Ken Mitchell, managing director of Coolmation Group company Climate4Hire Ltd, advises how data centres can make the right air conditioning hire choice and so reduce the risk of data loss
Data centres are feeling the strain. Servers are becoming ever more powerful while reducing in size allowing even more disks to be packed into towering racks of existing data centre footprints. However, despite servers’ physical reduction their power demands remain high. As a result, power and cooling demands are on the rise in data centres, with contractors on the search for solutions to reduce the risk of their clients’ servers being damaged by temperature rises.
Unfortunately, the ‘Barbecue Summer’ now being predicted will be an added burden to air conditioning systems already struggling to maintain design indoor temperatures just as a result of increased high density hardware. High external temperatures are likely to pose a severe threat to data centres’ sensitive internal environments further increasing the risk of overheating, repeated crashing and data loss.
Therefore preparation is everything, with contingency planning the key to ensuring that data centres reduce the risk of overheating and data loss. So where should data centre facilities managers start? To draw up a contingency plan the first step is to assess the risks and then discuss them with a specialist temporary air conditioning and chiller hire company. When choosing a hire supplier the ready availability of the right equipment is essential but this works hand-in-hand with applications and installation know-how and technical support - these are key to ensure the hire specialist can act rapidly to fit the right kit, in the right place, if emergency cooling is required.
The next step is an internal site survey by the specialist a/c hire company to ensure that if emergency cooling is necessary it can be installed as quickly as possible minimizing or better still, preventing downtime and protecting valuable data. Factors the site survey should establish to ensure the correct specification of the temporary air conditioning or chillers include:
determining potential cooling loads to ensure the correct size temporary air conditioning units are specified;
assessing available space for temporary equipment to ensure it can be quickly and easily installed;
mapping practical routes available to exhaust hot air, dissipate heat or connect indoor cooling units to outdoor water chillers depending on the equipment type and specification.
There are a number of temporary air conditioning solutions that can be specified for emergency cooling in data centres and it is essential to match the right solution to the right application. Small mobile air conditioning units might be adequate for smaller server rooms but in buildings with high heat loads a good solution are specially prepared chilled water air handling units (AHUs) or fan coils, positioned at strategic locations around the server rooms to deal with hot spots and to ensure adequate air circulation. As an additional measure, circulation fans can be positioned to keep air moving in areas where it might stagnate.
When specifying AHUs for this type of application it is important to consider they are the internal half of a chilled water system. Therefore when the site survey is conducted routes need to be found to connect AHUs to existing or temporary outdoor chillers via interconnecting small diameter flexible hoses. Since most server rooms have raised floors it is usually fairly simple to install hoses beneath them, but routes need be pre-mapped on a floor plan, to save time during installation.
Another option to consider is siting larger AHUs around the outside perimeter of the building, connected again to temporary or existing chillers. This type of system will only work if cold air can be accessed via flexible ducting through the outer fabric of the building into the space to be cooled. The easiest way to achieve this is to remove windows or doors and replace them with blanking pieces through which the flexible ducting can introduce the cold air.
However, if the data centre has an installed chilled water system, it is unlikely that all the indoor units would fail at the same time, but the chiller serving them could. Therefore in the site survey the hire specialist can identify the best position for a temporary chiller or chillers. At the chosen location, a set of water connections could then be pre-installed into the building’s chilled water ring main, saving time in an emergency. Provision could also be made for an electrical supply, either from the building’s own supply or by specifying the type and capacity of generator which could be hired when required. This all adds up to valuable time savings when it comes to installing emergency critical cooling. Further time savings can be made by opting for packaged chillers since they contain integral water circulating pumps and buffer tanks, which speeds up the installation process.
When drafting a contingency plan an external site survey is as essential as an internal survey. Unless a data centre is located on an industrial estate or has an adjacent car park or grassy area, space for siting packaged chillers could be severely restricted. An external survey will detail the number of chillers required, whether it is a single large capacity one or a number of smaller ones. Where site access is restricted or where water piping is tricky it is sometimes better to locate a number of smaller chillers around the building perimeter.
Whatever temporary chillers are specified, it is important to select chillers with an element of redundancy or certainly multi circuit ones. This limits the damage which could result from some failure of part of the temporary system. Similarly on the inside, some additional air handling units should be installed to insure the cooling capacity is available in case of unit failure.
As data centres are huge power consumers power supply availability forms a major part of a contingency plan, since managers want to avoid fines for exceeding their power limitations. Hire companies therefore need to determine whether indoor units and outdoor units can operate off the building’s electrical ring main or if temporary power supplies are required via a generator. For temporary power supplies, cable runs need to be mapped out in the plan and generator sites identified.
Once the cooling equipment has been selected the hire specialist can maintain a stock of it for the exclusive use of that particular data centre. Such an arrangement often requires the data centre to pay an ’annual fee’ which guarantees the availability of the equipment when needed. Alternatively, smaller mobile air conditioning units can sometime be kept on site, with data centres paying a ’pepper corn rate’ until a real need arises when an agreed hire rate is then charged. These arrangements can all be negotiated between the hirer and data centre as part of the contingency plan.
Key to implementing the contingency plan for critical areas, is a 24/7 technical and field support arrangement, with guaranteed on site call out times forming a major part of the plan. The very best equipment can malfunction and when it does a timely response is required.
The increase in high density hardware means data centres can pack more processing power into a smaller space but as a result rising heat levels are putting pressure on ac systems. It is therefore essential for data centres to draw up temporary air conditioning contingency plans to ensure if the worst happens they are prepared and so can protect their data and IT equipment. Specialist temporary air conditioning and chiller hire companies can give data centres the peace of mind that contingency plans will be implemented swiftly if needed.
-ends
