Intel Data Center Tour

Today we want to talk about modern server parks and data centers of Intel, which has come a long way before working out a specific strategy in this area. Historically, it was that in those places where the company bought or built buildings, data centers were also located for the needs of the development departments there. Experience has shown that it is not very effective. With this approach, the number of Intel data centers in the world has exceeded 150 and, accordingly, they have become a very expensive part of the corporation’s resources. Therefore, a number of decisions were made to reduce the total cost of ownership of data centers. One of the main directions was their consolidation in order to reduce the number of data centers and server rooms, as well as increasing the utilization of the computing resources themselves. It is obvious that several small server rooms with a load of about 40% in maintenance cost more than one large room with a load of 60-70%.

And if at the very beginning, the global goal was to reduce the number of data centers to the maximum, almost to 18-20 pieces (from 150, as we remember), over time it became clear that this is physically impracticable and still some kind of equipment should be located in remote locations, not just on large hubs. Because WAN channels at the moment do not allow to consolidate everything to such a high degree, and a lot of things prevent or make impractical such a concentration of server resources. Therefore, it was decided instead of such a global consolidation to move to an increase in utilization of data centers. After all, the main goal is to minimize capital investments in data centers, that is, not to build new ones, but to effectively use existing ones. Building a new data center is a huge amount of money, millions and millions of dollars, depending on its size.

This stage of the data center strategy took place at Intel a few years ago. A huge program of consolidation activities was implemented, which turned out to be very successful.
')
Consider the main points of the organization of data centers of the company on the example of the Russian data center. The first question that most likely arises from a habraiser is what unique can we find in a given place?

First, the division of the data center into zones with high and low density of computing resources. Since the dimensions of the office are not so gigantic, all this is combined in one room, where racks with low-density equipment are located separately from the racks with high-density equipment. These are, for example, racks with disk arrays, emitting about 4 kilowatts of heat (more such racks simply do not emit), or racks with network and communication equipment. Of course, it does not make sense for them to build cooling and power in the same way as for racks with computing servers. For low-density equipment, less electrical power is supplied to the rack, and if you look at the room from above, in the part where such equipment is assembled, one row of air conditioners is placed. For equipment with high density air conditioners are arranged in two rows and blow the racks on both sides.

Secondly, the free-cooling system used here is free-cooling. In fact, there are two technologies of free cooling - direct and indirect cooling. In the first case, the air from the street is chased directly to the cooling of the servers. With indirect free cooling, there are chillers that cool the water during the warmer months, but with a sufficiently low outside temperature, the external circuit, usually removing heat from the chillers, is directly operated through the heat exchanger to cool, and the chiller simply turns off. For Russia, it is more efficient to use an indirect cooling system for a number of reasons. In general, they will be described in more detail in one of the following materials on Intel's strategy for data centers.

The third point is the introduction of a heat recovery system as a pilot project, which was originally part of the overall environmental program of Intel. The system consists, relatively speaking, of a device that is connected to a hot pipe and produces hot water, using the difference in temperature between the hot and cold pipes. As with the cooling water for precision air conditioners, there is also installed a capacity in which hot water is accumulated. Its temperature is about 60 degrees. This water is used for the needs of hot water supply of the building (more precisely, its office premises). First of all directly, i.e. served in washbasins. Also in the building there are, for example, two cafeterias where dishes are constantly washed, or, for example, there is a shower, etc. In this building, the hot water from the centralized supply is not connected at all, so there are electric boilers. Turning them off, of course, saves money. The second goal in the implementation of this project is the supply of hot water to the heating system.
Intel is actually a pioneer in this business and the main task is to find out how much heat can actually be generated by recuperation, how best to use it in order to get the greatest savings, etc.

It is still too early to say how much this system is capable of reducing the cost of cooling the data center, and what specific savings can be obtained in meeting the needs of the building. The system was installed in December 2010 and launched in January 2011. The system works in trial mode and is gradually gaining momentum. It is still too early to judge the real level of cost reduction and the possible scale of deployment of such a recovery system - its results will be measured in the long term. However, today we can say that Intel has a completely unique approach to the organization of its data centers, using innovative methods and technologies.

You can read the full version of the interview with Leonid Shishlov following the link to the Intel community of professionals.