Although in the last years we delivered some important VDI (virtual desktop infrastructure) projects I still think VDI will remain a niche market and likely be soon surpassed by new technologies and different approaches to the problem.
Analysts (and vendors) forecasts actually differ from my point of view, many of them think 2011 will be the VDI year but I’m pretty sure it’s not a realistic expectation, here it is why:
the problem
Well, the problem is indeed a simple one: running PCs is a drama!
Hardware and software maintainance, security and power consumption are just the first issues coming to my mind. TCO of your PCs installed base is huge and sometime not very apparent!
VDI in a few words
Let me say that conceptually VDI is a travel back in time to the days of 3270 terminals: shared and centralized computational resources with distributed displays. The goal of VDI is to concentrate all the operating systems of your PCs in a centrally managed virtual infrastructure and replace each PC with the so-called thin-clients (ethernet connected graphic terminal).
VDI isn’t a panacea
VDI addresses only a part of the problem: hardware and power consumption can be successfully tackled with, but there’s much more left out to cut the overall TCO. An example? VDI leaves unchanged the whole PC software stack: OS, antivirus, SW provisioning and monitoring, applications, etc., etc. so you’ll probably have to be supported by the same helpdesk team.
long term ROI
If compared to servers virtualization, VDI does not result in any immediate return!
Before you see the light at the end of the tunnel, you’ll have to buy a quite big infrastructure, VDI management software, thin clients. All
in all, a quite long deployment process. The deployment can take a very long time: each client will usually be replaced at the end of its support life only (3/4 years) and most of your PCs are not equally old.
IT projects whose implementation phase will last several years are very risky business, especially when dealing with PCs: you’ll likely have to deal with different thin clients, to struggle to update your VDI software and to manage your original platform all the way until the end of deployment!
VDI isn’t pervasive.
Your installed base is heterogenous!
Sometimes not, but most of the times you own and have to manage Office PCs, industrial PCs, R&D workstatios, laptops/netbooks. Each of these hardware platform is targeted to a different user profile: a secretary won’t need the same software stack and degrees of freedom as a support technician, a field engineer notebook will surely be different than a manager’s!
Moreover, in my everyday job I see many companies moving fast toward notebooks (one of my customers has 2000 PCs and 1400 netbooks!) Many of above mentioned the profiles are very difficult to virtualize and maintain. VDI doesn’t like exceptions, if your environment comprises many of this exceptions (different kind of hw and user profiles), you’ll be forced to set aside part of your company from of your VDI project, and this will lead to failure.
the risk of a split
The main risk with VDI is to have two different kind of platforms (and teams?) to manage: a set of old PCs and the newly born virtualized clients. Although this is the worst scenario, it depicts the reality in many VDI deployments. A partial solution won’t be a solution at all, but some customers realize this bitter truth only when wading knee-deep through the project.
mobile devices adoption is growing fast
Smartphones and tablets can be more and more easily found in the enterprise. Even when in office, many users usually receive and send emails from their Blackberrys, it is fast and simple. Having a tablet/smartphone with you when travelling, enables you to open (some even edit) office documents and send them back in few seconds. Newer enterprise application clients for tablets are so terrific and easy to use that many users actually choose them to access their data… also when they are in their office!
One more big advantage is that, many times, phone and tablets aren’t company assets but are supplied as per their mobile operator supply contracts, they don’t need frequent hw/sw maintenance (sw updates are often automatically performed). Mobile devices usually allow the user to self-service software provision (via internal/external app stores!!)
cloud computing is around the corner
Tablets are cloud enablers and cloud is around the corner. Many enterprises are looking favourably to cloud computing, but at the moment many legacy constraints prevent them from moving forward. Tablets actually suffer from the lack of a good sized internal storage (we can consider it mora as a cache then a storage), they need to access a centralized storage, and usually this kind of storage is properly backupped.
Tablets sport good connectivity (wifi+3G), apps availability is getting wider and wider enabling them to work in and within the cloud (Google Apps is only the first example to think of).
If VDI is ok for you, then go that way, but if you have any doubt, take a 360 degree look before marrying VDI. Two options (a short term and a long term one) you should consider:
PCaaS
In a everything-as-a-Service era, why not consider buying the whole PC stack as a service?
While is not a viable option for every kind of company, in many cases it could be. Outsourcing assets and help desk services may help the company to cut the costs and finally have a clear view of SLAs you can guarantee to final users.
(future and) smooth migration to tablets
First generation tablets are alredy here, the next generation (the first one to address enterprise?) will soon be available: Apple iPad2, Cisco CIUS, Blackberry Playbook. I think you need to carefully study and evaluate this kind of products to better understand how they can fit your enterprise.
I’ve seen some interesting projects starting here in Italy, focussed on the conversion of some mobile (and intracompany mobile) users to tablets.
First impressions from users and IT managements are awfully positive.
