Why we need fog Computing?- A bridge between Cloud and IoT (Part 1)
IoT (Internet of Things) is one of the growing technologies that influence most of the domains including agriculture, healthcare, astrology, manufacturing, architecture, transportation, water management etc. It is a modern paradigm based on machine to machine communication between unintelligent and smart devices in a system such as sensors,actuators etc. Almost all social sectors are currently converting their infrastructure to support IoT to survive in the fast developing world.
Due to the high increase of devices connected in IoT, Cloud Computing became a necessity to overcome the issue in handling unprecedented volume and variety of data which is generated by IoT devices. Cloud computing enables ubiquitous access of data to the users and provides cost-effective, user-centric services.
Even though, a middleware became a necessity for IoT and the Cloud for the preprocessing of data before transfer it to the Cloud due to the following shortcomings of Cloud computing [1].
- Infinite data production
As the number of heterogeneous devices connected to the IoT network has increased drastically, data production is also increased. This resulted in need of more storage space and computation of data future usage.
- High Latency
Most of the IoT systems are mission critical and latency sensitive. For example, sensor based security systems, traffic control systems are very sensitive to communication latency. In addition, data transfer between the Cloud and edge devices is prone to delay because of the geographical distribution of the edge devices. Therefore, this latency can be a barrier for the systems which needed quick response to make decisions especially for the real-time systems.
- High transport cost
Due to frequent transmission of large volume data to the cloud, it has a high cost. This will be a problem when huge numbers of devices are connected in the network.
- Data security and location/data privacy
Cloud computing is based on Centralized data management which can be vulnerable for security attacks, as the data is centrally managed in Clouds. When user critical data like location and personal details are managed in clouds, the vulnerability of attacks on this data is higher inside clouds.
A network middle layer called ‘Fog’ is introduced to overcome these drawbacks of Cloud based networks. Fog computing is a modern concept proposed by Cisco in 2014 to decentralize the networking services between the end devices of IoT and the Cloud to compute and store data. Fog computing is the extension of the cloud computing paradigm, which overcomes most of the limitations of cloud computing by enabling the computation, storage and network services near to the end devices. In simple words, fog is an intermediate layer of network which contains core network devices (like routers) and Field Area Network devices (like intelligent gateways) to convey the services in between edge devices and the Cloud Data Centers.
SIMILARITIES AND DISSIMILARITIES BETWEEN CLOUD COMPUTING AND FOG COMPUTING
According to the above mentioned drawbacks of Cloud computing, fog computing differs from data clouds. When comparing to clouds, fog has more localized architecture, low latency, higher proximity to end nodes [2] and context aware computing. In an IoT based network, fog layer can be distributed within the core network which is near to the central Cloud or the edge network with higher proximity to IoT devices. Further, fog layer contains larger number of resourceful fog nodes that can provide processing.
Despite of these aspects, fog and cloud cannot be replaced by one another. Thus, the tasks between fog and cloud layers of a network can be classified based on application and can be changed according to the state of the network such as bandwidth, storage capacity, fault events, processor loads and threats. Therefore, Fog became an unavoidable resource for the working of IoT based networks.
The next part of the article explains the architecture of Fog network from a functionality based perspective and analyses the infrastructure of the software components of Fog nodes.
References
[1]. M. Aazam and E.-N. Huh, “Fog Computing: The Cloud-IoT/IoE Middleware Paradigm,” IEEE Potentials, vol. 35, no. 3, pp. 40–44, 2016.
[2]. E. Marín-Tordera, X. Masip-Bruin, J. García-Almiñana, A. Jukan, G.-J. Ren, and J. Zhu, “Do we all really know what a fog node is? Current trends towards an open definition,” Computer Communications, vol. 109, pp. 117–130, 2017.
