Plan 9: The future

[Vijay Gill vijay@gl.umbc.edu]

First of all: A minor gripe. I made the mistake of installing sam and
giving the plan 9 papers (all of them) to the programmers at the place
where I work.  After a month, I was approached by the manager of our
programmers asking if I could install plan 9 on the machines as he was
getting tired of hearing the constant demands for more plan 9 stuff from
them.  Apparently the set up that I am in charge of (3 6 processor
AViiONS, with 1 gig of ram, 36 gig Clariion disk arrays, optical juke
boxes etc with a lot of the smaller Aviions on the programmers desks + the
odd DEC 3000 Alpha box) maps well into the plan 9 model and it turns out
that the programmers seem to agree with what the designers of plan 9 had
to say.  They use sam exclusively now and want more of the same.  I was 
sitting in a meeting for 2 hours today trying to convince them that as 
yet, they could not get plan 9 and trying to calm down 8 irate 
programmers dosed on caffine is not a task for the faint of heart.  A 
heartfelt request from me.  Please make plan 9 available.  Please.  I 
don't do much programming, my forte is adminstration, but plan 9 seems to 
have hit our programmers where it counts and they are starting to make 
it a bit hot for me.  Our manager is willing to pay for plan 9 and pay 
well.  The crowning touch came when I showed them the letter Bob 
Kummerfield wrote about the time he attended INET93 and used his Compaq 
LTE in San Francisco to get the same exact environment (including 
CPU/file servers) there as he does in Australia.  Our people would pay 
kill for that.  Without saying more, I would venture that we pay AT&T 
well in the 6 figures per year for leased lines for our product installed 
in New Jersey ;)

Ok, so it was not a minor gripe.

On to more, bigger better things.  I was reading over Andy Tannenbaum's
papers on Amoeba, some papers on Chorus MiiX, and QNX (my other favourite
operating system) and I was struck by how easy they make it to have multi
computers with SSS (Single Site Semantics).  A group of non shared memory
cpu modules on a backplane acting as one computer.  This would grow to
include hundreds of CPU's, whereas the shared memory computers start to
run out of bandwidth after a few dozen or so.  Is there support for plan 9
for something like this? 

And finally: What next for plan 9?  What do the designers of plan 9 
envison next? 


--
Vijay Gill                         |The (paying) customer is always right.
wrath@cs.umbc.edu                  |                  - Piercarlo Grandi
vijay@gl.umbc.edu                  | Eagles may soar, but weasels don't get
These are my opinions only.        | sucked into jet engines.