Re: [j-nsp] VPLS design - dual homed

2012-10-30 Thread Luca Salvatore
Yes, the MX routers are PE. CE devices at each end will be two EX4500 in VC mode. One connection from each EX to each MX. From: Per Granath [per.gran...@gcc.com.cy] Sent: Tuesday, 30 October 2012 8:46 PM To: Luca Salvatore; juniper-nsp@puck.nether.net Subje

Re: [j-nsp] VPLS design - dual homed

2012-10-30 Thread Per Granath
Are those four MX your PE routers? Does your CE devices connect to one or two PE routers? > I have a question regarding dual VPLS links. My topology will look like this: > > MX1-darkfibre--MX2 > | | > |

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Doug Hanks
The MX can do this several different ways: 1) VPLS MH - it's a O(n) convergence problem, where n is the number of VPLS instances 2) MC-LAG A/S - now becomes a O(1) convergence problem On 10/29/12 4:19 PM, "Luca Salvatore" wrote: >Hi Guys, > >I have a question regarding dual VPLS links. My top

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Jonathan Lassoff
On Mon, Oct 29, 2012 at 8:03 PM, Luca Salvatore wrote: > Right. So BGP needs to be full meshed also for this to work right? Yes -- assuming you're using BGP for signaling. Since BGP is used to signal where the PE/edge ports are, once it comes time to use that to establish LSPs, the routers just

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Luca Salvatore
: Re: [j-nsp] VPLS design - dual homed > My topology is a bit different though... There is no 'service provider' cloud > between my MX routers. > They are directly connected via dark fibre. Does the split horizon rule come > into play here? It does - although the physical

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Ben Dale
> My topology is a bit different though... There is no 'service provider' cloud > between my MX routers. > They are directly connected via dark fibre. Does the split horizon rule come > into play here? It does - although the physical topology looks like a ring, remember that it will be all L3

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Luca Salvatore
: juniper-nsp@puck.nether.net Subject: Re: [j-nsp] VPLS design - dual homed Hi Luca, > MX1-darkfibre--MX2 > | | > | | > MX3-darkfibre--MX4 > > So above you see that there are dual links w

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Luca Salvatore
30 October 2012 11:17 AM To: Luca Salvatore Cc: juniper-nsp@puck.nether.net Subject: Re: [j-nsp] VPLS design - dual homed In VPLS, loop avoidance is arranged by the following rule: A PE never forwards a frame received from a PE, to another PE. The use of a full mesh combined with split horizon

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread Ben Dale
Hi Luca, > MX1-darkfibre--MX2 > | | > | | > MX3-darkfibre--MX4 > > So above you see that there are dual links which will create a loop. > > How doe

Re: [j-nsp] VPLS design - dual homed

2012-10-29 Thread chip
In VPLS, loop avoidance is arranged by the following rule: A PE never forwards a frame received from a PE, to another PE. The use of a full mesh combined with split horizon forwarding guarantees a loop-free broadcast domain. From: wikipedia http://en.wikipedia.org/wiki/Virtual_Private_LAN_Service#

[j-nsp] VPLS design - dual homed

2012-10-29 Thread Luca Salvatore
Hi Guys, I have a question regarding dual VPLS links. My topology will look like this: MX1-darkfibre--MX2 | | | | MX3-darkfibre--MX4 So above you