|
On 2002-01-14T23:14:45,
abhijeet <abhijeetmore@xxxxxxxxx> said:
> But yes, atleast at this point I feel that we're ready to do *real*
> hard work, if that's what's required.....once we have the correct
> problem that is :-(.
What do you think about this?
LVS is all about load balancing, however it only happens inbetween two tiers
and is somewhat limited in its knowledge.
Wouldn't it be nice if you had hierarchial load balancing - if a servers
weight didn't only include itself, but also all servers behind it. (Think
"tree" and "aggregrated weight" in complex server topologies)
And if the node value you computed did not only take the pure machine load
into account, but also network latency to the remote server, so it would scale
nicely to distributed environments.
And if your "local node" would sort of also detect whether you are running
CPU, IO or memory bound and would adjust its calculation based on that, that
would just so totally rule.
A side aspect of this would be to roll out the uptodate configuration
seamlessly to all nodes from a master node.
This would IMHO be worthwhile, and while it may in fact be quite a lot of
work, you might be able to implement most of it and get quite a few impressive
charts out of it for your final presentation.
What do you think, would that be something of interest to you?
Sincerely,
Lars Marowsky-Brée <lmb@xxxxxxx>
--
Perfection is our goal, excellence will be tolerated. -- J. Yahl
pgp5w1kjv5M0N.pgp
Description: PGP signature
|
