sensors-logo

Journal Browser

Journal Browser

Automation and Learning in MEC and Cloud Systems

A special issue of Sensors (ISSN 1424-8220). This special issue belongs to the section "Communications".

Deadline for manuscript submissions: closed (15 April 2022) | Viewed by 27017

Special Issue Editors


E-Mail Website
Guest Editor
Network Engineering Department, Universitat Politècnica de Catalunya (UPC), 08034 Barcelona, Spain
Interests: optimization; placement; SDN/NFV; edge computing; resource orchestration; IA/ML; DL; DRL; automation
Special Issues, Collections and Topics in MDPI journals

E-Mail Website
Guest Editor
Telematic Engineering Department, Universidad Carlos III de Madrid (UC3M), 28911 Leganés, Madrid, Spain
Interests: UAV; drones; 5G; NFV; resource management
Special Issues, Collections and Topics in MDPI journals

E-Mail Website
Guest Editor
Department of Signal Theory, Telematics and Communications, University of Granada, 18071 Granada, Spain
Interests: LPWAN networks and the Internet of Things; 5G mobile networks using SDN/NFV paradigms; 4G broadband wireless networks; wireless local area networks; multimedia communications and quality of experience
Special Issues, Collections and Topics in MDPI journals

E-Mail Website
Guest Editor
Communications Engineering Department, University of the Basque Country UPV/EHU, 48013 Bilbao, Spain
Interests: 5G; SDN/NFV; in-network computing; mobility; cyber security
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

Users will expect more global coverage, higher data speeds, and ubiquitous availability of new services and applications in the future. All these requirements along with the need to serve various vertical industries with ultrareliable and low latency connectivity are among the key drivers for 5G networks and beyond.

The 5G architecture integrates the network, control, operation, and maintenance layers with a novel service layer that is adaptable to various verticals (Industry 4.0, CCAM, etc.) and emerging applications and can support any environment (IoT, V2V, data, etc.). The 5G requirements are fostering the creation of a distributed architecture with a collective intelligence that will allow the automation and dynamic self-orchestration of end-to-end resources over a multitechnology, multiprovider, and multitenant infrastructure, suitable for verticals with stringent requirements, such as critical situations, high user density, increased bandwidth, security, isolation, and low latency.

This Special Issue aims to highlight advances in the development, testing, practical implementations, and modeling of smart ecosystems for 5G and beyond, formed by a set of new procedures, algorithms, technologies, and mathematical tools for advanced system configuration and automation that leads to the zero-touch paradigm. State-of-the-art reviews on these topics are also welcome.

The topics of interest include but are not limited to:

  • Network automation;
  • Zero-touch network and service management;
  • Dynamic multidomain and multitechnology service orchestration;
  • Resource-constrained NFV infrastructures;
  • Network slicing;
  • Smart network slicing management;
  • RAN smart management and orchestration;
  • Machine learning for networking;
  • AI-based VNF adaptive placement algorithms;
  • Collective intelligence;
  • In-network computing and programmable data planes;
  • Efficient and secure 5G core;
  • Energy awareness solutions;
  • SDN-based management planes for next-generation TWDM PONs and TSN.

Dr. Cristina Cervelló-Pastor
Dr. Francisco Valera
Dr. Jorge Navarro Ortiz
Dr. Jasone Astorga
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Sensors is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

  • smart networks and services
  • network automation
  • multidomain orchestration
  • placement
  • adaptive slicing
  • security
  • artificial intelligence
  • collective intelligence
  • machine learning
  • programmable data planes
  • SDN-based management planes
  • energy awareness
  • 5G and beyond
  • zero-touch network
  • vertical services

Published Papers (7 papers)

Order results
Result details
Select all
Export citation of selected articles as:

