Virtualisation
Paul Russell
Paul Russell is Head of Strategic Development at Servo (a Microsoft Gold Partner). Servo is dedicated to helping companies achieve competitive advantage through “IT Infrastructure Excellence”. As part of the Phoenix IT Group, Servo specialises in providing services across a wide range of industries to companies in the core and upper mid-market sectors within the UK.
Question: Our IT department manages a data centre and 180 desktop PCs. They are recommending we invest in virtualising our IT infrastructure over the next 18 months. Is there any evidence that it will represent a cost saving, or is this another technology fad? - Gareth, London
Answer: Hi Gareth, before we get to the specifics of your question, it’s worth explaining a little about what virtualisation means, because it changes the way we think about IT, how it works and what it can achieve for a business.
Virtualisation in a nutshell
For the past few decades, the way enterprise computing works has been pretty static. Applications have run on servers. Between applications and the servers on which they run is an Operating System (or OS) which manages the complex interactions between the application and the hardware. Because of their complexity, most servers have until now used one OS, and therefore been restricted in the applications they can run.
Most servers therefore run only one application (CRM, for example), irrespective of how heavy a load the application places on the server. As servers have become more powerful, there are plenty of instances in which servers are massively underused- wasting computing power and energy. To continue with the CRM example, if a marketing team only sends out marketing literature once a month, the CRM system may be using no more than 1% of its capacity for the other 29 days of the month.
Most large companies are therefore victims of “server sprawl”- lots of single-purpose servers (and the maintenance cost associated with them), all of which are rarely used to their full capacity.
Virtualisation essentially unties the relationship between the applications, the OS and the hardware. A virtualised server can run multiple operating systems, and multiple applications. The physical hardware represents itself to the company as a number of “virtual servers”, running as many applications as the hardware will allow and therefore massively reducing the number of servers required.
The benefits are significant:
| • | For any company concerned about its green credentials, an efficient virtualised data centre can massively reduce power consumption- and therefore energy bills too. |
| • | Virtualisation also reduces the cost of business continuity. To ensure 100% uptime and protection against data loss, companies have traditionally had to replicate their entire server base as a backup- a doubling of the hardware cost even though the backup system may only come into play in the occasional emergency situation. As you can imagine, fewer “live” servers means fewer backup servers- a huge saving. Furthermore, the administrative tools associated with virtualised servers mean that switching over to the backup system in the event of a drop in service is practically automatic. |
| • | A virtualised IT architecture by definition allows for faster and more economical deployment of new business services- and that means IT can better meet the needs of the business. You’ll hear IT evangelists talk about “moving from servers to services”- and that’s not just jargon. Because virtualised IT no longer relies on clusters of bespoke machinery, new business ideas can be implemented with new software on existing hardware. You can even try out a new software system on a virtualised server before making any commitment to roll it out across the company. Risk is reduced, expense is reduced, and so is the load on the IT department. |
Financial benefits
That’s the sales pitch for virtualisation: now to your question- does it represent a real cost saving? Well, we mentioned the energy savings above; but the most important savings can be understood by examining your IT department expenditure holistically.
Most companies typically spend most of their IT budget just standing still. Around 80% of the budget is spent keeping things ticking over- maintaining existing systems and solving problems. Only 20% is “useful money”, improving productivity or meeting the new needs of a developing business.
Early adopters in large roll-outs of virtualised systems are reporting an IT saving of between 10% and 20%. However, allocate that money back into IT and you have effectively doubled your “useful money”, with 30-40% of the IT budget delivering measurable business benefit.
These figures are expected to improve even further. The virtualisation principle can be applied to desktop PCs too. PCs can be virtualised, using tools like Microsoft’s SoftGrid, to run applications which are based on a central server, independently of their operating system. This increases the number of applications available to each desktop, reduces the total cost of ownership of each PC across its lifecycle, and also saves on software licensing costs.
Tips for successful virtualisation deployment
To gain all these benefits, you will need to engineer a major change in the way the IT is deployed throughout the business, though; and that requires that your IT team approach virtualisation with their eyes open. Here are some tips from my experience:
| • | Virtualisation requires an efficient IT infrastructure in the first place. Individual servers which don’t work properly won’t work any more effectively with a greater load. Your IT team needs to be running an efficient and mature systems architecture in the first place. If they have their hands full with the existing setup, then make sure decent IT management processes are routine before moving to a virtualised system. Microsoft provides an Infrastructure Optimisation Model which gives both business and IT managers guidance on effective platform improvement. |
| • | The IT team may also have to change in terms of in-house skillsets. Talent may need to be redirected from managing physical servers (i.e. using screwdrivers) to managing conceptual systems (using screens of information flows). For example, whilst it’s fairly hard to lose a physical server, inadequate recordkeeping can easily lead to a virtual server being forgotten about. Virtualisation requires a new type of rigour and discipline in the IT team. Invest in governance and management training if necessary. |
| • | Finally, you will never virtualise every machine (and even if you did, it couldn’t happen overnight). You’ll therefore require your new virtualised system to work seamlessly with older legacy systems. This is where tools like Microsoft System Center Operations Manager come in- allowing your IT team to build bridges between your older systems and your new, more manageable environment. Ensure that your new virtualised system is compatible with your existing technologies. |
| • | Watch out for server sprawl in the virtualised world. It is very easy to provision new virtual servers and without the legacy hardware issues that went before, organisations can find themselves creating just the sort of overcomplicated environment they were trying to escape from. |
The great thing about virtualisation is that the inbuilt flexibility of these systems makes for easy testing. At Servo, we think that 2008 is the year for proofs of concept in the virtualisation process; and good IT specialists are building test environments for their clients. Before signing up for a fully-fledged virtualisation rollout, ask your IT team to build a small-scale example- and get the administrative processes around it in place too. You’ll then be ideally positioned to see the financial benefits for yourself, as they apply to your specific operations.
--Paul
Related Links
•
More Microsoft Virtualisation resources
•
Microsoft Systems Center Operations Manager
•
Find out more about Microsoft SoftGrid
•
Read more about Infrastructure Optimisation and take one of the three IO assessments
