Microsoft offers the option to license using a Server+CAL model. This allows you to license users and/or devices that are connecting to your SQL Server and can be a lot more cost-effective option if you are able to quantify those.
Under this model, you will purchase a server license for each server, as well as, a client access license for each Device (Device CAL) and/or user (User CAL) accessing your SQL Server and/or any of its components. 1 CAL grants 1 user or 1 Device access to the SQL Server. Like a lot of Microsoft licensing, this works on a honor system so if you install SQL Server and use CAL licensing, you will not be prohibited from going over the amount you purchased, however, you would be out of compliance and could face financial consequences if discovered.
In order to license using this model, you must be running Standard edition and must keep track of what users and/or devices are assigned. However, this information is based on the SQL 2017 licensing guide, so if already have CALs on a previous version then you may want to speak to your licensing rep about whether these can be applied differently. You may be able to use these on editions other then Standard.
CALs are also version specific so you need to make sure whatever CAL you are assigning has access to that version of SQL. It should be at or above the version. If you have active software assurance on it, then you should be fine, this really comes into play when you do not.
Also, if something connects no matter how, it needs a CAL. Multiplexing is a term that Microsoft uses to refer to software or hardware to pool connections, reroute information, or reduce the number of devices or users that directly or indirectly access or use SQL Server. Be aware that multiplexing does not reduce the number of CALs. I suggest reading through this portion of the licensing guide which does a great job explaining this.
For more detailed information on this, check out the SQL 2017 licensing guide. https://www.microsoft.com/en-us/sql-server/sql-server-2017-pricing
SQL By Dustin 