Energy efficient Container as a Service in Software Defined Edge Computing

Abstract

Internet of Things (IoT) has converted the entire conventional technological paradigm newlineinto a data-driven plethora of smart and connected ecosystems. The data generated newlinefrom this revolutionary ecosystem needs to be transmitted at a high data rate and newlinethereafter processed or analyzed in real-time at a remote cloud data center or the newlineedge of the network using resource-constrained devices. For quite some time, the newlinedata centres (DCs) have been the hub of providing big data solutions in the IoT envi- newlineronment using cloud computing platforms. Cloud supports flexibility and scalability newlineby adopting the virtualization technology, that is used to create multiple copies of newlinethe virtual machines on the physical serves. However, the increase in load on the newlineservers can increase latency and energy consumption during processing the tasks. newlineTo handle these challenges, a lightweight technology such as Container-as-a-Service newline(CaaS) comes to the rescue of the Cloud. In CaaS, the computing resources reside newlineinside a container, which runs on a Docker engine rather than a hypervisor (as in newlinethe case of virtual machines). To execute time-sensitive applications, containers are newlineprobably suitable alternative to hypervisor-based virtualization. newlineHowever, nowadays, the need of processing data closer to the IoT devices has newlinenecessitated the provisioning of the big data solutions at the network edge. newlineIn newlinethis context, the Edge computing emerged as Cloud near to the ground , which newlineprovides computing devices closer to the location of the data source. Whenever newlinea request is generated by the users, it can be handled at the edge layer (mostly newlinelightweight workloads) instead of forwarding it to the core cloud DCs. It is more newlinefeasible to process the huge workloads at the Cloud and the delay-sensitive work- newlineiii newlineiv newlineloads ate the edge resources. This leads to the edge-cloud interplay that further newlineincreases the movement of data traffic from end-user applications to the edge-cloud newlineecosystem. Moreover, the surge in smart applications across the globe has escalated newlinet