[Spread-users] 1 problem with Spread.pm and 1 problem with spread daemon, was

[Theo Schlossnagle]

On Friday, January 18, 2002, at 01:20  AM, Guido van Rossum wrote:
> In our case, we're sending many (mostly) small messages, so we hit the
> 1000 message limit quickly.  We'll be rewriting our protocol to batch
> up things so we will be sending fewer large messages.  This is more
> efficient for spread anyway (per kilobyte payload), and the processing
> time in our app won't be affected.

Spread can pack messages.  I would agree with you though.  I would wager 
that your packing method would be more efficient than Spread's because 
Spread can't make the blanket assumption that all the messages are going 
to the same place.  The packet headers on Spread messages are big when 
you look at message packing.

There are many applications where it makes good sense to increase those 
"hard limits" in Spread considerably.  There are a lot parameters that 
can be tweaked and tuned to make it run better on LANs vs. WANs and more 
appropriately "fit" the needs of your application.  All of these require 
a recompile of the Spread daemon.

--
Theo Schlossnagle
1024D/82844984/95FD 30F1 489E 4613 F22E  491A 7E88 364C 8284 4984
2047R/33131B65/71 F7 95 64 49 76 5D BA  3D 90 B9 9F BE 27 24 E7