On Monday 07 July 2008 16:32, Daniel Cheng wrote: > On Mon, Jul 7, 2008 at 7:17 PM, Matthew Toseland > <toad at amphibian.dyndns.org> wrote: > [...] > > 1. Selective randomization. > > 2. Targeted swapping. > > 3. Opennet nodes and swapping. > > I think (3) deserve a higher priority then (1) and (2). > Location chunk have been a big problem for a long time. The change is > simple: disable swapping on opennet should relief the problem, result > is deterministic, no dark magic are involved...
Dark magic are most definitely involved in the interface between opennet and darknet on a hybrid network. We need simulations here IMHO. > > [...] > > 4. Non-convergent splitfile encryption. > > Should we introduce a new url type (CHK@/USK@/SSK@) for this? CHKs can be happily encrypted with any key, this is by design. > > [...] > > 1. CPU profiling. > > I have been doing this for some time. Most CPU time are spend on FEC > and client layers. I will do more test after db4o branch is merged. That was more or less my impression, but Ian or somebody was complaining about CPU usage for an idle node... -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080709/964bcfd2/attachment.pgp>