Re: Mystery hardware query

["Michael J. Mahon" <mjmahon@aol.com>]

David Empson wrote:
> Steven Hirsch <snhirsch@gmail.com> wrote:
>
>
> > David Empson wrote:
> >
> >
> > > > > At its peak, our network included about 17 Apple IIe workstations on the
> > > > > Nestar network.
> >
> > > > Just out of curiosity, what data bandwidth did the network support?
> >
> > > It was an 8-bit parallel bus, with handshaking. Somewhat like a parallel
> > > printer interface, with a data strobe and an acknowledge. I don't recall
> > > details about the protocol and I'm not sure if I ever analysed its
> > > performance in close detail.
> > >
> > > Each sector or block was transferred over the network as a single burst,
> > > and the average performance of the network was pretty bad with a lot of
> > > computers tring to use it at once, probably due to the contention phase
> > > and polling time.
> > >
> > > Some specialised applications were able to hog the network for short
> > > periods in order to transfer a larger amount of data. One example was
> > > the network login screen, which loaded a fairly large amount of data
> > > quickly.
> >
> > What you describe sounds very much like the original Corvus "flat cable"
> > network.  It was wired in a star configuration with each computer having a
> > 34-pin ribbon cable back to a device called a multiplexer.  This connected to
> > a hard disk and acted as an arbitrator.  I have a bunch of Corvus hardware and
> > software (probably just about everything they ever made), but haven't had it
> > setup for a while.
> >
> > The flat-cable networks were replaced in 1982 or thereabouts with the
> > twisted-pair OmniNet, which was a daisy-chain topology using a form of
> > collision-detect / backoff protocol.

For a software-driven network, it is quite time consuming to sense the
state of the bus to determine that another sender is interfering with
your transmission.  Therefore, I would expect that collision avoidance
techniques were used (all in the arbitration phase) and that collision
detection existed only in the sense that a failed transmission checksum
was detected and retried.

This approach is quite efficient in NadaNet, though it sounds like their
collision avoidance algorithm had some problems as the network became
busy.  NadaNet's approach is to assign fixed arbitration delays to each
machine, amounting to an arbitration "priority".  The resulting bias
is detectable under some heavily loaded cases, but it is simple and
effective, even under high loads.

> Nestar used a 16-pin ribbon cable in a multidrop bus configuration,
> physically wired as a daisy chain between adjacent computers. I'm not
> sure about the electrical interface - it was bidirectional, which would
> complicate the interface if each card had to buffer and retransmit.
> Individual computers could be switched off, which suggests passive
> signal forwarding rather than active retransmission.
>
> There were probably some multidrop pins dedicated to the arbitration
> process, and active transmission by the computer which is sending data,
> by enabling its line drivers.

Yes, I would expect that it is a full parallel bus (multidrop).  That
is the simplest and fastest configuration for small networks.

> Don't know how it dealt with fan-out problems or transmission line
> effects like signal reflections. I don't recall anything like a
> terminator as a separate component, or as a switch setting on the "end"
> cards.

At Apple II signalling speeds, 1 cycle is about 200 feet of cable, and
since handshaking involves multi-cycle delays between sensing "data rdy"
and sampling data lines, any reflections would be fully damped out.

> The configuration lent itself well to classroom layout, with all the
> computers around the outside wall (or in rows with a single "return"
> cable to get back to the wall) and not requiring a massive lump of
> cables arriving at a common point.

Right--tree wiring topologies are the most natural, and virtually
everything from plumbing to utilities use that topology.  Any other
topology must either "simulate" tree structure or require new wiring
ducts.  ;-)  Of course, an open classroom plan is more versatile.

-michael

NadaNet file server for Apple II computers!
Home page:  http://members.aol.com/MJMahon/

"The wastebasket is our most important design
tool--and it's seriously underused."