Scott Young ([EMAIL PROTECTED]) wrote:

> Another semi-related problem is that nodes sometimes get too much data
> in their sendqueues.  I think I remember someone on this mailing list
> saying that their sendqueue got up to 80 MB.  This is simply too large.
> 
> One solution to this would be to throttle incoming data to the same
> speed as outgoing data for transfers that are part of the same request. 

I don't think slowing things down *more* is the answer.  I'd rather
get that data as fast as I can, because then (if my data store is
less than 90% full), I'll *cache* it, and then I can send it out at
my leisure.

> Back on topic, why do we default nodes with only 8 to 20 k/sec upload to
> 512 connections?

Because keeping an open connection costs very little (with NIO),
and opening a new connection is very expensive (cryptographic
handshaking).  512 open connections isn't a problem; 512 open
connections all sending data *is* a problem.

I'd rather have 512 open connections, but only allow 10 or 20 of
them to be sending large chunks of data at a time.

> Then again, NGR should fix this, since nodes with too many open
> connections send stuff slower.  Therefore, I think your solution is not
> really needed.

We'll see.

-- 
Greg Wooledge                  |   "Truth belongs to everybody."
[EMAIL PROTECTED]              |    - The Red Hot Chili Peppers
http://wooledge.org/~greg/     |

Attachment: pgp00000.pgp
Description: PGP signature

Reply via email to