Five desktop virtualisation tips for IT project managers There’s a good reason why the word ‘strategic’ should drive fear into the heart of the most hardened project manager. Strategic decisions are generally wide-reaching, and therefore involve multiple stakeholders (which, incidentally, is another word in corporate bureaucracy’s dark lexicon). The more stakeholders there are, the …
Know why you're doing it.
As the article says, virtualisation is a strategic decision. The people responsible for company strategy and direction inhabit the boardroom (not the I.T. cupboard - if this is being driven by the IT dept, it's fundamentally running at the wrong level) and they should be able to complete the sentence:
"We are committing to Desktop Virtualisation in order to ....."
and that answers the "why are we doing this?" question. If you have the world's only talented IT director, that sentence will be followed by a qualifier "and we'll know it's succeeded when we can do <X> better/cheaper/faster/more reliably than we could before."
Rather an odd thing to ask, but don't you need to know what 'done' looks like?
The article appears to be written in management consultant mode; they don't like good requirements as it means a client that knows what they want so fewer change requests.
Why no mention of matching the tech to the possible given the constraints of budget, existing systems and hardware...?
Sage advice here from Robin, but one key element is missing. Before you consider any desktop transformation, virtualization included, you really need to build an inventory of the current desktop assets including endpoint devices, applications and end users. Plus you must understand which applications are being used, by whom and how often. Also understand if they are being used online/offline and what computing resource they consume. These factors all need to be understood with fact-based data BEFORE even considering desktop virtualization. Once you have this empirical data then you can determine the best end-user computing platform for your users, or group of users. End-user computing analytics are required for a successful implementation and in more cases than not will highlight a blend of application delivery is best fit - as one size does not fit all.
"If you’re an IT person going into a project feeling as though you’ve been lumbered, then you’re already setting yourself up for failure. How can you be powerful and effective when dealing with other IT functions on desktop virtualisation projects with a ‘not invented here’ attitude? Leave your reservations at the door and accept the goal you’re working towards, or you’ll become part of the problem."
Having read the post you linked to, it wasn't complaining about being lumbered, or "not invented here". It was about the merits, or lack of them, of VDI itself, from someone with direct experience of the problems it has.
It seems to me that it is your "It's been decided, therefore it must be a good idea" approach that is the problem here. Projects can't expect "buy in" automatically - they need to make the case that what is being proposed is actually the right thing to do. For VDI, that case seems conspicuously lacking. As a niche, for some special cases, fine. As a general-purpose desktop, no. It's just not the right model, for many many reasons. From the post you linked to:
"What DV is is overkill for the bottom end, underpowered for the top end, and less robust than a distributed solution. And, at least in my experience, it never costs less in the end."
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
