* Matthew Toseland <toad at amphibian.dyndns.org> [2008-07-07 12:17:33]:

> 3. Eliminate all checkboxes in the installer, move to the wizard, and
> create an advanced options button in the installer. (Toad)
> 
> This would streamline installation significantly: Paranoid users could
> click a box to turn UPnP etc on/off, everyone else doesn't need to worry
> about it.
> 

Heh, what's the story about the installer and the checkboxes?
Good to know that you have debated that... and reached some kind of
agreement; I wasn't even aware that those checkboxes were a problem!

That said I'm more than happy to hand over the maintainance of the
installer :p

> 5. THROTTLE EVERYTHING !! (Toad)
> 
> Not all traffic is throttled at the moment. We should subject all
> traffic, not only data transfer packets, to both congestion control and
> bandwidth limiting. This is related to the low-level streams proposal,
> but can be done without it, and should be easier. This is especially
> problematic for online gaming. See bug 2312. Note that a lot of Freenet
> runs on altruism, thus it is *essential* that it not severely disrupt
> a user's internet connection!
> 

I think that this one is important.

> 7. Automatic bandwidth limit calibration. (Toad)
> 
> Several other p2p apps implement this, we should too. Bandwidth is *the*
> scarce resource most of the time, we want to use as much of it as we can
> without significantly slowing down the user's internet connection (see
> above).
> 

I don't think that such a thing can reliably work. It might work in 80%
of the cases but will badly screw up in others.

> TO IMPLEMENT IN 0.9:
> 
> 1. Streams (low level optimisation)
> 
> At the moment, Freenet is quite inefficient in bandwidth usage: the
> payload percentage, especially on nodes with low bandwidth limits but
> also on nodes with high bandwidth limits, can be quite low, 70%, maybe
> 50%... this is unacceptable. A lot of this is due to message padding. If
> we split block transfers (and large messages) into streams which can be
> sliced up byte-wise however big we want them, rather than into 1K blocks,
> then we will need much less padding and so acheive a higher payload
> percentage. This also has significant side benefits for steganography and
> for connections with low MTU, since we can (on a link with live streams)
> construct useful packets of almost any size.
> 

I am not convinced that I understand the logic here... I can't believe
you guys plan on implementing (7) without that.

In case you didn't suspect it: here is a big news: 
"available and effective throughput depend on packet sizes amongst other
things". At the moment we have oversized packets with a per-link MTU...
which is poorly guessed and never adapted.

> DATA COLLECTION:
> 
> 1. Push/pull tests.
> 
> We should do push/pull tests regularly on the real network. On darknet this
> won't work well because of the difficulty of getting somebody you're not
> near to to fetch your data, although this is IMHO possible with FMS etc
> volunteers. On opennet we can simply select a different pair of peers every
> time to avoid the two nodes converging on each other as they are reused.
> 

What's the point here? Getting some data from the real network to run
the simulations? Testing availability of data, paths length, end
user experience?

I haven't found the old-fashioned threaded-probe-requests you were
planning on implementing in your sum up; isn't it something we could use
here?

NextGen$
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20080710/2161a539/attachment.pgp>

Reply via email to