Cloud advice

Scaling genomics workloads using HPC on AWS

AWS has made the recording available from their recent webinar, Scaling genomics workloads using HPC on AWS, presented by Aniket Deshpande and Evan Bollig. This is a great resource and I strongly recommend watching it, especially if you are relatively new to the idea of running research workloads on a public cloud platform like AWS.

The ouline agenda for the webinar is as follows (which I’ve done my best to annotate with approximate timings):

  • The case for HPC genomics workloads on AWS (3:00)
  • Trends from the trenches (5:22)
  • How AWS enables digital lab operations (17:18)
  • Data transfer and storage use-cases (21:25)
  • Architecting for HPC on AWS (30:33)

You can watch the recording here:

This webinar covers a lot of ground and includes several industry case-studies where the concepts outlined in the webinar are put into practice. Whilst it is billed as focusing on genomics, I think that most of the material covered here is relevant to a wide range of research use-cases. It covers approaches for moving data between on-prem and the cloud, tiered storage options, filesystems, EC2 instance type selection, architectural principles, HPC tooling, workflow management, pipeline processing and cost optimisation.

The presentation ends with a set of links to useful resources, repeated here:

If you are interested in knowing more about this area, why not come and talk to Jisc’s AWS Solutions Architects. We specialise in helping you to deploy research and education workloads of all kinds securely and cost-effectively on AWS.

Finally, if you are deploying research, or other, workloads onto AWS, Azure, GCP or other cloud platforms, remember that the OCRE Cloud Framework will probably offer you the most cost-effective way of buying your cloud services.

In both cases, contact if you want to know more.


By Andy Powell

Leave a Reply

Your email address will not be published. Required fields are marked *