There’s a perception that virtualisation has almost eliminated the need for capacity planning. But, Herman van Heerden, Managing Director of Starship Systems, a local technology company specialising in the design, supply and deployment of virtualisation technologies, says this is a misconception.
“Capacity planning should be a priority and an ongoing function in all data centres for all resources whether they’re physical or virtual. When there’s no proper and effective capacity planning, resources are either underutilised or there’s a desperate rush to add capacity when something goes awry. This reactive approach is not efficient and or cost-effective,” he says.
Van Heerden says that proper capacity planning is not only about ensuring that computing power and storage aren’t wasted. It’s also about ensuring that there’s enough space, hardware, software and processing resources to service an organisation’s IT needs now and over some future period of time.
“In virtualisation, efficient capacity planning is about making sure that current infrastructure can handle a failure of a host. Yes, wasted computing power and wasted storage are a waste of capacity, electricity and money. However, while running your servers to full capacity for instance might leave you feeling satisfied that wastage is being kept to a minimum, there’s always the risk that if something fails, there won’t be enough capacity left to carry the load, and hardware acquisition will stand in the way of restoring your working environment.
“A better solution is to plan for this kind of eventuality, leaving enough capacity free to carry a single failure and allow your people to continue working without interruption.
“From a storage perspective, effective capacity planning ensures that you have enough space left to accommodate that inevitable data influx.
“Ultimately the goal is to allow for new capacity to be added just in time to meet anticipated need but not so prematurely that resources are wasted.”
“Prudent capacity planning is therefore about evaluating what you have, optimising the use thereof and keeping the eye on the future so that reallocations and new purchases are catered for proactively so that business continuity is not interrupted by under-capacity.”
Van Heerden says the days when companies could purchase new technologies without planning first are long gone. Purchasing decisions nowadays need to be informed and in line with business requirements.
“Haphazard technology acquisitions are not cost effective because they just result in sprawling, uncontrolled infrastructure with storage and servers that are wasted on applications and services that are hardly ever used, if at all. Capacity planning in respect of new technology acquisitions makes the difference between millions of Rands well spent and millions of Rands wasted on a big white elephant.”
He concludes with these hints:
* Internally, a strong management team with the relevant technical knowledge should be able to create an understandable schematic of the existing infrastructure. This provides a snapshot overview of which components will carry which loads if any of the units go down.
* If you don’t have the skills in house, consult a third party expert who can do the audit for you and ensure that your IT team is up-skilled in the process.
* Document everything, and if anything changes, document the change and update the master schematic.
* Don’t see capacity planning as a quick task that an already overworked administrator can do on the side. Capacity planning should be an ongoing function carried out by a competent capacity planner, network architect or planning team.
* Do “human compression” of your data: manually look at the data you have and the servers you run. You will find before too long that some systems duplicate what others are doing, and no software program will be able to identify this.
* Restrict the access to the administrator section of your infrastructure. If your company has grown into silos, allowing every department’s IT expert to have his/her run of the central systems is a recipe for disaster.
* Have a central person or team that keeps tabs on what is required and is responsible for managing new allocations.
* Before rushing into any new technology acquisition, conduct a thorough audit of existing infrastructure and evaluate it against the company’s future needs. From there, evaluate your options to ensure that you get the most bang for your buck from new technology purchases.
