A manufacturing firm hosts a proprietary analytics application on two load-balanced Windows Server virtual machines. Around 825 employees have accounts, but usage logs show the peak number of simultaneous sessions tops out at 230. The vendor lets the IT team choose one licensing model for the software: per-server, per-core, per-instance, or per-concurrent user. To stay compliant while buying the fewest total licenses, which model should the systems administrator recommend?
The per-concurrent user model requires the organization to license only the maximum number of users connected to the software at the same time-in this case, 230. Licenses are checked out of a shared pool when a session starts and returned when the user disconnects. Because the firm never exceeds 230 simultaneous sessions, it can legally operate the application on both virtual machines with 230 concurrent licenses.
By contrast, a per-server model would obligate the company to purchase a separate license (or CAL set) for every server hosting the application, effectively doubling the cost. Per-core and per-instance models scale with the number of virtual CPUs or running copies of the software and therefore ignore actual usage patterns; these approaches would require more licenses than the organization truly needs. Consequently, per-concurrent user licensing is the most cost-effective and compliant choice for this usage pattern.