[Codel] [Cake] Proposing COBALTLike the issues with streaming video, which are potentially addressed by the fq qdisc. I wonder how much it would help for torrents. (Avoiding bursting the entire congestion window at the start of chunk 2, using pacing).

Noah Causin n0manletter at gmail.com
Sat Jun 4 17:21:37 EDT 2016


Steam uses a content-distribution system called Steampipe.

https://developer.valvesoftware.com/wiki/SteamPipe

I think they used to use UDP, but now they use http.  I think it helps 
clients get around restrictive firewalls.

On 6/4/2016 4:04 PM, Alan Jenkins wrote:
> On 04/06/2016, Noah Causin <n0manletter at gmail.com> wrote:
>> I notice that issue with Steam.  Steam uses lots of ECN, which can be
>> nice for saving bandwidth with large games.  The issue I notice is that
>> Steam is the one application that can cause me to have ping spikes of
>> over 100ms
> Am I right in thinking Steam uses a torrent download?  Torrent
> download (e.g. Transmission client) has the same effect on my
> connection with fq_codel.  See my last post on the bloat list and
> Dave's comment on it :).
>
> It amuses me because I used to think a) the main problem with torrents
> was due to ubiquitous upload bloat b) the uTP congestion control
> (LEDBAT) fixes it.  After monitoring uTP v.s. codel you see that's not
> the whole story.
>
> Dave's point was you can fix download bloat more easily on the ISP
> side of the bottleneck.
>
>
>> even though I have thoroughly tested my network using both
>> flent and dslreports.
>> I also notice that I get large sparse delays in the cake stats during
>> steam downloads.  The highest I can remember right now is like 22ms.
>>
>> On 6/4/2016 9:55 AM, Jonathan Morton wrote:
>>>> On 4 Jun, 2016, at 04:01, Andrew McGregor <andrewmcgr at gmail.com> wrote:
>>>>
>>>> ...servers with ECN response turned off even though they negotiate ECN.
>>> It appears that I’m looking at precisely that scenario.
>>>
>>> A random selection of connections from a packet dump show very high
>>> marking rates, which are apparently acknowledged using CWR, but a
>>> subsequent dropped packet (probably due to queue overflow) takes many
>>> seconds to be retransmitted (I’m using a rather high memory limit for
>>> observation purposes).
>>>
>>> Overall the TCP behaviour is approximately normal for NewReno on a dumb
>>> FIFO, and the ECN signalling is completely ignored.  This doesn’t rule out
>>> the possibility that it’s a different Reno relative, such as Westwood+ or
>>> Compound.
>>>
>>> There’s often more than one CWR per RTT.  This isn’t a consistent
>>> characteristic; some connections have normal-looking CWRs while others
>>> issue them every three packets, as if they’re fishing for “more accurate”
>>> ECN feedback.  It might vary by host; I didn’t keep track of that.  But
>>> this can’t be DCTCP; even that should back off in the face of a 100%
>>> marking rate, which is often achieved at my low bandwidth and with very
>>> persistent queues.
>>>
>>> Other servers respond normally to ECN signals, ruling out interference by
>>> my ISP. It’s possible the ECE flag is wiped and the CWRs are faked, but
>>> there’s no legitimate reason to do that.  The CWRs ultimately make no
>>> difference, since at 100% CE marks, every ack has ECE set anyway.
>>>
>>> Turning off ECN negotiation at the client results in a much better managed
>>> queue with similar throughput.  It’s not immediately obvious whether
>>> that’s due to a functioning congestion response or simply the AQM clearing
>>> out the queue the hard way.  It’ll be interesting to see what effect
>>> COBALT has here, when I get it to actually work.
>>>
>>> As for who these servers are: Valve Software’s Steam platform.  I did say
>>> they were large and popular.
>>>
>>>    - Jonathan Morton



More information about the Codel mailing list