Blockchain-Powered Service Migration for Uncertainty-Aware Workflows in Edge Computing
Journal article, Peer reviewed
Accepted version
Åpne
Permanent lenke
http://hdl.handle.net/11250/2636537Utgivelsesdato
2019Metadata
Vis full innførselSamlinger
Originalversjon
Communications in Computer and Information Science. 2019, 1123 217-230. 10.1007/978-981-15-1304-6_18Sammendrag
In edge computing, the workflow is used to simulate and manage computing tasks as well as information exchange for compute-intensive and data-intensive application, which is convenient for the various complex process to work orderly. However, the resource conflict among cooperative works of multiple mobile edge computing (MEC) nodes by workflow, together with the service failure and the performance degradation, bring about additional uncertainties of scheduling strategies. Consequently, such uncertainties delay the completion of tasks and spoil the user experience. To deal with that issue, we propose a blockchain-powered resource provisioning (BPRP) method to design policies for workflows in the edge computing environment. Technically, we use the directed acyclic graph to indicate workflows of each edge node and regard its scheduling strategy as an individual gene to adapt to the following algorithm. Then, we use the non-dominated sorting genetic algorithm-III (NSGA-III) to optimize the workflow scheduling strategies on the basis of tasks’ timely completion with good quality. A large number of experiments were carried out to verify the effectiveness of our method.