subscribe: Daily Newsletter

 

How to get value from virtualisation

0 comments

Despite its six syllables, "virtualisation" is a straightforward concept. It can enable your organisation to get more value not only from computer and storage hardware but also from the labor required to keep your systems up and running.

Virtualisation is an approach to pooling and sharing IT resources so the supply of resources-processing power, storage, networking and so on-can flexibly and automatically meet fluctuating business demand.
Virtualisation can improve the quality of your IT services, enabling more consistency and predictability of operational availability.
If you are considering a new or expanded virtualisation solution, HP has the technology and expertise to help your organisation achieve real business benefits from virtualisation today.
HP sees virtualisation as a key enabler of the transformation to an Adaptive Infrastructure, its way of delivering on the next-generation data center trend.
 
What's all the fuss?
Today, most of the IT infrastructure supporting an enterprise is provisioned and managed one machine at a time. PCs are typically dedicated to a single user; servers are typically dedicated to a single application.
When it comes to storage, the situation is similar. Most enterprise data is used by single applications, or written to storage in proprietary formats. Thus, an application-specific focus often permeates the way storage is attached, configured and provisioned.
This physical approach to isolating applications has the advantages of enabling peak-load processing capability and reducing potential confusion in systems management. But these advantages come at a price that can be measured in terms of cost of available capacity, flexibility and quality of service.
 
Delivering real benefits
By enabling an IT infrastructure to share its load intelligently, virtualisation delivers three benefits to corporations and institutions:
* It reduces IT costs across the board;
* It increases the agility of the IT infrastructure; and
* It delivers a higher quality of IT service with more consistent and predictable availability.
Organisations are grasping the benefits of virtualisation today. The University of Utah, for example, will save $2.5-million over a three-year period by using virtualisation to enable server consolidation.
A midsize manufacturer, Polaris Industries, boosted its ability to respond to business changes by slashing the time required to bring additional server capacity online. What used to take four to six hours prior to virtualisation can now be done in minutes.
German media giant Hubert Burda Media has achieved a new high in the quality of IT delivery, as measured by end-user service levels on key applications.
As a business, HP is also benefiting from virtualisation. To support our mission-critical BEA WebLogic applications, HP IT implemented a Shared Application Server Utility (SASU). This solution, based on the HP Virtual Server Environment, improved server utilisation by 300%. Total cost of ownership has dropped by 70%.
By streamlining the process of bringing new or modified applications online, the Shared Application Server Utility has shortened the turnaround time for needed changes and reduced errors.
In the development area, HP implemented a Shared Test and Development Utility based on VMware that gives developers self-service access to needed capacity.
Virtual servers are automatically allocated, configured and provisioned with the requested software. Developers now have a quick and easy way to get their code tested.
Even without considering the benefit of accelerating our development cycle, the financial value of the Shared Test and Development Utility is impressive: 75% savings on hardware and software; 90% reduction in data center space and annual support; and 60% reduction in system management costs.
 
Where to get started technically
The business benefits of virtualisation are compelling. And there are situations that lend themselves to particularly easy adoption.
For example, scheduled IT infrastructure refresh projects are a perfect opportunity to use virtualisation to reduce the costs of server acquisition, floor space and labour. This is also an excellent time to think about storage consolidation into a storage area network (SAN) or other network storage deployment.
IT consolidation projects are also excellent opportunities to introduce virtualisation. After all, once you have made the decision to modernise your infrastructure, you have the project team and the funding in place.
You are then ideally positioned to take advantage of virtualisation and make your investment go further and improve your ability to respond to business changes.
For example, you could start by consolidating the work of many older servers into a smaller number of servers that are partitioned into virtual servers.
If the development cycle is a hassle at your organisation, the Shared Infrastructure Utility for Development and Test is a straightforward way for your team to get acquainted with virtualisation technology and prove the benefits in your own environment.
The Shared Infrastructure Utility for Development and Test that HP originally developed for internal use was so popular that it's been packaged as a solution that delivers rapid results to HP's customers.
If your company is moving toward a shared services approach to delivering business applications and/or a service-oriented architecture (SOA) approach to developing and integrating applications, infrastructure flexibility is a mandatory requirement.
To achieve the promised financial benefits of sharing and reuse without losing control of the desired service levels, shared service utilities and SOA absolutely require virtualisation to handle changes and spikes in demand.
Finally, you can also begin experimenting with virtualisation simply to gain more value from your existing hardware investment. For example, if you've got lots of underutilised servers, consider consolidating using HP Virtual Server Environment or VMware to share server resources.
 
