Using 14.2R7.5, so far I've scaled my dual mx104 cgnat deployment to ~4,500
private ip customers with NAPT (nat overload) to a /25 of public address space
per MX104 (so /24 total)
I think in 14.2R6 there was a MS-MIC-16G reboot issue that was occurring! With
14.2R7.5 this no longer occurs :)
CPU seems like it's barely trying at ~55,000 nat translations per MX104.... our
end-state goal is entire DSL community to be behind the dual MX104 cgnat
boundary (~8,500 subs)
This cgnat is with MPLS L3VPN integration. One vrf (ri) for inside nat domain,
and another vrf (ri) for outside nat domain. I can one-arm/nat-on-a-stick my
MX104 PE's pretty much anywhere in my MPLS cloud and accomplish cgnat.
I also use AE's for the dual 10 gig, so 20 gig lag/lacp fat pipe, into the mpls
network, and works fine.
--------------------------------------------------------------------------------------------
agould@stlr-h-104> show chassis routing-engine | find Load
Load averages: 1 minute 5 minute 15 minute
0.00 0.00 0.00
agould@stlr-h-104> show services flows count
Interface Service set Flow count
ms-0/0/0 cgn-sset 57335
--------------------------------------------------------------------------------------------
agould@blcn-h-104> show chassis routing-engine | find Load
Load averages: 1 minute 5 minute 15 minute
0.02 0.02 0.00
agould@blcn-h-104> show services flows count
Interface Service set Flow count
ms-0/0/0 cgn-sset 54141
- Aaron
_______________________________________________
juniper-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/juniper-nsp
