|
>>>>> On Fri, 24 Mar 2000 19:01:19 +0000, Markus Bernhardt
>>>>> <mbernhardt@xxxxxxxxxx> said:
> Hi Andreas
>> The most important reason behind this is that we want to
>> avoid bloat on the LVS to make it as fast as possible so it never
>> becomes the bottleneck. Before adding features to it, one should
>> evaluate readily available solutions.
> I think it is not a good idea to see LVS as a stand alone network balancer.
> Its o.K. to say: "If you use it as simple load balancer, it will never be a
> bottleneck, but if you use sophisticated features it can be."
Fine with me.
>> ...., it's only drawback is that it's not as fast as LVS.
> I think the focus always has to be the whole system and LVS is a part
> of the system. It can happen that I am needing a quite complex algorithm
> for the job and integrating it into LVS delivers the fastest solution.
> So why not ?
I do not want to stop anybody from implementing extreme stuff. Of
course I welcome solutions that compete with a squid farm that does
both redirecting and accelerating. Please, do not let anybody stop
you. I only read your posting as a wish list and suggested an
alternative that is available today and that seems to work very well.
--
andreas
|
