'Customers are over provisioning their data centers'

While many organisations are adopting cloud with open arms, some are following a hybrid model by which they store critical information on data centers and non-critical on cloud. Either ways, it looks like legacy systems aren't planning to make a quick exit, but are evolving.

According to Gartner, the Indian data center capacity is poised to touch 6.6 million square feet by 2016. Companies here are looking to build energy-efficient data centers through better design, improved system architecture and right sizing. CIOs have realized that energy and operational costs can be saved by planning physical infrastructure, minor system reconfiguration and small process changes through modern data center physical infrastructure management software tools. The new data centers are not like the old ones and organizations should realise that innovation in data center design will help reduce capital and operating expenditure.

Schneider Electric's Gay Chi Sen, director of sales DCIM software (Asia Pacific, Japan, China) speaks exclusively to Data Quest about how CIOs can optimise datacenters with data center infrastructure management(DCIM).

What changes have you observed with the evolution of data storage?

Data centers are evolving rapidly with adoption of cloud and virtualisation. Customers are managing not just physical servers, but twice or thrice as much. There can be about five virtual servers for each physical server and people are challenged with capital and operational expenses to cut their op-ex and energy costs.

What should one do to increase a data center's productivity?

Organisations have invested a lot in the IT layers of data centers - in managing networks, servers and applications, but the most critical layer is power and cooling. This is an area where customers have little visibility. The segregation of duties in the data center has created 2 silos - the facility and IT. It is the responsibility of a facility but IT owns everything on top. DCIM can be developed to break this silo for the IT guys understand the facility, and the facility people understand the challenges in IT. There is a need for them to talk and understand each others' worlds, encourage them to look more cohesively. This by itself will optimise the data center and increase its productivity.

One can get information from the physical infrastructure about the UPS, cooling, power, networking equipments etc. When you put the traditional facility information and the IT information into a centralised repository, the data analysis process is capable of providing the right information to the right user in an organisation. The information is generated for managers, IT, facilities and financial users in the organisation. For instance, the facilities people want to know about how the power environment and how the cooling is, whereas IT guys want to know how the server is, or the best location for a new equipment at data center. The management and the CFO wants to know if they are using energy efficiently, and how they can save costs and op-ex in the data center. The data is crunched to provide information that different users require and is presented in the simplest form, enabling quick decision making.

What can one do to ensure energy efficiency at data centers?

It is all about understanding your data center. Many customers are over-provisioning their data centers assuming that they will run out of capacity. This also leads to over-provisioning of power and cooling. Often, users believe it has reached optimum storage even if it hasn't. They feel they have reached optimum storage but don't realise there are under-utilised fragmented capacities within a data center which are not visible to them since they don't use a sophisticated technology. DCIM can help customers identify these hidden capacities when the customer thinks he has run out of space, power or cooling and convert them into usable capacities. In many cases, DCIM can also extend a data center's life by another two years. It usually gives them enough time to create a budget for a new data center.

What advice do you have for those handling data centers?

In their entire life cycle of data center, the maintenance and operations phase is the longest. It is in this phase that people feel challenged about optimum utilisation and efficiency. Earlier, they were not aware but today, more people are evaluating DCIM. Those who haven't started should do so right now. DCIM's penetration will grow from the current 60% to 90% in three years. CIO's should be consider this for their 2013 budget.

What do you foresee in the future?

Today, there is human intervention in healing and remediation of datacenters, but what we see next is self healing datacenters. I would imagine this to be out in about 4 years or so.