the majority of battery operated computing devices have the capability to employ Dynamic Voltage Scaling (DVS) as an effective method to minimize the power used by their processors. Algorithms employing DVS in real-ti...
详细信息
ISBN:
(纸本)9780889866386
the majority of battery operated computing devices have the capability to employ Dynamic Voltage Scaling (DVS) as an effective method to minimize the power used by their processors. Algorithms employing DVS in real-time systems must guarantee the satisfaction of all tasks' timing constraints while minimizing power consumption. In this paper, we present a framework for scheduling and dispatching a distributed set of hard real-time tasks with relative timing constraints. In our approach, the scheduler employs the parametric real-time scheduling methodology to provide on-time completion guarantees for periodic hard real-time tasks. Moreover, the scheduler associates with each hard real-time task a feasibility range parameterized by timing parameters generated at run-time. the run-time dispatcher uses this information to calculate the system's dynamic slack time. It uses the slack time to set the CPU speed using DVS in the time interval till the next scheduling point. the dispatcher reclaims the additional slack generated when a job uses fewer execution cycles than its worst-case estimate to further reduce power consumption. the proposed power-aware dispatching method has running-time of O(1) with no need for off-line computations. Our simulation results show significant improvement in power savings over previously presented DVS-based scheduling methods.
the correctness of applications that perform asynchronous message passing typically relies on the underlying hardware having a sufficient amount of memory (message buffers) to hold all undelivered messages-such applic...
详细信息
ISBN:
(纸本)9780889866386
the correctness of applications that perform asynchronous message passing typically relies on the underlying hardware having a sufficient amount of memory (message buffers) to hold all undelivered messages-such applications may deadlock when executed on a system with an insufficient number of message buffers. thus, determining the minimum number of buffers that an application needs to prevent deadlock is an important task when writing or porting parallel applications. Unfortunately, boththis problem (called the Buffer Allocation Problem) and the simpler problem of determining whether an application may deadlock for a given number of available message buffers are intractable [1]. We present a new epoch-based polynomial-time approach for approximating the Buffer Allocation Problem. Our approach partitions application executions into epochs and intersperses barrier synchronizations between them, thus limiting the number of message buffers necessary to ensure deadlock-freedom. this approach produces near optimal solutions for many common cases and can be adapted to guide application modifications that ensure deadlock freedom when the application is ported. Lastly, we describe a space-time trade-off between the number of available message buffers and the number of barrier synchronizations, and describe how this trade-off can be used to fine-tune application performance.
We present our research findings on the practicality of using RFID for building context-aware homes in India. Our research is the result of an endeavor to bring this global trend of smart homes to India. Rather than s...
详细信息
ISBN:
(纸本)9780889866386
We present our research findings on the practicality of using RFID for building context-aware homes in India. Our research is the result of an endeavor to bring this global trend of smart homes to India. Rather than studying the effects of such an influence in its original form-building smart homes equipped with such high-end sensors as real-time, audio-video sensing, and sophisticated, environmental sensors as are popular in a typical smart home-we studied the viability of realizing ambient-intelligence using simple, low-cost, and robust RFID sensors, which we believe would be suitable to Indian dwellings in the near future. We describe the learning and prediction algorithm used in our study;we modeled time as fuzzy-sets and used Markov model for predicting the future locations of inhabitants wearing RFID bracelets. Our system provides context-aware reactive and proactive services. We modeled the uncertainties-involved in inferring users' activities from their location information-using Bayesian Belief Networks or BBNs. this improved the accuracy of prediction considerably. Our experience showed that it is possible to develop cost-effective, easily deployable, context-aware homes using only RFID in India in the short-term before high-end sensing technologies mature and become commercially affordable in the longer run.
the talk will address some of the key algorithmic and computational challenges associated withthe modelling of certain classes of biological networks (e.g. biochemical, signalling). this will involve several research...
详细信息
the way of collecting sensor data faces a revolution when the newly developing technology of distributed sensor networks becomes fully functional and widely available. Smart sensors acquire full interconnection capabi...
详细信息
the way of collecting sensor data faces a revolution when the newly developing technology of distributed sensor networks becomes fully functional and widely available. Smart sensors acquire full interconnection capabilities with similar devices, so that run-time data aggregation, parallelcomputing, and distributed hypothesis formation become reality with off-the-shelf components and sensor boards. this revolution started around in 1996, and now hardware and network are converging on the first convincing solutions. Exploring and exploiting this paradigm are a renovated challenge for the pattern recognition and data mining community. this paper attempts a survey on state-of-the-art of wireless sensor technology, with an eye on data-related problems and technological limits. Although the possibilities seem promising, the limited computational resources of individual nodes hamper the elaboration of data with computationally-intensive algorithms. New software paradigms must be developed, both creating new techniques or adapting, for network computing old algorithms of earlier ages of computing
暂无评论