Description
Date depot: 13 novembre 2019
Titre: Dynamic management of IoT data stream analytics in the edge-fog-cloud continuum
Directeur de thèse:
Nikolaos GEORGANTAS (Inria-Paris (ED-130))
Directeur de thèse:
Vassilis CHRISTOPHIDES (ETIS)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Systèmes et réseaux
Resumé:
IoT data stream analytics applications continuously process unbounded data streams generated by multiple geographically distributed sources. Cloud-based analytics is the prevalent processing approach, where all generated data items are transmitted to the cloud. However, at high data rates and/or stringent time constraints, accessing the cloud via limited-bandwidth, high-latency wide-area network connections becomes a serious bottleneck incurring delays and/or significant costs. Fog computing, as an alternative processing approach, partially handles data items near their sources and enables reducing wide-area network utilisation and application response times. However the fog has limited processing resources and requires deciding which part of the analytics application should be performed there. For applications dealing with highly variable data stream rates, such as real-time road traffic monitoring, continuous dynamic management of data streams is required. This PhD research project aims at proposing new models and algorithms for deployment and runtime adaptation of IoT data stream analytics applications on a distributed resource space spanning the edge/fog and cloud. The proposed algorithms will be evaluated against state-of-the-art solutions via simulation and system prototype implementation and evaluation.
Doctorant.e: Ntumba Wa Ntumba Patient