SDN-enabled resource orchestration for industrial IoT in collaborative edge-cloud networks
|Author:||Okwuibe, Jude1; Haavisto, Juuso2; Kovacevic, Ivana1;|
1Centre for Wireless Communication, University of Oulu, 90570 Oulu, Finland
2Center for Ubiquitous Computing, University of Oulu, 90570 Oulu, Finland
3VTT Technical Research Center of Finland, 02044 Espoo, Finland
|Online Access:||PDF Full Text (PDF, 3.4 MB)|
|Persistent link:|| http://urn.fi/urn:nbn:fi-fe2021101150543
Institute of Electrical and Electronics Engineers,
|Publish Date:|| 2021-10-11
Effective, long-lasting Industrial IoT (IIoT) solutions start with short-term gains and progressively mature with added capabilities and value. The heterogeneous nature of IIoT devices and services suggests frequent changes in resource requirements for different services, applications, and use cases. With such unpredictability, resource orchestration can be quite complicated even in basic use cases and almost impossible to handle in some extensively dynamic use cases. In this paper, we propose SDRM; an SDN-enabled Resource Management scheme. This novel orchestration methodology automatically computes the optimal resource allocation for different IIoT network models and dynamically adjust assigned resources based on predefined constraints to ensure Service Level Agreement (SLA). The proposed approach models resource allocation as a Constraint Satisfaction Problem (CSP) where optimality is based on the solution of a predefined Satisfiability (SAT) problem. This model supports centralized management of all resources using a software defined approach. Such resources include memory, power, bandwidth, and edge-cloud resources. SDRM aims at accelerating efficient resource orchestration through dynamic workload balancing and edge-cloud resource utilization, thereby reducing the cost of IIoT system deployment and improving the overall ROI for adopting IIoT solutions. We model our resource allocation approach on SAVILE ROW using ESSENSE PRIME modeling language, we then implement the network model on CloudSimSDN and PureEdgeSim. We present a detailed analysis of the system architecture and the key technologies of the model. We finally demonstrate the efficiency of the model by presenting experimental results from a prototype system. Our test results show an extremely low solver time ranging from 0.47 ms to 0.5 ms for nodes ranging from 100 to 500 nodes. With edge-cloud collaboration, our results show about 4 percent improvement in overall task success rates.
|Pages:||115839 - 115854|
|Type of Publication:||
A1 Journal article – refereed
|Field of Science:||
213 Electronic, automation and communications engineering, electronics
This work was supported in part by the Academy of Finland through the projects: 6G Flagship and DigiHealth under Grant 318927 and Grant 326291, and in part by the AI Enhanced Mobile Edge Computing Project through the Future Makers Program of Jane and Aatos Erkko Foundation and Technology Industries of Finland Centennial Foundation. The work of Ijaz Ahmad was supported by Jorma Ollila Grant.
|Academy of Finland Grant Number:||
318927 (Academy of Finland Funding decision)
326291 (Academy of Finland Funding decision)
© The Authors 2021. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.