|
On Wed, Jul 30, 2003 at 11:31:04AM +0200, Nicolas Niclausse wrote:
> >>>>> "horms" == horms <horms@xxxxxxxxxxxx> écrivait:
>
> horms> The other reason that persistance is a good idea relates to
> horms> session resumption. This allows subsequent connections to be set
> horms> up much faster if an end-user connects to the same real
> horms> server. Some Layer 4 Switching implementations allow persistance
> horms> bassed on session Id for this reason. LVS doesn't do this. And
> horms> it is a bit hard to put into the current code (when I say a bit,
> horms> I mean more or less impossible).
>
> Does ktcpvs can (or will) do this ?
I don't believe so, but I will stand corrected.
> There is an alternative to persistance: you can share the session IDs on
> the real servers side with distcache http://distcache.sourceforge.net/
>
> (I haven't try it myself)
Nice, sounds very interesting.
> horms> Note that using different real-servers will not cause
> horms> connections that try to use Session IDs to fail. They will just
> horms> use the slower version of the handshake.
>
> Indeed, it will be MUCH slower. I've made a few benchmarks, and https
> with renegociation is ~20 times slower.
That wouldn't supprise me if a client is making many small connections.
--
Horms
|
