Pareto-Optimised Fog Storage Services with Novel Service-Level Agreement Specification

Abstract

(1) Background: Cloud storage is often required for successful operation of novel smart applications, relying on data produced by the Internet of Things (IoT) devices. Big Data processing tasks and management operations for such applications require high Quality of Service (QoS) guarantees, requiring an Edge/Fog computing approach. Additionally, users often require specific guarantees in the form of Service Level Agreements (SLAs) for storage services. To address these problems, we propose QoS-enabled Fog Storage Services, implemented as containerised storage services, orchestrated across the Things-to-Cloud computing continuum. (2) Method: The placement of containerised data storage services in the Things-to-Cloud continuum is dynamically decided using a novel Pareto-based decision-making process based on high availability, high throughput, and other QoS demands of the user. The proposed concept is first confirmed via simulation and then tested in a real-world environment. (3) Results: The decision-making mechanism and a novel SLA specification have been successfully implemented and integrated in the DECENTER Fog and Brokerage Platform to complement the orchestration services for storage containers, thus presenting their applicable value. Simulation results as well as practical experimentation in a Europe-wide testbed have shown that the proposed decision-making method can deliver a set of optimal storage nodes, thus meeting the SLA requirements. (4) Conclusion: It is possible to provide new smart applications with the expected SLA guarantees and high QoS for our proposed Fog Storage Services.

Keywords

• Fog Storage Services
• Quality of Service
• Service Level Agreement specification
• decision making
• Pareto