I am trying to write / research about fog computing platforms or benchmarks and perform experiment to (1) understand fog computing platforms (2) build a very basic platform (3) build a very basic mobile app in fog computing based system connected to cloud to inform users of natural disaster in real time. However, I have following questions that I am looking to hear from you.
1. How can I design / implement fog computing environment with basic devices like raspberry pi devices?
2. What platforms are available for fog computing environment to experiment?
3. Is there any platform/tool to develop fog computing mobile app using higher level languages (C#, JAVA)?
The important point is the platform, how can I design very basic fog computing platform?
FOG COMPUTING: AN OVERVIEW Fog computing is “an architecture that uses one or a collaborative multitude of end-user clients or near-user edge devices to carry out a substantial amount of storage (rather than stored primarily in cloud data centers), communication (rather than routed over the internet backbone), and control, configuration, measurement and management (rather than controlled primarily by network gateways such as those in the LTE (telecommunication) core)”. The fog computing paradigm does not necessarily stick to one architecture, instead it represents a notion that supports to push data analytics towards leaves (i.e., edge nodes). The important message to understand is ‘towards leaf /edge node’. Therefore, fog computing does not suggest that all analytics needs to be done on leaves, instead it promotes to use leaf nodes as much as possible within realistic limitations. Leaf nodes also represent the closeness to the users. Further, it also suggests that multiple nodes close to the leaves would work in a cooperative manner to accomplish a task.
Fog networking constitutes two planes, namely, 1) control plane and 2) data plane . The responsibility of the control plane is to configure and manage the network (e.g., sync routing table information, efficient traffic control). The data plane is responsible for transferring data from the originator to the destination. based on the routing information provided by the control plane logic has highlighted main differences between Cloud and Fog.
— Store most of the data at or near the edge instead of sending all the data to the cloud all the time. Features such as Mobility Context Discovery and Awareness , and Data Analytics facilitate these characteristics.
— Use local networks (i.e., closer to the leaf nodes) to transfer data to processing nodes instead of using backbone networks (e.g., Internet) to route the data. Features such as Multi-Protocol Support: Communication Level , Multi-Protocol Support: Application Level , Mobility , Security and Privacy , and Cloud Companion Support facilitate these characteristics.
— Process substantial amount of data at the edge devices (e.g. leaf nodes) instead of using cloud computing infrastructure. Features such as Semantic Annotation , Data Analytics and Context Discovery and Awareness facilitate these characteristics.
Hi Abdul,
you may find some useful information in this survey paper (Article Fog Computing for the Internet of Things: A Survey
) on Fog Computing for the Internet of Things. In the last section, there is an overview of some of the available platforms in this field. Specifically, platforms are classified in three main categories: (i) software platforms; (ii) development platforms; and (iii) hardware platforms. This may help you to choose the software and hardware components for your solution.In another paper ( Article Container Migration in the Fog: A Performance Evaluation
), I built up a Fog Computing platform using Raspberry Pi 3 Model B as Fog nodes. The purpose of this platform was to evaluate and compare the existing container migration techniques in a Fog Computing environment. I installed and configured hostapd and dnsmasq on each Raspberry to let it behave as a Wi-Fi access point. Besides, I installed runC as container runtime (other container platforms can be used, e.g., Docker).I hope that this can help you.
Regards,
Carlo
Hi Abdul, you maybe interested in the DeFog project (Fog computing benchmarks) developed by my team at Queen's University Belfast. Repository is here: https://github.com/qub-blesson/DeFog. Watch the space, research article will be available soon. Hope that helps.
Further to previous response, please find a paper entitled, "DeFog: Fog Computing Benchmarks" accepted to ACM/IEEE Symp. on Edge Computing: http://bvarghese.staff.cs.qub.ac.uk/papers/DeFog-SEC2019.pdf. Source code is available here: https://github.com/qub-blesson/DeFog with six benchmarks.
Please go through the following work developing FogBus framework [1] under CLOUDS Lab, University of Melbourne, Rajkumar Buyya. It has been published in of the leading journals namely Journal of Systems and Software (JSS), Elsevier. It has been used to develop various other platforms like EdgeLens [2] (object detection in Fog) and HealthFog (paper coming soon)
[1] http://www.buyya.com/papers/FogBus-JSS.pdf
[2] http://www.buyya.com/papers/EdgeLensAnekaCloud2019.pdf
Source code is available here: https://github.com/Cloudslab/FogBus
Please check this repository https://github.com/dream-lab/riot-bench. Anshu Shuklai and Yogesh Simmhan had several studies on IoT benchmarks. With respect to platforms, I would the IoT-Lab (https://www.iot-lab.info/) initiative. In our paper (Article Distributed Data Stream Processing and Edge Computing: A Sur...
), you can find some issues related to future directions in data stream processing on fog computing.- Badges
- Science topic
- Similar topics
- Solid State Physics
- Devices