If my PROCR is assigned to the intervening NR250 and I have h.323 and SIP trunks to other CM's/SM's what should the far-end NR be set at? It seems default on NR1. Perhaps I don't know exactly the purpose of that field, but what if the site or resources associated to NR1 were not available?
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations derfloh on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Signaling Group - Far-end NR
- Thread starter trilogy8
- Start date
It's strictly for deciding what codec set to use. Per the screen reference, if unspecified, you'll use the NR of the near-end node - so, 250.
Either way, all your DSPs should be directly connected to NR250. So, if CM needed a DSP on a call toward NR1, as long as all your regions with DSP are directly connected to 250 AND all your regions are connected to all other regions with intervening region 250.
And even then, CMs decision on network region is more complicated. It uses the real media IP to determine where the bandwidth is actually nailed up.
So, for example, you have 5 sites with SIP phones off of 1 SM and 1 CM. Obviously all the SIP phones will come thru 1 SM and possibly 1 trunk group to CM for features. It may or may not have a far-end network region.
If those 5 sites subnets are in the network map as 5 regions, CM can still peel out what region is using bandwidth. It uses the IP of the oldest/furthest down via header in the SIP message - which is that of the endpoint. Same thing delivering calls to those SIP phones over the same single trunk to a single SM - it'll know based on the responses in signaling which IP/region the call is ultimately being set up for so you can tinker with your per-region bw controls.
Either way, all your DSPs should be directly connected to NR250. So, if CM needed a DSP on a call toward NR1, as long as all your regions with DSP are directly connected to 250 AND all your regions are connected to all other regions with intervening region 250.
And even then, CMs decision on network region is more complicated. It uses the real media IP to determine where the bandwidth is actually nailed up.
So, for example, you have 5 sites with SIP phones off of 1 SM and 1 CM. Obviously all the SIP phones will come thru 1 SM and possibly 1 trunk group to CM for features. It may or may not have a far-end network region.
If those 5 sites subnets are in the network map as 5 regions, CM can still peel out what region is using bandwidth. It uses the IP of the oldest/furthest down via header in the SIP message - which is that of the endpoint. Same thing delivering calls to those SIP phones over the same single trunk to a single SM - it'll know based on the responses in signaling which IP/region the call is ultimately being set up for so you can tinker with your per-region bw controls.
- Thread starter
- #3
- Thread starter
- #4
Also, when dealing with the intervening and other NR's what should the intra/inter and audio hairpinning be set at? In reading about these it seems reliant to some degree on Medpro resources. What if the Medpro resources or that entire NR wasn't available and all other sites were using G450's?
for CM1's sig group between procr CM1 and procr CM2, all that FE NR means is what codec set CM1 will pick.
I'm sure there's a newer version of this doc, but it explains it pretty well -
And the selection algorithms might have been updated over time, but the short answer is you're network regions will try to find DSPs in directly and indirectly connected places to service the call.
I'm sure there's a newer version of this doc, but it explains it pretty well -
And the selection algorithms might have been updated over time, but the short answer is you're network regions will try to find DSPs in directly and indirectly connected places to service the call.
- Thread starter
- #6
That was an informative read, thank you. Since it's rather old it references G650's and Medpro resources etc. How does that play with just H.248 GW's or even in a case such as mine where I do have 1 PN at one site, along with H.248 GW's? Or is it basically as long as you define your IP-network-map subnets into specific NR's the phones will use what's locally available to it?
Mostly the 2nd one. As long as you're modern enough - so, no fiber connected port networks, and no CLANs with G450s registering to a CLAN in a PN thru a IPSI to CM, etc, then it's all pretty well the same.
If you need a DSP, start by searching in your NR. If it doesn't have one, start with the next hop on your way.
So, in an example with NR1 with phones and DSP, if NR1 phone calls NR1 phone and conferences NR1 phone, look in NR1. If the gateway in NR1 is busy/down, then look in all directly connected network regions - that should only be procr/250. Procr's NR should never have DSPs. So we look in NRs connected to procr that do have DSP. NR2 has a gateway at the other site. Oh great! we'll use that!
If you need a DSP, start by searching in your NR. If it doesn't have one, start with the next hop on your way.
So, in an example with NR1 with phones and DSP, if NR1 phone calls NR1 phone and conferences NR1 phone, look in NR1. If the gateway in NR1 is busy/down, then look in all directly connected network regions - that should only be procr/250. Procr's NR should never have DSPs. So we look in NRs connected to procr that do have DSP. NR2 has a gateway at the other site. Oh great! we'll use that!
- Thread starter
- #8
So what if PROCR was in NR1 and the DSP's in NR1 were offline would that be problematic for other NR endpoints or same rules apply as far as CM finding any DSP's? Or better explained.. NR 2-5 with h.248 GW's all directly connect to NR250 and the endpoints register to PROCR in NR1, but NR1 DSP's are all offline.
Procr should be in NR250, and not have any DSP.
