[rescue] Bus Speed v I/O rules of thumb?

[Joshua D Boyd]

On Fri, May 17, 2002 at 01:13:33PM -0700, Francisco Javier Mesa-Martinez wrote:

> The problem is that you are thinking of a bus bandwith as a divisible
> commodity, which would be somewhat true if the bus was fully switched. The
> problem is as follows, most devices can not theoretically saturate the
> bus, however every device does take the whole bus once it is writing to
> it. Thus the behavior of the bus is no where to be as linear as your
> calculations show... so besides bus overhead you also have to take into
> account the actual inefficiency of the devices... several studies proved
> that the behavior of a simple shared bus depended more on the slower
> devices attached to buses than on the higher performance ones.

I didn't know that.  So, a single PCI bus is also unlikely to do what I want
then.  Darn, and I really wasn't expecting to be pushing this machine hard.
This was just supposed to be an upgrade for the NAS, and later (when I'm
rolling in dough), I'd worry about how to push things to the limits (while
still serving more than one machine at 60megs a second).

Well, I can cut back on the network interfaces to one FDDI and one 10baseT.
I still need the FWD cards for the arrays for capacity reasons.  But other
than getting the missing FWD, I probably shouldn't invest much in this
project until I can get a dual bus PCI Sun.

-- 
Joshua D. Boyd