Containerised solutions offer the data centre owner an efficient way to keep control of cooling costs, says Chris Smith
In today’s ‘always-on’ business environments, company owners and managers face unprecedented pressure to scale up, meet fast-changing IT needs, rationalise sprawling server estates, enable hybrid cloud options – and all the while keeping a lid on costs.
In trying to square this performance circle, company managers are now starting to take another look at a previously unlikely answer to this demanding brief: modular data centres or prefabricated modular units (PFMs) that are capable of transforming organisations’ computer room operating, particularly cooling costs.
Cloud computing needs are increasingly redefining organisations’ computing capabilities.
As IT teams increasingly deploy server estates and storage that are ‘virtualised’ on the same infrastructures, analyst Gartner predicts that half of enterprises will have some form of hybrid cloud deployment by 2018.
Even the Data Center Users’ Group predicts that two-thirds of data centre computing will be carried out in the cloud in 2025.
Innovation around computing infrastructures has tended to focus on server manufacturers’ breakthroughs for their customers, led by the rise of products such as the ‘V-Block’ class of units.
But this focus on computing needs alone fails to address the other parts of the data centre equation: these converged units are beyond many organisations’ operating budgets – and often their power supplies.
What are the options for smaller enterprises and mid-range firms that urgently need more capacity or to set up their own hybrid cloud?
Container units can redefine organisations’ costs because their server racks, cooling, UPS, generators are all built and tested in controlled environments prior to deployment.
They provide smaller firms with optimum performance for virtualised servers, storage environments and cheaper cooling that was previously the preserve of just the bigger enterprises that could afford to specify and build dedicated data centres.
New containerised datacentres enable different configurations for their power and cooling equipment as well as their new converged computing infrastructures.
As a result, UK businesses can now design in lower CAPEX and OPEX costs before deploying their hybrid cloud or extra compute capacity options.
New cooling possibilities
Most significantly of all for an expanding business’s long-term computer room operating costs, pre-fabricated containers are specifically designed to be self-contained – with modular elements, server racks, UPS and refrigeration units that can all be swapped around while maintaining required hot/cold air flows, even where the container unit is added to an existing building.
Because container designs ensure efficient airflow, whatever the room set-up, they open up the possibilities for two further options that could secure long-term cooling cost savings.
Firstly, buyers can specify very efficient direct air technologies and low-cost direct or indirect evaporative (adiabatic) cooling systems as an integral part of a container data centre.
Today’s adiabatic units regulate room temperature and relative humidity to ASHRAE A1 standards.
With few components, these units are very reliable and they use fewer than 100 litres of tap water per hour.
Secondly, because container data centres are engineered with optimum cooling capacity, they open up the possibilities of free cooling, which provides further energy-efficient operation depending on local ambient temperatures.
The pre-testing and pre-assembly of containers means there is no need for the prolonged and expensive in situ temperature testing and modelling and the retrofitting of expensive-to-run computer room air conditioning (CRAC) units.
With these different cooling options, data centre professionals are at last starting to take control of their cooling costs.
What do these advances mean in practice? One UK supplier estimates that a containerised datacentre could make projected ten-year Opex savings of approximately £21,000 for an entry-level 6 kW, 2-rack 42U/600/1000 container alone, or £175,200 for a 50 kW, 16 Rack 42U/600/1000 unit (based on a PUE of 1.1 versus 1.6 and a cost of £0.08/kWh).
This, combined with a 60 per cent Capex saving (against Uptime Institute’s 2013 cost model), is surely compelling.
Small footprint containers can also rapidly be added to existing offices or car parks, thus dealing with expansion demands.
They are also overcoming the perception that they are aren’t robust enough for harsh environments: today’s units are meeting security and fire resistance to EN1047-2, and water & dust protection to EN60529.
The suppliers can provide insulated units ensuring fire detection and suppression to EN12094-1.
All this means that the containers are now likely to survive the type of fires that would destroy entire office buildings. Many of the UK’s bigger companies undoubtedly see converged infrastructure and V-Blocks as the way forward.
However, the standardisation of pre-built datacentres can combine converged infrastructure with the power and cooling performance uplift – advances that upgrading an existing data centre cannot realistically match.
In most computer rooms, scaling up or rationalisation means extended server, power and cooling unit reconfiguration, leading all too often to open-ended running cost calculations.
But today’s containerised units are not only transforming UK companies’ computing capacity, they are
also delivering predictable cooling costs.
Chris Smith is sales director for data centre management specialist on365