The promise of server optimisation

Despite significant investments in IT infrastructure and advancements in technology, server utilisation still remains a somewhat downplayed issue that is yet to be resolved.

It is no longer enough to guarantee server availability. Companies must now manage server performance and costs, while meeting the dynamic demands of business across the enterprise.

To set the stage for the server optimisation process, it is important to understand the problem and how it originated.

Through the 1980s and 1990s, when a new business application was needed, it was created in an independent system environment - contained within its own server and storage infrastructure.

Utilisation then was planned for peak periods, which worked well in most cases. However, today the growing complexity of IT infrastructures has led to more and more applications, with companies believing they have to purchase new servers or overprovision capacity for existing ones.

Subsequently, it is estimated that many servers have excess capacity sitting idle of more than 75% - a very costly way of doing business.

The solution to the above is an application-and-vendor-agnostic on-demand server utilisation system that can pull from an available pool of resources to automatically self-heal, ultimately providing ongoing availability and performance to end-users.

And here are a few steps that will put companies well on their way to an on-demand server utilisation system.

In order to minimise skyrocketing costs asset costs, it is critical to perform an analysis of hardware and software configurations. Indeed, asset management helps organisations to answer the following types of questions: "How do we license software?" "What changes are needed?" and "What are the projected costs?"

Through asset management companies get to know their environments, and more significantly, how to contain costs. From a consolidation perspective, asset management identifies servers, memory, CPU, software running on each box and the associated costs.

It is key that companies define what a certain service means to the organisation and to determine the current service levels by using clear, quantifiable metrics.

Many organisations have defined SLAs (service level agreements), yet they often lack the methodology needed for measurement, or they track application performance manually.

If companies can begin to establish baselines by tracking the response times of business-critical applications such as payroll and human resources (HR) it will give them an insider`s insight into how their infrastructure is performing.

And it need not be overwhelmingly complex, the metrics can be as simple as looking at weekly/monthly reports or calculating the number of open desk issues. This all will enable companies to react quickly to negative changes before business is affected.

At this stage, companies can leverage the information obtained at levels one and two and execute informed, intelligent and dynamic techniques to optimise server efficiency and reduce costs.

These many include tuning, clustering, server consolidation and/or virtualisation - management of service levels over this period are again critical.

Companies might at this stage choose consolidate compatible applications with opposing peaks and valleys, therefore resulting in consistent server utilisation levels.

In the ideal on-demand world, a server can automatically request resources such as extra CPU, and a server with extra capacity can automatically provide it.

By following steps one to three, companies can now begin to optimise server utilisation across the enterprise. However, in order to this they need to follow cyclical, repeatable processes - this is where other consolidation projects often fall short.

Dynamic, self-managing provision needs to be the end goal. Automated processes must recognise and function within the context of the business in order to make the notion of on-demand a reality.

This all may even involve changes to procedures and methodologies within the business organisation itself. It will also take a gradual and granular approach, applying specific solutions to a manageable problem or opportunity within the business.

Ultimately, companies will then achieve server utilisation - adhering to the direct needs of their business while also establishing an on-demand environment for the future.