Cellular Network Infrastructure » osmo-ePDG - VoWifi Evolved Packet Data Gateway

I started work on this, WIP in osmo-ttcn3-hacks.git branch pespin/epdg.

lynxis I cleaned up all the stuff you had in "lynxis/wip" branch in osmo-epdg.git and submitted a bunch of patches here:
https://gerrit.osmocom.org/c/erlang/osmo-epdg/+/34726/

What I'd need you to do: Get the osmo_ss7 and osmo_gsup patches you have in branches merged into master.

Using those patches (osmo-epdg.git branch "pespin/master), I can use the osmo-ttcn3-hacks.git/docker-playground.git TTCN3 testsuite (branch "pespin/epdg") and I can get to the point to connect over GSUP to osmo-epdg.
I can send an AuthInfoReq to it, but it will be rejected because the Diameter conn EPDG->TTNC3 is not yet done by osmo-epdg.

The Diameter SWx conn is not being done because I don't seem to be able to properly change the Diameter remote IP address through the config file in osmo-epdg, it is not applied. I need to continue looking into why this happens and fix it.

I submitted a bunch of patches to osmo-epdg gerrit to improve and fix the configuration of several IP addresses, ports, etc.

with that and improved code in osmo-ttcn3-hacks.git/docker-palyground.git (branch=pespin/epdg), I can run a full GSUP+SWx AuthInfo ping pong.

Only thing missing now is check why the AuthInfoResult received from osmo-epdg doesn't match exactly our expectancies:

received from osmo-epdg:
{ { tag := OSMO_GSUP_IMSI_IE (1), len := 8, val := { imsi := '262420000000001'H } },
 { tag := OSMO_GSUP_AUTH_TUPLE_IE (3), len := 104, val := { auth_tuple := {
 { tag := OSMO_GSUP_RAND_IE (32), len := 16, val := { rand := '0B35AE485BBD9BE29655EE50E7030E29'O } },
 { tag := OSMO_GSUP_IK_IE (35), len := 16, val := { ik := 'FBB94F5765F454A1CE764E02D779CC34'O } },
 { tag := OSMO_GSUP_CK_IE (36), len := 16, val := { ck := '3180B12B70E09965621E60654CCC27DE'O } },
 { tag := OSMO_GSUP_AUTN_IE (37), len := 16, val := { autn := '0E828188919015E068FD89996D461637'O } },
 { tag := OSMO_GSUP_RES_IE (39), len := 30, val := { res := '0B35AE485BBD9BE29655EE50E7030E2958E8A4B295AFB8EC5D82A62E2408'O } } } } } } with

expected:
{ { tag := OSMO_GSUP_IMSI_IE (1), len := ?, val := { imsi := '262420000000001'H } }, *,
{ tag := OSMO_GSUP_AUTH_TUPLE_IE (3), len := ?, val := { auth_tuple := {
{ tag := OSMO_GSUP_RAND_IE (32), len := ?, val := { rand := '0B35AE485BBD9BE29655EE50E7030E29'O } },
{ tag := OSMO_GSUP_SRES_IE (33), len := ?, val := { sres := ? } },
{ tag := OSMO_GSUP_KC_IE (34), len := ?, val := { kc := 'A33CF3ADE8BDBC4C'O } },
{ tag := OSMO_GSUP_IK_IE (35), len := ?, val := { ik := 'FBB94F5765F454A1CE764E02D779CC34'O } },
{ tag := OSMO_GSUP_CK_IE (36), len := ?, val := { ck := '3180B12B70E09965621E60654CCC27DE'O } },
{ tag := OSMO_GSUP_AUTN_IE (37), len := ?, val := { autn := '0E828188919015E068FD89996D461637'O } },
{ tag := OSMO_GSUP_RES_IE (39), len := ?, val := { res := ? } } } } }, * }

• iK and cK are swapped in position. This is probably a bug in osmo-epdg.
• SRES and KC are missing. Not sure where this should come from. I guess I need to feed them from our emulated HSS somehow.

We have ~10 tests running nightly already, all passing, so closing this ticket.