netdev - Netem 'caveat' clarification

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Mon, 6 Apr 2020 11:12:09 +0000 (UTC)
From:   Jason Forster <jasonajf@...nternet.com>
To:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Netem 'caveat' clarification

Hi all,
I'm looking for clarification to the statement on https://wiki.linuxfoundation.org/networking/netem, namely;

Caveats
When loss is used locally (not on a bridge or router), the loss is reported to the upper level protocols. This may cause TCP to resend and behave as if there was no loss. When testing protocol reponse to loss it is best to use a netem on a bridge or router

What are the limits to this statement? If I run on a router on a separate piece of h/w with netem, then clearly there'll be no issue. Is it an invalid configuration (which will in turn lead to misleading results) to run netem in a container and implement the router functionality inside the container?

I've created a setup using a container configuration on which netem impairments were applied, routing the tcp iperf from the host to the container and encouragingly saw the remote iperf server dealing with the lost packets. However, in an attempt to prove the container approach was not giving misleading results, I removed the container from the ip path, applied the losses locally and got the same results as those when running with the container. I was expecting the 'caveat' to be hit since these would be classed as local losses and the iperf server would not be seeing any packet loss. 

Any pointers or info would be gratefully received.

Apologies if I've arrived at the wrong form - if so - please advise the correct address to post to.

Cheers,
Jason.