The start of a new year is a time when resolutions are made – and generally broken within a few days. However, from a datacentre facility perspective, it is time to make some resolutions that can be stuck to – not on an individual basis, but at a level that can help the IT infrastructure better serve the business.
An organisation’s datacentre infrastructure is critical, regardless of how it is provisioned – it could be completely self-owned and operated, sourced from a co-location facility, or even procured via on-demand services from cloud service providers.
Ensuring that the overall IT platform remains fit for purpose and supports the business is imperative, so here are six resolutions that datacentre managers should make – and keep – to ensure this is the case.
1. Find those lost items
Like searching down the back of the sofa to find that lost change, it is amazing what IT teams can find lost in a datacentre infrastructure.
Research carried out by analyst firm Quocirca shows that it is common for an organisation’s asset database to be out by +/-20% on server numbers alone. So, if you have a datacentre with 1,000 servers, there could be some 200 assets that are missing or wrongly identified – and so are over or under-licensed.
Tracking down these lost items can help organisations save costs in licensing, as well as avoid duplication. Datacentre managers should carry out a proper asset audit. The best way to do this is to implement an automated asset tracking system so it can be carried out on a continuing basis, rather than as a one-off, high-cost activity on an ad-hoc basis.
2. Shed a few pounds
In many cases, the way datacentres have been run over the long term has led to massive inefficiency in how equipment is utilised. Quocirca research shows that many servers are running at less than 10% of their potential capacity, and storage systems are often less than 30% utilised.
Consolidation of applications and virtualisation of IT platforms can drive usage rates up markedly. Even if IT executives set and achieve a target of 50% for servers, that could free up 80% of existing physical servers.
If nothing else, these can be turned off to save electricity. Better still, IT executives can decommission them and sell them on, saving on licensing and maintenance costs. They could even keep some of the more modern servers mothballed so that new server purchases can be put back for a while.
3. Exercise more control
Organisations that have consolidated and virtualised still find that things can get out of control.
The biggest promise of virtualisation is that it is easier to provision new images of applications and functions than it was before. However, this is also its biggest issue, as developers and even system administrators in the run time environment can find it very easy to provision a new virtual image – and then forget to decommission it after it has been used.
Such virtual sprawl can lead to false reading as to overall systems utilisation, as CPU and storage resources used by these images are perceived as being part of the “live” load when they are carrying out no useful work. On top of this, every live image is using up licences that could either be used elsewhere or not bought in the first place.
Putting in place application lifecycle management (ALM) tools will help in ensuring that such virtual sprawl is controlled and avoided.
4. Get out more
The self-owned and operated datacentre is no longer the only option. Co-location facilities and cloud computing have expanded the options for how IT functions can be provisioned and served. The mantra for the IT department should no longer be “how can we do this within the datacentre?” but should instead be “how can this be best provisioned?”.
In many cases, this will mean that new applications and functions will be brought in from outside third parties, and this will mean that overall network availability has to be more of a concern.
Multiple connections to the internet are becoming more the norm, ensuring that overall systems availability is not compromised through the network connection being a single point of failure when connecting to the outside world.
5. Be more friendly
IT can be seen as the group that likes to say “no” with a “don’t fix it if it ain’t broke” attitude. Make 2013 the year where IT professionals embrace change and become better at saying “yes”. IT must put in place systems that allow their organisation to embrace (bring your own device (BYOD) rather than just treating it as part of shadow IT.
Datacentre managers must ensure that they are aware of how cloud computing works both in the datacentre and as an external platform, and be able to advise the business on the best direction.
6. Be more flexible
With all the changes that are going on in the general economy and the way IT systems are deployed and used, IT departments need to be far more dynamic and flexible in how rapidly and effectively they respond to the needs of the business. As IT and the business embrace the needs of areas such as cloud and big data, the datacentre will need to be more flexible, both at a facility level and at the platform level.
IT must avoid plans to implement monolithic components within the datacentre. Instead, they must adopt a modular strategy when it comes to uninterruptible power supplies (UPSs), backup generators, datacentre chillers and so on. This will make it easier to add – or remove – incremental capability as required, as the datacentre grows or shrinks, to reflect the organisation’s needs.
As with any resolution, the key is to make each datacentre action plan achievable and identify the value in sticking to it. From an IT point of view, a more flexible, efficient and controlled overall IT platform can be implemented by adopting relatively small changes as mentioned above. If the large scale of the value of changes can be demonstrated to the business, it will reflect well on the IT department.
Here’s to 2013 – a year that is likely to remain challenging from an economic viewpoint, but one where IT has the capability to put itself where it needs to be – at the heart of the business. Just be resolute!
Clive Longbottom is a service director at UK analyst firm Quocirca