Our consultants are experienced in assessing the suitability and practicality of either bursting to the cloud or deploying and running an HPC system in the cloud. With the public cloud offering access to virtually unlimited resources and the ability to deploy resources quickly and in the geographic region of your choice, certain workloads and use cases are very well suited to running in the public cloud.
Run in the cloud when your on-premise HPC is over-subscribed or you need resources not available to you e.g. GPUs
Deploy a cluster in the cloud and start using it the same day – no long procurement
Grow and shrink your HPC environment based on workload demands
If your workload demands are variable, with times of high demand and other periods where the demand is low, managing an infrastructure based on paying only for what you use in the cloud could deliver a better return on investment when compared to having to size your on-premises infrastructure to cope with the peak workload demand.
Alternatively, you could be running an on-premises HPC system which copes with your day-to-day HPC workload demands, but occasionally experiences periods of high demand leading to extended queues and delayed time to results for users. In this instance, bursting to the public cloud could provide the extra capacity you need to clear the backlog of delayed workloads from your on-premises HPC system. A hybrid cloud solution provides a controlled and managed mechanism to extend your HPC to the cloud.
Our consultants have worked extensively with customers who have implemented cloud based HPC either as a hybrid deployment or a standalone HPC in the cloud platform, and can provide guidance, planning, professional services and support to customers looking to embrace the cloud for HPC workloads.
Pros and Cons of HPC in the Cloud
We have created a compilation of all relevant aspects to consider when looking at moving HPC workload to the cloud – head over to HPC in the Cloud – Pros and Cons.
Any questions just let us know and get in touch for a deeper dive on HPC in the Cloud.