Virtualization Technology News and Information
Six Strategies for Moving Toward a Virtualised Environment

Quoting SDA Asia

How many industry-standard x86 servers in your company are unused more than 70 percent of the time?  Many commodity platform servers fall into this category, since they are relatively inexpensive. Additonally, since many application teams work in silos they tend to buy their own servers when needed, because they erroneously believe it is an inexpensive way to increase capacity. Because capacity analysis is a time-consuming process, IT organisations typically make decisions for managing capacity requirements based on CPU utilisation, instead of doing a comprehensive analysis of the resources needed to sustain service through changing workloads. To ensure availability 24/7, organisations often use dedicated, redundant back-up servers, storage devices, network components, and other resources to take over in the event that primary resources fail. All of these reasons combine to cause extremely low IT resource use, which ironically costs businesses a lot of money.

While the cost of servers may be decreasing, the expenses related to maintaining them continue to rise. Costs scale directly with the number of existing servers when you add the costs of licensing, maintenance, software, and personnel. Other costs, such as floor space and cooling, actually scale faster geometrically than the total number of servers. In all cases, costs escalate independently of whether the servers are actually being used. These expenses are several times the price of a server, which means that simply adding more servers is not a cost-effective approach to getting more computing power. Organisations instead should consider server virtualisation, which enables IT to deliver sufficient resource capacity, flexibly and dynamically, without over-provisioning IT resources.

Server virtualisation involves running multiple images, termed virtual servers, on a single physical server, so it appears to the application environment that each virtual server is a separate dedicated physical server. This enables an organisation to turn a single physical server into 10, 20, 30, or more virtual (logical) servers. You can run different operating systems and multiple applications on these machines (one per virtual server), and control them from a single interface. This strategy delivers increased server utilisation, consolidates legacy applications onto one server, and increases business agility.

Most people think of virtualisation at the system level, but the entire infrastructure and related supported services can benefit from virtualisation. If you are the end customer, for example, do you really care whether the company has 30 physical servers or 30 virtual servers running on six physical servers? You likely are more concerned that these servers deliver high-performance services, giving you the availability you need, and can respond quickly to changing business requirements.

The following section talks about strategies to identify ways to move toward a virtualised environment.

Assess Resource Utilisation — Where You are and Where you Need to Be

Start by identifying how resources relate to the business services you are supporting. How are those resources being utilised? You may want to learn more details about your organisation, how technology is allocated to users within the organisation, and how processes are used. Consider implementing solutions that help you discover how to identify those resources and how they have been used over time, to ensure that your planning is based on complete understanding of business needs. The most valuable solutions will give you modelling capabilities to help assess the impact of possible changes. They will also assure that consolidated and virtualised environments are sized and configured accurately, while keeping aligned with evolving business needs.

Manage Environment from a Business Context

It’s important to allocate resources based on business priorities. In typical virtualised environments, the focus should be on mapping virtual resources to business processes. All too often decisions are based on CPU utilisation and memory consumption. As a result, resource decisions do not consider whether a service is impacted.

Today, many decisions are made reactively. With virtualisation, you can make the right decisions, proactively, and faster. If you are not managing IT based on business priorities, you could be making the wrong decisions. Resource decisions should be made based on established objectives. Base your automated provisioning decisions on intelligent analysis of capacity and performance requirements, not on simple resource trending and monitoring 'triggers' alone. Provisioning, therefore, comes from intelligent decision-making, and not automating a 'reaction'.

One company, for example, knew it was grossly under utilising its servers. By using a virtualisation solution that provided the necessary analysis, the company was able to eliminate 40 percent of its redundant fail over servers, saving over USD 100,00 per year. Ultimately, this company made the right decisions based on managing what matters.

Predictability is Critical

It’s important to predict business cycles so that resources can be aligned to these cycles. For example, all companies have a process for payroll, and they may run applications to support payroll twice a month. Therefore, it’s wise to use the payroll servers to run another service when the payroll is not being run. Thus, with the right strategy, you can reduce costs and increase agility. Predictability requires an understanding of performance and resource use over time. This discipline is known as capacity management. Without reliable capacity management, anything else is a guess. Guess low, and services are unavailable. Guess high, and resources are wasted.

Don’t Try to do Everything at Once

When you think about virtualisation, begin with one area, such as change management, and then progress to another related area, such as capacity management. For example, one company I’m familiar with did not have a change management process and lacked decision-making processes for any new applications its IT organisation developed. The company had been purchasing an average of 100 servers a month to support new applications, without considering if capacity already was available.

After this company implemented solid change management processes, however, its IT staff used capacity management strategies to search the environment and appropriately allocate resources from discovered resource capacity. As a result, the IT organisation stopped purchasing servers reactively just to meet development needs. No new servers were purchased for the entire year that followed. Thereafter, new purchases were approved only after enterprise needs had been evaluated via predictive capacity management technology. This enabled the company to drive up its overall resource usage rates.

Select an Open, Pragmatic Solution

As companies or customers move to virtual environments, they still need effective applications and tools to manage them. No IT organisation will want to use different methodologies for managing virtual and non-virtual environments. Further, it is impractical to replace the entire infrastructure just to transition to a virtual environment. Look for a solution that is more pragmatic, can work in an environment that includes products from various vendors, allows for incremental implementation, and can leverage previous investments in resources, technologies, personnel and procedures.

Take Virtualisation to the Next Level — Application Containerisation

Rather than just taking a systems-level approach, try virtualising the applications themselves. This strategy manages everything from an application perspective. For example, if your HR applications are all loaded onto one server, the container approach identifies all the elements related to each application on the server – the hardware, operating environment, management tools, and such. This strategy is enticing because, rather than needing to manage everything at a system level, you can manage functions at the application level and automatically align resources.

Making the Right Decision

You can choose from dozens of virtualisation solutions in the market. However, most importantly, the technology you select should provide intelligence. A service-oriented resource management approach should be considered as you begin to virtualise your environment. This methodology offers a predictive, adaptable environment that automatically provisions new server and software resources based on dynamically changing business requirements.

Organisations can begin to experience significant benefits by implementing this methodology, which secures just-in-time delivery of IT resources to applications based on business importance. Because this strategy leverages an organisation’s current IT infrastructure, this approach is cost-effective and relatively easy to implement. Organisations can then easily adopt new technologies as they emerge, continually gaining more value as they progress in their transition to virtualisation.

Read and comment on the original article, here.

Published Wednesday, September 06, 2006 8:45 AM by David Marshall
Filed under:
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<September 2006>