[c-nsp] IPerf alternative

[CiscoNSP List]

Much appreciated Saku!


Thanks.


________________________________
From: Saku Ytti <saku at ytti.fi>
Sent: Monday, 7 August 2017 5:37 PM
To: CiscoNSP List
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] IPerf alternative

On 7 August 2017 at 10:25, CiscoNSP List <CiscoNSP_list at hotmail.com> wrote:

Hey,

> Looking for an alternative to IPerf for link testing - Having some "reliability" concerns with IPerf results (Primarily with simultaneous data transfers) - Any recommendations are greatly appreciated (10G capabilities would be fantastic, but not 100% necessary) - Has anyone used/tried Ostinato ? Primarily after a tool that can provide load testing results (simultaneous bidirectional, one direction, and UDP/TCP are primary requirements...Happy to look at a commercial product/appliance if it ticks all the boxes, but not wanting to spend a fortune :)

I am not aware of not-broken-by-design UDP performance measurement
software. iperf,  Netperf at least are entirely broken as they use OS
UDPSocet, which does not perform. AF_PACKET might be able to reach 1GE
on small packets, which I currently view as absolutely minimum for
testing.
TCP makes very little sense, as I want to test the network, not TCP
settings of end-hosts. Friend of mine and I started writing AF_PACKET
based testing tool with rust, and initial results show that 1.5Mpps is
entirely reasonable on typical laptop, but 15Mpps is not. For 10GE
you'd need something like netmap or DPDK based solution, which also
means you'd need to have dedicated NIC for this, which to me is
entirely another class of software, these exists, you might want to
check out https://trex-tgn.cisco.com/.
TRex<https://trex-tgn.cisco.com/>
trex-tgn.cisco.com
TRex Realistic traffic generator . TRex is an open source, low cost, stateful traffic generator fuelled by DPDK. It generates L4-7 traffic based on pre-processing and ...




If you're looking for commercial kit, I'd say go with Spirent or Keysight.

--
  ++ytti