Service Provisioning in Edge-Cloud Continuum: Emerging Applications for Mobile Devices

Abstract

Disruptive applications for mobile devices can be enhanced by edge computing facilities. In this context, Edge Computing (EC) is a proposed architecture to meet the mobility requirements imposed by these applications in a wide range of domains, such as the Internet of Things, Immersive Media, and Connected and Autonomous Vehicles. EC architecture aims to introduce computing capabilities in the path between the user and the Cloud to execute tasks closer to where they are consumed, thus mitigating issues related to latency, context awareness, and mobility support. In this survey, we describe which are the leading technologies to support the deployment of EC infrastructure. Thereafter, we discuss the applications that can take advantage of EC and how they were proposed in the literature. Finally, after examining enabling technologies and related applications, we identify some open challenges to fully achieve the potential of EC, and also research opportunities on upcoming paradigms for service provisioning. This survey is a guide to comprehend the recent advances on the provisioning of mobile applications, as well as foresee the expected next stages of evolution for these applications.

Distributed Learning in the IoT–Edge–Cloud Continuum

“…Five review articles [2][3][4][5][6] and three survey articles [7][8][9] were found that research similar aspects of the IECC. Kampars et al [3] reviewed the application layer communication protocols that can be used for interconnecting the layers of the IECC.…”

“…The survey article from Gkonis et al [8] combined the current technological advancements of the IECC, focusing on challenges as well as identifying open issues and limitations. Oliveira et al [9] researched the mobile device technologies that could support the development of Edge-Cloud continuum systems by analyzing their mobility, characteristics, and popularity.…”

“…None of the found review and survey articles focus on the distributed learning methods for the IECC. Some mention federated learning [4,5,8,9] and transfer learning [8] However, they do not deeply research the innate challenges and problems related to this direction, such as security and privacy enhancing methods for DL in relation to the IECC domain. For this reason, this article focuses exactly on this subject in relation to the IECC.…”