Welcome to the fifth and final part of this series on choosing storage for your high-performance computing (HPC) solution. In the previous blog posts, I talked about speed, reliability, simplicity, and scalability. In this post, I will cover a key consideration upon which many business decisions are made: cost.
By nature, HPC environments are complex and require some of the highest-performing hardware available on the market today. In some cases, like large research laboratories, the HPC environment is constructed with highly customized systems. Depending on the number of nodes and the speed of the processors, deploying an HPC solution can cost anywhere from $10,000 for an entry-level system to $600 million (or more) for the fastest supercomputer.
However, when considering the cost of an HPC solution, the initial investment in hardware and software is just the beginning. To more accurately assess cost, you need to look at the total cost of ownership (TCO) for the solution. That means that in addition to the deployment costs, you also need to consider the costs of operating the solution once it is deployed, including the cost of:
- Energy to power and cool a high-wattage system
- Data center space to house the system
- Training your IT staff to operate the system
- Managing and maintaining the system
- Scaling the infrastructure as business needs grow
The NetApp® HPC solution built on the NetApp E-Series storage architecture delivers industry-leading performance and reliability while lowering TCO. Ultradense drives enable you to store more data on fewer drives, helping to reduce rack space by up to 60%, lower power costs by up to 40%, and reduce cooling costs by up to 39%. Nondisruptive scaling of performance and capacity reduces downtime costs. A single platform that is easy to manage, automation of common tasks, and proactive monitoring and support help reduce management overhead. All of these features add up to significantly lower TCO.