11

In essence I want to know if there is any way that we can use extended ping or any other command variants to prove to the customer that that the circuit is able to achieve the minimum CIR. The CE routers are Cisco ISRs 881,19x1 & 29x1 running 15.1, and the PE routers are Juniper MXs.

In an ideal world we would just plug a laptop into a LAN port and use iperf to a remote destination behind the CE routers, but the WAN / LAN deployment is being done on separate days and therefore have been instructed to leave the LAN ports shutdown.

We need to contractually prove that 80/20MBps VDSL2 circuit meets the minimum 20/4MBps CIR requirements.

For reference: On the Juniper MX side, we can drop down into the shell and and run multiple flood pings with the DF flags set, but the problem is currently with the Cisco ISRs.

NOTE: RE-Protect will be disabled for the bandwidth test period and therefore wont be a factor. and were potentially deploying 12 sites a day so simplicity is a key success factor.

  +---------------+          +---------------+             +-------------+
  |               |+--20mb-->|               |             |Cisco ISR CE |
  |Juniper MX PE  |          |Carrier network|<------4mb---|             |
  +---------------+          +---------------+             +-------------+
  • Hi, could you add model numbers of the Cisco ISR, and how much bandwidth you're trying to generate in each direction? A quick ASCII-art diagram might help – Mike Pennington Aug 22 '13 at 10:51
  • So the Juniper generates 20Mbps towards the ISR with pings... what is the Juniper pinging? If it is the ISR, how much traffic is already generated from ping responses? – Mike Pennington Aug 22 '13 at 14:07
  • If this is a common task, is there a reason why setting up a small dedicated Linux box on each end with packet generations wouldn't be practical? It sounds generally easier to plug in a preconfigured testing box than to touch the routers every time. – chrylis -cautiouslyoptimistic- Aug 23 '13 at 8:54
  • 1
    I immediately thought of Cisco's IP SLA, but adding Juniper to one end, nulls it. I am pretty curious about any solution you might find. – laf Aug 23 '13 at 12:15
  • @MikePennington sending traffic from PE to CE created 1.47 Mbps outbound & 1.45 Mbps inbound. – DrBru Aug 27 '13 at 11:34
8

IOS includes ttcp, albeit it might not be supported officially by Cisco it can come in handy in situations like this.

JUNOS does not support ttcp as far as I know, but it's probably not too much hassle adding one central Linux machine connected to the PE that you can do measurements with.

On IOS, you simply run 'ttcp', like so;

LAB-C7600-1#ttcp
transmit or receive [receive]: transmit
Target IP address: 1.3.3.7
calculate checksum during buffer write [y]: 
perform tcp half close [n]: 
send buflen [32768]: 
send nbuf [2048]: 
bufalign [16384]: 
bufoffset [0]: 
port [5001]: 
sinkmode [y]: 
buffering on writes [y]: 
show tcp information at end [n]: y

ttcp-t: buflen=32768, nbuf=2048, align=16384/0, port=5001  tcp  -> 1.3.3.7

Now I don't have a second end point, but you get the idea :)

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.