As I write this, today (4-Mar-2008) is primary day here in Texas, so it got me thinking.  Over the past 20 or so years, I’m sure that numerous political party strategists have mused something along the lines of, “If only we had a credible, recognizable woman candidate that we could run for president...”  Likewise, many of those same party luminaries as well as some different ones probably had the thought, “If only we had an African-American politico with mainstream appeal that we could run for president...”

Well, be careful what you wish for, because right now we’ve got both.  I’m not saying anything about policies or their relative merits, but rather commenting on the inevitable:  as in all contests, someone is going to win, and someone is going to lose.  …and whoever is on the losing side is probably going to set back the opportunity for another person with the same characteristics in a future election by several years.  In other words, no matter who wins the nomination, the pundits will inevitably conclude that America wasn’t ready for a [black or woman] candidate for president.  Obviously, we’ll see how it goes.

But wishes so often have some kind of downside—witness the numerous genie-in-a-bottle and leprechaun jokes that abound (sorry, couldn’t think of one suitable to publish here, so do your own search).  On a side note, I have always wondered why you couldn’t wish for 3 more wishes as your third wish and keep the ball rolling.  But, back to the subject, just a few years ago there were organizations of all sizes looking at their infrastructure and wishing for a solution to what was called “server sprawl.”

This so-called sprawl was brought about because of the desire to have a stable and sustainable environment, so the practice of hosting a single application per server was deemed the best practice to avoid the dreaded “blue screen of death” since applications couldn’t be trusted to interoperate together.  A side effect of this is that since Moore’s Law kept pushing processor performance higher and higher, the 1-application-per-server rule meant that average utilization would keep dropping with each successive generation of server.  Now all of a sudden what started out as a technology policy had implications on the financial side, and that is that the return on IT assets didn’t look very good.

Think of it this way:  if you operated a dry cleaner and owned several washers, dryers, pressing machines and whatever is used to do “dry” cleaning (and just what is “Martinizing” anyway?) and they were in use less than 30% of the time—equivalent to what Gartner Group and others report average utilization of servers to be—how long would it be before your accountant commented that you needed to either improve your usage of those assets or look to downsizing?

Virtualization rode in on its white horse to save the day. Now there was an effective method to stack multiple instances of an operating system on a single server to improve utilization. Server consolidation projects became all the rage in the wave of “virtualization 1.0.”

What’s interesting is how many customers I’ve chatted with that have said that their server consolidation project was undertaken with no more analysis than “We have too many servers. Reduce them by half. (By next Tuesday…)” Along the way, another phenomenon occurred. It became so easy to deploy a virtual machine that now we had to confront a new horror: virtual server sprawl. You see, even when virtualized, a server has to have an operating system and applications and connections to the outside world. And those components have to be configured, secured, administered and monitored. And it’s the same amount of work whether it’s a physical host or a virtual one.

Virtualization isn’t a technology magic bullet. If you move from physical to virtual and use the same processes, procedures and tools that you always have, your administration costs are going to remain the same. That’s why it’s so important that as part of a well-considered virtualization strategy that you implement time-saving tools such as HP Systems Insight Manager, in which an IDC report documented that administrator-to-server ratios could be doubled afterward (yes, I know I’ve blogged about this before). There’s even a more recent report that shows that the addition of the Insight Control Environment (which includes specific management for virtualization from VMware and Microsoft) can provide even more cost savings when implemented.

Okay, I’m off my soapbox for the week.