🚀 go-pugleaf

flaps@dgp.toronto.edu (Alan J Rosenthal) writes:

>ho@cs.arizona.edu (Hilarie Orman) writes:
>>I've always relied on the notion of information flow - if the writer
>>of the transmitter program did not specify the information as an
>>output, and if a receiver program can detect the information, then you
>>have a covert channel,

>But the usual covert channel demonstration involves having a transmitter
>program which was written just for the purpose of transmitting the data.
>This is surely output, even if it's not output to a file.

The definition in "Computer Security Basics" [published by O'Reilly & 
Associates] is a good one:

	"A communication channel that allows a process to transfer information
	in a way that violates a system's security policy."

I think the key points here are "communication channel" and "violates
security policy."  Some process has to communicate (eg write to disk, send
over phone line, etc.) data to another process, and that communication 
has to violate the security poilicy of the system (eg not to write out
classified data to processes that are not trusted to read it).

It may be a process that intentionally writes data for another process to 
read, or one that does it inadvertantly.  Similarly, the reader can read the
data intentionally or not, it doesn't matter.

hope that helps,
--f