Subject: Re: SDL-News: Broadcast
From: John Dalton (j.dalton#virgin.net)
Date: Mon Aug 03 1998 - 15:44:44 GMT
The originator of this message is responsible for its content.
-----From John Dalton <j.dalton#virgin.net> to sdlnews -----
tariq#erg.abdn.ac.uk wrote:
> For example the protocols we are develpoing in Aberdeen require a message to be
> transmitted to all the process and a process makes a decision either message is
> addressed to it or not. I belive it is very similar to a packet which transmitted
> over the internet and each reciver makes a decision wheather the packet is addressed
> to it or not. If packet is addressed to it, the reciver accept a packet and if not
> it ignores it.
This is not how Internet works! If it did it would be a logjam. Routers decide which
ports (channels) to send messages out of depending on their address headers. Only a few
trial messages get multicast to find out which channel is fastest/closest, and then only
along the list of available ports. Stuff on your local network is broadcast to all
receivers on the network, but it doesn't all go out of the internet gateway!
The LAN activity is really layer 1 stuff, a characteristic of the
transmission medium implemented in hardware. It's normally really hard to describe
physical layer stuff in SDL. Things like collision avoidance, time-delay propagation
protocols and so on are impossible in SDL.
I agree broadcast would be convenient at times, but to be practical it would also have
to be limited as well. How else could you terminate the broadcast? The echos could go
on forever......
________________________________<http://www.reflex-tech.co.uk>
>From John Dalton of Reflex Technology, High Wycombe UK 01494-465907
SDL tools, real-time OS tools and debugging hardware (01308-863795 direct)
-----End text from John Dalton <j.dalton#virgin.net> to sdlnews -----
For help, email "majordomo#sdl-forum.org" with the body of your email as:
help
or (iff this does not answer your question) email: owner-sdlnews#sdl-forum.org
This archive was generated by hypermail 2a23 : Sun Jun 16 2013 - 10:41:40 GMT