|
Hi!
> Have you thought about the algorithm that I posted?
I'll do that today.
> I don't think consistent hashing scheduler can work well. Static mapping
> will overload a cache server, and you never know which IP address is
> "hot" when packets are redirected.
I already wrote about it:
- There are commercial products that work like this
- The basic scheduler had only slight imballances when using a read
trace
- The hot spot solution will remove even those (I hope)
I implemented the hot spot solution in the scheduler and in a simulating
program on user level. And the results with the same workload are
encouraging. Further tests have to be done.
> LBLC's strategy is to assign a new destionation to least loaded server,
> and keep sending packets for the same destination to the same server,
> unless the server is overloaded. Then, allocate the least loaded server
> for this destination.
Hmm. I think I have to look at your algorithm before I ask some stupid
questions.
Thomas
|
