TY - JOUR
T1 - A task-efficient sink node based on embedded multi-core SoC for Internet of Things
AU - Qiu, Tie
AU - Zhao, Aoyang
AU - Ma, Ruixin
AU - Chang, Victor
AU - Liu, Fangbing
AU - Fu, Zhangjie
PY - 2018/5/31
Y1 - 2018/5/31
N2 - With the increase of collected information, the computing performance of single-core sink node for Internet of Things (IoTs) cannot satisfy with the demand of large data processing any more. Therefore, the sink node which based on embedded multi-core SoC for IoTs and maximizing its computing performance has brought into focus in recent years. In this paper, we design a multi-core Task-Efficient Sink Node (TESN) based on heterogeneous architecture and the Weighted-Least Connection (WLC) task schedule strategy has been proposed to improve its efficiency. There are two types of cores in the sink node, master core and slave cores. The master core deals with tasks allocation and the seven slave cores deal with data processing. All of the cores are communicating with each other through mailbox. By considering each core’s real-time processing information and computing performance, the proposed WLC can balance each core’s load and reduce network congestion. The platform is used to evaluate the performance of WLC and Round-Robin (RR) algorithms for multi-core sink node. The experiment results show that the WLC strategy improves the processing speed obviously, achieves load balance and avoids large scale congestion of sink node in the sensor networks of IoTs.
AB - With the increase of collected information, the computing performance of single-core sink node for Internet of Things (IoTs) cannot satisfy with the demand of large data processing any more. Therefore, the sink node which based on embedded multi-core SoC for IoTs and maximizing its computing performance has brought into focus in recent years. In this paper, we design a multi-core Task-Efficient Sink Node (TESN) based on heterogeneous architecture and the Weighted-Least Connection (WLC) task schedule strategy has been proposed to improve its efficiency. There are two types of cores in the sink node, master core and slave cores. The master core deals with tasks allocation and the seven slave cores deal with data processing. All of the cores are communicating with each other through mailbox. By considering each core’s real-time processing information and computing performance, the proposed WLC can balance each core’s load and reduce network congestion. The platform is used to evaluate the performance of WLC and Round-Robin (RR) algorithms for multi-core sink node. The experiment results show that the WLC strategy improves the processing speed obviously, achieves load balance and avoids large scale congestion of sink node in the sensor networks of IoTs.
UR - http://www.scopus.com/inward/record.url?eid=2-s2.0-85009291777&partnerID=MN8TOARS
U2 - 10.1016/j.future.2016.12.024
DO - 10.1016/j.future.2016.12.024
M3 - Article
SN - 0167-739X
VL - 82
SP - 656
EP - 666
JO - Future Generation Computer Systems
JF - Future Generation Computer Systems
ER -