What is Azure HPC Cache?
Let's start by understanding the key concepts within Azure HPC Cache. This overview helps you understand how the cache service works.
What is a storage target?
A storage target is where your data is kept. Often, these targets are network attached storage appliances. There are specific storage targets that HPC Cache supports like Azure NetApp Files, physical NetApp appliances, EMC Isilon appliances, or Linux storage servers.
Most high-performance applications use NFSv3 for the file system protocol. HPC Cache uses this efficient protocol to transfer data among the storage target, the cache, and the clients.
Azure HPC Cache definition
HPC Cache is a managed service that accelerates file-based storage for compute-intensive workloads.
Caching data close to compute clients allows those clients to run more efficiently and complete high-performance tasks faster. When compute clients are waiting for storage responses, they aren't computing. Compute cycles are wasted, and charges continue to be incurred when the data can't get to the clients fast enough. Rather than waiting for responses from slower storage or from high-latency WAN links, compute clients can quickly access the needed data from the cache.
How to know if your workload is cacheable
HPC Cache accelerates cacheable workloads. If your HPC workload is cacheable, or one that reuses most of the same data, a caching service like HPC Cache might be a good fit.
Cacheable workloads have a static set of data that several clients access at the same time.
For example, running a stock market simulation using historical stock market data would be a perfect use for a caching service. The dataset doesn't change and you can have hundreds of compute clients running thousands of scenarios.