Companion fog computing: supporting things mobility through container migration at the edge

TitleCompanion fog computing: supporting things mobility through container migration at the edge
Publication TypeConference Proceedings
Year of Conference2018
AuthorsPuliafito, C., E. Mingozzi, C. Vallati, F. Longo, and G. Merlino
Conference NameProceedings - 2018 IEEE International Conference on Smart Computing, SMARTCOMP 2018
Pagination97-105
PublisherInstitute of Electrical and Electronics Engineers Inc.
Conference LocationTaormina, Italy - 18-20 June 2018
ISBN Number9781538647059
KeywordsCarrier mobility, Computing platform, containers, Docker, Fog, Fog computing, Internet of Things, Migration, Mobile devices, Network architecture, Proof of concept, Reference architecture, Resource Constraint, Topological distance, Topological proximity, Topology
Abstract

Due to their intrinsic resource constraints, the mobile Internet of Things (IoT) devices are not able to provide intensive services by just relying on their own facilities. Fog Computing effectively helps overcome this hurdle. Indeed, it extends the Cloud toward the network edge, distributing resources and services of computing, storage, and networking close to the end devices. This topological proximity is the key enabler of several advantages that are essential in many emerging ICT domains. Nonetheless, the mobility of an IoT device compromises such benefits as it increases the topological distance to the serving Fog node. Therefore, the Fog service has to be migrated in order to be always close enough to the served IoT device. We name this Companion Fog Computing (CFC), since the Fog service behaves as a 'companion' of the correspondent application on the mobile device. In this paper, we present a Fog Computing Platform that performs stateful container (i.e., Fog service) migrations in order to enable CFC. Specifically, we introduce a CFC model from which we derive a reference architecture comprising all the functionalities required in a platform to make migration decisions and carry them out. Moreover, we demonstrate the soundness of the proposed reference architecture by discussing a proof-of-concept implementation based on the Stack4Things (S4T) platform, and we report a set of conducted experiments to show the feasibility of stateful container migrations. © 2018 IEEE.

URLhttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85051489867&doi=10.1109%2fSMARTCOMP.2018.00079&partnerID=40&md5=4a8069123a5f19d5e842a86b2e9163f4
DOI10.1109/SMARTCOMP.2018.00079