By Vimal Kaw, Head, Data Center Services, NTT Ltd. India
The Covid-19 pandemic has made digital a huge priority. Every business now has a touch of digital or is forced to adopt a digital strategy. Given the pace that most businesses operate today, it is extremely vital that organisations redesign their data centers to cope up with the new realities required by a rapidly changing digital ecosystem. Modern data centers have to be redesigned keeping in mind that users can be based anywhere in the world, and not restricted to the headquarters or static locations.
An optimal data center design is extremely critical in ensuring that enterprises can scale up quickly and cost effectively. If this is not done right, it can lead to disjointed systems with inefficient processes and sub-optimal performance. If the data center design is engineered by taking into account some of the best practices, it can lead to significant cost savings with a greater ability to scale. There are several lessons to be learnt from hyperscalers, who are redefining how data centers must be built. Firstly, hyperscalers are designed with a blueprint or a framework that promises massive scalability, extremely good reliability and reduced risks.
Taking a cue from hyperscalers, all data centers have to incorporate a modular approach. For example, in a hyperscale data center, individual physical components can be replaced without any impact or downtime. This modular approach helps data center players in progressively advancing to a design that is more flexible and customised to their short-term needs, while keeping the long-term goals in mind. This modular approach enables data center players to maintain a high level of system availability as there is no single point of failure.
From a data center design perspective, power and cooling is an extremely important area. With many servers packed in a small area, today data centers have a greater probability of being overheated, which can further lead to downtime. Data center players hence need a well-planned data center design that helps in efficient air flow and heat management, which can keep existing machines running smoothly. One can also consider exploring newer cooling management techniques such as liquid cooling.
A good data center design should also enable organisations to automate as many processes as possible. Automation templates help in deploying the IT infrastructure in a consistent way across many data centers. It is also equally designed for a data center that is low in latency, given the increased demands of remote working and rise of edge data centers. As data center needs are growing extremely fast, it is important to provision for space, as there may be a need for increasing the number of servers, racks and aisles for expansion. An efficient layout plan can help in planning for short term as well as long term needs, which can take into consideration an improved layout and also better optimise air flow.
To ensure redundancy and a higher level of reliability, the data center design must provide for incorporating a minimum of two separate power sources from different service providers. There must also be an option of tapping power from renewable sources such as wind and solar energy plants. Similarly, water conservation will also play an extremely important role, which needs to be planned at the design stage.
The pandemic is a good opportunity to relook and redesign your data center for the future. By following some of the best practices in data center design by hyperscalers, organisations can look at optimising their data centers which can ultimately lead to better cost savings, a better ability to scale and reduced environmental impact.
