[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: (Network Connections)
- Subject: Re: (Network Connections)
- From: david@wraith.cs.uow.edu.au (David E A Wilson)
- Date: 1996/10/22
- Newsgroups: comp.sys.apple2, comp.sys.apple2.programmer
- Organization: University of Wollongong, NSW, Australia.
- References: <32546DD2.7CF3@wctc.net> <32600EC2.54AB@wctc.net> <markem-1410961903120001@hpm2-153.flash.net> <falcon-1510961130540001@cda0-15.iea.com> <540p8a$n08@darla.visi.com> <543t19$315@cisu2.jsc.nasa.gov>
mark@sun (Mark Manning/Metrica) writes:
>Secondly, I have to agree with what Nathan is saying. If
>you had:
>SCSI Disks<--->Computer<--->SCSI CABLE<--->Computer<--->SCSI Disks
>(basically the SCSI disks were hung off of each computer on
>a separate line from the SCSI line going between the
>computers), then what Ross proposed would not be all that
>bad.
You mean each Computer has two SCSI cards (aka Host Adapters)
>But if you had:
>Computer<--->SCSI Disks<--->Computer
>then you are asking for trouble. This is because reads and
>writes are being left up to the SCSI firmware which (I
>believe) is not intended to handle multiple computers
>trying to read/write information.
Unless you have a SCSI card with firmware designed for this.
Cirtech made their cards work like this using the following:
1) The cards use the SCSI arbitration phase so that they do
not step on data transfers already in progress.
2) The software allows for partitions on a per machine basis
(read only for other hosts, read write for the owner host)
Cirtech's system was really quite flexible except that it does not read
Apple standard partition tables...
--
David Wilson Dept CompSci Uni Wollongong Australia david@cs.uow.edu.au