Virtualisation staffing: Plan now to prevent disaster
- 19 June, 2008 22:00
There are a lot of things that can sink a technology project, but none of them are more effective than ignorance and insufficient preparation.
That's at least as true for virtualisation projects as any others but - because a virtual infrastructure will underpin and become a part of every other technology project you run-basic training in virtualisation is more important that high-level training in many other technologies. And virtualisation skills-both high-level expert integrator and low-level VM admin-are scarce.
The disaster potential in a virtualisation project is comparable to an ERP project, both for the breadth of the technology the new systems touch, and the level of knowledge [of both the business and specific technologies] IT people have to bring to the table.
And ERP is famous for its potential for disaster.
The city of Portland, Oregon, for example, has had to fire one set of consultants, hire another, increase the budget almost 40 per cent, and push out the completion date for a wide-ranging IT integration project, largely because the city's IT staff didn't know how to manage the technology or the consultants, and the consultants weren't up to doing the job on their own.
"Government incompetence," is the general culprit, according to Barry Brunetto, VP of IS at lumbering and building-tool manufacturer Blount which, like the City of Portland, has standardised its business applications on SAP, but far more quickly and at far lower cost.
The US$500 million/year company's IT infrastructure is far smaller than Portland's, though its 13 international locations and the complications of tweaking SAP's software to deal with so many sets of national business regulations, tax rules and financial disclosure requirements make its project equally complex.
With only 25 people in IT and a cost of less than $16 million for the total project, Blount finished its migration more quickly, less expensively and with an ongoing increase in efficiency that cut 30 percent from both existing IT operations and new projects as well, Brunetto says.
The key to his success was in good planning and execution, which, in turn, depended on the quality of the IT team managing the project and the outside contractors.
"I believe in creating a team of top-notch people," Brunetto says. "I'd rather have 10 top-notch people than 30 schlocky ones, even if the salaries are that much higher for the good ones. If I were looking for a server administrator, I could get a general one off the street or get a top-notch one for 10 or 20-thousand [dollars per year] more, but the top-notch one can probably to twice as much work."
What's that mean in virtualisation?
It doesn't mean that you have to hire a ton of virtualisation specialists. You do, but not as many as you would have to in order to bring a high level of virtualisation expertise to every part of your IT operation.
As with other technologies, you only really need a core of high-level experts with certifications from the major vendors, to design, launch and manage virtual server farms, oversee service providers, and act as a source of expertise to other areas of the business and the IT department.
The sticking point comes not with how many virtualisation experts you hire, but how expert the rest of the IT staff is in virtualisation.
Would you hire someone who's great at designing or Web-based apps but who didn't know how to use a PC well enough to fix minor configuration glitches, do regular updates or other routine maintenance?
Not everyone has to be an expert, but they do need to know the basics of how a PC works - how it connects to the network, how to do basic repairs and how to adapt their own work to the strengths or weaknesses of a particular type of hardware.
It would be difficult to get to that point, in fact; developing even basic web skills require enough time on a PC to become basically competent with it.
Virtualisation is new enough, however, that a lot of LOB application developers, DBAs, networking gurus and other must-have IT experts don't know much about it. Virtual servers, storage, applications and desktops will become part of their projects by fiat, sliding in with the server hardware, changing the capabilities, risks and benefits of a layer of the application stack the specialists thought they already understood.
Without a basic level of understanding about virtual servers, though, those experts can't build that knowledge into their own work so the systems they build don't break because of some fully forseeable conflict between how a virtual machine works and how untrained IT people imagine it would.