To be honest, I'd run them as a P instead and use the MX as your PE. For large L2 flows over MPLS there's no support or plans (or maybe even the ability afaik) to support any sorts of FAT-PW or entropy label for the box. As an LSR they work great.
There are some issues in the current 14.1 code with any traffic that *might* get an ICMP redirect (even if no-redirects is turned on) that causes some awesome behavior (such as duplicate packets where one gets forwarded by the RE, and another by the BCM asic).. Outside of that and a couple of other issues 14.x has been fine so far on it. -- Tim On Thu, Dec 18, 2014 at 4:07 PM, Joe Freeman <[email protected]> wrote: > Before we go out and spend a large amount of money to replace some gear > we're not happy with in our network, I thought I'd check to see what > opinions I could get on the QFX5100. > > We are looking at using them in an MPLS PE role, with the new code release > that has added support for L2VPN's (according to our SE). Each of these > 5100's would be connected to at least one MX router (preferably two for > redundancy) via one or more 10Ge LAG groups. Our CO's are too far apart for > 40Gbe at the moment. > > Any thoughts or opinions would be helpful. > > Thanks- > Joe > _______________________________________________ > juniper-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/juniper-nsp _______________________________________________ juniper-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/juniper-nsp
