Switches off of 1552e APs

Answered Question
Mar 12th, 2012

I am designing a wireless mesh network for a client using several 1552 access points that will be bridged together. Only the RAP has a wired connection to their existing network. The client has informed me they want to hang switches off of the MAPs to connect existing wired devices. I found this in the install manual, "To forward bridge traffic, add a switch between the power injector and controller" but I will not be using power injectors (just AC power). Will adding a switch off of the AP's impact their wireless bridging abilities? Thank you.

I have this problem too.
0 votes
Correct Answer by Steve Rodriguez about 2 years 1 month ago

That should work fine, so long as you follow the MESH design rules.

  • 1
  • 2
  • 3
  • 4
  • 5
Average Rating: 5 (4 ratings)
airframes Wed, 03/14/2012 - 21:37

Torrance,

I recently did an extensive lab to test the capabilities of bridging over the mesh network. Things work pretty well when you just bridge one VLAN (the access VLAN of the RAP(s)), but when you get into VLAN tagging, the configuration starts to look pretty weird on the switches and MAPs. I'm not sure if it's because the code is buggy or if I was just reading the documentation wrong (which I read at least a half dozen times, each time a little more puzzled).

In the end I got it working and stable, and my results don't look much like the docs, so post back here if you reach that point and want some help getting it going.

Also, exercise caution about which VLANs you bridge. If you have large client networks (e.g., wired workstations running chatty protocols or applications) and you start bridging those, all your broadcast traffic is going to hit the backhauls, which could degrade throughput on the mesh.

Justin

yuriy-sokhan Wed, 03/21/2012 - 13:02

Justin,

I have 2x 1552E bridges and I need to bridge (ethernet to ethernet) about 6 VLANs over to the other side. This is a point to point. Can you provide a little insight as to what configuration issues you've experienced and how you got it to work? I have both sides connected to switches with the respective switchports configured as trunks with native vlan 700.

I guess I should add the WLC is a 5508 and running 7.2.103.0

Thanks.

airframes Wed, 03/21/2012 - 13:54

Yuriy,

Yes, I have a collection of lab notes that I will post for you in a few hours. The vlan tagging is a little unconventional in terms of configuration and a bit temperamental operationally. I'm at a point that it's stable and working reliably, but the Cisco documentation comes up short, unfortunately.

Justin

... typd on tny kybrd.

yuriy-sokhan Wed, 03/21/2012 - 19:11

Thank you. I'm really curious what your notes look like, as I've been bashing my head against the wall on this for a good part of my day.

airframes Wed, 03/21/2012 - 21:06

Yuriy,

Attached are my notes. Sorry it took me so long to post them.

They are lab notes, with a minor amount of cleanup for readability. If something doesn't make sense, ask me and I'll clarify.

Please do post back and let me know how it goes. It seems that there only a handful of us out there doing this, and we'll benefit by comparing notes and experiences.

Justin

mihelcic.dinko Thu, 03/22/2012 - 09:35

Justin,

nice notes!

I think that catch is in defining RAP's port as normal, not trunk.

Dinko

yuriy-sokhan Thu, 03/22/2012 - 09:41

The WLC won't let you set the backhaul interface (e0) as anything but normal. Which is good I suppose. I hope Cisco updates their documentation regarding mesh wireless bridging. Lots of holes and little bits of info spread throughout 3-4 different guides.

airframes Thu, 03/22/2012 - 12:15

Yuriy,

What version of code are you running? My 7.0.220.0 WLC lets me set the RAP wired ethernet backhaul (g0) as Trunk. It does throw a warning, but I get the warning no matter where I set it (backhaul or non-backhaul wired port on RAP or MAP).

That's been part of the confusion.

Justin

yuriy-sokhan Thu, 03/22/2012 - 12:24

It is possible that when I tried it, the new setting didnt stick. I've gotten the warning, went back and it was still normal. I didn't get the warning when changing e0 on the MAP though. I believe cisco documentation says the backhaul of the RAP will do pretty much whatever you need as is without any further configuration...pass tags etc.

My code is: 7.2.103.0

yuriy-sokhan Thu, 03/22/2012 - 09:48

Justin - thanks for the notes. I also ran into the PVID mismatches on the trunks, so it does seem like a glitch in the mesh code or a combination of a certain switch IOS and the mesh bridges.

Currently I have both bridges up and trunking about 5 VLANs I had to manually add on the MAP and set its e0 as trunk. I disabled VLAN transparent bridging on the WLC as well.

The RAP config is fairly straightforward with switchport set as trunk with native vlan 700 and RAP backhaul e0 always set as normal. Both MAP and RAP have ethernet bridging checkboxes enabled as well.

So far everything looks good and I'm optimistic that it will work. I haven't had a chance to test it, as the site is 3 hours away from me, but should have confirmation later tonight.

Thanks again.

alzhou Tue, 05/01/2012 - 20:11

Yuriy,

I realized it's been a while. Just wondering when you were doing this Ethernet Tagging on a MAP for bridging over multi-VLAN traffic back to its RAP, was the MAP still providing wirelss client access on its 2.4GHz radio at the same time? Any noticeable performance impact?

Thanks

yuriy-sokhan Tue, 05/01/2012 - 20:25

Unfortunately no, the MAP and the RAP were set as infrastructure only and did not serve wireless clients. I don't see how this would cause much of a performance hit since the backhaul link is over the 5Ghz radio.

Actions

Login or Register to take actions

This Discussion

Posted March 12, 2012 at 7:38 AM
Stats:
Replies:15 Avg. Rating:5
Views:1221 Votes:0
Shares:0
Tags: No tags.

Discussions Leaderboard