The inherent design scalability of modern cloud computing makes possible subscription models for each of the High Performance computing architectures, either within a single infrastructure or across several as a multi-cloud subscription.
If an organization uses on-premise traditional computing, or has developed a private or hybrid cloud architecture for its computing needs, but needs higher performance as part of an exceptional event, a periodic surge in compute requirements, or a new digitalization initiative, high performance cloud computing offers many advantages.
Machine learning is one example of a short term high demand compute power requirement that is suited to executing in the cloud. Providers such as IBM, Google and AWS have developed offerings to host supervisor or unsupervised machine learning models in their clouds using their systems on client data sets. Public Financial sector workloads are another ideal case for cloud based high performance computing.
