I recently became involved with the configuration of our UBR10012 CMTS for adding another 48 channels to our RFGW1. (We currently had 48 already, 40 of which are being used.) After several hours of trying to figure out what the heck I was doing, I finally got the additional 48 channels (40 of which are being used... so we have 16 unused channels on QAM 6/1 and 6/2 of the RFGW1) up and working.
I'm VERY new to a lot of this, but have some way (through a miracle, no doubt) figured out how to make it all work. This is a brand new world to me, so I can't quite grasp all the concepts that I need to just yet. Hang in there with me, I'm gonna post a lot of information, because I have a question I want to ask and get some guidance on.
- - - - - - - - - - -
Our hardware:
2 - UBR10-MC5X20H-D (C5/0/0 & C5/1/0)
4 - SPA-24XDS-SFP (1/0/0, 1/1/0, 1/2/0, and 1/3/0)
1 - RFWG1 (EQAM running 96 channels)
At the time of deployment (which I was not involved), our MC interfaces were setup like this on our primary DS:
C5/0/0 = MC1/0/0 rf channel 0-3
C5/0/1 = MC1/0/0 rf channel 4-7
C5/0/2 = MC1/0/0 rf channel 8-11
C5/0/3 = MC1/0/0 rf channel 12-15
C5/0/4 = MC1/0/0 rf channel 16-19
C5/1/0 = MC1/0/0 rf channel 20-23
C5/1/1 = MC1/1/0 rf channel 0-3
C5/1/2 = MC1/1/0 rf channel 4-7
C5/1/3 = MC1/1/0 rf channel 8-11
C5/1/4 = MC1/1/0 rf channel 12-15
We we turned up the additional 4 channels per DS, we utilized the additional 1/2/0 and and 1/3/0 SPAs to add the channels. So it looks like this currently:
interface Cable5/0/0
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 0-3 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 0-3 upstream 0-3
interface Cable5/0/1
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 4-7 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 4-7 upstream 0-3
interface Cable5/0/2
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 8-11 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 8-11 upstream 0-3
interface Cable5/0/3
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 12-15 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 12-15 upstream 0-3
interface Cable5/0/4
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 16-19 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 16-19 upstream 0-3
interface Cable5/1/0
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 20-23 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 20-23 upstream 0-3
interface Cable5/1/1
downstream local upstream 0-3
downstream Modular-Cable 1/1/0 rf-channel 0-3 upstream 0-3
downstream Modular-Cable 1/3/0 rf-channel 0-3 upstream 0-3
interface Cable5/1/2
downstream local upstream 0-3
downstream Modular-Cable 1/1/0 rf-channel 4-7 upstream 0-3
downstream Modular-Cable 1/3/0 rf-channel 4-7 upstream 0-3
downstream local upstream 0-3
downstream Modular-Cable 1/1/0 rf-channel 8-11 upstream 0-3
downstream Modular-Cable 1/3/0 rf-channel 8-11 upstream 0-3
interface Cable5/1/4
downstream local upstream 0-3
downstream Modular-Cable 1/1/0 rf-channel 12-15 upstream 0-3
downstream Modular-Cable 1/3/0 rf-channel 12-15 upstream 0-3
- - - - - - - - -
There are 4 modular cable interfaces for each of those definitions, so each downstream has 2 sets of 4 modular, but they are across two different spas.
I'm beginning to think that this may not be the best setup and that I would be better off taking all 8 channels of those DS frequencies and grouping them together on the same interface, having 8 channels across the wideband interface instead of just 4.
