Re: About interactivity.
- From: Elizabeth D Rather <eratherXXX@xxxxxxxxx>
- Date: Wed, 12 Sep 2007 17:05:09 -1000
Cesar Rabak wrote:
kenney@xxxxxxxxxxxxxxxxxxx escreveu:In article <46E73576.8050406@xxxxxxxxxxxx>, csrabak@xxxxxxxxxxxx (Cesar Rabak) wrote:Philosophically, *all* is a matter of design not <whatever> until you find the complexity of doing the design with some form of <whatever> makes damn hard to do implement the design.
Try to keep a backup of Telephone switch and transfer the calls to another place to "avoid single point of failure"¹
Well that was done with Strowger and other mechanical exchanges. It is a matter of design not language.
Back to my example: How in the Earth did you make a change in the equivalent to changing the engineering (firmware) of a telephone switch without interrupting the connections (calls) going on?
Or rather a better example: how did you with Strowger or Cross{point,bar} switches did you swapped exchanges without dropping the connections they were serving?
20 years ago the design of the control system for the King Khaled International Airport provided for full backup for every component, including the provision for hot switchover of all systems without information loss. This included over 30 conceptual subsystems (power management, security system, fuel farm, water wells, runway lights, ...) involving a network of over 500 computers spread over 100 square miles controlling and monitoring over 30,000 points (digital in, digital out, analog in, analog out, communications lines, etc.). Each subsystem had its own issues, but virtually all were critical in some way to safe continuous operation of the airport. The system was completely programmed in Forth, and has been in successful operation ever since (although the computers themselves were hopelessly obsolete many years ago... last I heard the airport was cannibalizing some machines to keep others running because they couldn't figure out how to replace the whole thing).
Basically, it was done with dual shared memories to store the state of each node. Both copies were updated by every event, so that one could switch over to the backup node seamlessly. A complete software update was achieved by downloading the new program into the backup node and then requesting a switchover.
To be sure, that involved hardware in the design, but if you have a really critical system that cannot be allowed to go down you need to do that anyway.
Cheers,
Elizabeth
--
==================================================
Elizabeth D. Rather (US & Canada) 800-55-FORTH
FORTH Inc. +1 310-491-3356
5155 W. Rosecrans Ave. #1018 Fax: +1 310-978-9454
Hawthorne, CA 90250
http://www.forth.com
"Forth-based products and Services for real-time
applications since 1973."
==================================================
.
- Follow-Ups:
- Re: About interactivity.
- From: Cesar Rabak
- Re: About interactivity.
- From: brennie
- Re: About interactivity.
- References:
- Re: About interactivity.
- From: Cesar Rabak
- Re: About interactivity.
- From: kenney
- Re: About interactivity.
- From: Cesar Rabak
- Re: About interactivity.
- Prev by Date: Re: About interactivity.
- Next by Date: Re: Intercepting the text interpreter.
- Previous by thread: Re: About interactivity.
- Next by thread: Re: About interactivity.
- Index(es):
Relevant Pages
|