Where do you want your IT infrastructure to go (and how can virtualization help you get there)?
Where do you want your IT infrastructure to go (and how can virtualization help you get there)?
By Robin Crewe published: Tuesday, October 10 2006
“If a man does not know what port he is steering for, no wind is favorable to him”
Seneca
Where is corporate IT infrastructure today?
IT Infrastructure is what, in a different context, Herzberg1 would have described as a hygiene factor, not a motivator. Like roads, railways, and utilities, users seldom appreciate it when it works but complain vociferously when it doesn’t. According to Gartner2 47% of total IT spend is on infrastructure. In other words, roughly half of all the money companies spend on IT goes on networks, PCs, storage etc and support for those items. No wonder CIOs have a tough time convincing their employers they should be on the board – half of the services they manage are not unique to their industries, let alone their organizations.
On top of that, like its physical counterparts, corporate IT infrastructure isn’t flexible. If you want your corporate applications delivered to you somewhere else (at home or on the road for example), your employer typically has to extend its infrastructure to do this. IT infrastructure costs are not transparent either. There is no equivalent of the electricity meter ticking along in the background, recording what you’re using and charging you accordingly. Internal IT departments still tend to allocate costs rather than charge for services.
Although the consumer technology experience is getting better, corporate IT infrastructure keeps getting more complex. When an organization introduces a new technology (Vista for example), it very seldom de-commissions an old one. The problem is exacerbated by the proliferation of PCs, thin clients, PDAs, and mobile phones that can be used to access corporate information services.
In summary, today’s corporate IT infrastructure is expensive, inflexible and opaque.
Is it all the fault of the IT department?
Before concluding that the IT department is to blame, consider the rate of change in the IT industry. The last 40 years have seen incredible innovations in information technology. Most CIOs have done a great job of introducing wave after wave of new technology with very little disruption to service. They have had to master new skills while retaining expertise in technologies that are several generations out of date – how many of their business colleagues can claim the same?
From this perspective, the contribution of most IT departments is massively under-valued.
Where the IT community can perhaps be criticized, though, is in not highlighting the trade-off between short-term responsiveness to new technologies and long-term complexity. Too often organizations pin their hopes on ‘the next big thing’. Client-server, thin client, web services and the Internet have all been hailed as panaceas for corporate IT infrastructure problems (and there is a danger that SOA will be the latest technology to fall into this category). Many large infrastructure upgrade projects have been initiated as a result but the underlying problems of complexity remain remarkably persistent – the new technology is introduced but most or all of the old ones remain.
The reality is that most organizations will have hybrid IT environments for the foreseeable future. It is time to find ways of managing these mixed environments efficiently while in parallel simplifying the infrastructure and moving it in a direction that can embrace further technological change.
‘Where do you want your infrastructure to go?’ is therefore a management question, not a technical, one. Although some organizations have tried to sidestep the question by outsourcing their IT infrastructure, it’s a question that cannot be avoided. While some infrastructure outsourcing deals have been successful, all too the contracts are tightly worded around the organization’s current, complex infrastructure. Scope changes then erode the cost benefits and the underlying complexity is still there when the infrastructure is re-insourced or passed along to another provider.
So where do you want your IT infrastructure to go?
Non-IT people are often criticized for not being able to articulate their IT requirements. While it would be perfectly reasonable to ask a company’s management where they want the company’s IT infrastructure to go, it may not be necessary at the outset. A shortcut is to look at the characteristics of the IT services people use when they’re not at work. Nobody is obliged to use Google, Flickr and other similar services. The fact that millions do is testament to the fact that these services (of which infrastructure forms a part) are delivered in a way that customers appreciate.
From an infrastructure perspective these services are:
Secure – contrary to popular perception, the Internet, through https, offers an extremely high level of security for transactions that require it
Consistent – you access Google mail and it does the same things regardless of where or how you access it
Ubiquitous – you can reach services on the Internet like Flickr from almost anywhere (the exception ironically usually being your office)
Reliable – sometimes Google is down but not very often and this despite very rapid growth and variable processing loads
Variably-priced – admittedly, thanks to advertisers, a lot of consumer Internet services are free to use; that certainly makes pricing easier. Those that aren’t, though, such as back-up services, are typically priced according to usage and service level
Immediately available3 – you can sign up for a service and starting using it immediately. There is no lengthy approval or provisioning process.
