Data Center Consolidation

Data Center Consolidation

Today a common data center looks like a chaotic assembly of systems purchased at different times, and as a rule each separate server carries out a single specialized business task using only a small portion of its performance resource.

According to different estimations, an average server load comprises not more than 30%. However, despite of only small portion of resources being used, the hardware still requires the full time care and support, which results in unreasonable spending of budget on data center maintenance. This is for a reason that every piece of hardware requires individual attention.

This problem was designated a long time ago, and even some solutions had been developed. But only now we can state that there exist technologies, solutions and opportunities for implementing those ideas in practice for a scale of the entire data center, as opposed to implementation on the level of separate elements.

Data center consolidation concept
As we introduced the term data center consolidation, we can speak about a large united “organism”, which incorporates peer hardware and software resources, constantly controls execution of business applications and reallocation of computation power depending on requirements (See also "data center virtualization" term ).

In order to build such well-organized mechanism, both classical standardization approach and newest virtualization technologies should be used.

Standardization
Standardization is beneficial in regard to ownership value owing to minimization and unification of maintenance processes as well as introducing interchangeability of physical components.

Virtualization
Virtualization allows flexible redistribution of used resources among applications, providing for balance load, while introducing a centralized control mechanism for carrying out tasks such as updating and backup.

Based on practical experience, the most demanded mechanism of server standardization is purchasing blade systems, which allow enlarging the total computation resource of the data center to the needed level. Besides, IT virtualization tools currently offered at the market feature flexibility in balancing the load between these “bricks” of the data center.

When using data center consolidation, you must clearly know all items of expenses and the affecting factors in order to estimate IT costs. The first step in assessing these costs is listing the inventory of IT resources.

The total cost is comprised of all IT-related expenses summed up:

  • expenses spent on hardware and software purchases,
  • systems maintenance,
  • administration and engineering costs,
  • end user related expenses,
  • losses due to failures and other.

While planning IT budget, both maintenance expenses and IT development costs should be considered.

With hardware and software costs being clear enough, the costs of other components (services, for example) are controversial. You should know not only the cost of the service, but also have the means to control its quality.

These tasks can be carried out by implementing IT Infrastructure Library (ITIL). This approach gives an opportunity to estimate costs of IT services, to control their quality and possibly switch to another provider thus controlling the service cost.

Company’s IT department turns from being a completely unprofitable subdivision to a business department with a well-organized system of internal mutual settlement of expenses and costs control.

It’s important to understand that if you save money today by refusing to create a scalable solution, then in future you will face the need of additional investments for business extension.

In case the service cost offered by third party provider is lower than the cost of own IT department, the service can well be outsourced.
Benefits of outsourcing include optimization of company’s expenses spent on IT department and own infrastructure, since when a provider’s infrastructure is used there is no need to maintain own one.

Therefore, company saves the money that otherwise would be spent on hardware, electricity, rooms, software and support.

Also, virtualization solutions allow cutting costs of electricity, improving controllability of the system, reducing recovery time, and in many cases reducing software licensing expenses.