How to get started organisationally

While the business benefits of virtualisation are undeniable, some companies have found that it is more difficult than they expected to catalyse the organisation into action.
HP's experiences with customers and its internal operations lets it share insights on how your organisation can move forward smoothly in your early efforts to adopt virtualisation.
First, moving to pooled and shared equipment in the data center requires just that – sharing. Sharing means that a line of business, a department or an application team no longer has sole control over its own dedicated infrastructure. In a virtualised world, the physical server is shared and the dedicated server exists in essence or effect, though not in actual fact.
The new breed of virtualisation software, such as HP Virtual Server Environment or VMware, creates a virtual server on a just-in-time basis.
Business managers, and the IT staff that supports them, may initially be skeptical of the shared infrastructure. They want to see virtualisation proven before they entrust business-critical applications to a new environment.
Their concerns are practical in nature:
* How can our complicated environment be made simpler to operate?
* How will a virtualised system actually perform when two or more business groups or applications are experiencing peak-load requirements at the same time?
* What new skills will be required?
* Will the old approaches to troubleshooting work?
* Will new approaches be adequate?
* Will we be able to maintain our security and compliance practices? If not, can we adjust them to be compatible with a virtualised environment?
These questions must be answered before the organisation can move forward. Customers report that a test period is frequently required for the skeptical pragmatists in the organisation to monitor how the virtualisation technology works and accept that the automation will perform correctly.
Even within HP IT, there was some initial skepticism about virtualisation. The clearly documented financial benefits of virtualisation were not sufficient to get the application teams to automatically accept virtualisation as the new mode of operation.
The teams had a strong sense of responsibility for keeping the businesses up and running. They wanted to see the technology proven every which way. Extensive tests were conducted to validate that mission-critical applications would perform as expected in the virtualised world.
Even when the technology is tested sufficiently to be deemed trustworthy, established metrics may get in the way. Some companies, for example, measure labour productivity in terms of the number of servers managed by each system administrator.
Moving to a virtualised world reduces the number of physical servers that must be managed by each administrator. So administrators appear (erroneously) to be less productive when they preside over a virtualised infrastructure.
Given HP's positive experience with virtualisation, it took a different approach to justification. HP IT has established that deploying new applications on to virtualised infrastructures can be done in a much shorter time at less than a third of the cost of dedicated approaches inside HP.
Therefore, the company is no longer asking business teams to justify why they should be in a shared services environment. Instead, they're being asked to justify why a shared virtualised environment will not work for their project.
 
Paradigm shift ahead
Some early adopters of virtualisation have confided the presence of another factor behind the discussions of how to prove and justify the move to a pooled and shared infrastructure approach.
Moving from an environment of physical servers managed by people to virtual servers controlled automatically means moving to a new work paradigm. In the trenches, IT operations staff can be daunted by the requirement to tackle new learning requirements and grow into a more highly evolved role. It's human nature to shy away from change.
These concerns are not unique to the move to a virtualised infrastructure. If your organisation is moving to adopt shared services or SOA, you will have seen, heard or felt similar fears.
Organisational and cultural resistance to automation and shared resources is something that the world has seen before. It's a problem the world has solved before, simply by aligning technology with the organisation.
 
