Edge computing for intelligent applications.

Author: Lanin George, Faculty of Computer Science, Higher School of Economics, Russia.

Email: gmlanin@edu.hse.ru

INTRODUCTION

In the rapidly evolving landscape of technology, the demand for real-time data processing and analysis has become paramount. As traditional cloud computing models struggle to meet the ever-growing needs of latency-sensitive applications, a new paradigm, known as edge computing, has emerged to address these challenges.

Edge computing decentralizes data processing by moving it closer to the source - on local devices or edge servers - thereby reducing latency, enhancing data security, and improving overall service efficiency. This essay delves into the transformative potential of edge computing for intelligent applications, exploring how it not only complements existing cloud infrastructures but also enables innovative solutions in fields such as smart cities, autonomous vehicles, healthcare, and the Internet of Things (IoT).

LITERATURE REVIEW

Over the past few decades, the topic of edge computing has increasingly become the subject of scientific discussion. A total of 10,872 are reported on EC (Edge Computing) from Scopus in English language till date[1]. And most authors identify the following advantages of this technology: “It offers several benefits in terms of security and privacy, fewer transmission delays, decreased network burden, and improved inference time.”[1]

It is also worth mentioning that this topic is often considered in conjunction with the concept of industry 4.0, smart cities, self-driving cars and artificial intelligence.

Recent studies have expanded upon this foundational work, exploring various applications of edge computing. In the realm of IoT, Weisong Shi, Jie Cao, Quan Zhang, Youhuizi Li, Lanyu Xu (2016) discuss how edge computing enhances the performance and scalability of IoT networks by decreasing the dependency on centralized data centers[2]. This is echoed by Cisco's reports, which highlight edge computing's role in supporting the massive growth of connected devices expected in the coming years.

Furthermore, edge computing's applications in smart cities have garnered significant attention. Wuhui Chen and colleagues (2024) provide insights into deployment scenarios where edge computing facilitates real-time data analytics for traffic management, energy distribution, and public safety, thereby optimizing urban living[3].

Autonomous vehicles are another area where edge computing is making substantial progress. Research by Shaoshan Liu, Liangkai Liu, Jie Tang, Bo Yu, Yifan Wang and Weisong Shi (2019) demonstrates the critical role of edge computing in enabling fast decision-making processes, essential for the safe navigation and operation of self-driving cars. By processing data locally, these systems can react promptly to environmental changes, a necessity for ensuring passenger safety and reliability.[4]

In addition, several review papers have investigated the integration of AI in edge-based applications. The paper “At the Confluence of Artificial Intelligence and Edge Computing in IoT-Based Applications: A Review and New Perspectives” briefly describes the use of AI at the edge of the network and common AI applications in these networks (“data preprocessing (aggregation, filtering, imputation, and reduction), data analytics (prediction, classification, visualization, and decision-making), resources management (task scheduling, and load balancing), and intelligent sensing (data collection, and data transmission)”.[5]

METHODOLOGY

This essay is the result of a comprehensive review of existing work on advanced computing and an analysis of EC trends.

RESULTS

So, there are several factors that contribute to the development of the trend towards edge computing:

Year by year, the computing power of modern devices has grown significantly in accordance with Moore's law. This makes it possible to perform calculations directly on users' devices or directly on the network of these devices. In the context of deglobalization, high-tech computing resources are becoming less accessible, supply chains are collapsing due to sanctions restrictions, which slows down the development of cloud infrastructure and entails resource optimization. Now, during the rise of AI, many companies are redirecting all available resources to training and operating their ML services. This leads to a shortage of electricity. For instance, many experts are already talking about the growing need for nuclear power plants for such computing clusters.[6] Because of this, supporting your own data centers is becoming even more expensive, and the growth of cloud infrastructure is slowing down.

In the context of the development of intelligent systems, edge computing is already being actively implemented. In the last decade, the data driven approach to decision-making has become particularly active. The key value is data that can be processed, analyzed, and provided with valuable information. So any intelligent system now partially uses the power of its clients and/or intermediate nodes (for example, local servers of smart homes, which are already placed in small smart speakers, for example). At a minimum, data is preprocessed and analyzed on edge devices. If the infrastructure of the intermediate nodes is established, the rest of the analysis and decision-making will soon be delegated to these nodes.

DISCUSSION

The exploration of edge computing for intelligent applications has highlighted several pivotal advantages, but it also unveils ongoing challenges and areas for potential improvement. The deployment of edge computing technologies demonstrates substantial benefits, such as reduced latency, enhanced autonomy for IoT devices, and improved reliability and efficiency across various sectors, including smart cities, autonomous vehicles, and healthcare.