So, instead of having the following:
interface Cable5/0/0
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 0-3 upstream 0-3
downstream Modular-Cable 1/2/0 rf-channel 0-3 upstream 0-3
I would end up something like this:
interface Cable5/0/0
downstream local upstream 0-3
downstream Modular-Cable 1/0/0 rf-channel 0-7 upstream 0-3
interface Wideband-Cable1/0/0:0
cable bundle 1
cable rf-channel 0 bandwidth-percent 30
cable rf-channel 1 bandwidth-percent 30
cable rf-channel 2 bandwidth-percent 30
cable rf-channel 3 bandwidth-percent 30
cable rf-channel 4 bandwidth-percent 30
cable rf-channel 5 bandwidth-percent 30
cable rf-channel 6 bandwidth-percent 30
cable rf-channel 7 bandwidth-percent 30
1. Is this possible?
2. Is there any benefit of me doing this?
I'm thinking the way it is setup, that each CM can only tap into 1 group of the 4 channels and not pick from the 8 channels across the board. But, that's just my guess... again, a lot of this is new to me. If this is doable, I'm sure I can figure it out (having to figure out how to just turn up those 48 channels was enough teach how to configure modular and wideband interfaces...hahahah) But I don't want to start configuring anything if it can't be done.
I'll try to answer any questions you all my have. Thanks for taking the time out to read this small novel. :D
On Cisco, you can create logical channels using different groups of the physical channels, so you would have a logical 4-channel group and a logical 8 channel group, in order to have modems lock onto different group sizes. A 4-channel modem, for example, won't lock onto more than 1 channel if you don't have a 4-channel logical group... It sets up pretty much just like a sub-interface on any other cisco interface, at least it did on the 7225 we used to have...
so you would have
interface Wideband-Cable1/0/0:0
and
interface Wideband-Cable1/0/0:1
and
interface Wideband-Cable1/0/0:2 etc etc.
Also, you want your bandwidth percents up as high as possible- its not going to reserve that amount of channel for just that group, its how much of that channel that group will be allowed to access when available.
With your method 1, you can create 4X bonding-group on the original SPA, plus another 4X bonding-group on the 2nd SPA. The CMTS presents both BG to the modems but randomises the order, so that you should end up with even spread of modems on each.
However if you want to be able to create 8X bonding-groups, you need to go with your method 2.
With method 2, if you have a mix of 4X and 8X modems, then you will need to do what capm suggests, and for each wideband create an 8X subinterface, plus probably 2 x 4X subinterfaces.
If you only have 8X (or larger) modems, then don't bother making the 4X BGs.
Thank you both for your responses! We tried to implement this over the weekend and while we were able to get modems to bond 8 channels, we ran into exactly what you have described: 4 channel only modems only locked on to a single channel, while 8 channel modems locked into all 8. We also found out some other issues we need to address, so we reverted our changes for now. But, I do have a few follow up questions.
1. To use all 8 channels, do the 8 channels HAVE to be on the same SPA? Can I split 4 channels up between 1/0/0 and say, 1/1/0 and have a single cable modem utilize all of those? Or, do they need to be off the same SPA unit? My assumption is that they have to be on the same SPA (see below)
2. Can I create overlapping fiber node groups that allow modems to use tap into either a 4 channel group, or an 8 channel group? Example would be:
cable fiber-node 1
downstream Modular-Cable 1/0/0 rf-channel 0-3
downstream Modular-Cable 1/0/0 rf-channel 4-7
downstream Cable5/0/0
upstream Cable 5/0 connector 0
cable fiber-node 2
downstream Modular-Cable 1/0/0 rf-channel 0-7
downstream Cable5/0/0
upstream Cable 5/0 connector 0
Two fiber nodes, 1 for the two 4 channel blocks, but then 1 defined (same DS and US connector) for an 8 channel block. Would this work in the same way, or do I need to look into the sub interfaces as you have suggested instead? Assuming that it would work with the fiber-nodes, I would imagine that this is why my channels would have to be on the SAME SPA using the above method, correct?
Again, thanks so much for your responses. We are slowing getting an idea of how this all works and the more we mess with it, the more we are learning on what goes on inside this magical UBR10k CMTS. :)
Edit: Also want to point out that we converted our bandwidth percentages to 'dynamic bandwidth sharing' as well, so that should help with any excess bandwidth that wasn't being used.