netdev - Re: rtt vs. rcv

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>] [<thread-prev] [day] [month] [year] [list]

Date:	Wed, 10 Sep 2014 11:06:46 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Robert Engels <rengels@...ionscity.com>
Cc:	netdev@...r.kernel.org
Subject: Re: rtt vs. rcv_rtt ?

On Wed, 2014-09-10 at 12:15 -0500, Robert Engels wrote:
> When using 'ss', both rtt, and rcv_rtt are reported.
> 
> My understanding is that rcv_rtt is an estimate of the receivers rtt ?? Is
> this true? My assumption is that differences in rtt vs. rcv_rtt would be
> due to asymmetric networking connections ???
> 
> Here is a sample 'ss' output:
> 
>  cubic wscale:2,9 rto:437 rtt:101.625/84 ato:40 cwnd:10 ssthresh:15 send
> 991.9Kbps rcv_rtt:18291 rcv_space:151142
> 
> Can someone explain why the rcv_rtt would be so much higher than the rtt?
> Is it that the client is not reading the data fast enough? Other reasons?
> 
> Thanks so much for your help.

rcv_rtt is not the rtt as you think it is ;)

rtt is the smooth rtt of delays between sent packets and received ACK.

But rcv_rtt is the time to receive one full window.

Its used by DRS (Dynamic Right-Sizing) 
Check tcp_rcv_rtt_measure() & tcp_rcv_rtt_update() for details.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html