Advertisment

'Data center efficiency curves becoming flatter'

author-image
CIOL Bureau
Updated On
New Update

BANGALORE: The data center is one zone where challenges keep surfacing one after the other. As more companies join the ‘green’ bandwagon optimizing efficiency seems to be top priority. For some others, given the current economic slowdown curtailing operational costs of the data center could be a key concern. Amod Ranade, India Country manager, American Power Conversion (APC), discusses some of the major innovations in its product portfolio and ways of maximizing energy and operational efficiency in an interview with Sigi Achappa of CIOL.

Advertisment

Excerpts.    

CIOL: Many organizations are looking to modernize their data centers. What are the trends that are impacting data centers?

Amod Ranade: The main trends from the infrastructure perspective is in the physical layer – the UPS, power distribution, racks, access control. The biggest trend is that our customers look for is improve efficiency in all the systems that is the top most concern. In the past, reliability of the solution was the major requirement but today all most all vendors are reliable that is a given, but efficiency is the most important issue. We have also seen that if we division the life cycle of a data center into four parts –the  building of a data center, operating a data center and carrying out changes in the data center. What we have  been trying to do is build products for each of these stages and help customers modernize these levels, thereby increasing efficiency.

Advertisment

CIOL: Increasing efficiency and reducing operational costs of data centers is a key concern for most companies these days. How is APC innovating its products to cater to this need?

AR: Yes, first we have introduced a new architecture for designing data center called the  –InfrastrXture – it is a modular design for high density and high efficiency predictable infrastructure.

We have lot of products and services that ascertain the present level of infrastructure and asses how we can improve for benchmarking, for data center audits or energy audits or availability.

Advertisment

We have software to remove human error and speeden the design processes. The turn around time to design a data center has been compressed with our techniques.

Second, is the actual building of the data center we have new technologies that are hot pluggable, where a large amt of engineering work is eliminated. It is like a kit on the site which is already tested to be compatible with the other components.  The only engineering part left is the input cable everything else comes pre-integrated.

Another thing we have done is divide the four components which help reduce operational costs  of a  data center .

Advertisment

First, we address the components in power and cooling arena where the component itself is more efficient. For example, the expected UPS efficiency is normally between 90 percent  to 91 percent but our products give 93 to 97 percent efficiency, which helps increase overall efficiency. The electricity bill reduces and because it more efficient it produces lesser heat  and requires lesser sir-conditioning. 

Since the UPS systems are more efficient when running with a larger load, our components have been designed up to 30 percent of capacity availaibility where its efficiency will be maintained.

Efficiency curves are becoming flatter, upto 30 percent, after that it will start reducing a little. 

Advertisment

Another thing is we have a new architecture – Symetra where the whole UPS is modular, the same UPS chassis can increase capacity or reduce, while the UPS is on. In India IT refresh cycle is longer compared to companies in the west. If normally it is 3 years in India it is five years and expect three refreshes. But you only have to put in a part of the component in modules to increase, so we have 6 KV to 1.6 MW in modular UPS systems.

The second area of innovation is in the power distribution part - how to size the output distribution of power. Converting 50 KV chassis to 100 KVA chassis basically for  powering more IT to increase out put. So, if I have a modular UPS and I put in one more UPS but  I need 10 output but I cannot change the panel while the old ones are running- so we have innovated on modular power distributor where even the output circuit breakers can be added, changed or removed without the need to power the electrical panel. So the second important opportunity is cooling – we have a product called Inrow-a cooling architecture and thermal containment techniques help customers in reducing energy for cooling by close to thirty to fifty percent.  We have targeted three main areas: the first area which effects the refrigeration or cooling is the air - we avoid all possibilities of mixing hot air and the air which goes into AC is hotter so, AC functions better. We avoid mixing the hot air, so air in its hot state goes into the AC and improves the efficiency of the system and capacity.

With inrow cooling we have reduced the average part required for hot air to travel from 80 feet to 8 ft – ten times lesser.  

Advertisment

The third thing is humidity control. All our ACs have a sensible heat ratio of 1, where the humidity in the room is maintained. 

In the operations phase  our infrastructure solutions are automated for remotely managing to same extent as an NMS (Network Management System). The management we use has the same protocol as NMS - TCP IP. So the customer can have a single window view of the whole data center.

CIOL: What are some of the tools available to gauge the results of the optimization efforts?

Advertisment

AR: We have capacity management software and change management software to get a real time view of the capacity utilization and management of the data center up to particular racks –so with the software you can choose whether to shift capacity between racks and  also find out where you want to put the next server. We can also simulate the position of the servers. These tools are sold separately and can integrate with other vendor equipment.  

CIOL: How has virtualization impacted the dynamics of the data center?

AR: There are certain challenges of virtualization which we address. With virtualization the CPU usage of the main server which hosts multiple applications will go—hence it will geberate more heat - rising density . We make it easier for the customers by addressing such challenges.

CIOL: Cloud computing seems to be gaining momentum lately, curtailing the need for companies to expand their data center—what bearing will it have on APC’s business and strategy? 

AR: Cloud computing from where we are seeing it, the total computing requirement of the world is increasing at a very high rate—almost doubling (storage) so more computing means more data center requirements. With cloud computing the only difference is instead of many small data centers customers are seeing that it is better to have less number of large data centers.

But, when we say cloud, we call the cloud because it is not very visible, there  has to be a data center somewhere.  So, we are observing it for a long time, ever since the computing waves have been changing in the market and we are participating in it. Irrespective of these waves there are applications which demand small server rooms or many small data centers and there are many factors which influence that decision. So, because of this trend there is more focus toward providing solutions,  than components. Earlier an organization had to source components  and put in engineering effort. With a solution approach customers are looking for an operational data center- they are more interested in turn key solutions. So we are able to give the full benefit of the technology..