Re: [Bug-gnupedia] Architecture Questions

[Rob Scott]

1)Nobody's really sure if GNU's servers can/want to
cope with the traffic/db/webserving.   We also
shouldnt get too far into this, i dont think we can be
saying things like your britannica comment quite yet. 
 We can probably adapt to the situation better when it
arises, rather than preplan now, making hasty guesses
as to what the problems will be.   We can solve the
problem much better when we know what it is.

oh i think thats given my opinion on 2 3 and 4 too.

5)Apache.   Who knows we may even get sponsorship to
set up our own server farm, we just dont know that
yet.
If done in Perl, and mod_backhand, i dont think theres
any finite limit on concurrent threads.
I definitely think we should minimise CLIENT side
work.   I mean, us having servers can adapt, but i
think we want to cater even for people using lynx on
an old 386.   I cant see encarta doing that.


IMHO of course



--- Bob Dodd <address@hidden> wrote: > I was
wondering, has much been done to consider
> database sizing issues
> and/or response times? I hope *something* has… 
> 
> Assuming we are aiming for an encyclopedia of such
> breadth/depth that
> Britannica will pale into insignificance, this
> really is something we
> need to consider before leaping to any architecture
> decisions, let
> alone choosing particular database products to
> support us.
> 
> So, my first questions are:
> 
> 1) Where will our master database live? I'm assuming
> GNU servers… But
> if our encyclopedia takes off, will GNU survive, or
> will so many
> "normal" users (as opposed to the limited number of
> developers who use
> it regularly)  querying the encyclopedia look start
> to like a denial of
> service attack? :-)) How many hits can it really
> take? Will we be
> limited to a maximum throughput rate, after which we
> must
> reject/redirect queries?
> 
> 2) What quantity of data can we put on this "master"
> server before the
> owners start knocking on our door and asking for
> their disk back
> please?
> 
> 3) What kind of update rates are we expecting on our
> "core" data? I
> know it can only be an informed guess, but our
> sizing work has to start
> somewhere... And ina similar vein, what hit rate are
> we guessing at?
> 
> 4) How much processor load are we allowed for
> handling the database
> queries? Is there a limit to the number of
> concurrent threads etc?
> 
> 5) What web server will we be sitting behind? Apache
> I assume... What
> feature of that server are we allowed to use? Is
> there a requiremet to
> maximise client-side work?
> 
> /Bob Dodd
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Auctions - Buy the things you want at great
> prices. 
> http://auctions.yahoo.com/
> 
> _______________________________________________
> Bug-gnupedia mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/bug-gnupedia


____________________________________________________________
Do You Yahoo!?
Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk
or your free @yahoo.ie address at http://mail.yahoo.ie