[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Steve Willis writes:
> So how do we get the Sym and the Sun to interact?? There are a couple of
> possibilities as we see it, but we are naive, and would appreciate more suggestions.
> 1) Use some sort of RPC call. Any ideas on how to do this would be helpful.
> 2) Use some sort of daemon on the Sun with the daemon checking for the presence of
> some file with the program to run and data descriptors. These seems similar
> to the way the present line printer spooler works.
> 3) use standard unix spooler technology somehow.
We have been using home-grown software which, among other things, does
what you need. It is probably not the most efficient approach, but it
does work, with minimal effort. It is called "Common Memory", and
mimics a shared memory scheme. An application declares a common memory
variable as a writer and/or a reader. Thus, the Symbolics can declare
a variable as a writer, and a Sun process can declare the same
variable as a reader (inside some process running on the Sun). The way
we use it, sometimes one variable has several writers, and often has
many readers (all residing on different computers). Usually, each
process has an "in box" and an "out box", i.e. two variables - one
declared as a writer, one as a reader.
The software comes in two pieces - a server and a client. The server
is called the Common Memory Manager, and is written in C for a Sun.
The client software comes in a C version for a Sun, Lucid Common Lisp,
Franz Lisp and Franz Allegro Common Lisp for a Sun, and Symbolics
Common Lisp for the Symbolics. The software is provided without any
claims as to reliability, support, maintenance, etc, etc, but we have
had much success with it. You can get the Sun code via anonymous ftp
from durer.cme.nist.gov in directory /pub/cm. Should you choose this
route, send me mail and I will mail you the Symbolics client code.
P.S. A paper describing Common Memory appears in the Usenix Winter
'85 conference, "User-Level Shared Variables" by Don Libes.
Steven R. Ray
National Institute of Standards and Technology
Building 220, Room A127
Gaithersburg, MD 20899
- [no subject]
- From: Steve Willis <bruno%oz.geology.washington.edu@RELAY.CS.NET>