Re: [Codel] [PATCH 1/2] codel: Controlled Delay AQM

[Jim Gettys <jg@freedesktop.org>]

On 05/08/2012 09:57 AM, Kathleen Nichols wrote:
> I think there are two different kinds of "ECN work" going on: one to
> work on the algorithm and the other to see how widely deployed
> it is.

Yup.  Steve Bauer and others have been studying ECN deployment, along
with the problems they have encountered.  Figuring out if we can safely
enable ECN is the interesting question right now (the server side
deployment is going quite well at this date.
                    - Jim

>
> On 5/8/12 3:25 AM, Jim Gettys wrote:
>> On 05/08/2012 01:10 AM, Eric Dumazet wrote:
>>> On Tue, 2012-05-08 at 00:14 -0400, Jim Gettys wrote:
>>>
>>>> Too late tonight to try to resurrect what I remember from Steve.
>>>> Barring unforeseen events, he should be able to join the list then and
>>>> better coming from him anyway.
>>>>
>>> I never heard of Steve, yet we did a lot of ECN related work on linux
>>> kernel to fix various issues (or allowing a safe ECN activation) these
>>> past years.
>> Steve Bauer works for Dave Clark at MIT.  He's been studying ECN
>> deployment in the global internet the last couple years, by probing of
>> order 1,000,000 sites (and the network paths to them).
>>                     - Jim
>>
>>> commit bd14b1b2e29bd6812597f896dde06eaf7c6d2f24
>>> tcp: be more strict before accepting ECN negociation
>>>
>>> commit e4ae004b84b315dd4b762e474f97403eac70f76a 
>>> netem: add ECN capability
>>>
>>> commit ddecf0f4db44ef94847a62d6ecf74456b4dcc66f
>>> net_sched: sfq: add optional RED on top of SFQ
>>>
>>> commit b5c5693bb723a019deac3cd1345f3e7233c8a67e
>>> tcp: report ECN_SEEN in tcp_info
>>>
>>> commit 7a269ffad72f3604b8982fa09c387670e0d2ee14
>>> tcp: ECN blackhole should not force quickack mode
>>>
>>> commit 5173cc057787560c127c6e9737f308c833dc4ff3
>>> ipv4: more compliant RFC 3168 support 
>>>
>>> commit ca06707022d6ba4744198a8ebbe4994786b0c613
>>> ipv6: restore correct ECN handling on TCP xmit
>>>
>>> commit 4a2b9c3756077c05dd8666e458a751d2248b61b6
>>> net_sched: fix ip_tos2prio
>>>
>>> commit 6623e3b24a5ebb07e81648c478d286a1329ab891
>>> ipv4: IP defragmentation must be ECN aware
>>>
>>> commit 172d69e63c7f1e8300d0e1c1bbd8eb0f630faa15
>>> syncookies: add support for ECN
>>>
>>>
>>>
>>> Yes, some middlebox might be broken.
>>>
>>> But far more users will be stuck in July when FBI stop DNSChanger
>>> servers for good.
>>>
>>> Far more users will have problems at IPv6 launch in June
>>>
>>> And so on.
>>>
>>> Of course we could try to focus on AQM problems for the time being ;)
>>>
>>>
>>>
>> _______________________________________________
>> Codel mailing list
>> Codel@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/codel
> _______________________________________________
> Codel mailing list
> Codel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/codel