Nice video and very good and clear explanation. Quick question, if you have multiple APN's that have the same way of PGW selection I guess you can do it as well with the same CNAME aliases principle like with the TACs?
Perfect video, I watched it several times in the past and use it as a reference. A quick question. how do you define the IMS apn on the DNS for the sip signaling & voice calls.
Hi Bulutunga, 3GPP doesn't actually draw a distinction between an IMS-serving APN and any other APN at a DNS level. While TS 29.303 does use "ims" in the names of APN-NI's throughout their examples, it doesn't have logical significance. For IMS-specific procedures in the EPC, such as P-CSCF restoration, generally speaking MME's that support such features will recognized the reserved APN-NI name "ims" and handle it accordingly. It's possible that MME's may support manual definition of APNs as supporting or not supporting IMS-related logic, but unfortunetly this isn't given to the MME by DNS records. A good reference for this (heavy dependency on the literal name "ims" as an indicator for IMS service logic in EPC) is GSMA IR.88 section 6.3. Even in the current writing of the 5GC (TS 23.501 section 5.16.3.4), the 5G core uses the literal name of the APN (or DNN in that architecture) to determine whether it is an IMS-related PDN connection or not.
I came for topoff/topon! you kept saying I'll get to that in second... when we reach time for delivery, it wasn't clear.... so anoying How the hell a node going to understand "closeness" of a network topology?!! and what's the longest match?!!
With "topon", you read the name from right to left, where a "closer" node is one with more matches until they stop matching. For example, this SGW interface name has five levels of hierarchy (mcc234 > mnc123 > east > sgw1 > eth1): topon.eth1.sgw1.east.mnc123.mcc234. Of the following two PGW interfaces, the first will win on topological closeness, because it has three common hierarchy levels (mcc234 > mnc123 > east) while the second only has two levels (mcc234 > mnc123). Here "east" is said to be "closer" than "west", but by "closer" we just mean more names match through the right-to-left hierarchy as we read the FQDN: topon.eth2.pgw1.east.mnc123.mcc234. topon.eth1.pgw1.west.mnc123.mcc234.
it is a master piece. thank you a lot for your efforts
Very informative and elaborate, thank you
Nice video and very good and clear explanation. Quick question, if you have multiple APN's that have the same way of PGW selection I guess you can do it as well with the same CNAME aliases principle like with the TACs?
Perfect video, I watched it several times in the past and use it as a reference. A quick question. how do you define the IMS apn on the DNS for the sip signaling & voice calls.
Hi Bulutunga,
3GPP doesn't actually draw a distinction between an IMS-serving APN and any other APN at a DNS level. While TS 29.303 does use "ims" in the names of APN-NI's throughout their examples, it doesn't have logical significance.
For IMS-specific procedures in the EPC, such as P-CSCF restoration, generally speaking MME's that support such features will recognized the reserved APN-NI name "ims" and handle it accordingly. It's possible that MME's may support manual definition of APNs as supporting or not supporting IMS-related logic, but unfortunetly this isn't given to the MME by DNS records. A good reference for this (heavy dependency on the literal name "ims" as an indicator for IMS service logic in EPC) is GSMA IR.88 section 6.3.
Even in the current writing of the 5GC (TS 23.501 section 5.16.3.4), the 5G core uses the literal name of the APN (or DNN in that architecture) to determine whether it is an IMS-related PDN connection or not.
@@QuadraticSquared Thank you for the reply and referring additional documents for reading.
*!!!DING!!!!*
:)
Good video, turn off the system beep!
Oh yes, I'm sorry about that. There were a couple other videos I completely re-did because of sound issues that were even worse. :)
I came for topoff/topon! you kept saying I'll get to that in second... when we reach time for delivery, it wasn't clear.... so anoying
How the hell a node going to understand "closeness" of a network topology?!! and what's the longest match?!!
With "topon", you read the name from right to left, where a "closer" node is one with more matches until they stop matching.
For example, this SGW interface name has five levels of hierarchy (mcc234 > mnc123 > east > sgw1 > eth1):
topon.eth1.sgw1.east.mnc123.mcc234.
Of the following two PGW interfaces, the first will win on topological closeness, because it has three common hierarchy levels (mcc234 > mnc123 > east) while the second only has two levels (mcc234 > mnc123). Here "east" is said to be "closer" than "west", but by "closer" we just mean more names match through the right-to-left hierarchy as we read the FQDN:
topon.eth2.pgw1.east.mnc123.mcc234.
topon.eth1.pgw1.west.mnc123.mcc234.