Provoking IT from Good to Great
Desktops R Us

Your desktop superstore
In the past few months I’ve talked with lots of customers, partners and Cisco colleagues about VDI and how the operational model isn’t keeping up with new desktop technologies.
Let’s be clear: the major number one cause of VDI failure is people and process, not technology. Static work practices stifle technology adoption.
This isn’t because people don’t realise things need to change, and it’s not because they don’t want them to change. The causes of inertia are complex, real and need a lot of energy to displace.
This is likely to get worse because everyone can see the current technologies struggling to get a foothold (e.g. pooled desktops don’t work everywhere, neither does thin provisioned storage), never mind the fantastic technologies to come like application virtualization and client hypervisors.
What’s the opportunity cost of not adopting these new technologies?
- Not being able to close down expensive real estate and move users. This is millions.
- Not being able to offer the work force mobility. This affects your productivity, including not being resilient to disaster.
- Paying double or more to build and operate your desktop infrastructure. What if your competitors are enabling their workforce better for half the price?
If you have more than a thousand desktops then I am probably talking about you.
And just when you thought things weren’t difficult enough, there’s an arms race going on between Citrix and VMware to produce the best VDI solution and customers feel compelled to choose one over the other, only to know that their choice will only be in front for a few months when the other catches up and overtakes. Ad infinitum.
So, two problems then – operational, and technological. What’s the answer?
If you accept that you are always going to have multiple desktop technologies in play at once, and you accept that your operational model cannot be static, then you need a way of accepting this and exploiting it, and hiding all the ever-evolving complexity from the end user so they only have to worry about solutions that fits their requirements (not specification). As a desktop provider to corporate end-users you should recognize these statements:
- Technology: You will have physical desktops, laptops, virtual desktops, terminal services and more all running at once to match different end user requirements. Your blend of these technologies should change over time.
- Operations: You will have different deployment and management operational processes for each technology and possibly end-user type. These practices should change over time.
The only way to achieve this is to evolve to a Desktop as a Service.
Use something like newScale to build a Desktop Shopfront to connect your end-users to the desktop offerings, like Amazon does.
Develop a Desktop Factory on the back end with multiple production lines for each end-user solution (e.g. virtual pooled desktop).
These changes will separate the end-user solution from the actual technology to give yourself the options of swapping between XenDesktop and VMware View to suit your business.
Develop automated procedures for unified (all components under same system of control) systems and service management. Measure it like a business, with a marketing plan, a pipeline forecast, know where your break even point is, assign resources that make most business sense in your Desktop Factory. The key here is provide multiple technologies managed as one cohesive, unified system.
Here are two use cases:
- Organic (e.g. new hire). HR / Manager uses the Desktop Shopfront to order the new hire’s complete computing needs based on their job role, which is linked to requirements, which is matched to a solution. Costs are charged to the manager’s cost centre, an order is sent to the Desktop factory, you are notified of the successful order processing and delivery date. Delivery is physical or software/download or email/logon details. One user might get multiple deliveries (e.g. laptop + virtual desktop).
- Bulk Buy (e.g. acquisition). If your org has acquired another org, you will have a bulk transition to do. You should be able to “import” their desktops into the Desktop Factory, and/or issue users with new desktops as required.
Is anyone out there doing this today? I can’t find anyone, so if you know of something like this I’d love to hear from you!
Related posts:
| Print article | This entry was posted by Steve Chambers on 15 October, 2009 at 16:47, and is filed under DaaS. Follow any responses to this post through RSS 2.0. You can leave a response or trackback from your own site. |
about 10 months ago
Thanks for this well written post, Steve.
Newscale certainly looks interesting. Let’s catch up soon!
about 10 months ago
Steve,
We have many newScale customers who are doing this today.
They use our RequestCenter product to provide an IT storefront with all the proper options available to those roles. The catalog contains not just desktop images, but also access management – because it’s no good to have a desktop you can’t log on to.
Some customers also automate software distribution to those desktops.
The caveat is that for the most part (but not all) they are doing this for physical configurations still or for terminal access to hosted applications.
The good thing about this is that they have really worked on standardizing and locking down configurations. Also some are very specialized, for example some in banks, the desktop includes things like check readers.
As for bulk moves, they happen a lot more than we thought. Companies are always re-organizing, and the bosses like their people nearby. Or they are project oriented – and you need to move 20 people for a couple of months.
As you point out, getting there is not trivial. The two key things that we see getting in the way are lack of standardization and lack of centralized process. These go hand in hand. If you don’t have a set of standards, then the process is more like a bespoke model than an industrial one.