I noticed the MSC test that wants osmo-msc to repeat a Paging to the BSC and hNodeB.
After fixing the Iu tests, I wanted to get this last ttcn3-msc-test working, so I was about to implement repeated Paging from osmo-msc, but the more I think about it, the less it makes sense to me.
The commit log in osmo-ttcn3-hacks says:
Repeating will improve the reachability of MS when a Paging is lost or not received because the MS is moving between states.
This reasoning seems flawed to me, because the BSC / hNodeB is between the MSC and the MS, and the BSC *does* repeat Paging. That should cover MS moving between states.
The link between MSC and {BSC,hNodeB} is considered reliable, and AFAICT nothing really gets better when repeating a Paging request to the BSS.
It could make sense to maybe repeat Paging with a larger/more general Cell Identifier List? But why not send the full list in the first place?
Also 3GPP TS 48.008 3.1.10 Paging says:
A single PAGING message across the MSC to BSS interface contains information on the cells in which the page shall be broadcast.
I interpret the "A single" as: there is no repetition of Paging requests toward the BSS, and that's also what makes most sense to me infrastructurally. Is there any spec indicating repeated Paging from the MSC?
I would actually remove the test TC_lu_and_mt_sms_paging_repeated. If that's the wrong call, we should specify how osmo-msc should repeat Paging. Before that, having the test makes little sense...
What do you guys think?
~N
Hi Neels,
I think the main issue is that there appears to be differences between 2G and 3G here, at least with those systems that we use for 3G (mostly nano3G).
While according to fixeria, repeating of paging is required on the Iu/3G side to make things work reliably, we wshould not require or do it on the A/2G side - for reasons you have covered in your e-mail.
I didn't review the Iu paging related specs. IF they agree with 2G and paging should only be sent once, then maybe we're just looking at an implementation flaw of the nano3G?
One could repeat that test with other 3G hNB hardware such as sysmoCell 5000, and compare the results. In the worst case, we could add some kind of vty command, or we could also think of repeating the pagign messages in the hnb-gw, *if* it's only an implementation flaw.
Regards, Harald
On Mon, Nov 04, 2019 at 06:12:36PM +0700, Vadim Yanitskiy wrote:
Hi all,
While according to fixeria, repeating of paging is required on the Iu/3G
side [...]
I have no idea about the Paging requirements in Iu/3G, so most likely Harald meant somebody else ;)
That would by lynxis.
Ok, thx, now it makes sense where this is coming from. We could remove the paging_repeat test from MSC_Tests and keep it in MSC_Tests_Iu...
~N
-
Harald Welte -
Neels Hofmeyr -
Vadim Yanitskiy