Hi Neels,

On Fri, Nov 24, 2017 at 04:28:46AM +0100, Neels Hofmeyr wrote:

Looking at https://osmocom.org/projects/cellular-infrastructure/wiki/Point_Codes

which I btw crated from the compile-time information I found in the code, but apparently not saw all related pieces.

OsmoMSC currently uses 0.23.1 for A-and-Iu, and 0.23.2 for Iu only if Iu is on a different cs7 instance than A (i.e. practically never).

Ah, this particular behavior seems to have slipped my reading of the code.

Either way that doesn't really make sense to me. If OsmoMSC is connecting to the same STP for A and Iu, it is sufficient to have one point-code for the two SSNs for A (BSSAP) and Iu (RANAP) and hence use one cs7 instance for both.

Yes, it's sufficient. Nevertheless, people do use "secondary point codes" for similar reasons people are using "secondary IP adresses" in IP networks, e.g. separate IP addresses for individual services so later on you can run them on separate machines without having to reconfigure your routing or all of the peers.

If it is connecting to two different STPs, the point codes do not collide anyway,

this is only correct if those two STPs are in separate MTP-level networks, again similar to the IP sphere: If you have two STPs in one company network (or one national network, or the international network) then the addresses are global across that network. Only if you decide to have two networks that don't direclty route on point codes you can have identical point codes in both networks. You could not easily transport e.g. ISUP across such a boundary withouta proxy, but you could use a STP with Global Title Translation to use the SCCP address to get the new point code while transitioning a gateway between the two networks.

and it makes no sense to use a different one for Iu. It only confuses the defaults, e.g. the wiki page was wrong until I fixed it just now.

I would rather have just 0.23.1 for the MSC for all SSNs.

fine with me. Whoever wants it differently can always configure it differently.