Introduction  

In this current COVID 19 times, we can more than ever see how everything technologically based need to be autonomous. IoT devices and applications are ever-growing and according to sources, the number of these would be massive of more than 50 billion of these IoT devices will be in use around the world, creating an immense maze of interconnected devices traversing between everything from smartphones to smart kitchens. Adding complication to this ever-growing number of IoT devices, time-critical devices would need a more reliable computing capability to have the best possible outcome or result.

Cloud Computing, conventionally supporting the IoT devise might not be able to serve that purpose alone, one of the main concerns would be its characteristics like their resource centralization. The data and their connectivity, with the increased separation amongst the devices and their cloud, would lead to delays or lags leveraging the network latency and jitters. (M. Caprolu, 2019)

There are evolutions to the traditional computing norms i.e. cloud computing to assist minimize the networking issues, Edge and fog computing have been budding as complimentary architecture to the traditional cloud to offload the ever-increasing inclusion of IoT devices and application.

Finding

Massive data volumes are generated due to the adoption of IoT (Internet of Things) with the exponential rate of growth in the number of connected devices. The current network of the IoT systems doesn’t adequately support this growth and most researchers seem to be overlooking the fact that it requires a proper and appropriate design for the IoT Network. There are shortcomings of the current network methodologies to support this influx of the IoT Devices and Application and to address this shortcoming a relevant network methodology must be implemented (S. Verma, 2017). Apart from the traditional IoT devices which have its own shortcoming, there are these Time-Critical IoT devices and application which would require more robust and accurate network methodologies to support it.

Cloud Computing has been the service when you think the requirement and the processing aptitude for IoT devices and the growth of cloud have brought forth the Everything-as-a-service era. Cloud Computing simplified the requirement to interconnect lots of sensors and other devices but would make it reliant with a high-speed network and interconnections. Central Date center would be called for but transferring those to and fro would, though convenient, would bring in latency and other similar issues. In the case of traditional IoT devices, this could be compromised but time-critical IoT devices would not and require more attention with the data transfer. This situation gave rise to the approach of pushing data collection and processing them back to the devices (end-user) which could be termed as Edge computing. There are different forms of these from fog computing, cloudlets to mobile edge computing. (Gusev, 2018)

The Computing paradigm which is in close proximity to all the IoT related devices and hiving this computing capability placed at the edge of the internet would improve the latency, bandwidth, trust, and survivability. (Satyanarayanan, 2017)

Result

The current stunning upsurge of the IoT (Internet of Things) and the accompanying swell of the data surge motivated the materialization of Edge computing which was different from the traditional centralized Cloud and having processing as close to the data sources as possible. Though it does minimize latency with improved bandwidth and trust but would be a challenge to see how the performance of an application would improve (or degrade) when having processing closer to the edge. There could be trade-offs and these could see a combination of computing to deal with the trade-off. Such combination could lead to Hybrid deployments like an edge-cloud which might bring the best of two to have a positive impact on performance. (P. Silva, 2019)

Furthermore, IoT with its mammoth interconnected intelligent devices which makes smart communities and maintaining them could be a resource-draining affair not only draining the processing capabilities at the data center or servers but also having high latency issues and network congestions. More than others, the time-critical IoT Devices such as smart cars, Early warning systems, or critical medical devices cannot compromise on these issues. Though edge computing diminishes these issues, it has its limitation, like limited computing resources of the edge computing devices and having uneven load among these devices. Therefore, efficient task management in edge computing networks would be absolutely necessary to recognize the potential of edge computing to support time-critical IoT devices and IoT devices in general.  An example of this would be to have QoS (Quality of Service) requirement set while edge computing and trying to achieve the desired standard to maintain the best possible edge computing network. (Xue, 2017)

Edge computing can be looked at as a promising alternative to cloud infrastructure where the resources are provided at the edge IoT Devices as the traditional Cloud has its limitation with regards to its centralized computing, storage, and networking. This limitation was mitigated with the advent of cloud-bits, fog computing, and data centers but edge could be the ultimate answer. (J. Pan, 2018) Though Fog Computing is more or less like edge computing is but the major difference is where the processing lies and in case of fog computing the process lies at the edge but not of the device but that of the network. Just to understand the sheer size of data being generated with the IoT devices, an Autonomous vehicle could generate a massive amount of data estimated at around 1GB per second (Mearian, 2013). This could be details like the car route and speed, car conditions like wear and tear, the geographical environment the car, the weather condition, and also their video feed for traversing or for safety reasons. This would entail for very expensively high network bandwidth to have the data transfer to the cloud and back to the device which would though be possible but might have high latency issues considering the hardware requirements are in place. Edge computing enables the processing to be at the edge i.e. the device network itself which would ensure better reliability in the shortest possible time.

