1760 have enough Horsepower for Multilink ppp 2 T1s bonded?

[smccollum]

Router: Cisco 1760 81920/16384

Config: 12.3-17

Mem: 32768 system flash

My datacenter and I configured matching multilink ppp.. I opened a ticket with TAC to confirm my config was solid. Tech verified it was OK.

2 ptp T1s, i have 1 nbar policy matching citrix protocol...

I have seen a few message board posts that hint 1720/1760 do not have enought processing power for multilink... Your opinion? would it be better to ip cef load-share?

After we setup the configs, our ping times were all over the place with an occasional time out... times were like 4ms, 11ms, 486ms, 4ms, 1234ms, 754ms, etc..etc..

new T1 circuit checked out OK...

[paolo bevilacqua]

Hi, I've seen issue with multilink and tunnels even on routers faster than the 1760.

So it's always more conservative and safer to avoid mlpp and stay with cef load balancing.

[Joseph W. Doherty]

When you ping, is this through a Citrix client or, if not within a Citrix client, is the ping in the same traffic class as Citrix? Is your Citrix NBAR looking at all Citrix traffic, including Citrix printing and disk copying? QoS configured both directions?

PS:

For performance, unless you need the additional aggregate bandwidth for individual flows, per-destination CEF would be optimal.