Automatic Application Placement and Adaptation in Cloud-Edge Environments

Abstract

Edge computing describes a paradigm for combining computational resources at the edge of the network with the cloud. Even though complementing the cloud with these resources provides benefits, e.g., low latency, it also introduces new challenges to the operational staff. Such challenges can be: deciding if the applications should be placed in the cloud or at the edge, and monitoring them at runtime to ensure that all the application requirements are met. This becomes more challenging when using microservices due to the complexity of the resulting placement problem. To mitigate such concerns, we introduce an automatic deployment framework along with a prototype implementation, called D-DAD. This framework provides a transparent (to the operational staff) way to deploy applications with respect to all their requirements-including the non-functional-using mechanisms for monitoring and adapting the deployments to the available resources in a cloud-edge environment. For evaluating our framework, we provide results from a series of experiments which show how the adaptation mechanism meets the application requirements, including a ∼90% reduction of CPU utilization violations, compared to using only the local resources.

Towards a Programming Paradigm for Reconfigurable Computing: Asynchronous Graph Programming

“…However, Horde already supports the full AGP semantics and model of computation, as well as parallelization (implemented through multi-threaded execution). 1 https://github.com/paulofrgarcia-carleton/Horde-Public-release-The current version of the Horde interpreter was primarily built as a testing environment for the AGP semantics; hence, it is not optimized for performance. AGP graphs are implemented as multiply-linked linked lists which the interpreter reduces throughout execution.…”

Relating Edge Computing and Microservices by means of Architecture Approaches and Features, Orchestration, Choreography, and Offloading: A Systematic Literature Review

“…Raspberry Pi 1 [94] Raspberry Pi 2 [78,79,114] Raspberry Pi 3 [13, 18, 21, 28, 29, 41, 49, 57, 77-79, 84, 88, 93, 95, 106, 113, 121, 136, 139, 143, 146] [147, 149] Raspberry Pi 4 [15,20,49,73,88,103,123,146,148] Beaglebone Black (BBB) [140] Custom Hardware [47,71,76,83,95] Intel NUC [93,140] Lenovo ThinkCentre 920x [55,95] Nvidia Jetson TX2 [20,72,111,148] Nvidia Jetson Xavier [123] Google Coral Dev AI Board [133] Simulated (Amazon EC2) [74] ATMega 2560 [78,79] ThinkerBoard [29] Smartphone [152] Table 14: Selected Studies -Edge Node -Devices 4.2 RQ1.1 -Based on the microservice architectural approaches and features, are there advantages of using microservices in the Edge? The selected studies highlight several advantages, primarily because Edge Computing is a decentralized paradigm next to the users' devices.…”

“…It impacts IoT applications committed to a very low response time (e.g., smart car) or a minimum QoS level to their users (e.g., real-time video surveillance). For instance, an IoT application that controls a servo motor in a factory could not handle a variable or high response time [71,74].…”

The FORA Fog Computing Platform for Industrial IoT

“…This model is implemented by orchestrating containers running the single applications. Similarly, Meixner et al [61] present a framework for automatically placing applications in fog environments, especially aiming at industrial settings.…”