Project

General

Profile

Bug #4892

gbproxy2: Route BSSGP-STATUS based on "Erroneous PDU IE"

Added by laforge 3 months ago. Updated about 1 month ago.

Status:
New
Priority:
Normal
Assignee:
Target version:
-
Start date:
12/08/2020
Due date:
% Done:

0%

Spec Reference:

Description

NS-STATUS can only be routed as follows:
  • PTP downlink: Simply route by NS-BVCI
  • PTP uplink: route based on "Erroneous PDU IE"
  • SIG downlink + uplink:
    • don't route (but locally terminate) if optional BVCI IE, as that one only occurs for BVCI blocked / BVCI unknown whihc is a gbproxy problem, not one with a remote peer
    • route messages without BVCI IE but with "PDU in error" IE based on the latter
If we route on the "Erroneous PDU IE", we must consider
  • the contained BSSGP PDU might be truncated, and hence our TLV parser might not be playing well along
  • we should look for TLLI + TMSI (and route based on NRI)
  • if we cannot find any routing informatoin, we treat it like NULL NRI?
  • fast path for single-SGSN case: simply route to "the" SGSN?

Related issues

Related to osmo-gbproxy - Feature #4472: Intra-domain connection of OsmoGBPROXY to multiple SGSNs (pooling)In Progress03/29/2020

History

#1 Updated by laforge 3 months ago

  • Related to Feature #4472: Intra-domain connection of OsmoGBPROXY to multiple SGSNs (pooling) added

#2 Updated by laforge about 1 month ago

There are now TTCN3 tests for both uplink and downlink direction in https://gerrit.osmocom.org/c/osmo-ttcn3-hacks/+/22257

#3 Updated by laforge about 1 month ago

BTW: we should make sure that no STATUS messages related to BLOCK/RESET/UNBLOCK are proxied/forwarded, as the local FSMs originate the messages.

#4 Updated by daniel about 1 month ago

The tests seem to be for BSSGP-STATUS PDUs while this ticket is about NS-STATUS PDUs

Regarding routing we will need to do something similar for BSSGP STATUS on signalling bvc (ptp can just be passed through)

#5 Updated by laforge about 1 month ago

On Mon, Jan 25, 2021 at 10:42:07AM +0000, daniel [REDMINE] wrote:

The tests seem to be for BSSGP-STATUS PDUs while this ticket is about NS-STATUS PDUs

thanks for catching that. I'm not sure what I was thinking 2 months ago exactly,
but I think I may have been thinking about BSSGP even back then.

In terms of NS-STATUS: I'm currently not entirely sure if we should route them
at all, or not. After all, we really do terminate the NS layer completely
in gb-proxy. The NS-VCI / NSEI are going to be different on both sides.

Hence, I think it's best to rename the ticket to BSSGP.

#6 Updated by laforge about 1 month ago

  • Subject changed from gbproxy2: Route NS-STATUS based on "Erroneous PDU IE" to gbproxy2: Route BSSGP-STATUS based on "Erroneous PDU IE"

Also available in: Atom PDF

Add picture from clipboard (Maximum size: 48.8 MB)