...@beamspeed.commailto:bl...@beamspeed.com
Date: November 2, 2010 23:59:16 PDT
To: WISPA General List wireless@wispa.orgmailto:wireless@wispa.org
Subject: Re: [WISPA] PPPoE Concentrator Redundancy
Reply-To: WISPA General List wireless@wispa.orgmailto:wireless@wispa.org
I assume you're using MikroTik
: Thursday, November 04, 2010 12:36 PM
To: wireless@wispa.org
Subject: [WISPA] PPPoE Concentrator Redundancy
I'll try to paint the picture a little more now on this. We are getting a
new feed at a separate location from where our current fiber is at. The idea
is to traverse our existing network and utilize
Cross posting from another list for different opinions..
We're looking to have more than one PPPoE Concentrator available so that if one
goes down due to catastrophic failure, the customers associated to that
concentrator will rollover to the next one. However, the concern is that
because the
Scott,
Typically in this scenario I would recommend one of two things:
1) Use an MPLS VLL/L2-psuedowire with a secondary failover endpoint on one
side.
Only some equipment vendors implement this (Juniper being one of them)
2) Use a PPPoE pado delay. Set one BRAS to be some decent amount
I assume you're using MikroTik.
You can run multiple PPPoE servers on a single Ethernet segment. The client
will send its PPPoE Active Discovery Initiation (PADI) packet, and both servers
will reply with their PPPoE Active Discovery Offer (PADO). The client will then
select which AC it wants
Thx both of you for the replies. We're using ImageStream routers. I've
considered the two running in parallel and whoever responds first thing- but it
seems like a router reboot or equipment failure or whatever would totally throw
off the load balancing aspect of things from which its never
