Locations and Bandwidth - Need help

[voip7372]

Hello.

Please take a look and let me know if I'm understanding this correctly so I don't configure it incorrectly.  CUCM 10.5.

We have a few large offices that act as the hub for many things like voice mail, email, etc.  The CUCM servers are located in 2 of these 3 main locations (it's one cluster though).  Between those main sites, we have 5000 kbps of bandwidth with QoS for VoIP usage.  Between the main sites and the WAN links to all other sites, there is also another 5000 kbps of bandwidth available for VoIP with QoS.   The all other sites which I'll refer to as remote sites or 'all other' have varying amounts of bandwidth for VoIP/QoS, depending on the size of the office and WAN link.  My question is related to how to configure the Locations in CUCM so that the correct amount of bandwidth is allocated to each site based on which site the call will ultimately go to.  

I have two examples on my drawing attached to this posting.  

Example 1:  Phone at location 6 (Loc 6) calls a phone at location 9.  I think this is pretty straight forward.  Each of those sites are remote sites and each of them has 1536 kbps available on it's WAN link to our network.

Example 2 (this is the reason why I'm asking about this ;-)):  Phone at Loc 4 calls phone at Loc 1.  Loc 4 only has 1024 kbps available on its WAN link, but it's going to one of the main sites and that main site has 5000 kbps of bandwidth available for it's WAN link.  I'm assuming I would configure the Locations screen to say Location 1 only has 1024 kbps to Location 4, correct?  

Continuing with my thoughts...looking at my drawing, Location 1 should be configured to have 5000 kbps to locations 2 and 3, correct?  

Currently, there is no 'hub none' location define in our CUCM.  I think it might have been renamed with the correct name we use for Location 1.  

As always, I appreciate your help :-)

[Vivek Batra]

Can you please clarify in your topology, if there is a call from Loc 9 to Loc 6, will it also utilize the Loc1 bandwidth?

[voip7372]

Call from Loc 9 to Loc 6 should not have to go through Loc 1.  They can talk directly to each other via the WAN.  Maybe my drawing is not as clear as it should be.  

Loc 1, 2 and 3 basically have 2 links to the WAN.  One link to the 'backbone' WAN which has more bandwidth between those 3 sites and they also have a link to the WAN that goes to all the other remote/branch office sites and that link also has a lot of bandwidth.   But, the remote/branch sites just have a single connection to the WAN and those links vary in bandwidth.

[Vivek Batra]

Ok, in that case I think traditional location approach should work for you instead of defining explication bandwidth locations between each site (/location). You can link all locations with 'hub_none' with their own bandwidth available. Like for Loc 1 to hub_none, you will configure 5000 Kpbs and Loc 4 to hub_none, you will configure 1024 Kbps. Now when there is a call from Loc 4 to any other site, appropriate bandwidth will be deducted from 1024 Kbps (depends on codec being used). If there is a call between Loc 4 and Loc 1, appropriate bandwidth will be deducted from Loc 4 (out of 1024 Kbps) and Loc 1 (out of 5000 Kbps). Same will be applicable for rest of the locations.

[voip7372]

Thanks.  So if we already had renamed or removed 'hub_none', I should just go ahead and create a new location called hub_none and configure it as you described?

[Vivek Batra]

It should be ok, you should try with two sites and see everything goes fine as per your expectation.

- Vivek

[voip7372]

I just realized I have another question when thinking about this scenario.  I have a separate device pool created just for voice mail since our voice mail servers only want G.711 (the Region for this uses G.711 as the first choice)...so I have a different device pool and region setup for this...but I also have a different Location assigned to this.  I assume with the new way you're suggesting, I should set the 'location' of the voice mail device pool to be the same location assigned to one of the main sites where the voice mail servers actually reside, so it can be figured in with the bandwidth we know we have at the main site?

EDIT:  Thinking about this some more.  Since the voice mail servers are located at both Loc 1 and Loc 2, I think it would be reasonable to only link the Location for voice mail (on the Locations page) to Loc 1 and 2 and set the bandwidth to Unlimited between the voice mail Location and Loc 1 and 2, since the call would have to pass through either Loc 1 or 2 first and those locations would already have the amount of bandwidth to the WAN (hub_none) defined, so once that call passed through Loc 1 or 2 on its way to voice mail, we could let the BW be unlimited since the voice mail servers are on the local LAN at Loc 1 and 2.  Make sense?  I attached an image of what I'm talking about.

[Vivek Batra · Accepted solution]

Although you have different device pool for voicemail because of difference codec requirement, but IMO with respect to locations, still you need to treat voicemail server in the same manner as any other phones in respective location. What ever location you decide for phones to put in that location, you can keep voicemail in the same location. Anyway, you will always keep intra-location bandwidth set to unlimited for all locations. In this way, you will be able to control the bandwidth between the same locations and between different locations.

[voip7372]

Sorry, I had almost no help when our CUCM was first configured so I did most of this myself and my understanding of how Regions, Locations and Device Pools should be configured for different reasons was limited.  

So...basically, for voice mail servers I can keep a separate device pool and REGION (region to force the codec I want), but use the same LOCATION for both the phones and the voice mail servers if these are both in the same physical location, correct?   

[Vivek Batra]

Hi,

Yes, that is what I meant.

- Vivek

[voip7372]

Thanks.  See attached image and if that type of config is what you're talking about (look at location 1 and 2), I'll go ahead and reconfigure it right now before it gets too late in the day.  The system is still reasonably small so now is the time to get this setup right (before it grows too big to change easily)

[Vivek Batra]

Seems good.

Yes, that will be good idea to work on these aspects in early stage of its usage.

- Vivek

[voip7372]

Thanks so much for your help and quick replies.  I appreciate it.  I made the changes.  When I look at the Serviceability page and check the effective paths for all these sites, they all show Hub_None sitting in between them, so I think I'm ok now :-)  

[Vivek Batra]

Happy to help :)

- Vivek