On Sun, 22 Feb 1998, Michael Sokolov wrote:

>    Jacob H. Suter <jsuter@intrastar.net> wrote:
> > The problem is with farms is that one thing goes down, its
> > broken.  Lets say your central NFS server for /home and /var/mail
> > die...  You're screwed.
>    
>    Why can't the same happen with a single box?
>    
	With a farm you have multiple points of failure:
	Comparing 1 box with a 95% reliability rate to a farm of 10
	with the same rate (per box:

	    			Probablility of failures
	    One box		  1 - 0.95
				= 0.05 or 5%

	    Ten boxes		  1 - .95*.95*.95*.95*.95*.95*.95*.95*.95*.95
				= 1 - 0.60 
				= 0.40 or 40%

	This is rather simplistic as it assumes all failures are equally
	disruptive, but it does illustrate the point.

> > Well "new crap" isn't that bad of a thing.  Forcing a school full of
> > people to use outdated stuff is pretty shitty IMHO.  Students *do* pay
> > for proper Internet access.
>    
>    But what if the "outdated stuff" is _BETTER_ than new crap? I claim that
> proper ARPA Internet access is that through a VAX running Berkeley UNIX(R).
>    
	Since your machines are locked away from the users all they see is
	the software interface, be it on a Vax or an Alpha.

	You are arguing two points here - the hardware and the software.
	The hardware _only_ matters to you - the users only care about it
	if it is slow or down. (The software has been covered in a different
	email)

		David/absolute

		     He who laughs, lasts...