Re: [Bismark-users] coping with conflicts

[Nick Feamster <feamster@cc.gatech.edu>]

I think we are both interested in:

- measuring the current state of affairs (as a baseline)
- designing techniques to try to improve latencies/problems
- measuring how well respective techniques improve latencies/problems

On Mar 15, 2011, at 2:54 PM, Dave Täht wrote:

> Nick Feamster <feamster@cc.gatech.edu> writes:
> 
>> I agree.
> 
> To what? :)
> 
> Would you consider your existing study "ground truth" and the upcoming
> one "revealed truth"?
> 
> When I talked to de-heisenberging below, the goal would be to eliminate
> or at least compensate/understand the factors introduced by the
> measurement box.
> 
> Measuring cable modem and DSL latencies/problems and the internet
> itself, remain a goal, but, for example, the end-user experience of
> using the box itself with some/much of the new stuff is important to
> us. 
> 
>> 
>> On Mar 15, 2011, at 2:43 PM, Srikanth Sundaresan wrote:
>> 
>>> It seems to me that they can coexist; to study the impact of the
>>> changes, we do need to know the existing ground truth, no?
>>> 
>>> - Srikanth
>>> 
>>> On 03/15/2011 02:42 PM, Dave Täht wrote:
>>>> 
>>>> 
>>>> I have some conflicts that I would like to clearly resolve as we ramp
>>>> up.
>>>> 
>>>> 1) The bufferbloat project is trying to fix the internet
>>>> 2) you are trying to study the internet as it exists
>>>> 
>>>> While some of the bufferbloat fixes in the pipeline will de-heisenberg your
>>>> measurements, the two ideas do conflict somewhat as does your schedule
>>>> vs ours. 
>>>> 
>>>> The wndr3700, which 6+ people have now, is looking to become a focus of
>>>> our efforts for item 1, and may ultimately become a model for the
>>>> "perfect router". (Although I would bet on the upcoming 32MB flash stuff
>>>> coming out later this year as a better model)
>>>> 
>>> _______________________________________________
>>> Bismark-users mailing list
>>> Bismark-users@lists.bufferbloat.net
>>> https://lists.bufferbloat.net/listinfo/bismark-users
>> 
>> _______________________________________________
>> Bismark-users mailing list
>> Bismark-users@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/bismark-users
> 
> -- 
> Dave Taht
> http://nex-6.taht.net