[Tech] The effect of slow nodes

[toad]

On Fri, Dec 08, 2006 at 12:17:23PM +0000, Michael Rogers wrote:
> toad wrote:
> >> "failed (rnf)" means the search ran out of nodes before running of hops. 
> >> However, this could be caused by timeouts, because if we time out 
> >> waiting for a peer to accept the search we move on to another peer.
> > 
> > Right. But the timeout is fixed on each hop, so very often the first
> > node will timeout by the time the later node times out.
> 
> IIRC the search timeout is longer than the accepted timeout, so once 
> you've accepted the search it's possible to get several accepted 
> timeouts from downstream peers before the upstream peer gets a search 
> timeout from you.

Okay, that's true. Does this happen a lot in the simulation?
> 
> > The implemented algorithm is that we increase the rate of sending
> > requests if we complete without timing out ourselves, or receiving a
> > timeout notification from another node involved in the request (these
> > are propagated back to source).
> 
> Cool, I think that matches what I've simulated.
> 
> > By the way, did you try implementing the throttle with RTT, and
> > comparing to throttle without?
> 
> Not yet but I will.

Looking forward to that. We tried to simulate TCP as closely as
reasonably possible in the original AIMD implementation, while tweaking
the numbers a bit, and leaving out slow-start. It's probably not the
best metaphor, especially with pre-emptive rejection.
> 
> Cheers,
> Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://emu.freenetproject.org/pipermail/tech/attachments/20061208/59c740c6/attachment.pgp