The documentation specifies a heavy forwarder. Is this truly the case, or can a Universal Forwarder work with this TA?
Yes, you need a search head or heavy forwarder to run DB Connect and the Add-on at: http://docs.splunk.com/Documentation/AddOns/latest/MSSQLServer/Hardwareandsoftwarerequirements
Note that this doesn't mean you should install HF's on all your MS-SQL Servers, we're assuming you'd have remote DB Connect access to the databases and then use UF's to get the log files, like so: http://docs.splunk.com/Documentation/AddOns/latest/MSSQLServer/Configuremodularinput
Yes, you need a search head or heavy forwarder to run DB Connect and the Add-on at: http://docs.splunk.com/Documentation/AddOns/latest/MSSQLServer/Hardwareandsoftwarerequirements
Note that this doesn't mean you should install HF's on all your MS-SQL Servers, we're assuming you'd have remote DB Connect access to the databases and then use UF's to get the log files, like so: http://docs.splunk.com/Documentation/AddOns/latest/MSSQLServer/Configuremodularinput
Let's say there are 20 MS-SQL servers. Are you suggesting to monitor sql audit log files of binary format using UF on all MS-SQL servers and stream that binary data to a heavy forwarder on a port? How will heavy forwarder be able to decide which specific MS-SQL server instance to connect to for interpreting binary events coming in from multiple MS-SQL server instances on a single port? Can you please share some sample configurations to elaborate your answer?
I also notice that platform of heavy forwarder is mentioned to be windows. So I guess even if you install windows UF on MS-SQL servers, you need another windows machine to create heavy forwarder?
binary audit logs are different, and they do need db connect.