Crypto farms are big news these days. Some question the value that can be gained from cryptocurrencies such as Bitcoin and Ethereum. A few go as far as to say it is a whole lot of hot air.
In one sense they are correct. Crypto farmers cram massive amounts of computing density into tiny spaces. These banks of servers generate hot air in large quantities the successful ones transform it into a lucrative revenue source.
What determines success? How well a space is laid out to pack as much computing muscle into one place while facilitating the inflow of cold air and the removal of hot air. Unfortunately, there is no single room layout or design to help crypto farmers achieve this.
Traditional data centers were often subjected to careful design and planning all the way down to the construction of a new facility. Best practices improved over the years about the use of redundant computers and electrical systems, raised floors, hot aisles and cold aisles, backup generators, batteries, and the deployment of air handling units (AHUs). This resulted in some wonderfully efficient data centers. But there was one big problem: Cost. These facilities were expensive to equip and run.
Crypto farmers don’t have the budget, the margins or the time to take the traditional route. They typically use whatever space is available, cram in as many multi-core servers and Graphics Processing Units (GPUs) as possible, and get to work mining cryptocurrency. Some use a room in a house, a small office space or an unused basement. Others utilize large warehouses where the hot air can dissipate easily. More than a few find containers an ideal place to stack hundreds of servers in their crypto farms.
Most of these spaces, however, soon face challenges with regards to overheating, rising electricity costs and low efficiency. It’s all very well to pile banks of servers into a small office or container. But if the expelled air can’t escape rapidly, the equipment will eventually shut down.
The best way to solve these issues is simulation. It is possible to simulate room conditions to optimize the position of servers and determine how best to set up the airflow. As every crypto farm is unique, this has to be done on an individual basis.
Some may argue that there is no time for lengthy design, simulation and optimization efforts. The crypto mining rush is on and there is no time to lose. However, simulation is fast and has a rapid return on investment. Those in too much of a hurry to simulate their farms may regret it when their equipment overheats. In some cases, they spend twice as much fixing the problems caused by poor room layout.
A simple simulation can rapidly determine the optimum amount of computing density for a specific square footage. It calculates the temperature, the air velocity and the interchange of hot and cold air in the facility. A heat map is produced which highlights equipment hot spots that are likely to result in a server meltdown. This is accompanied with recommendations on how to modify the room configuration to optimize crypto mining efficiency, performance and cost.
In this fiercely competitive field, time is of the essence. But those willing to spend the time on simulation, find their mining efforts enhanced, their equipment no longer subject to sudden shutdowns and their profitability rising. By simulating their spaces, they see where hot air may build up, how best to channel it away from hotspots, and how to import cold air from outside to cool the environment. Done right, this can sometimes be achieved without the need to purchase AHUs, coolers, large fans or other equipment.
Maya HTT helps crypto miners and engineering firms to optimize cooling efficiency and compute density. These simulations validate room design during construction without the need for delay. They can also take advantage of free cooling and natural ventilation to eliminate some of the biggest crypto farming expenses.
Maya HTT has worked with over one hundred data centers and crypto farmer sites. Maya HTT’s expertise in Computational Fluid Dynamics and simulation helped them establish highly productive, efficient and cost-effective computing spaces.