[Bloat] fixing bufferbloat in 2017
Benjamin Cronce
bcronce at gmail.com
Wed Nov 23 13:05:27 EST 2016
On Wed, Nov 23, 2016 at 11:56 AM, David Lang <david at lang.hm> wrote:
> that doesn't even do 5GHz, so your wifi performance will be cripped by
> interference and the lack of available bandwidth.
>
>
> On Wed, 23 Nov 2016, Noah Causin wrote:
>
> There is a company called Netduma which sells a product called the Netduma
>> R1 Router. It's main feature is reducing lag. It does this through QOS
>> and GEO-IP Filtering. (Limiting available servers to your local region =
>> reduced RTT)
>>
>> It seems relatively popular in the gaming world, especially console.
>>
>> It is based on OpenWRT Chaos Calmer: https://netduma.com/opensource/
>>
>> It has an advanced QOS system that already uses FQ_Codel.
>>
>> Here are the hardware specs:
>>
>> https://netduma.com/features/hardware/
>>
>> I assume it has an ath9k.
>>
>> Maybe they could implement the ath9k fq_codel and airtime patches.
>>
>> The user base that buys this product seems like they would be more
>> familiar with setting up routers than the average person.
>>
>> On 11/23/2016 12:31 PM, Mikael Abrahamsson wrote:
>>
>>> On Wed, 23 Nov 2016, Benjamin Cronce wrote:
>>>
>>> If there is a simple affordable solution, say Open/DD-WRT distro based
>>>> bridge that all you do is configure your up/down bandwidth and it applies
>>>> Codel/fq-Codel/Cake, then all you need to do is drive up awareness. A good
>>>> channel for awareness would be getting in contact with popular Twitch or
>>>> YouTube gaming streamers. But I wouldn't put much effort into driving up
>>>> awareness until there is a device that people can easily acquire, use, and
>>>> afford. At first I was thinking of telling people to use *-WRT supporting
>>>> routers, but changing the firmware on your router requires too much
>>>> research, and many people care about bleeding edge features. You need
>>>> something that works in tangent with whatever they are using.
>>>>
>>>
>>> If Comcast sells you 100/20 (I have no idea if this is a thing), you set
>>> your upstream on this box to 18 meg fq_codel, and then Comcast
>>> oversubscribes you so you only get 15 meg up part of the time, then you're
>>> still bloated by the modem. This is not a solution.
>>>
>>> I don't think "buy $thing, install *WRT on it, configure it like this"
>>> is above most gamers, but I'm afraid we don't even have a working solution
>>> for someone with that kind of skillset.
>>>
>>
I would be curious to know what the 80/20 rule is. Can we reach it with
what I described? The other way to handle the situation you mentioned is to
tell the end users they can trade more bandwidth for a less likely chance
of having high latency, depending on the stability of their ISP.
There is also the strange issue of crazy high bursts from video streaming
services. I know Netflix is working on the packet-pacing problem with
FreeBSD, but I've done packet-dumps from several streaming providers and
the issue seems to be with TCP with transient activity and data transfers
that with in the TCP transmit window. A 5Mb/s average really turns into a
40Gb/s burst of 256KiB of data 3 times a second. Since the buffers are
large, they don't drop anything. The bigger issue is the end-user sees an
"average" ping that is low, but they get constant transient oddities while
gaming and can't figure out why someone streaming 5Mb/s is hosing their
100Mb connection.
Most people only have a 1Gb network link, so a 40Gb burst won't get through
anyway, but they will see a 1Gb burst dragged out 40x longer, giving a
bridging device time to drop a packet or two and signal TCP to back-off.
Looking at my WAN port, I actually see back-to-back packets at 1Gb
line-rate from Netflix, Hulu, Youtube, and Twitch for long lived
connections that have periodic activity.
>
>>>
>> _______________________________________________
>> Bloat mailing list
>> Bloat at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/bloat
>>
>> _______________________________________________
> Bloat mailing list
> Bloat at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/bloat/attachments/20161123/3b250367/attachment-0002.html>
More information about the Bloat
mailing list