A Security layer in an OS is only as good as the testing that is done with this system, so trying to implement a security system specialized for the embedded field would be counter productive since it would be deployed on a small number of systems only. Relying on a widely deployed system like Linux, ensures a high bug detection rate, and a minimum time for bug-fixes to be available. It might not always be the most elegant solution that is safe, as an example here is how to talk to an RT-FIFO remotely via an ssh-based "tunnel".
tigger: > ssh -l USER HOST \
cat rtf0 >local_data_file
tigger: > echo "start" | ssh -l \
USER HOST "cat > rtf1"
A few points to keep in mind when designing the embedded systems app with respect to security are in the following list, it is hard to give any strict rules and this list is no more than a general ''checklist'' for security
MiniRTL requires some RTLinux related security considerations, notably the requirements that users operating these systems need root privileges at least in an indirect way. This requires some consideration for application design.
Designing logging on an embedded system must be done very carefully since there are storage limitations in most cases, a solution to this problem can be off-site logging. Also writing large log-files is something that calms down administrators but it is rather senseless to write 100MB of log-files if there is no way of detecting faults and attacks with a reasonable time-budget, so the key to all logging activity is putting meaningful log messages into the apps and give proper consideration to the priority of messages.