Local IT twigs on to tight budgets

It is with a mild sense of panic that I have watched the recent decline of the rand, as I am sure we all have - even more so those of us who rely on the IT industry to keep the wolves from the door.

The reason I have been concerned is we import a large percentage of all IT products and a lot of our skills based on the US dollar. Up until now, in many of the meetings I have had with enterprise concerns around lowering costs and reducing IT footprint, I have been met largely with indifferent attitudes and disbelieving stares. Quotes about space, cooling and maximising usage are not a priority for IT departments at the moment.

I do understand the logic behind this thinking, as a lot of the enterprise IT departments are still focused on providing stability to the business on what is already implemented in their large, complex and sometimes cost-prohibitive systems. Often this is through no fault of their own, but rather the effect brought on by the monolithic technologies that have, for a long time, been considered the safe and mainstream choices.

I have, in the last couple of weeks, noticed a huge swing in attitude, which has surely been triggered by the economic meltdown and, to a lesser extent, by the power issues we have been having.

I have been preaching the message of increasing utilisation and reducing costs - whether via green products or innovative technologies - for some time now. I have been trying to get people to think differently and attempt to do more with less. The CFO and financial teams have been demanding this for a while, but it has often been a hard task to get this implemented by the IT teams, either because of budgets or the inability of products on the floor to cater to these requirements.

There is good news, though, as the technologies to do this are available and flourishing, and already tried and tested in the international market and global corporations. Some methods specifically relating to increasing storage utilisation and effectiveness are discussed below and should be investigated and tested in order to further stretch the budgets made available to us.

Most organisations are only consuming between 20% and 30% of their purchased physical capacity, and this estimate could be on the generous side.

Adam Day is product manager at SYSDBA.

All major industry analyst reports show that most organisations are only consuming between 20% and 30% of their purchased physical capacity, and this estimate could be on the generous side. Local information backs up these figures, with an average usage of 24.7% in customers we have audited.

There are a number of reasons why this happens. The first is that we always buy for the next several years to cover growth, and while there is no problem planning ahead, the one thing we know is technology gets cheaper.

We also overprovision to reduce workloads for ourselves. If a company has ever had to grow an application beyond its current capacity, it will understand it is a complex and sometimes risky task, so the solution up until now has been to ask for more than is required.

With these and other factors consuming our capacity as opposed to effectively utilising it, thin provisioning is a viable solution. We can overcome the issues relating to having to reconfigure devices simply by virtually provisioning what we need for the next three years, fooling the application and users into believing they have it, and then only adding as we really need it. This will drive down our capex costs, as well as boosting “green” benefits as we don't have to power and cool or house the equipment we won't be using for the next few years. Also, if the rand co-operates, we will be able to procure this equipment at lower costs as disks become cheaper, bigger and faster.

This involves determining the actual value of the data to the organisation, as all data is not created equal and some data is more important to the business than others.

What we can do here to reduce costs is to store data in an area with a cost and availability equal to its importance. All disk arrays should support a level of disk protection such as RAID; nearly all disk systems support replication; and all decent storage systems will support a different class of disk within the array, such as high speed and highly available disk and low cost, high capacity but less performing disk.

These facilities need to be used by IT departments and data should be moved to the area in our storage that best meets the cost and availability requirement. Some technologies allow us to migrate dynamically between protection levels as well as disk types without impacting the application. This allows simplicity of management and better use of our resources, once again reducing expenditure and delaying purchases as we can free up costly equipment.

There are many other ways of lowering costs and better utilising our infrastructure investments, and the abovementioned are just a few options to be considered. As always, when looking at making an investment in a new technology, ensure you do your research and take cognisance of what the analysts have to say.

With all next-generation technologies, some will become mainstream and some will disappear, but there is no doubt that thin provisioning and data life cycle management - with quality of service data tiering - have been around for a while and are, in my opinion, here to stay.

* Adam Day is product manager at SYSDBA.