Facilitating Dynamic RT-QoS for Massive-Scale Autonomous Cyber-Physical Systems

David W. McKEE  Xue OUYANG  Jie XU  

Publication
IEICE TRANSACTIONS on Communications   Vol.E101-B   No.8   pp.1760-1767
Publication Date: 2018/08/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.2017ADI0001
Type of Manuscript: INVITED PAPER (Special Section on Autonomous Decentralized Systems Technologies and Approaches Innovation through Structure Change of Society and Life)
Category: 
Keyword: 
cloud,  SOA,  services,  autonomous systems,  real-time,  straggler,  IoT,  IoS,  simulation,  

Full Text: FreePDF(1.7MB)


Summary: 
With the evolution of autonomous distributed systems such as smart cities, autonomous vehicles, smart control and scheduling systems there is an increased need for approaches to manage the execution of services to deliver real-time performance. As Cloud-hosted services are increasingly used to provide intelligence and analytic functionality to Internet of Things (IoT) systems, Quality of Service (QoS) techniques must be used to guarantee the timely service delivery. This paper reviews state-of-the-art QoS and Cloud techniques for real-time service delivery and data analysis. A review of straggler mitigation and a classification of real-time QoS techniques is provided. Then a mathematical framework is presented capturing the relationship between the host execution environment and the executing service allowing the response-times to predicted throughout execution. The framework is shown experimentally to reduce the number of QoS violations by 21% and provides alerts during the first 14ms provide alerts for 94% of future violations.