However, the integration of edge computing into existing infrastructures is not without its hurdles. One major concern is the complexity of effectively managing and orchestrating distributed networks, which necessitates sophisticated resource management and coordination mechanisms. There are a lot of approaches to manage it, one of them is L6C, described in the paper “Learn to Coordinate for Computation Offloading and Resource Allocation in Edge Computing: A Rational-Based Distributed Approach”[7]. Compatibility and interoperability between diverse edge devices and platforms pose additional challenges that could impede seamless integration and scalability.

Security and data privacy remain top priorities as edge computing processes sensitive data closer to its source. Although edge computing can minimize data exposure by reducing the need for centralized data transmissions, it still requires comprehensive security frameworks to protect against localized cyber threats and ensure robust data protection. Research must continue to develop and implement advanced encryption and security protocols tailored to the unique architecture of edge environments. This is very clearly shown in the article by Yongchuan Niu and his colleagues.[8]

Furthermore, as edge computing continues to evolve, ensuring infrastructure scalability while maintaining energy efficiency is crucial. Innovations in edge device architecture and software optimization are essential to enhance performance without disproportionately increasing energy consumption. Ongoing efforts to standardize edge computing practices and develop benchmarking tools are vital in addressing these concerns and fostering wider adoption.

CONCLUSION

In conclusion, edge computing represents a transformative evolution in the landscape of intelligent applications, addressing the limitations of traditional cloud computing by providing localized, efficient, and responsive data processing solutions. The findings of this essay underscore the significant improvements in latency, autonomy, and reliability facilitated by edge computing, marking it as an essential component of modern computing ecosystems.

Despite its challenges, the continued advancement of edge computing technology presents an array of opportunities for further innovation and application development. By overcoming obstacles related to network management, security, and scalability, edge computing can fully realize its potential to enhance intelligent applications across varied sectors.

Ultimately, the success of edge computing hinges on collaborative efforts between academia, industry, and policymakers to create supportive frameworks, stimulate research and development, and establish best practices. As edge computing becomes increasingly integral to the fabric of intelligent systems, its impact will likely redefine how we approach data processing and intelligence in an interconnected world. This ongoing evolution promises to yield significant benefits, driving progress toward more efficient, secure, and responsive technological environments.

REFERENCES

1) Garima Nain, K.K. Pattanaik, G.K. Sharma. “Towards edge computing in intelligent manufacturing: Past, present and future”. In Journal of Manufacturing Systems Volume 62, Pages 588-611, January 2022. URL: https://doi.org/10.1016/j.jmsy.2022.01.010

2) Weisong Shi, Jie Cao, Quan Zhang, Youhuizi Li, Lanyu Xu. “Edge Computing: Vision and Challenges”. In IEEE Internet of Things Journal, Volume: 3, Issue: 5, October 2016. URL: https://doi.org/10.1109/JIOT.2016.2579198

3) Wuhui Chen, Zhen Zhang, and Baichuan Liu. “Smart cities enabled by edge computing”. In Book “Edge Computing: Models, technologies and applications”, July 2024. URL: https://doi.org/10.1049/PBPC033E_ch15

4) Shaoshan Liu; Liangkai Liu; Jie Tang; Bo Yu; Yifan Wang; Weisong Shi. “Edge Computing for Autonomous Driving: Opportunities and Challenges”. In Proceedings of the IEEE, Volume: 107, Issue: 8, August 2019. URL: https://doi.org/10.1109/JPROC.2019.2915983

5) Amira Bourechak, Ouarda Zedadra, Mohamed Nadjib Kouahla, Antonio Guerrieri, Hamid Seridi, Giancarlo Fortino. “At the Confluence of Artificial Intelligence and Edge Computing in IoT-Based Applications: A Review and New Perspectives” in Journal Sensors, Volume 23, Issue 3. URL: https://doi.org/10.3390/s23031639

6) Emily Waltz. “Big Tech Backs Small Nuclear: Google and Amazon invest in small modular reactors to power data centers”. In IEEE Spectrum, December 2024 URL: https://spectrum.ieee.org/nuclear-powered-data-center

7) Zhicheng Liu; Yunfeng Zhao; Jinduo Song; Chao Qiu; Xu Chen; Xiaofei Wang. “Learn to Coordinate for Computation Offloading and Resource Allocation in Edge Computing: A Rational-Based Distributed Approach”. In IEEE Transactions on Network Science and Engineering,Volume: 9, Issue: 5, October, 2022. URL: https://doi.org/10.1109/TNSE.2021.3136942

8) Yongchuan Niu; Jiawei Zhang; An Wang; Caisen Chen. “An Efficient Collision Power Attack on AES Encryption in Edge Computing”. In IEEE Access, Volume: 7, pages: 18734 - 18748, January 2019. URL: https://doi.org/10.1109/ACCESS.2019.2896256