Back to the main page

Mailing List Logs for ShadowRN

From: Todd Montgomery <tmont@****.WVU.EDU>
Subject: Re: CPU Controllers
Date: Mon, 9 Aug 1993 09:56:10 -0400
> From: Robert Watkins <bob@**.NTU.EDU.AU>

> A decker with a high Hacking pool, and a brilliant deck, should be able to
do
> stuff really well. Most systems are open to a good decker. My post reflected
> this.

A decker with a high pool and a great deck already does stuff really well!

> I was trying to emulate something that would be the case in RL, and is ruled
> out by the SR rules. That is: the CPU controls EVERYTHING. If the system can
do
> something, then the CPU can ask it to do it. In some cases a SPU will be
> sitting in the way, but all this means is that you've got another place ot
make
> the request from.

Is it really correct to asumme that Massively Parallelled machines from
2054 have a controlling CPU? Difficult call. Yes, even though their is
a CPU node that does not mean that the system would have an all controlling
node. I can think of several reasons why it would not:
1: The cost in hardwiring all this would be enormous. And
unnecessary to allow the system to perform its actions.
2: Security! This would be one reason why I would not want to
design a system that would have a central point of
control. This is never a good tactic. ANd with
systems needing the security they have (IC and such)
why make it easier for an intruder by making everything
centralized?
3: What are the SPUs used for then. Why not beef up the CPU
to handle everything? Less expensive that paying
for mutiple SPUs to handle the traffic. Good example,
check the price of the CMs as they go up in the number of
processors.

Another flaw I found in the CPU controller was the assumption that a
sysop had such control from the outside. In my 6 years as sysop for
various boards I never once did not include this trick. Upon login,
check modem, if hot, then flag login as remote. If login is sysop
access level, then check login flag, if remote, disallow login. With this
and a blind for priviledged users it was impossible for an intruder to
login as sysop without me being able to stop it before they did any harm.
But that is neither here, nor there. I really don't think a sysop of
SR would allow their system to be so insecure as to make ALL actions
of the entire system, central to one point. I think that is a very important
thing to consider when designing a system, its security, don't you?

Now about the frame, Nightstalker, I do think that is a good idea.

-- Quiktek
-- Todd Montgomery
tmont@****.wvu.edu
tmont@***.wvu.edu
un032507@*******.wvnet.edu

Disclaimer

These messages were posted a long time ago on a mailing list far, far away. The copyright to their contents probably lies with the original authors of the individual messages, but since they were published in an electronic forum that anyone could subscribe to, and the logs were available to subscribers and most likely non-subscribers as well, it's felt that re-publishing them here is a kind of public service.