Information Technology

Every IT concern has a data center which can be viewed as a server location, square footage and the CPU cores but is the data center evaluated on solid metrics to determine the value or performance of the data center. Now, this is a big question as the value of the data center indicates the power utilized by the users at the data center and the costs associated to it.

Measuring the data center metrics is an issue because most of the current data centers are built and expected to deliver the peak demand levels. An argument has been put up to build the datacenters away from the dense urban areas and rather locate where there will be less utilization of power and less costs. A lot of waste has been noticed so definitely there needs to be certain out-of-box solutions to have and maintain the perfect data center. A need is to rewind servers and resources which can save up to 25-30% of power and to use tools and virtualization so that servers can be reallocated. Some network administrators also advocate that the architecture of the network should be as flat and simple to have lower latency and higher capacity.

The management of data center is essential in order to improve upon the metrics and if the data center is correctly measured, it will allow the data center managers to bring specific changes and make the center stay under control. Metrics can be called as an important tool to judge the uptime and reliability of the data power center.

A data center being a vital part of an overall IT infrastructure is responsible for the growth of any enterprise. The enterprise application software like ERP, a host of CRM applications, BI etc. have enabled the need to lift up the data centre. It is to be understood that whenever a data center will encounter any problem, the business in any enterprise will suffer because all these enterprise applications are seated here.

A real measurement of data center can be done by measuring the power and the cooling and a balance needs to be struck between these two. Costing is another factor which needs to be kept in mind when evaluating the data center. Other metrics like environmental, physical access, network performance, server performance are also important. Many a time, the peak and the average of the CPU utilization and disaster recovery utilization is measured for backup and testing purposes. Data center metrics in such a way enable the data center manager and often the CIO too to judge the performance of the data center and thus add value to the business.

Apart from all these, what other things a data center manager needs to look for as other parameters of assessment:

* Network bandwidth/connectivity.
* Server horsepower to assess the efficiency.
* Supply and cooling of power.
* Data storage capacity.

IT enterprises such as Wipro, HP, IBM and sun Microsystems use variable metric systems to keep their data centers under control and to boost up their performance.

For instance, Wipro keeps it’s network bandwidth utilization below 70% and the temperature to be near 22 degree centigrade. CPU utilization is allowed up to 80%. The standard uptime with regard to network availability is 99.5%. The storage capacity is measured and also the employee’s mailboxes are tracked to limit the storage capacity. Wipro has now also started measuring the temperature and power of it’s data center. Utilization of the space or rather location of the data center is another very vital issue which is being addressed by the datacenter.

What is the solution to address the problems faced by the data center?

* The need is to go for environmental monitoring solutions which can proactively monitor factors like temperature, humidity, and other environmental hazards or natural disasters. If an enterprise has these kind of solutions, then one can be alerted beforehand and keep the physical security under control. An enterprise can also then remodel it’s data center.
* High-density server racks and blade servers should be incorporated into the IT infrastructure.
* The use of PUE (Power Usage Effectiveness) and DCE (Data center Efficiency) should be measured on a continuous basis.
* Shifting of power to a three-phase will allow the data center to operate more efficiently since higher voltage will reduce the requirement of high amperes.
* Virtualization of servers is one another good solution so that there will be more runtime on the same servers and the servers can be subjected to repeated uses. A few enterprises such as Cisco and VMware have been particularly keen on this aspect.
* Energy efficiency benchmarks should be used in order to know what kind of servers should be used for different kind of workloads.
* Strict adherence to security norms is also important since it will allow cutting down on costs and investments on power, equipments and also human resources.

By having a clear data center metric system, a enterprises can upgrade their data centers and boost their productivity.