, or a purveyor of theories, or a consultant with something to sell: I’m not talking to you!

This shout goes out to the IT folks who are running or thinking of moving to a Unified Computing System.  Note: I’m not just talking about Cisco UCS, so please read on if you’re from HP

If two people stepped forward to claim the prize for the best Unified Computing System – whether Cisco, IBM or HP customers – on what basis might they make such a claim.  How would the judges decide who was best?  Who’s qualified to be a judge?  Who judges the judges?

Why even bother to measure a Unified Computing System, and why bother to compare two systems to each other?

1. IT providers want to measure how efficient, effective and well governed their Unified Computing System is so they can work out how and where successful they are, and likewise where they need to get better.
2. If IT providers can compare their system to another then they can learn and improve through ideas from others who are better in some places than you: and if they are altruistic, they can share your successful approaches with others.

What is this Unified Computing System?  I’m not talking about Cisco UCS specifically: I recently talked about a seven layer Unified Computing System, and as I can’t function without pictures, I sketched one up:

Note that this kind of high-level, patterned approach is a technique to compare two complex systems.  There are those out there that will say “It’s impossible to compare two Unified Computing Systems that are deployed in different IT organizations, there are too many variables.”  I respectfully disagree with that opinion, I do believe the variables are finite and numbered in the hundreds instead of thousands, but I don’t have the full list.  Yet.

The fact that something like ITIL, love it or loathe it, has risen out of the minds of hundreds of IT folks from different backgrounds and different parts of the worlds – and, therefore, different cultures – means that there is some hope for common patterns in IT that can be measured and compared.

Even better, systems like COBIT and ISO have very high level patterns and definitions, but these aren’t detailed enough on their own to connect to solutions like VDI (and why should they, you ask? – well, wouldn’t it be great to show how brilliant one VDI system was compared to another?).

COBIT and ISO don’t go enough into the detail to apply to today’s Unified Computing Systems, they weren’t intended to and probably never will.

An example is deploying a server: this release management procedure has a consistent pattern despite it’s varied implementation. For example, to deploy a server there is a common Request -> Approve -> Schedule -> Deploy -> QA -> Go Live -> Archive -> Decommission lifecycle.  Hey, you might have one or two more, or one or two less, steps in that lifecycle but I’d bet one of my homemade Yorkshire Puddings that it looks familar.

You can measure each of those steps in similar ways:

• Elapsed Time from start -> end of each step, and total for each iteration
• Level of Effort for each step and each complete iteration
• Complexity in terms of steps, roles involved and how serial the procedure is vs. complex decisions.
• Quality in terms of failures to meet standards, including adherence to the process as well as of the end product.

A universal measurement of success for this release management procedure is one of efficiency (least effort employed), effectiveness (high quality) and governance (ticks all the compliance boxes).

If we can define a pattern for releasing a server, then all we need to do is try and write the same description for all the other IT procedures: that would be tantamount to defining a framework of operational patterns that can be measured, controlled and compared.

And then we’d have a way of deciding who, of our two suitors, had the best system. In fact, we’d have the beginnings of an industry-wide system to measure and compare IT systems and solutions that was more prescriptive and actionable and solution-specific than Cobit, ISO and ITIL.

Before you say “that would be too restrictive”, let me throw a number at you. VMware claims to have over 135,000 customer IT organizations. Each one of those is a potential user of any proposed framework, and that number includes all of the top companies around the world, including a huge chunk of SMBs.

I think all of those customers would benefit from such a system ONLY if it was designed by them and NOT by the same consultants that wrote ITIL and all the other high-level best practice frameworks. It certainly can’t be written by the vendors because they don’t have the required customer perspective (believe me!). So who, out there, in customer land, would step up and try to do such a thing?

I would happily spend money from my own pocket to meet up with people who are interested in this for a day long discussion and see what popped out of the other end.