search

jagt / clumsy

clumsy makes your network condition on Windows significantly worse, but in a controlled and interactive manner.
http://jagt.github.io/clumsy/
Other
5.14k stars 503 forks source link

Buffer bloat simulation #103

Open reductor opened 2 years ago

reductor commented 2 years ago

It would be good to have the ability to simulate buffer bloat either with a total number of packets or buffer size, this makes it easier to design and implement things which are a little more resilient to networks which suffer from buffer bloat.

reductor commented 2 years ago

Looking at the lag implementation it might be possible by adding a "drop buffered" and making "KEEP_AT_MOST" tweakable it might be easy to reuse it to achieve this

Tester47c commented 2 years ago

Looking at the lag implementation it might be possible by adding a "drop buffered" and making "KEEP_AT_MOST" tweakable it might be easy to reuse it to achieve this

That is a great idea, you have great ideas where could I contact you? I have a solution for that

reductor commented 2 years ago

That is a great idea, you have great ideas where could I contact you? I have a solution for that

Just reply here, I'm not interested in being contacted on any other medium, additionally just asking that makes me sus of your account being used for scam/spam, especially when you have done the same in another issue.

Tester47c commented 2 years ago

That is a great idea, you have great ideas where could I contact you? I have a solution for that

Just reply here, I'm not interested in being contacted on any other medium, additionally just asking that makes me sus of your account being used for scam/spam, especially when you have done the same in another issue.

Scam how? Im not going to sit here in a thread explaining, it's fine. I suggest you just do it your self not hard at all

reductor commented 2 years ago

Scam how? Im not going to sit here in a thread explaining, it's fine. I suggest you just do it your self not hard at all

The reasons which make it sus to me are:

I am not saying it was your intention but it was my first thought with the wording of your post it made be instantly think that, unfortunately github is not immune from spam and scams.

Also as you suggested I have done a mock implementation of this, using exactly what I mentioned in my comments and it worked 'alright' but not as ideal as I hoped, but am currently rethinking it in terms of throughput and buffer size which is more realistic.

Feel free to put any other suggestions in this ticket, I will not be reachable on anything but Github and related to this ticket will not respond on anything but it and any relevant PRs, unless a good reason is raised to do otherwise. (E.g. there is a gitter or discord for the project)