Alternative : Is there really a cost justification for Virtual Workspaces?

From the sales and marketing perspective, we spend a lot of time looking at using Virtual Workspaces to support cost optimisation by decreasing hardware costs, increasing end-user productivity, driving mobility and providing greater flexibility for applications and operating systems.

The million dollar question of course is whether or not this actually adds up to a better deal for businesses - and I'm going to talk about how you can work that out.

These benefits are certainly very convenient when you are trying to justify spend, but unlike server virtualisation, the business case is often not so clear cut.

Hardware costs are unlikely to be reduced because of the cost of servers, storage and the thin client devices required to make the virtual workspace fly, not to mention licensing for operating systems and hypervisors. In truth the best you're likely to do is break even on capital costs compared to buying new PCs for 'not-so-virtual' workspace deployments.

Demonstrating cost savings beyond the soft benefits of enhanced mobility and security needs a different angle. Two good hunting grounds for this are the areas of energy savings and opex reduction.

Energy Savings

Let's look at energy savings first.

'Fat' PCs use between 65-250 watts of power per hour. Conversely, thin-client devices use 5-14 watts of power per hour due to the fact that they have no hard drives, no moving parts, low-energy processors and relatively small amounts of memory. Whilst I appreciate that many organisations will not go 'fully thin-client', it is clear that even if only a sub-set of the estate gets on the thin-client diet there are some decent savings to be had.

If we assume that thin clients and PCs are in use about 9 hours a day 230 work days a year (excluding weekends, holidays and other times where computers aren't needed) and that the rest of the time they will be on standby to conserve energy.

The hawk-eyed amongst you will have also realised that there are also some servers somewhere that are needed to host all these virtual workspaces, so let's throw those into the mix for good measure - an average server hosting virtual workspace workloads will consume 450-650 watts of power per hour, and should comfortably host 60-100 virtual workspaces. We must also consider that servers are in active use 24x7x365, since one of the key benefits of the virtual workspace ideal is anywhere, anytime access.

Assuming an electric rate of 14 pence per KWH, the energy costs for 1,000 PCs would be about £18,837 (based on the lowest 65watt per hour desktop devices).

(Number_of_PCs * PC_Energy_Use * Hours_Desktop_Non_Sleep_Mode * Number_Work _Days_Per_Year * Cost_per_KWH/1000).

By comparison, the energy costs for 1,000 virtual desktops (including servers) would be about £12,485. Over the 3-year lifespan of a PC, that's a saving of almost £37,455 in favour of VDI.

(Number_of_Virtual_Desktops * Thin_Client_Energy_Use * Hours_Desktop_Non_Sleep_Mode * Number_Work_Days_Per_Year * Cost_per_KWH/1000) + ((Number_of_Virtual_Desktops/Desktop Density)) * Server_Energy_Use * 365 * 24 * Cost_per_KWH/1000))

So, £37k is no bad saving, but isn't exactly going to provide you with the career defining moment where the c-suite give you a standing ovation for making the CFO cry with happiness.

Let's dig into operational expenses.

OpEx covers a wide area of course, so it is worth drilling into the elements most likely to benefit as a result of virtual workspace deployment. These are generally as follows:

• Windows Image Management
• Application Delivery
• Workspace Break/Fix

Windows Image Management

The patching of operating systems is a major cost for most IT organisations due to the sheer number of images that must be updated. Whilst it would be nice to have a single image omnipresent across the entire estate, 'image sprawl' is almost always the reality.

Furthermore, in physical PC environments, images often 'drift' due to the differences in PC hardware configurations, even if every device is the same model number due to modern component procurement processes. The lack of a standardised image engineering methodology can also have an adverse impact.

Regardless of the reason why image sprawl happens, let's look at the overheads. Assuming an approximate ratio of 1 image for every 100 desktops and a timeframe of 3 hours to deploy and test the Windows update on each image, (10 different images for the 1,000 desktops), and that patches are applied once a month (12 times a year), then 360 hours are spent on patching per year.

Using a PC administrator's hourly IT staff rate of £40 per hour, this could translate into as much as £14,400 per year.

In a virtual workspaces scenario, image sprawl is finally eliminated, a single Windows OS layer can be used for all workspaces and applications can be packaged separately from the base OS layer. Without any apps 'baked in' to Windows, extra images aren't needed to reduce app licensing costs, and patching costs can be greatly reduced.

Therefore, if it takes 3 hours to deploy and test Windows updates on the 1 OS layer, and if patches are applied once a month on Patch Tuesday for a total of 12 times a year, then patching time is reduced from 360 hours to just 36 hours. Using the same hourly IT staff rate of £40 per hour, this translates into a total cost of only £1,440, a saving of £12,960 per year.

Application Delivery

Application delivery is another other activity that represents a significant challenge for organisational IT departments. The most common way to deliver software on physical PCs is through PC configuration management or electronic software distribution software. The way this manifests varies of course, but usually the process entails the following steps:

• Install the application on a test machine
• Identify prerequisite updates for deployment
• Create a software distribution package (source files, paths, icons and settings)
• Create a program that defines the command line parameters to run during setup
• Distribute the package to the desktops and schedule execution
• Check the logs to determine how many installations completed
• Diagnose and repair the desktops where installations failed

It is estimated that this process typically requires an average of 3 hours per application. Assuming 100 Windows desktop applications in total and that they are updated, on average, twice per year, that equates to 600 hours of IT staff time per year. Using the same £40 hourly rate, that equates to an annual cost of as much as £48,000.

