In data-intensive real-time applications, e.g., transportation management and location-based services, the amount of sensor data is exploding. In these applications, it is desirable to extract value-added information,...
详细信息
ISBN:
(纸本)9781479989386
In data-intensive real-time applications, e.g., transportation management and location-based services, the amount of sensor data is exploding. In these applications, it is desirable to extract value-added information, e.g., fast driving routes, from sensor data streams in real-time rather than overloading users with massive raw data. However, achieving the objective is challenging due to the data volume and complex data analysis tasks with stringent timing constraints. Most existing big data management systems, e.g., Hadoop, are not directly applicable to real-time sensor data analytics, since they are timing agnostic and focus on batch processing of previously stored data that are potentially outdated and subject to I/O overheads. To address the problem, we design a new real-time big data management framework, which supports a non-preemptive periodic task model for continuous in-memory sensor data analysis and a schedulability test based on the edf (Earliest Deadline First) algorithm to derive information from current sensor data in real-time by extending the map-reduce model originated in functional programming. As a proof-of-concept case study, a prototype system is implemented. In the performance evaluation, it is empirically shown that all deadlines can be met for the tested sensor data analysis benchmarks.
The problem of hybrid scheduling of hard period task and soft aperiodic task is important part of the research of real-time scheduling. To analysis the idle and movable time of period task set scheduled by earliest de...
详细信息
The problem of hybrid scheduling of hard period task and soft aperiodic task is important part of the research of real-time scheduling. To analysis the idle and movable time of period task set scheduled by earliest deadline first scheduling algorithm, two definitions scheduling and converse scheduling are given. These two definitions are used to calculate the max movable time of period task set. Using the max movable time, an algorithm named idle stealing algorithm (ISA) is given to decrease the response time of aperiodic task. ISA makes full use of the max movable time of period task set. It can largely decrease the response time of aperiodic tasks while guaranteeing the deadline of period tasks. The ISA algorithm provides the shortest response time of aperiodic task and is proved to be optimal by experiments.
In real-time task scheduling, response time and time complexity are two of important requirements that draw many attentions. Virtual release advancing (VRA) [Tanaka, K. (2015, June). Virtual release advancing for earl...
详细信息
In real-time task scheduling, response time and time complexity are two of important requirements that draw many attentions. Virtual release advancing (VRA) [Tanaka, K. (2015, June). Virtual release advancing for earlier deadlines. ACM SIGBED Review, 12(3), 28-31] is an effective technique for shorter response times in the Earliest Deadline First scheduling [Liu, C. L., & Layland, J. W. (1973). Scheduling algorithms for multiprogramming in a hard-real-time environment. Journal of the Association for Computing Machinery, 20(1), 46-61], but not adaptive to precise systems due to its high time complexity. In order to mitigate the time complexity, a new technique, called enhanced VRA, is presented in this paper. Applied to the Total Bandwidth Server context, the enhanced technique significantly improves the time complexity while guaranteeing the responsiveness and schedulability. The technique is implemented on an ITRON real-time operating system running on an ARM Cortex-A9 processing core with a field programmable gate array. With supporting of accelerator hardware, the new algorithm shows that the maximum additional runtime overhead per tick is reduced by up to 30% compared with that of the original (software) VRA.
暂无评论