Seven disastrous data centre mistakes -- and how to avoid them

By Ravi Raj Data centres are almost never perfect at the beginning; there is always something to learn from your first data centre experience. At the same time, a mistake can prove to be expensive and disastrous for your business, as well as your customers’. Paying attention to some aspects of data centre management can help you get closer to that ‘perfect’ and efficient data centre. Understanding these common mistakes will help you to better manage and plan the data centre. [caption id=“attachment_2013921” align=“alignleft” width=“380”] Pic[/caption] 1. Isolating operations team from facility design: When the operations team is excluded from facility design, an organisation tends to spend more time and money on modifications and repairs. Instead, an organisation should include the operations team while determining the design criteria and performance characteristics. When you include the operator right at the beginning - in the design phase - you will build considering problems that may arise in the future. 2. People power: Organisations often underestimate the complexity of a data centre and compromise on the skills required to do the job. Provide your employees with proper training and guidance so that they can react to issues and concerns, and solve them effectively. Re-enactments and role playing scenarios are useful strategies which can help to equip employees to deal with unexpected challenges. 3. Believing there’s only one standard way to design a data centre: Each organisation faces a different challenge when it comes to data centres. There is no ‘standard model’ that serves as an example for the most cost-effective solution. Data centres are generally similar, but it is up to the organisation how they put it together, based on the philosophy of the company and what works best for it. Create your own standards that provide the results you want to achieve. Understanding challenges which are unique to your data centre will enable you to invest wisely in suitable power and cooling solutions. 4. Overcrowding the rack: Poor planning will lead to rack overcrowding- which is unnecessary. It may seem tempting to overpopulate the space in a rack, as it may save your organisation a few pennies. However, it proves inefficient in a long run - ask the employees who are responsible for plugging and unplugging components. You might even experience an outage because of accidently unplugging a server due to lack of space between systems. Instead, rack space can be used efficiently by using virtualization, consolidation and efficient arrangement. 5. Underpowered Facility: We often come across situations where a data centre has more space but no more power to offer. This is a result of poor planning for the facility. The amount of power required must be estimated at the facility - planning stage and arrangements should be made accordingly. Here too. Virtualization and server consolidation can bring some relief but are short-term fixes for a larger problem that must be looked into as soon as possible. 6. Inefficient Cooling Practices: Ensure that your data centre is at the right temperature. This basic step can help to operate them very efficiently. There is no need to freeze the systems. When you walk through your data centre, if you feel it’s comfortable for you, it’s the right temperature for your servers. In fact, it needs air flow more than arctic temperatures. Besides, about half of the energy is consumed by essential support systems like power, cooling and lighting at data centres. On an average, the cooling system consumes 37 percent of the data centre electricity, making it the second biggest power consumer after technology systems. Poor implementation of cooling technology will lead to utilizing even more power to run the IT equipment. CIOs have realized that cooling is the second highest opportunity for saving costs in IT energy, after reducing the IT equipment load. On IT Rack Infrastructure, it is recommended to have the cold aisle containment and cabinets with plinth. If the racks are on casters, ensure it is covered with panel, use of brush tile or grommet at cable entry area in raised floor, usage of air seal kit and blanking panels in the rack. All this in turn eliminate cold air losses and increase the efficiency of data center, power saving and better PUE. 7. Failure to overlay and implement documented process and procedure: Documentation creates a foundation for measuring results, taking corrective action and for continuous improvement. Reporting information about aspects like equipment lists, maintenance schedules etc., is vital for your changes organization. Data centre changes can only be implemented with a formal set of processes and guidelines. (The author is brand head, director sales & support at NetRack)