Resources "made to glow" for COVID research

In this article you will read about,

• how T-Systems simulated maximum operation at the new Open Telekom Cloud site,
• how the test simultaneously supported the Folding@home initiative for COVID research,
• and why vulnerabilities were uncovered much earlier than usual.

Understanding COVID-19 and developing new therapies – that's what the Folding@home initiative is working on. To this end, it has created what it claims is the world's most powerful supercomputer. Its enormous computing capacity will help to decode and study SARS-CoV2 and the human proteins with which it interacts. And in doing so help to find an effective, patent-free drug. The same applies to diseases such as Alzheimer's, Huntington's, and cancer, which the initiative also addresses. 

This modern supercomputer is by no means located in the data center of a renowned research institute. Folding@home is a cloud-funding initiative based on the principle of distributed computing that uses computing power anywhere in the world for the project. Any PC user with an Internet connection can participate – from private users to large corporations. The initiative provides special software for this purpose. Once installed, the participating client receives data packets for complex calculations at times when its resources are not being used elsewhere. The computer performs these in the background and then automatically sends the results back to the relevant research institutions via the Internet.

The more computing power, the faster the research progress. It was in this context that T-Systems also participated in the global initiative – and temporarily contributed its new state-of-the-art twin-core data center and the power of all the resources there to the project. The Open Telekom Cloud’s new data center in Amsterdam was already fundamentally ready for use and was due to be tested on a large scale before the official launch. "So why not combine the necessary with the useful," says Thomas Wetzler, Senior Architect Public Cloud, T-Systems. The T-Systems IT expert built a powerful infrastructure in the cloud data center, including a large number of CPUs (central processing units) and virtual machines, with the help of Terraform, the infrastructure-as-code software. "Virtuality and automation made it possible to flexibly ramp up and ramp down the IT services." In the meantime, T-Systems contributed resources from the Open Telekom Cloud to the initiative equivalent to the performance of around 11,000 computers – helping to run Folding@home's highly complex scientific computing operations.

"This allowed us to ramp up operations in our new public cloud data center to almost full load. This was a unique opportunity for T-Systems to put the data center through its paces before it went live," says Wetzler. "We put all systems to the test – from the entire infrastructure to the quality of the technical equipment and the air conditioning. For an IT guy like me, it was a unique experience," says Wetzler, who at the same time faced an additional challenge: namely, to score as high as possible in the points system that Folding@home uses to evaluate the efficiency of its total of around 225,000 participants. "The higher the efficiency in the project, the faster you get assigned new task packages for your contribution to the research." With its computing power from Amsterdam, Telekom has in the meantime reached 47th place.

The cloud services in the Netherlands passed the load test. "We have proven that we can even handle such high-load situations," says Wetzler. Even if is unlikely that the cloud provider's workload will approach this maximum operation in everyday life. There’s a positive side effect of the project: Errors or difficulties that came to light during the approximately two-month collaboration on the cloud-funding initiative could be addressed before the launch of the Open Telekom Cloud in Amsterdam and are currently being rectified. This is an ideal case, because "weak points usually only become apparent months or even years later, when the number of customers and thus the utilization of the data center gradually increase.”

For more technical background, read Thomas Wetzler's article in the Open Telekom Cloud Community. Get in touch with our Cloud Architect and ask any questions you may have.