A Frontier: Dependable, Reliable and Secure Machine Learning for Network/System Management
- PDF / 1,166,356 Bytes
- 23 Pages / 439.37 x 666.142 pts Page_size
- 85 Downloads / 156 Views
A Frontier: Dependable, Reliable and Secure Machine Learning for Network/System Management Duc C. Le1 · Nur Zincir‑Heywood1 Received: 14 November 2019 / Revised: 15 January 2020 / Accepted: 20 January 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Modern networks and systems pose many challenges to traditional management approaches. Not only the number of devices and the volume of network traffic are increasing exponentially, but also new network protocols and technologies require new techniques and strategies for monitoring controlling and managing up and com‑ ing networks and systems. Moreover, machine learning has recently found its suc‑ cessful applications in many fields due to its capability to learn from data to auto‑ matically infer patterns for network analytics. Thus, the deployment of machine learning in network and system management has become imminent. This work provides a review of the applications of machine learning in network and system management. Based on this review, we aim to present the current opportunities and challenges in and highlight the need for dependable, reliable and secure machine learning for network and system management. Keywords Network and system management · Reliable and dependable machine learning · Secure machine learning
1 Introduction Networks are growing at exponential pace and becoming more and more diverse, not only connecting people but also machines and digital objects. The vast collec‑ tions of network devices, end user devices and heterogeneous links are also grow‑ ing, both in terms of numbers and types of devices. Naturally, this results in many opportunities as well as challenges in the process of managing such networks, services and systems. Furthermore, recent network developments, although creat‑ ing tremendous potential applications and greatly enhancing network capabilities * Nur Zincir‑Heywood [email protected] Duc C. Le [email protected] 1
Dalhousie University, Halifax, Canada
13
Vol.:(0123456789)
Journal of Network and Systems Management
and user experiences, bring with them new challenges for network and system management (NSM). For example, the proliferation of 5G networks has been anticipated to open several new opportunities. This next generation mobile net‑ work technology greatly increases data transfer rates, while reducing latency and energy usage. Essentially 5G will enable Internet of Things (IoT) and many other use cases, such as smart transportation and high-performance edge analyt‑ ics. Another example of network expansion and diversification is smart cities and homes. These in return create challenges in managing networks and ser‑ vices by introducing new heterogeneity and diversity, as well as cybersecurity concerns. Analyzing operational data and network traffic data generated by those networks for troubleshooting and detecting anomalies/faults/intrusions would be overwhelming to human analysts, given the sheer amounts of data they create. Similarly, Network Function Virtualization (NFV) and Softw
Data Loading...
