Aerial view of Microsoft data center in San Antonio, Texas (43660 m2 DC area)
Paul Slater is one of Microsoft’s most involved telecommunications workers. He is one of the corporation’s employees in charge of corporate data center strategy. In doing so, Paul believes that automation of data centers, with maximum replacement of human workers by robots, is a matter of the next 10 years.
In addition, Paul Slater sees some other directions for data center development, which he recently outlined to the audience at the conference Data Center World which was held in Las Vegas. These trends, or principles of data center efficiency are outlined in the sequel.
It’s worth noting that Microsoft has quite good data centers that many experts call the most efficient data centers in the world.
1. Designing a data center according to local conditions
The answer to the question "what makes a data center efficient" is quite complex, complex. This is one of the reasons why providers of data center solutions offer different equipment, which often differs significantly from their analogues according to one or another principle.
According to Microsoft, one of the main factors to take into account when creating a data center is local conditions. For example, a data center can be built in a place where land and electricity are very cheap. In this case, it is possible to build large data centers with high power consumption without much cost. If a company needs a data center in the middle of Manhattan, that’s a different story, every square meter and watt of energy matters.
All of this may seem obvious, but many companies build data centers based on typical solutions without considering local conditions at all. In that case, you should not expect high returns at low operating costs.
2. Standardization and unification
The most efficient data center will be the one whose space is 100% utilized. The corporation tries to fill the space of the data centers it owns completely, as quickly as possible. In addition, the equipment the company uses is standardized.
In this case, the operating costs of maintaining the equipment (which can fail suddenly) are lower than in the case of using any special equipment and elements.
The company tries to manage its data centers using tools such as Data Center Infrastructure Management.
3. Projecting the possibility of change
Over time, the data center and the requirements for it will change. Therefore, when creating a data center, it is worthwhile to allow for the possibility of change in advance. After all, technology changes so quickly that few have time to adapt to them immediately. Technological solutions used today may have to be changed in six months. And right now, you can’t even imagine what will change.
As an example, replacing conventional disks with SSDs has led to some changes in data center thermodynamics. It was very difficult, if not impossible, to assume right away.
Process automation is most effective in the case of a homogeneous, homogeneous environment. If everything is homogeneous, as recommended above, introducing automation processes will be a simple task (all relative, of course).
Two years ago, Microsoft tried to standardize all its equipment and processes in data centers. As a result, it was later possible to quickly deploy some solutions to automate a number of processes. At the same time, Microsoft tries to actively participate in the Open Compute Project, using its own know-how and that of its partners.
Openness is another important factor.
5. Integrated Environment
If you want an efficient data center, Slater says, it pays to plan from the top down. You should first consider the capabilities of the processes and applications that the data center equipment needs to support, and then you should look at the equipment itself.
If everything went right, the design will result in a highly integrated system where every element is used in operation, supporting one application or service or another.
6. Reliable, flexible software
Good, reliable software pays for itself quite quickly. The software can be used to "squeeze" the maximum possible efficiency out of the equipment, even if the equipment is somewhat obsolete.
Microsoft uses cloud solutions, and the company has also started offering cloud solutions to its customers instead of separate software licenses. So now the company can scale its work, to a very large extent.
One of the most prominent examples of cloud-based software adoption is the Microsoft Office 365 service.
7. The data center must respond quickly to changes
The faster the new service will be put into operation, the better. At the same time, the service must be supported by the equipment of the data center itself. Accordingly, the initial design of the data center should include the possibility of such changes and support for new equipment. In this way, Slater believes, you can "saddle up Moore’s law ".
At the same time, he emphasizes the need for the standardization of equipment and services described above. This allows the most efficient use of your DC’s resources.