David.Smith@esa.int writes:
> The initial plan was to set up a main INN server, and run NNTPCache on the
> site machines. but reading the documentation I have come to believe that
> it would be possible to do this using only NNTPCache. Am I right? If so,
> would we be better to use a central cache server and four site caches,or
> could we dispense with the central server altogether?
Provided you don't need local groups, this model is fine. I suggest using
a central cache server, because it will take additional load off your
upstream, and give a little more reliability if the up-stream goes down
(nntpcache can serve what it has in the cache if the up-stream becomes
unavailable). Because nntpcache tries very hard to keep 1:1 article numbering
with the up-stream, you can have a fall over-mode for the central nntpcache
server too.
> Can you also give me an idea of what size machines we're looking at for
> the caches? We'll be using SUN hardware running Solaris 8.
>
> Thanks in advance,
Given the page sizes of ultras, 1MB ram per concurrent user is more
than adequate. (half is that is probably fine, but if you want to play
it safe..)
CPU generally isn't an issue, disk seek speed can be. 3 seeks per article
is an upper estimate.
Cheers,
Julian.
--
Julian Assange |If you want to build a ship, don't drum up people
|together to collect wood or assign them tasks and
proff@iq.org |work, but rather teach them to long for the endless
proff@gnu.ai.mit.edu |immensity of the sea. -- Antoine de Saint Exupery
