Abstract
In many real-time processing systems for the Internet of Things (IoT), the correctness of real-time data objects that model physical world entities, such as the status of mobile robotics, depends not only on the functional correctness, but also on the temporal consistency. Maintaining temporal consistency of real-time data while reducing energy cost is of critical importance when designing such IoT systems. In this paper, we formulate the energy-aware real-time data processing problem on multicore platforms and prove it to be NP-hard. In view of the intractability of the problem, we adopt a divide-and-conquer strategy. We first propose a per-CPU solution, which can result in significant power savings. Next, in order to save energy in a fine-grained granularity, we propose an efficient per-Task solution by adopting the per-CPU solution as a building block. Finally, by developing new energy-aware mapping techniques, we further explore energy savings on multicore platforms. Extensive simulation results show that the proposed methods offer remarkable performance improvement in terms of energy savings, as compared to the state-of-the-art schemes.
Highlights
The Internet of Things has received great attention in recent years, examples of which include traffic monitoring [1], flight control [2], health care devices [3], and industrial automation [4]
Zhou et al.: Energy-Aware Real-Time Data Processing for Internet of Things (IoT) Systems control technique is Dynamic Voltage and Frequency Scaling (DVFS) [7], which is especially suitable for certain data-intensive applications where the workload of the systems would be comparatively high, e.g., during the rush hours in transportation management
We introduce the architecture of processing prototype scheduler, which can handle both user transactions and update transactions, while supporting real-time validity requirements of data processing in IoT systems
Summary
The Internet of Things has received great attention in recent years, examples of which include traffic monitoring [1], flight control [2], health care devices [3], and industrial automation [4]. Control technique is Dynamic Voltage and Frequency Scaling (DVFS) [7], which is especially suitable for certain data-intensive applications where the workload of the systems would be comparatively high, e.g., during the rush hours in transportation management In such scenarios, compared with energy-aware strategies by shutting down cores, a neveridle method that enables to scale the running speed in the system, such as DVFS, is more desirable. The proposed technique differs from the existing ones in three aspects It provides an efficient off-line schedulability analysis for dynamic IoT systems based on wireless sensors which has a pseudo-polynomial time complexity. We propose a per-CPU DVFS method called ML-CS, which satisfies temporal validity constraint of real-time data, to achieve energy savings on unicore platform.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have