• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Layer 2 Latency testing

nweaver

Diamond Member
I'm looking to do some L2 testing (Latency) to make sure that a linux solution can handle the load I am going to through at it...

similar to this: link

I am looking for a way to emulate/replace that IXIA machine, using open source tools if possible. I assume this is possible with Perl, i.e. send a packt from one interface, to the other interface, and measure latency, but can't seem to dig much up on CPAN. I'd rather not have to write my own or purchase S/W for this (small project).

Any ideas?

Cliffnotes: Looking to measure L2 latency through switches, using 2 interfaces on 1 machine.
 
You could try (what used to be called) "Tethereal: (the "T" is not a typo)

It is a variant of Ethereal (WireShark) that is capable of synchronization between analyzers so the transmitter and receiver's (real-time) clocks are in sync for an accurate latency measurement.

Some of our staff have used it to get good measurements across WAN links.

There might be a flavor of QCheck for Linux, I haven't looked lately.

The part about "one machine, two interfaces" just hit me ... I'm not sure you can do it without using VLANs and a router. If you put both interfaces in one subnet, the host would probably realize thast you are sending to the same box and it would likely just hairpin the traffic through the stacks.

If one interface is promiscuous, you'd still need to aim the transmitted traffic to a remote source and measure the round-trip time (using something like Ethereal/WireShark).

The ideal, though not likely to be the cheapest/free-est, solution would be to rent a SmartBits tester. In additon to latency, it'll give you throughput, packet-drop, head-of-the-line blocking & some other occasionally useful stuff.

If you can do it with two machines, that would make things a bit easier.

Good Luck

Scott





 
Originally posted by: ScottMac
You could try (what used to be called) "Tethereal: (the "T" is not a typo)

It is a variant of Ethereal (WireShark) that is capable of synchronization between analyzers so the transmitter and receiver's (real-time) clocks are in sync for an accurate latency measurement.

Some of our staff have used it to get good measurements across WAN links.

There might be a flavor of QCheck for Linux, I haven't looked lately.

The part about "one machine, two interfaces" just hit me ... I'm not sure you can do it without using VLANs and a router. If you put both interfaces in one subnet, the host would probably realize thast you are sending to the same box and it would likely just hairpin the traffic through the stacks.

If one interface is promiscuous, you'd still need to aim the transmitted traffic to a remote source and measure the round-trip time (using something like Ethereal/WireShark).

The ideal, though not likely to be the cheapest/free-est, solution would be to rent a SmartBits tester. In additon to latency, it'll give you throughput, packet-drop, head-of-the-line blocking & some other occasionally useful stuff.

If you can do it with two machines, that would make things a bit easier.

Good Luck

Scott

The problem with 2 machines is ensuring clocks work...I need to measure time from the packet leaving to the packet arriving, and it's hard to measure. I've heard of Tethereal, I may check that out, and just use 2 boxes.
 
Back
Top