B5c ~ Hot Standby and/or Redundant Link


#1

Having the ability to deploy Hot Standby and/or Redundant Link ability with the B5c radios would great.

  1. 1+1 Hot Standby links. These links are designed to provide full redundancy in the event of a single radio failure at one or both ends of the link.

  2. 2+0 Links. Redundant links in a ring or mesh configuration plus non-redundant links.


#2

If you use Mikrotik Routers this is pretty easy to implement with their interface bonding stuff. We have it on one of our Towers and it works pretty well. I just reset it up so if you are interested I could show you our getup.
Dunno how a B5c could be setup to do this, simply, some sort of connection between the two that reports status info to the standby link??? Not easy.


#3

We are a B5c and Mikrotik shop and have kicked this around some in the past. We’d like to hear more, if you are still offering. We’d assume you are running two pairs of B5cs. Or one pair of B5cs and something else? And running in different channels to avoid self interference?


#4

Sure no problem,
As far as the B5 stuff goes, it kinda depends on your situation, antenna, radio spacing, usable spectrum, etc.

On the Mikrotik side, there are a couple different ways to do the same job. (as is normal) If all you are trying to do is have a hot backup of a wireless link (in which case you probably want your B5s to be as different as possible) then it is really simple.

This is what we have configured at the moment. It works super fast and is really easy to setup. the 10.0.3.15 address is one of our main switches on the other side of the link. I have considered lowering the arp interval (mii-interval), but 100ms seems to work alright. I would also consider having a >1 second up-delay just in case the link is faulty, but keeps coming back online.
image

Here is an example setup from one of my test routers.
interface bonding add name=“testBonding” mode=“active” primary=ether9 slaves=ether9,ether10 link-monitoring=arp arp-interval=100ms arp-ip-targets=10.0.3.15 up-delay=10

Now, this only tests for when the link (ether9) goes down, and uses ARP to do the testing. When the link goes down all traffic is swapped over to the other link (ether10 in this case). The router keeps testing primary link and will switch back to it when it sees it has been up for 10 seconds.

If you want to do monitoring of the link speed itself, then you are going to get hairy pretty quick especially if you are like me and have different backhauls from various companies. Honestly at that point I am still trying to get something reliable. SNMP “should work” but I have not found a reliable way to pull that into the Mikrotik scripting. (I gave up pretty quickly tbh, so if someone else is in the know, please enlighten me)

If you want to aggregate traffic accros two links. Then change your bonding type to “ballance xor” and using a “Transmit Hash Policy of Layer 2 + 3” is what tested best for us. We have not use this in the field yet, but it was pretty promising. Though you will want a good bit of beef for the processing.

If what you are doing is a ring setup like #2 of WFF’s post, then you may want to do BGP, or at least you are going to have to understand link distances in your routing table and/or be running VLANs.

Something like that as an example: We started offering to connect some of our customers to two separate towers so they could have redundant routes in-case of a tower outage (rural Colorado has stock traders to my amazement.) We gave some of the first customers a deal because we install a Mikrotik that only we have access to and we can VPN to it if all the backhauls to a tower have gone down. (helped me diagnose a bad power supply before our techs arrived onsite)

This is the Mikrotik at the customer’s location. The key portion is that the second route has a distance of ‘2’ whereas the default is ‘1’. In the default route I have the Check Gateway set to “Ping” this works out pretty well especially when we have to do maintance on a tower sector with these customers. We can swap them over the the backup link and they never see any down time and we don’t need to wait till 2AM to make changes to their connection. If the first route fails it becomes disabled and the system goes to the backup link.


#5

Thank you for putting this together. It is very much appreciated.

David Coudron
david.coudron@advantenon.commailto:david.coudron@advantenon.com | Mobile: 612-991-7474

Advantenon, Inc. [cid:image001.png@01CEE562.60FF8FC0] http://www.linkedin.com/company/Advantenon [cid:image002.png@01CEE562.60FF8FC0] http://www.twitter.com/Advantenon [cid:image003.png@01CEE562.60FF8FC0] http://www.facebook.com/Advantenon [cid:image004.png@01CEE562.60FF8FC0] http://blog.advantenon.com/
info@advantenon.commailto:info@advantenon.com | 3500 Vicksburg Lane N, Suite 315, Plymouth, MN 55447 | www.advantenon.comhttp://www.advantenon.com/ | Phone: 800-704-4720 | Local: 612-454-1545

[/uploads/mimosa/original/1X/d84432586b6c962e5bfd4a671106164d5f988a16.jpg]