FYI: I2O Architecture (fwd)

Robert Glamm (
Tue, 19 Nov 1996 10:35:19 -0600 (CST)

> I haven't seen anything on this mailing list about I2O, so I thought I'd
> drop a pointer.
> The I20 architecture is suppose to allow the distribution of I/O functions
> over multiple processors. A lot of major companies appear to be behind this.
> Unfortunately, to receive the spec. you need to sign a NDA (no idea how
> restricting it is).
> Have a look at;

Looks horribly inefficient to me (as opposed to what they claim). I am
unconvinced that a "message-passing interface" to device driver/OS
interaction can provide "high-performance I/O systems." Why stick
what essentially amounts to another protocol stack sandwiched between
the hardware-specific points and the O/S layer? Eesh.

The only slightly cool thing that could arise out of this would be a
truly heterogeneous multiprocessor; with a little extension, one machine
could call OS services on another machine via the protocol layer. This
might be desirable for distributed-memory systems except for the fact that one
bad kernel call could bring down all the machines in the cluster.

"Honestly, it's like shooting  | Bob Glamm  H: +1 612 6239437 W: +1 612 6268981 
 fish in a barrel. Twice. With | URL:
 an elephant gun.  At point    +-----------------------------------------------
 blank range.  In the head." -- from the BOFH files, part 6