Conference Paper

Delay Analysis in Temperature-Constrained Hard Real-Time Systems with General Task Arrivals.

Michigan Univ., Dearborn, MI
DOI: 10.1109/RTSS.2006.16 Conference: Proceedings of the 27th IEEE Real-Time Systems Symposium (RTSS 2006), 5-8 December 2006, Rio de Janeiro, Brazil
Source: DBLP

ABSTRACT In this paper, we study temperature-constrained hard real- time systems, where real-time guarantees must be met with- out exceeding safe temperature levels within the proces- sor. Dynamic speed scaling is one of the major techniques to manage power so as to maintain safe temperature lev- els. As example, we adopt a simple reactive speed con- trol technique in our work. We design a methodology to perform delay analysis for general task arrivals under re- active speed control with First-In-First-Out (FIFO) sche- duling and Static-Priority (SP) scheduling. As a special case, we obtain a close-form delay formula for the leaky- bucket task arrival model. Our data show how simple reac- tive speed control can decrease the delay of tasks compared with any constant-speed scheme.

1 Bookmark
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we investigate a method of scheduling aperiodic tasks with a given periodic schedule, in a thermally constrained embedded real-time system. Considering thermal constraints is important for current and future processing systems because of the rapid rise in the power densities of processors, especially in multicore environments. Therefore, if thermal constraints are not considered, processor temperatures may increase to levels at which the reliability of the underlying semiconductor devices is compromised. As part of the proposed approach, an accurate and non-extensive temperature estimation scheme is presented. Based on the temperature estimation scheme, a scheduling scheme for aperiodic tasks in the presence of periodic tasks is proposed. The primary advantage of the proposed scheme is that it can account for temporal variations in task power consumption. The results show that our approach can provide significant improvements in aperiodic task response times, over schemes that do not allow tasks to have temporal variations in power consumption.
    VLSI Design and 2013 12th International Conference on Embedded Systems (VLSID), 2013 26th International Conference on; 01/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: We consider a real-time system running periodic tasks with probabilistic execution times. In this system, temperature behavior of the processor is affected by the jobs characteristics and scheduling algorithm. In turn, the processor temperature affects its reliability and power consumption. Moreover, the maximum speed by which the system can operate correctly and still avoid temperature runaway is dependent on its stochastic thermal behavior. Therefore, performance, power consumption, and reliability of this system are unavoidably dependent on its thermal behavior. In this paper, we present an analytical method to extract a fundamental temperature measure, namely temperature probability density function, to represent thermal specification of the system. Also, we show that how this specification could be used by the designer to perform temperature-aware analysis of reliability and power consumption of individual tasks.
    Proceedings of the 21st International conference on Real-Time Networks and Systems; 10/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Real-time scheduling in systems with energy or power constraints is challenging. Especially when a mixture of real-time and best effort tasks exist, it is difficult to guarantee that all deadlines are met and at the same time that the system does not run out of energy. This is the case for industrial instrumentation for hazardous areas, such as explosive atmospheres. A frequently used method of protection against explosion is intrinsic safety. That means, the power supply as well as the energy that is stored in the device is kept below a critical threshold. As a result, energy is a much scarcer resource than processing time in this class of systems. Therefore, it is appropriate to base the scheduling decision on the available and the consumed energy instead of the processing time. In this work, we adapt the Earliest-Eligible-Virtual-Deadline-First algorithm (EEVDF) for energy-driven scheduling using dynamic power management. The resulting system is hard real-time capable, takes the energy consumption of peripherals and sensors into account and utilizes slack energy efficiently and predictably. Since the scheduler guarantees the availability of sufficient energy for real-time tasks, the design of the system is significantly simplified.
    Real-Time and Embedded Technology and Applications Symposium (RTAS), 2013 IEEE 19th; 01/2013


Available from