Mixed criticality systems offer the opportunity to integrate system components with different levels of assurance that previously may have been placed on different nodes. While the vast majority of mixed criticality work features a HI and a LO criticality level, LO criticality tasks should not be mistaken for tasks with little value. Such tasks might contain mission critical functionality and are still vital for the correct and efficient operation of the system. A large portion of earlier work immediately suspends all LO criticality functionality upon a criticality change. It is clear that suspending tasks at all is highly undesirable, let alone all tasks of a single criticality level at the same time. In this work we consider this issue, we propose a scheme to maintain the operation of lower criticality tasks for as long as possible, even when the system is executing in a higher mode. We introduce the notion of importance as a means of deciding which tasks are suspended first. This is done with the aim of allowing the system designer to make these decisions and have greater control over the way their system degrades. We conclude that by using essentially the same analysis and functionality that facilitates multiple criticality levels we are able to provide improved lower criticality performance.

BibTex Entry

@inproceedings{Fleming2014,
 author = {T. Fleming and A. Burns},
 booktitle = {Proc. 2nd Workshop on Mixed Criticality Systems (WMC), RTSS},
 editor = {L. Cucu-Grosjean and R. Davis},
 pages = {33-38},
 title = {Incorporating The Notion of Importance into Mixed Criticality Systems},
 year = {2014}
}