Conclusion

The hype with the IoT development and deployment, we are more than certain that it would generate data of epic proportion, which would in turn make storing and processing theses days a challenge in itself. Cloud computing plays a vital role in addressing this issue with its massive storage availability, heavy-duty computation capacity, diverse geographical coordination and wide-area connectivity.

Fog computing could be key to edge resource pooling where the service required is user-centric and real-time processing but even that might not be enough for some time-critical IoT devices and their data. Though Fog computing brings or takes the processing at the edge of the network, there could be further moved towards the end devices or users which would bring down the turn-around time altogether.

Edge computing could be that answer but although it is beneficial to time-critical devices, it has its own demons to overcome. Some edge devices though have the edge computing possibility but might not have the capability or the hardware to support it. Though edge computing could minimize latency and other networking issues such as bandwidth, it could work best with some combination with other computing paradigms.

Though that could be left for some other time to research, I conclude that with the current trend of IoT devices and time-critical IoT devices, the cloud has been the primary source for managing the resources and processing and fog and other computing paradigm mitigating the limitation cloud computing presented. Though there are possibilities of having a hybrid computing paradigm to address the challenges with the connectivity of IoT devices, I feel that Edge Computing a way forward for time-critical IoT applications and devices.

Lesson Learnt

Primarily with the reason to learn about Edge computing, the basic lesson to learn was the fundamental of Edge Computing but have learned more than what was anticipated.  From the basic understanding of what and IoT devices and how it is different from a Time-critical IoT device, while understanding how the exponential growth would bring in challenges.

Evolution from the computing paradigm where the sensor data was processed in a standalone manner without much analysis required to having Clouding computing supporting the growth of IoT. Here I learned that with this massive centralized storage and processing, increases latency which is crucial pertaining to time-critical IoT devices. Though cloud has a high computational capability, latency is not only determined but that but could also be the long WAN (Wide area network) delay.

Though I have understood that Edge computing would be a good option to reduce latency by having computational power closest to the devices but have also learned that there could be a combination or hybrid computing paradigm which could be a better option.

Reference

Omoniwa, R. H. (2019). Fog/Edge Computing-Based IoT (FECIoT): Architecture, Applications, and Research Issues. IEEE Internet of Things Journal, 4118-4149.

Martín Fernández, M. D. (2018). An Edge Computing Architecture in the Internet of Things. 2018 IEEE 21st International Symposium on Real-Time Distributed Computing (ISORC), 99-102.

Suciu, M. A. (2017). IoT time critical applications for environmental early warning. 2017 9th International Conference on Electronics, Computers and Artificial Intelligence (ECAI), 1-4.

Gusev, M. D. (2018). Going Back to the Roots-The Evolution of Edge Computing, An IoT Perspective. IEEE Internet Computing, 5-15.

Pan, J. M. (2018). Future Edge Cloud and Edge Computing for Internet of Things Applications. IEEE Internet of Things Journal, 439-449.

Skirelis, D. N. (2017). Edge computing in IoT: Preliminary results on modeling and performance analysis. 2017 5th IEEE Workshop on Advances in Information, Electronic and Electrical Engineering (AIEEE), 1-4.

Toka, B. L. (2017). A resource-aware and time-critical IoT framework. IEEE INFOCOM 2017 – IEEE Conference on Computer Communications, 1-9.

Caprolu, R. D. (2019). Edge Computing Perspectives: Architectures, Technologies, and Open Security Issues. 2019 IEEE International Conference on Edge Computing (EDGE), 16-123.

Mearian, L. (2013, July 24). Self-driving cars could create 1GB of data a second. Retrieved from Computer World : https://www.computerworld.com/article/2484219/self-driving-cars-could-create-1gb-of-data-a-second.html

Schulz, M. M.-T. (2017). Latency Critical IoT Applications in 5G: Perspective on the Design of Radio Interface and Network Architecture. IEEE Communications Magazine, 70-78.

Silva, A. C. (2019). Investigating Edge vs. Cloud Computing Trade-offs for Stream Processing. 2019 IEEE International Conference on Big Data (Big Data), 469-474.

Verma, Y. K. (2017). A Survey on Network Methodologies for Real-Time Analytics of Massive IoT Data and Open Research Issues. IEEE Communications Surveys & Tutorials, 1457-1477.

Satyanarayanan, M. (2017). Edge Computing. IEEE, 36-38.

Xue, Y. S. (2017). An Approach to QoS-based Task Distribution in Edge Computing Networks for IoT Applications. 2017 IEEE International Conference on Edge Computing (EDGE), 32-39.

Zhang, M. C. (2016). Fog and IoT: An Overview of Research Opportunities. IEEE Internet of Things Journal, 854-864.