With virtual workspaces all apps can be layered as their own read-only virtual disks, eliminating the need for detailed up-front application analysis. File system and registry virtualisation technologies enable applications in separate layers to interoperate, so there is no need to plan packaging sequences or stack apps that need to cross-communicate in the same layer. As a result the total time needed to layer an application can be reduced to as little as 20 minutes.

If there are 100 Windows desktops apps in total and they are updated, on average, twice per year, this means that only 66 hours of app packaging time is needed per year, driving annual cost down to £2,640, a saving of £45,360 per year.

Desktop Break/Fix

Our last major operational overhead associated with the management of non-virtualised PC deployments is desktop break/fix. Whilst it would be reasonable to hope that PC environments are more robust than ever, a percentage of desktop users will still call the service desk with DLL conflicts, hardware driver issues, suspected malware or virus issues. Service desk incidents are also common the day after agent-based software distribution tools apply the latest patches, since not all patches successfully deploy.

While the solutions are usually simple, resolving these issues almost always requires escalation to Tier 2 or 3 IT support since helpdesk staff rarely have the expertise or the tools to resolve the issue themselves. The time it takes IT administrators to visit a desktop, even remotely, analyse file system and registry changes and manually back out or re-install the affected files or registry keys is usually between 2 and 3 hours.

If we assume (which we have done a lot of today) that 10% of the user base calls once a year (100 incidents) and that each incident takes 2 hours of an IT support engineer's time to fix, then 200 hours per year are spent on break/fix activity.

At an hourly rate of £40 per hour, that equates to an annual cost of £8,000.

Virtual Workspace deployments allow the tracking of all versions of application and persistent personalization layers, enabling easy rollback if something needs fixing. When a user calls in with a problem the helpdesk staff are better positioned to resolve the incident in minutes simply by reverting back to a previous layer version, undoing any problems.

As a result, costly escalations are eliminated, saving the £8,000 per year in break/fix costs, not to mention reducing the required volume of L2 and L3 engineering support.

Still hungry for savings? Time to turn to cloud

OK, so we have had a reasonable stab at finding savings and managed to find a three year saving of just under £200k over 3 years of operation of virtual workspaces in energy and OpEx optimisation.

Not enough?

OK, let's look at sticking Virtual Workspaces in the cloud, creating the whole 'freeing up your IT teams to focus on strategic initiatives' thing. Whilst golden marketing fodder for the cloud PR machine, the sceptics amongst you might see it as nothing more than a softer way of saying 'sack your IT team and let us do it', but it's more than that.

Sure, there is always an opportunity to reduce physical overheads when using a service that has support included, but you are probably looking at a range of other IT projects with potentially strong returns. Even when streamlined, quality infrastructure technical staff invariably find themselves way too busy with everyday activities.

Projects are delayed as a result because the team are constantly being pulled into fire-fight live service issues, and that takes priority, which means that moving the virtual workspaces element of you IT solution to a trusted cloud service provider can make a lot of sense. Doing things this way also removes the need to organically grow (or quickly recruit) virtual workspace infrastructure skills in the team which instantly reduces a significant overhead.

Alongside not having to create or recruit new skills into the team, using a cloud service provider can also drive reductions in out of hours working for IT staff, which leads to a reduction in overtime costs and improves staff retention and an improvement in staff morale due to a reduced number of mundane enquiries and better exposure to 'exciting' new projects and technologies.

Looking at the overheads associated with getting access to virtual workspace skills (at least 2 deployment and support engineers, probably commanding a £40k PA salary as a minimum) and the cost of replacing existing staff due to excessive working hours and a lack of time to develop themselves through exposure to new technologies and strategically important projects (let's say 1 head a year on £50K with a 3-6 month ramp up into the role the current guy knew like the back of his hand), there is a compound argument for going down the virtual workspaces 'in the cloud' route.

It could also make you the person who took the company from where it was, to a fully functioning virtual workspace-based desktop and application delivery model that is a significant catalyst in helping your organisation realise new working models, better staff productivity and greatly enhanced security. Not a bad thing to have on the CV!!?

In Summary

OK, clearly this is a ridiculously simplified cost analysis, but the bottom line is that there are a whole bunch of benefits associated with on-premise and cloud-based virtual workspace solutions that can be realised over and above the pure commercial case.

The models I have used are deliberately simplified, but generally speaking a more detailed cost analysis (as produced by one of our consultants) is a valuable step in identifying how you can realise a significant commercial benefit over a 3-5 year lifetime in your business, irrespective of the soft benefits such as those listed above.

If the cloud service model sounds like the safest way to get your business into the virtual workspaces slipstream, there are a couple of important things to consider. Make sure that the service offering being proposed integrates easily with your existing IT support functions, and the provider's business is one that you can trust. Due diligence is also the name of the game here. There are many providers out there, all with great marketing engines that look slick. Look beyond these, checking finances, reputation, accreditations, size of technical organisation and strategy. I would also demand a pilot scheme for a sub-set of indicative users (not just the IT guys) to establish how easy the solution is to live with.

Hopefully I have managed to show you some of the ways in which virtual workspaces can deliver some of the returns from an operational expense perspective. I have deliberately not covered the benefits of improved security or the value of delivering a great user experience (another time, another blog) but these also deliver direct and indirect benefits to the bottom line. Providing you select the right partner, Virtual Workspaces, Virtual Desktop Infrastructure, Desktop as a Service, Hosted Desktop, Desktop Cloud, OnlineDesktop whatever you want to call it work very, very well.

Alternative would love to be that partner!