Taking the fad out of data centre best practice

Efficiency Drive
From spiralling energy costs to increasing data demands, data centres have crept up the corporate agenda in recent years. There has been a growing focus on data centre optimisation strategies, with organisations looking not only to reduce energy usage but also increase flexibility and support business change.

Given the lack of efficiency within most data centres, the opportunities for improvement are clear. Research analyst, Quocirca, estimates that overall energy efficiencies are as low as 1.3%. This means just 1.3% of the energy that comes into the data centre is actually available to drive useful data centre work - due to over provisioning of IT equipment, low utilisation rates and dependencies on the national grid for energy transmission and distribution.

As a result there has been any number of new initiatives and technologies designed to improve data centre performance. From the energy efficient promises of vendors and the use of virtualisation, consolidation and rationalisation, to the introduction of hot and cold aisle containment, more effective cooling provisions, such as free air and water cooling and even better cabling – it would appear that data centre managers are now overwhelmed with opportunities to improve efficiency and agility.

Or are they?
So many of these products and initiatives look good on paper and may win the Chief Information Officer (CIO) plaudits from the board, but how effective are they on the ground? What is the reality for the data centre manager tasked with implementing the latest strategy: is it best practice or simply a fad that will be disproved next year? While your gut feel may be strongly for or against, how can you prove the case; or present a viable alternative strategy?

Fact or Fantasy
Such decisions should always be based on fact. In reality, they are based on generic industry averages and the equipment power efficiency plate values touted by vendors. Neither of which have any relation to the actual experience in the vast majority of data centres that have not only grown organically but are home to equipment that ranges from brand new to 20+ years old. These facilities may have been expertly designed once; but since then have fallen prey to the quirks of fashion and haphazard investment strategies. For all but the lucky few running a state of the art facility, the reality is that simply keeping the data centre cool, dry and clean is a major achievement.

So what is the option when the CIO demands free air cooling? Or hot aisle containment?
No data centre manager wants to just implement the latest strategy and sit back; the issue is not simply whether energy efficiency improves – or not – but to understand the overall impact on system performance and the risk of downtime. Data centre managers need visibility and control; they need to be able to proactively manage the environment, understand the impact and make changes if required. Critically they need to be able to say: is this strategy saving money? If not, why do it?

So how can a data centre manager achieve both control over the environment and visibility of performance – and gain an accurate picture of the impact of any new strategy and/or technology on efficiency? A Data Centre Infrastructure Management (DCIM) solution is the obvious starting point. Yet while a complete list of every component in the data centre is essential for assessing whether or not there is space to add another server, it does not help with efficiency monitoring or capacity planning. To really understand whether or not a new design, cooling system or power model is working requires more than knowing there are four Dell servers in the data centre and their average power consumption figures: you need to know exactly how that equipment is affecting the overall environment and vice versa.

Environmental Insight
Real-time environmental monitoring is the only way to accurately track and monitor the impact of new data centre designs and strategies. Rather than relying on multiple sources of information, from in-cabinet power display units (PDUs), chillers, air handling units and UPS systems, the data centre manager needs a single, consolidated view of environmental performance. The ability to view information from any environmental meter, irrespective of age, type or design, is the key to attaining an accurate and in depth understanding of how any changes to the infrastructure is affecting overall performance and efficiency.

Combining this monitoring and half hourly data logs with the complete asset register delivers the platform for far more intelligent data centre management. The data centre manager can focus in on specific areas, such as the introduction of cool aisle containment, and gain immediate visibility of the impact on efficiency. For example, making small changes to temperature and tracking the resultant change in energy consumption provides rapid insight into the optimal temperature for that environment.

Data centre innovation is essential – but not all ideas work in every environment; the key is to provide the data centre manager with the ability to both monitor and track performance and demonstrate to the board the true value of changing performance.

Conclusion
The data centre has seen unprecedented change over the past few years. Yet the reality is that while the shift towards the Cloud, co-location and virtualisation and the cost of energy/ carbon reduction issues may have raised the profile of the data centre at board level, the real innovation and new thinking occurs on the ground. However the case has to be proved: from applying for carbon credits to justifying new equipment and improving procedures, vague promises of greater efficiency just won’t cut it. When it comes to calculating the value of innovation, accurate and trusted environmental metrics are essential.

What is the difference between fad and best practice? Without real-time environmental monitoring, that question may just have to stay unanswered.