1. Introduction
Regarding Brazilian investments, even with federal government initiatives, through international partnerships on bilateral agreements between plans for AI, most of the incentives still seem to come from the private sector and the market, especially regarding multinationals. The initiatives of Meta and Qualcomm companies to bring AI to Smartphones seem to be relevant, and the plan is for 2024. NVIDIA IT also participates in this Innovation process with important technological data. This initiative is far from data stored in the Cloud.
The growing proliferation of devices connected to the Internet, the Internet of Things (IoT), and the demand for real-time services have driven the development of new approaches in computing. Among those approaches, Cloud, Edge, and Fog Computing stand out.
1. Definition and Concepts
Fog Computing refers to an emerging paradigm that focuses on the transfer, processing, and storage of data in local devices, such as routers, switches, and gateways, present at the edge (edge) of the network. The main idea behind Fog Computing is to reduce Cloud Computing overhead by decentralizing processing and allowing the execution of tasks in real-time. Edge Computing, in turn, is an even more decentralized approach, where processing takes place directly on IoT devices without involving intermediate infrastructures. On the other hand, Cloud Computing has sophisticated centralizing resources in remote data centers, providing large-scale storage, processing, and analysis services.
2. Latency and Response Time
One of the main arguments regarding Fog Computing over Cloud Computing and Edge Computing is the issue of latency and response time. Cloud Computing can impose significant delays in data transmission, especially when devices have geographically dispersed. On the other hand, Edge Computing, although it reduces latency compared to Cloud, can still be limited in touching processing resources and storage capacity. Fog Computing, by being close to the devices and, at the same time, using additional infrastructure resources at the edge, can offer lower latency and faster response times for applications.
3. Bandwidth Savings and Efficient Use of Resources
Cloud Computing requires significant bandwidth to transfer data between remote devices and data centers. In scenarios with many IoT devices generating large amounts of data, this approach can overwhelm the network infrastructure. Edge computing improves on this aspect but still requires significant data transfer between devices and processing points. Fog Computing, however, minimizes data traffic by performing part of the processing locally at the edge, saving bandwidth and optimizing resource utilization.
4. Security and Privacy
The centralization of data in Cloud Computing can raise concerns related to security and privacy, especially when sensitive data has transferred over the Internet. Edge computing can mitigate some of these issues, but it still faces challenges by concerning cybersecurity and threat management. Fog Computing provides an intermediary layer between IoT devices and the Cloud, allowing sensitive data to be processed locally, reducing exposure to external threats.
5. Conclusion
Based on the academic, formal, and conceptual analysis presented, we believe that Fog Computing is a superior approach over Cloud Computing and Edge Computing. Its ability to provide low latency, reduce network overhead, improve resource use efficiency, and offer greater security make it a more advantageous choice for the growing demand for Internet of Things services and real-time computing. However, it is crucial to emphasize that the choice between these approaches should depend on the specific needs of each application, considering factors such as system scale, response time requirements, and sensitivity to security and privacy aspects. Furthermore, it depends on the intended industrial application.
Based on the article: Meta and Qualcomm team up to run big A.I. models on phones – NBC4 Washington
References
Cassiolato, J. E., Bernardes, R., & Lastres, H. Innovation Systems in the South.
Das, R. and Muhammad Muhammad Inuwa (2023). A review on fog computing: Issues, characteristics, challenges, and potential applications. [online] 10, pp.100049–100049. doi:Redirecting.
Rzepka, M., Piotr Borylo, Dias, M., Lason, A. and Laurent Lefèvre (2022). SDN-based fog and cloud interplay for stream processing. [online] 131, pp.1–17. doi:Redirecting.
Bigelow, S.J. (2021). What is edge computing? Everything you need to know. [online] Data Center. Available at: What Is Edge Computing? Everything You Need to Know [Accessed 20 Jul. 2023].
Junsheng, S. (2014). Data Security and Privacy in Cloud Computing - Yunchuan Sun, Junsheng Zhang, Yongping Xiong, Guangyu Zhu, 2014. [online] International Journal of Distributed Sensor Networks. Available at: https://journals.sagepub.com/doi/full/10.1155/2014/190903 [Accessed 20 Jul. 2023].