But, in that case, so long as NR1 can talk to 250 and 250 can talk to 2-5, you'd get the same result and fish DSP in 2-5 if DSP in NR1 was down.
Procr being in NR1 has very little to do with it. That just means h323 phones registering to it inherit NR1 if their IPs are defined in the network map as belonging to other regions.
You can still have extra regions with nothing in them. DSP selection is ultimately based on hop count - that's why you can have many intervening regions between two NRs.
Practical example;
Customer has national presence. SIP trunks come into data center east and data center west. Call them NR 10 and 20.
All agents are at branch sites with no gateways.
All calls are recorded.
All calls delivered to DCwest are for west sites. Same for east. Say the west branches are NR 11-19 and east are 21-29
I want all my west people to use west DC DSP only, until they run out, or are down, then they can use east. Same rule for the east guys.
Now sure, based on the calls coming in the west sbc, with a media IP in the map in the west DC region, thru to agents in those other spoke regions would pick west DSP first. We wouldn't have to tweak the NRs much to accomplish that.
But what about when west agent calls west supervisor and conferences another west person - where does the DSP come from? Say 3 callers conference each other within NR11.
All NRs are direct to 250.
I make a NR 249 with nothing in it, also direct to 250, and 10 and 20.
11 reaches 10 thru 250
11 reaches 20 thru 250 and 249
21 reaches 10 thru 250 and 249
21 reaches 20 thru 250
Had I just made every NR hang off 250, then when 3 people conference within NR11 needing a DSP, CM would look '11 to 250 and all gateways are then equally distant. Let me do some quick math about which gateway is least used right now, I'll pick that'.
In my example with 2 intervening regions, 11 asking for DSP would have CM say 'well, this guy knows how to reach 10 and 20 where the dsps are at, but 11 to 10 is 2 hops - 11 to 250 to 10. 11 to 20 is 3 hops - 11 to 250 to 249 to 20. I'm going to math out the freest DSP in NR 10 before I offer this 3 way call of 3 phones in NR11 anything out of NR20. NR10 would need to be out or down before this call could potentially take NR20.
So, back to your example, as long as your network region matrices aren't blank between NR1 and 2-5 - to say, if NR1 has some defined path to NRs 2-5 and a codec set defined, CM will try to find a way.
There are all sorts of designs that don't have CM controlled H248 gateways on site. Drop a cheap audiocodes mp112 or a IPO as a 3rd point of SIP registration. It can all still leverage core CM network design despite not having local h248 cm managed DSP.
But, in that case, so long as NR1 can talk to 250 and 250 can talk to 2-5, you'd get the same result and fish DSP in 2-5 if DSP in NR1 was down.
Procr being in NR1 has very little to do with it. That just means h323 phones registering to it inherit NR1 if their IPs are defined in the network map as belonging to other regions.
You can still have extra regions with nothing in them. DSP selection is ultimately based on hop count - that's why you can have many intervening regions between two NRs.
Practical example;
Customer has national presence. SIP trunks come into data center east and data center west. Call them NR 10 and 20.
All agents are at branch sites with no gateways.
All calls are recorded.
All calls delivered to DCwest are for west sites. Same for east. Say the west branches are NR 11-19 and east are 21-29
I want all my west people to use west DC DSP only, until they run out, or are down, then they can use east. Same rule for the east guys.
Now sure, based on the calls coming in the west sbc, with a media IP in the map in the west DC region, thru to agents in those other spoke regions would pick west DSP first. We wouldn't have to tweak the NRs much to accomplish that.
But what about when west agent calls west supervisor and conferences another west person - where does the DSP come from? Say 3 callers conference each other within NR11.
All NRs are direct to 250.
I make a NR 249 with nothing in it, also direct to 250, and 10 and 20.
11 reaches 10 thru 250
11 reaches 20 thru 250 and 249
21 reaches 10 thru 250 and 249
21 reaches 20 thru 250
Had I just made every NR hang off 250, then when 3 people conference within NR11 needing a DSP, CM would look '11 to 250 and all gateways are then equally distant. Let me do some quick math about which gateway is least used right now, I'll pick that'.
In my example with 2 intervening regions, 11 asking for DSP would have CM say 'well, this guy knows how to reach 10 and 20 where the dsps are at, but 11 to 10 is 2 hops - 11 to 250 to 10. 11 to 20 is 3 hops - 11 to 250 to 249 to 20. I'm going to math out the freest DSP in NR 10 before I offer this 3 way call of 3 phones in NR11 anything out of NR20. NR10 would need to be out or down before this call could potentially take NR20.
So, back to your example, as long as your network region matrices aren't blank between NR1 and 2-5 - to say, if NR1 has some defined path to NRs 2-5 and a codec set defined, CM will try to find a way.
There are all sorts of designs that don't have CM controlled H248 gateways on site. Drop a cheap audiocodes mp112 or a IPO as a 3rd point of SIP registration. It can all still leverage core CM network design despite not having local h248 cm managed DSP.
- Status
Similar threads
