We worked with our financial services customers to develop an open-source, scalable, cloud-native, high throughput computing solution on AWS — AWS HTC-Grid. HTC-Grid allows you to submit large volumes of short and long running tasks and scale environments dynamically. In this first blog of a two-part series, we describe the structure of HTC-Grid and its objective to provide a configurable blueprint for HPC grid scheduling on the cloud.
The Financial Services industry makes significant use of high performance computing (HPC) but it tends to be in the form of loosely coupled, embarrassingly parallel workloads to support risk modelling. The infrastructure tends to scale out to meet ever increasing demand as the analyses look at more and finer grained data. At AWS we’ve helped many customers tackle scaling challenges are noticing some common themes. In this post we describe how HPC teams are thinking about how they deliver compute capacity today, and highlight how we see the solutions converging for the future.