|
On Fri, 28 Jul 2000, Julian Anastasov wrote:
>
> Hello,
>
> On Fri, 28 Jul 2000, Joseph Mack wrote:
> > >
> > > For UDP the picture is different. We can remove the implicit
^^^
ah, we can (in principle, if we wanted to) remove
> > > persistence for the UDP connections and to select different real
> > > server for each packet.
> >
> > I thought that UDP connections to the same service from the same client IP
> > all were scheduled to the same real-server, unless there were no
> > connections for a timeout period which i thought was about 15mins. I
> > remember someone trying to load balance ntp and he was not getting good
> > load balancing on a short time scale and this was the reason.
>
> Yes, this is the current handling. My thought was to implement
> a new feature: schedule each UDP packet to new real server. But it
> is not possible for TCP. I.e. something like timeout=~0 for UDP
> as service flag.
why was it done the way it is now? To save time making and deleting
hash table entries?
Joe
--
Joseph Mack mack@xxxxxxxxxxx
|