Of course, having the right infrastructure does not guarantee results. It is a necessary but not sufficient condition for an Internet service to be successful. There is also no guarantee that this list of characteristics is required in every corporate IT infrastructure. However, as boundaries between personal and corporate computing blur, it seems a reasonable place to start.
How do you get there?
If your corporate IT infrastructure doesn’t satisfy the above list of characteristics, how do you change it so that it does? If you assume there is no silver bullet that will transform your infrastructure from being expensive, inflexible and opaque to being cheap, flexible and transparent, the question becomes one of how best to manage a continuous migration process.
This is where virtualization4 comes in. Virtualization introduces an abstraction layer that enables things to be de-coupled. For example, if you virtualize a network you can decouple the configuration of the network from the physical wires and devices of which the network is composed.
Virtualization has been moving up the IT stack. It started with networks, moved through storage and servers and has moved in to desktops. Each of these steps has increased flexibility but there is one remaining layer that is critical to a successful infrastructure migration program: user virtualization.
If you virtualize user IT services, you can de-couple the delivery of those services from the underlying physical or virtual infrastructure. You can then migrate IT services to simpler, more efficient delivery platforms without impacting end users. The result is a consistent, secure user experience and a simpler, cheaper, more flexible IT infrastructure. Often this infrastructure will be centralized and will itself comprise virtualized platforms but, importantly, it will not matter to the end user where or how applications are delivered – they simply click on the service they require and it is delivered seamlessly and securely to them.
Practical Applications
If this all sounds too theoretical, here are three practical applications of user virtualization to think about:
1. Access Devices
How many organizations allow employees to bring their own pens to work yet how many adopt the same policy for PCs, PDAs and mobile phones? The rationale for the difference is usually ‘security’ (the PC needs to be locked down) or cost (we can get cheaper prices) but are these reasons valid? Pens can be used to write checks but the sensible point of control is the checkbook, not the pen. If an employee brings their own PC to work that they have paid for, why is a corporate discount relevant? In many cases, the same companies that insist employees use a company PC in the office on a private network allow at least some staff to use their own PCs on the Internet when they are at home or traveling. With user virtualization, companies can focus on access devices that do the job and can get away from the mindset of continuously refreshing PCs; such an approach also points to the natural control points (access to services) and the possibilities opened up by centralization.
2. Networks
People are accustomed to using the Internet to access services in their private lives so why not when they are at work? The Internet is ubiquitous and, by design, reliable. It is bandwidth constrained and cannot guarantee quality of service but the reason usually cited for an organization having its own network is again ‘security’. So although most security violations are internal, network traffic is unencrypted inside most organizations but encrypted outside. And although organizational boundaries are becoming blurred, most companies try to build rigid boundaries between themselves and the outside world. User virtualization enables organizations to choose networks based on service levels and availability rather than security, immediately simplifying the task of network engineering and increasing flexibility in the process.
3. Software as a Service
Internal IT departments rightly complain that organizations like Google have hundreds or thousands of staff looking after a handful of applications whereas they have a handful of staff looking after hundreds or thousands of applications. The corporate applications they have to look after were not written to run as Internet services and re-writing them would take many man-years of effort. All true. However, with user virtualization, the delivery platform is hidden by the virtualization layer. As a result, corporate applications, regardless of the platform they physically run on appear as secure services accessible over the Internet. Combine user virtualization with simple usage monitoring and, with very little effort, you can turn a portfolio of legacy applications into a suite of Internet software services.
In summary
Corporate IT departments have done a fantastic job responding to technological advances. The cost of this responsiveness has been an increase in complexity. Now is the time to ask where you want your infrastructure to go. We believe consumer-focused Internet services provide a useful starting point for this assessment. As consumer and corporate IT services blur, the characteristics of consumer services – secure, common, ubiquitous, reliable, variably-priced and immediately available – will be needed in corporate IT infrastructures. Getting to such an infrastructure is a continuous migration process. Virtualization, in particular user virtualization, represents an exciting way to get there with many immediate applications.
1 Herzberg, Frederick, The Motivation to Work, 1959.
2 Gartner Inc., “Realizing the Benefits of Project and Portfolio Management”, January 2005
3 We prefer this phrase to ‘on demand’ which tends to get confused with utility computing
4 We define virtualization as doing in software things that used to be done manually or in hardware
