Departmental Papers (CIS)

Document Type

Conference Paper

Date of this Version

12-2009

Comments

Copyright 2009 IEEE. Reprinted from:

Linh T.X. Phan , Samarjit Chakraborty , Insup Lee. Timing Analysis of Mixed Time/Event-Triggered Multi-Mode Systems. Proceedings of the 30th IEEE Real-Time Systems Symposium (RTSS 2009), Dec. 1-4, 2009. Washington, D.C., USA

This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Pennsylvania's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to pubs-permissions@ieee.org. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.

Abstract

Many embedded systems operate in multiple modes, where mode switches can be both time- as well as event-triggered. While timing and schedulability analysis of the system when it is operating in a single mode has been well studied, it is always difficult to piece together the results from different modes in order to deduce the timing properties of a multi-mode system. As a result, often certain restrictive assumptions are made, e.g., restricting the time instants at which mode changes might occur. The problem becomes more complex when both time- and event-triggered mode changes are allowed. Further, for complex systems that cannot be described by traditional periodic/sporadic event models (i.e., where event streams are more complex/bursty) modeling multiple modes is largely an open problem. In this paper we propose a model and associated analysis techniques to describe embedded systems that process multiple bursty/complex event/data streams and in which mode changes are both timeand event-triggered. Compared to previous studies, our model is very general and can capture a wide variety of real-life systems. Our analysis techniques can be used to determine different performance metrics, such as the maximum fill-levels of different buffers and the delays suffered by the streams being processed by the system. The main novelty in our analysis lies in how we piece together results from the different modes in order to obtain performance metrics for the full system. Towards this, we propose both – exact, but computationally expensive, as well as safe approximation techniques. The utility of our model and analysis has been illustrated using a detailed smart-phone case study.

Share

COinS
 

Date Posted: 15 December 2009

This document has been peer reviewed.