The simplest way may be to run the code to access the IXIA as a subprocess so that it can use Tcl 8.4 without disrupting the version used by Python (which would have a lot of other consequences). Now Tcl 8.4 is no longer supported at all, but the most recent version of Tcl is 8.4.20, and came out only a few years ago so it is largely compatible with current build systems. (You can get the 8.4.20 source code from SourceForge.)
When you are building Tcl, assuming you are targeting Unix, you can configure where it will install using the --prefix= option to configure; the default location is /usr/local (e.g., the tclsh8.4 binary goes in /usr/local/bin, the support libraries go in /usr/local/lib, the documentation goes in /usr/local/man, …). If you have Tcl packages elsewhere that you want to use, the TCLLIBPATH environment variable can be used to say where they are. Also, if you are running as a subprocess, you may want to tune the level of buffering used on standard output:
# The default with terminals is line. The default with pipes is full.
fconfigure stdout -buffering line
In general, you will probably be substituting runTCL.tk.eval in your Python code for printing the code to run to the subprocess pipe and reading the result back. There's going to be a need for a bit of work to make that work neatly, but that sort of thing has been discussed in other questions here. The only major complication is really that you're working with an unsupported version of the code for your subprocess.