[Bloat] quick review and rant of "Identifying and Handling Non Queue Building Flows in a Bottleneck Link"

David Lang david at lang.hm
Thu Nov 1 14:15:25 EDT 2018


re-sending since it bounced the first time

On Thu, 1 Nov 2018, David Lang wrote:

> On Thu, 1 Nov 2018, Michael Welzl wrote:
>
>>> On 29 Oct 2018, at 05:02, Dave Taht <dave at taht.net> wrote:
>> 
>>> 
>>> 
>>> Dear Greg:
>>> 
>>> I don't feel like commenting much on ietf matters these days
>>> but, jeeze,
>> 
>> (snip)
>> 
>> There seems to me to be a disconnect here, the core of which is this 
>> comment:
>> 
>> 
>>> Did I rant already that the vast majority of flows are non-saturating?
>> 
>> That's a bug, not a feature - and you seem to treat it as an unchangeable 
>> fact.
>
> Why would you think that saturating flows should be common? A very large 
> percentage of Internet traffic is streaming audio/video and that should never 
> saturate a link, it should be pacing the data to the rate of the content.
>
> DNS queries are not going to be saturating.
>
> queries to check cache validity are not going to be saturating.
>
> microservices calls (including most IoT data) and their replies are not going 
> to be saturating, in part because they don't have much to say, and in part 
> because even if they do have more to say, they aren't going to ramp up to 
> high packet rates before they run out of data to send.
>
> It's only bulk transfers of data that are possibly going to be saturating, 
> and they are only going to saturate their allowed share of the slowest link 
> in the path. On all other links they are going to be non-saturating.
>
> As links get faster, things that would have been saturating years ago fail to 
> saturate the new, faster links.
>
> So what would the Internet look like if it didn't have the vast majority of 
> flows being non-saturating?
>
> David Lang
>



More information about the Bloat mailing list