Enterprise Nucleus Cache Server is an HTTP daemon and its resource usage should be based on expected load.
CPU: enough to cover basic networking needs; Cache doesn’t perform any CPU intensive work
RAM: cache itself uses a minimal amount of RAM; filesystem cache may be used but should be monitored
Disk type/size: depends on bandwidth served. Ideally, it should be balanced with network speeds and RAM size (a slower disk will require more RAM for a larger filesystem cache)
Packages and Access
Nucleus Enterprise Cache Server is shipped, configured, and installed via Docker (the same way Nucleus Enterprise Server is packaged) and should be administered in the same manner. Cache Server documentation is available in the config files within the Docker artifacts.
Please contact your NVIDIA representative to obtain access to the Docker artifacts.
For more information, see Nucleus Enterprise Server Installation Documentation