Inefficiencies are costing data centers time and money, while 43 percent of data centers rely on manual methods for capacity planning and forecasting, according to an Intel survey of 200 data center managers across the U.S. and U.K.
The manual approach is not limited to smaller data centers by any means; the proportion was found to remain the same even among the larger data centers (with above 1,500 servers).
Despite its limitations, MS Excel emerges as a popular tool and nearly one in ten resort to walking around a data center with a tape measure. Only just over half are able to benefit from using data center infrastructure management (DCIM) tools.
When asked why manual methods were employed, 46 percent said it was because they felt that the alternatives would be too expensive, and a further 35 percent feared they lacked the resources to implement a more automated approach.
“We believed many data center operators had manual processes for monitoring IT device power and thermals, but we were surprised that it was as high as 40 percent,” Jeff Klaus, general manager of data center solutions at Intel DCM, told eWEEK. “As I speak at data center conferences around the world, we continue to be the first to educate operators and cloud service providers about the existing capabilities inside the servers they have already purchased – where they can leverage powerful agentless tools to expose and orchestrate telemetry data and provide simplified analysis on their environment.”
The survey revealed 56 percent of manual planners need to devote more than 40 percent of their time, every month, to capacity planning and forecasting, indicating some people may be locked in a vicious circle–lacking the time and resource to implement a DCIM tool because so much of their time is being wasted on tasks that a modern DCIM is designed to perform automatically.
The report also found that 63 percent were using DCIM analytics to help optimize cooling efficiency, with other methods used including rack sensors and spreadsheets and hot spot audits.
The constraints within which data centers have to operate were widely acknowledged–floor space constraints (75 percent) and power constraints (63 percent) were the most common issues.
Nearly a third (32 percent) of survey respondents said they lacked enough actionable data to be able to make day-to-day decisions or do long-term planning.
“As IoT comes to the data center, more and more instrumentation becomes available and finer tuning of the environment is possible. Datacenters requirements will continue to grow as requested from the business units and the market overall,” Klaus said. “Addressing these new models where compute on demand has become a minimum cost of being in the data center business. This will continue to lead us toward tools, which help identify necessary management solutions, which some DCIM tools provide.”