The lessons of history
Twenty years ago, industrial corporations were contemplating a move to intelligent equipment and automated control. No longer would machine tools be manual and single purpose. Intelligent "Computer Numerical Control" machines would be able to perform many different tasks in many different sequences.
No longer would automotive assembly lines be dedicated to a single model. Production capacity could be pooled and shared. Plant-floor workers would be relieved of repetitive labor. They would be required to become information workers. Work could be moved around the factory dynamically to adapt to changes in market demand.
The beginning of industrial automation was a time of enormous potential to gain business benefit. In 1987, manufacturers hoped to get product to customers on time more often, to get product through the plant in less time than ever before, and to cut total elapsed time from design to ship. In 2007, IT executives are testing virtualisation's ability to deliver greater agility for the IT infrastructure.
In 1987, manufacturers planned to cut cost via increased capacity and labor utilisation, lower work-in-process and raw materials inventory, and decreased costs of scrap, rework, downtime and maintenance. In 2007, IT executives expect virtualisation to reduce costs across the board by improving capacity utilisation and reducing redundant effort.
In 1987, manufacturing executives looked forward to making the business perform better by producing a wider variety of products with more consistent quality, while cutting down the number of customer complaints. In 2007, IT executives are looking at virtualisation as a means to deliver a higher quality of IT service with more consistent and predictable availability.
In the early days of industrial automation, enterprises were looking to the new technology to help them do business better, cheaper or faster. What actually happened was something far more powerful. For example, with the Corolla, Toyota proved that high quality and low cost were no longer a contradiction in terms. Industrial automation made it possible to make things better, cheaper and faster.
Despite the compelling business benefits of industrial automation technology and the growing realisation that the benefits compounded over time, some organisations were slow to adopt.
In the days before open standards became pervasive, it was important to make the right technical decision to avoid vendor lock-in. Even so, more than seven out of 10 organisations reported that technology was not the main problem in planning for industrial automation.
Twenty years ago, enterprises trying to harness the power of industrial automation were primarily bedeviled not by technical planning issues but by organisational issues. The benefit of hindsight shows us that survey respondents assessed the issue correctly. Very few companies failed to make the transition to the new approach because they chose the wrong technology.
Almost universally, the companies who didn't survive the change were the ones who could not make the needed organisational changes. These companies were acquired and change was forced on them, or they simply ceased to exist.
The organisational structure itself was the first barrier to planning for industrial automation. Why? The big idea was to automate the flow of information and materials from product development through detailed design through production into distribution and sales, all customised and timed to coincide with and adapt to local demand around the world.
Yet most companies had no appropriate forum even to discuss such an idea across the silos, let alone to make detailed plans.
Today, enterprising IT staff members face an analogous challenge in harnessing the benefits of virtualisation, shared services or SOA. The big idea common to these approaches is to share resources across silos.
The requirement for sharing may come at a level unanticipated by the current organisational structure. If this is the case, the IT staff requires executive support to create the appropriate planning and budgeting forum. Absent a forum for governance, it is hard to progress beyond tactical efforts at virtualisation.
With appropriate organisational support, IT staff can create a meaningful strategy for virtualisation, justify the approach, establish the metrics for evaluating success and gain the appropriate resources.
 
The lessons of experience
HP has lived through the industrial automation story. Its product design process, supplier network, factories and delivery systems have been transformed over the past two decades. From inventing whole new product categories and supply chain models to off-shoring and outsourcing, it's done it all.
HP expects the transformation to the next-generation data center to happen faster than the transformation to industrial automation.
Virtualisation and automation technologies are here today and are advancing rapidly, with strong benefits demonstrated both internally and at customer sites.
The lessons of history are there to teach us how to move forward by engaging the organisation.
HP knows from experience that organisational transformation is built not bought. Companies cannot reap the dramatic benefits from virtualisation, shared services and SOA without thought and attention to change management.
That's why HP is working actively to align technology with the organisation. This is a key to success.
HP recommends a governance approach involving:
* Standardisation of infrastructure configurations to facilitate sharing
* Education, collaboration and enforcement
* Strong IT executive support to make the shared infrastructure the default choice.
Governance plays an essential role in empowering the IT organisation to achieve the benefits of virtualisation, shared services and SOA.
 
* From an HP White Paper