[c-nsp] ipsla - latency - related to cellular backhaul

[Aaron]

Thanks Adam,

sh lpts pifib hardware police location 0/0/cpu0

....shows all 0's in the drop column, but at the bottom it shows...

RP/0/RSP0/CPU0:9k#sh lpts pifib hardware police location 0/0/cpu0 | in drop
Mon Apr 29 08:22:55.180 CDT
Packets dropped by deleted entries: 71429

...any idea what that is ?

Aaron


-----Original Message-----
From: Adam Vitkovsky [mailto:adam.vitkovsky at swan.sk] 
Sent: Monday, April 29, 2013 3:31 AM
To: 'Aaron'; cisco-nsp at puck.nether.net
Subject: RE: [c-nsp] ipsla - latency - related to cellular backhaul

Hi Aron,
Well I believe that any type of active probe that has a responder at the
other end is a valid (not necessarily dead accurate though) delay/jitter
measurement approach. 
By a responder I mean a process that will time-stamp the probe-packet before
and after processing allowing for processing delay elimination. 
And assuming the routers are well time-synced you can get accurate
delay/jitter measurements. 

So this includes IP SLA as well as Y.1731 for CFM. 
So while you can use IP SLA between two (PE routers) PW endpoints you can
use CFM between two of yours customer demarcation switches for L2VPNs. 



LPTS (Local Packet transport Service) it's like a routing process for the
router itself directing which packet needs to go to which RSP or Line Card
and can be used for per-line-card CoPP like functionality. 
As there's a default set of flows and their respectful policers specified
you may need to adjust the value for let's say MPLS OAM in order to avoid
drops in your PW pings. 

You can check for the drops using cmd: 
sh lpts pifib hardware police location 0/0/cpu0 

you can adjust the values with:
lpts pifib hardware police location 0/0/CPU0  flow mpls oam rate 500

-don't forget to do it per line-card


adam
-----Original Message-----
From: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of
Aaron
Sent: Thursday, April 25, 2013 7:02 PM
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] ipsla - latency - related to cellular backhaul

I have seen some latency (measured using ipsla icmp/udp/mpls pw pings)
beyond my agreements with some of our cellular backhaul customers..

 

We are concerned that if/when they ask to see their sla measurements for
their cell towers that we won't be looking very good

 

Cisco Tac is telling me that pings of any type (icmp/udp/mpls pw) are not
the way to truly measure the network and that I should be using CFM/OAM type
stuff.

 

I showed cisco tac a mpls pw ping test I did and it shows drops/time-outs
occasionally (!!!!!!.!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!
etc)

 

Tac says that this drop and the latency seen using various ipsla pings is
expected since all pings are treated less than everything else and could be
getting policed by LPTS (I don't know what LPTS is)

 

What do y'all think about all this?

 

Aaron

_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/