Data center operations are becoming more critical to the success of businesses around the world. Between 2014 and 2018, the construction of new data centers grew 21% each year. In keeping with the need to have better ways to handle and manage data, many more companies are making the switch to blade servers. There are a few reasons for this move. Blade servers have a much higher level of efficiency than can be had with more traditional servers. Another reason is that blade servers offer better solutions for server virtualization. The main drawback to moving from traditional servers to blade servers is the amount of high density cooling. It would not be an overstatement to say that the correct cooling plan is a real part of a company’s critical infrastructure management.
The good news is that with the right plan, a company can have a data center design that has enough cooling capacity and power. It takes some planning and effort but it is possible. Here are a few things your company can do to solve your cooling and power needs.
- Spread out the load. This is a self-explanatory bit of advice but it is important enough to repeat and elaborate on. Use multiple racks to store your blade servers and any of your 1U servers. There is often an impulse to put everything together on the same rack but that is an impulse that you should resist. When you spread out your servers, you prevent any particular rack from going over the power density that the manufacturer designed for the rack. With all of the space that is left, you will need to use blanking panels to make sure the cooling performance does not decline. In terms of high density cooling, this is the solution that is the most popular around the country.
- Allow shared cooling. You should leave enough room so that your servers can be powered and cooled at a level that is actually cooler than is needed. Once you have achieved this, you can set up a system that allows racks to donate cooling capacity to nearby racks when they have extra capacity that is not in use. This helps when you have some racks that draw less power and need less cooling than others. This also makes it possible for the overall power usage to be kept lower.
- Bring in some additional high density cooling. When you add extra high density cooling equipment to your managed infrastructure service, the value added is very high. This gives you some flexibility in how you configure the space and in how it is used. You may find that the design you have configured works for a while but when you have additional cooling equipment you have the option to increase it as you need to.
- Plan ahead. When you are configuring your space for your servers, blade and otherwise, it helps to know which racks are going to need more cooling and more power going in. That will allow you to set aside a specific area for the racks that will draw more power and require more cooling. When you are able to put a barrier between the racks that need a lot and the racks that do not, you can more effectively move power and cooling to the right areas. If you have already configured your space, this may not be something you can do but if you can keep the needs of all of your racks in mind when you set the space up, you will be in better shape.
- Keep the whole place cool. Keeping the entire space where all of your racks are housed may seem like a no-brainer but it is one thing that many companies fail to do. There are some reasons for this. It can be a lot more costly than not doing it. This is the main reason companies do not do this. This is also sometimes considered to be wasteful. If you can implement a system that keeps all of your racks cool, you will be in a better position.
At the end of the day, it may not be possible to do all of these but even one can really help.
