Subject: Re: SDL-News: Modelling of unreliable transmission medium in Tau SDT
From: Allan Weber (awmi00#control.auc.dk)
Date: Thu Jan 01 1998 - 05:13:07 GMT
Become an SDL Forum Society member <http://www.sdl-forum.org/Society/members.htm>
The originator of this message is responsible for its content.
-----From Allan Weber <awmi00#control.auc.dk> to sdlnews -----
We are running in to serious trouble with our model right now.
We are still trying to model an unreliable wireless medium (meaning
possibilty for loss of connection between two communicating nodes).
Right now we have an input from the environment into our medium, that
may trigger a "spontaneous" connectionloss. Furthermore, on every input
signal (which is data being sent through the medium) we use the "any"
descission to model whether the input should be lost or transmitted
successfully. When either an input from the environment or an "any"
descission causes a lost connection, we inform out to the other blocks
(the protocol layers above the medium on both nodes) in the system of
this situation.
The above layers have internal states, which is dependant of the state
of the wireless medium. It means, that if the medium becomes
disconnected, the state of the protocol layer above should also change
state at the same time. Unfortunatly, the other blocks (above lying
protocol layers) become highly unsyncronized, since the signals
informing them of connectionloss in the media gets heavily delayed,
because it uses the same input queue as "ordinary messages", and these
queues might contain several, signals before the "Connection lost"
signal. This result in that the "signal lost" signal is not received by
the above layers immeidately after a connection loss, which will be the
case in "real life" - and thats our main problem. We have given all
signals the same priority, as this models "the reality" the best.
We think a solution to our problem could be to use a global variable
telling each block whether or not there is a connection in out physical
medium. However, it is our impression that global variables don't
excists in Tau SDT. Is this correct?
Any other ideas on how to make a realistic model of an unreliable
wireless medium, where the connection can be lost both spontaneuosly (by
using input from environment), and when a signal i sent (by using any)?
And finally, is there any way to give "control signals" higher priority
that "ordinary messages", other than using prioritised inputs, which we
already use to model "ordinary messages" and "urgent ordinary messages"?
Is it possible to define an extra input queue for control signals etc?
Best regards
Allan Weber
--End text from Allan Weber <awmi00#control.auc.dk> to sdlnews ---
For extra SDL Forum Society benefits join at <http://www.sdl-forum.org/Society/members.htm>
This archive was generated by hypermail 2a23 : Thu May 09 2013 - 16:05:50 GMT