When using Cloudmersive Managed Instance and Cloudmersive Private Cloud, infrastructure is provisioned per core. As a general benchmark, each compute core can process two concurrent requests. Depending on how long the requests take, more or less requests can pass through the system. Request duration, or response time, will vary based on the API being used, the complexity of the transaction, including the size and content of the input and its parameters. Note that if a request takes less time, more throughput can be achieved, while if a request takes more time, then less throughput can be achieved, in the same compute environment and with the same number of cores. We recommend that all customers estimate the number of simultaneous transactions that they expect to require in their production environment and then license the corresponding number of cores needed. If you need help, reach out to our accounts team who can assist you in estimating the capacity needed based on your business needs.