Research

20 pages, 1312 KiB  
Article
Multi-Agent Reinforcement Learning Based Fully Decentralized Dynamic Time Division Configuration for 5G and B5G Network
by Xiangyu Chen, Gang Chuai and Weidong Gao
Sensors 2022, 22(5), 1746; https://doi.org/10.3390/s22051746 - 23 Feb 2022
Cited by 1 | Viewed by 1716
Abstract
Future network services must adapt to the highly dynamic uplink and downlink traffic. To fulfill this requirement, the 3rd Generation Partnership Project (3GPP) proposed dynamic time division duplex (D-TDD) technology in Long Term Evolution (LTE) Release 11. Afterward, the 3GPP RAN#86 meeting clarified [...] Read more.
Future network services must adapt to the highly dynamic uplink and downlink traffic. To fulfill this requirement, the 3rd Generation Partnership Project (3GPP) proposed dynamic time division duplex (D-TDD) technology in Long Term Evolution (LTE) Release 11. Afterward, the 3GPP RAN#86 meeting clarified that 5G NR needs to support dynamic adjustment of the duplex pattern (transmission direction) in the time domain. Although 5G NR provides a more flexible duplex pattern, how to configure an effective duplex pattern according to services traffic is still an open research area. In this research, we propose a distributed multi-agent deep reinforcement learning (MARL) based decentralized D-TDD configuration method. First, we model a D-TDD configuration problem as a dynamic programming problem. Given the buffer length of all UE, we model the D-TDD configuration policy as a conditional probability distribution. Our goal is to find a D-TDD configuration policy that maximizes the expected discount return of all UE’s sum rates. Second, in order to reduce signaling overhead, we design a fully decentralized solution with distributed MARL technology. Each agent in MARL makes decisions only based on local observations. We regard each base station (BS) as an agent, and each agent configures uplink and downlink time slot ratio according to length of intra-BS user (UE) queue buffer. Third, in order to solve the problem of overall system revenue caused by the lack of global information in MARL, we apply leniency control and binary LSTM (BLSTM) based auto-encoder. Leniency controller effectively controls Q-value estimation process in MARL according to Q-value and current network conditions, and auto-encoder makes up for the defect that leniency control cannot handle complex environments and high-dimensional data. Through the parallel distributed training, the global D-TDD policy is obtained. This method deploys the MARL algorithm on the Mobile Edge Computing (MEC) server of each BS and uses the storage and computing capabilities of the server for distributed training. The simulation results show that the proposed distributed MARL converges stably in various environments, and performs better than distributed deep reinforcement algorithm. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

16 pages, 1736 KiB  
Article
Effective TCP Flow Management Based on Hierarchical Feedback Learning in Complex Data Center Network
by Kimihiro Mizutani
Sensors 2022, 22(2), 611; https://doi.org/10.3390/s22020611 - 13 Jan 2022
Viewed by 1790
Abstract
Many studies focusing on improving Transmission Control Protocol (TCP) flow control realize a more effective use of bandwidth in data center networks. They are excellent ways to more effectively use the bandwidth between clients and back-end servers. However, these schemes cannot achieve the [...] Read more.
Many studies focusing on improving Transmission Control Protocol (TCP) flow control realize a more effective use of bandwidth in data center networks. They are excellent ways to more effectively use the bandwidth between clients and back-end servers. However, these schemes cannot achieve the total optimization of bandwidth use for data center networks as they do not take into account the path design of TCP flows against a hierarchical complex structure of data center networks. To address this issue, this paper proposes a TCP flow management scheme specified a hierarchical complex data center network for effective bandwidth use. The proposed scheme dynamically controls the paths of TCP flows by reinforcement learning based on a hierarchical feedback model, which obtains an optimal TCP flow establishment policy even if both the network topology and link states are more complicated. In evaluation, the proposed scheme achieved more effective bandwidth use and reduced the probability of TCP incast up to 30% than the conventional TCP flow management schemes: Variant Load Balancing (VLB), Equal Cost Multi Path (ECMP), and Intelligent Forwarding Strategy Based on Reinforcement Learning (IFS-RL) in the complex data center network. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

29 pages, 1249 KiB  
Article
5G Infrastructure Network Slicing: E2E Mean Delay Model and Effectiveness Assessment to Reduce Downtimes in Industry 4.0
by Lorena Chinchilla-Romero, Jonathan Prados-Garzon, Pablo Ameigeiras, Pablo Muñoz and Juan M. Lopez-Soler
Sensors 2022, 22(1), 229; https://doi.org/10.3390/s22010229 - 29 Dec 2021
Cited by 14 | Viewed by 4953
Abstract
Fifth Generation (5G) is expected to meet stringent performance network requisites of the Industry 4.0. Moreover, its built-in network slicing capabilities allow for the support of the traffic heterogeneity in Industry 4.0 over the same physical network infrastructure. However, 5G network slicing capabilities [...] Read more.
Fifth Generation (5G) is expected to meet stringent performance network requisites of the Industry 4.0. Moreover, its built-in network slicing capabilities allow for the support of the traffic heterogeneity in Industry 4.0 over the same physical network infrastructure. However, 5G network slicing capabilities might not be enough in terms of degree of isolation for many private 5G networks use cases, such as multi-tenancy in Industry 4.0. In this vein, infrastructure network slicing, which refers to the use of dedicated and well isolated resources for each network slice at every network domain, fits the necessities of those use cases. In this article, we evaluate the effectiveness of infrastructure slicing to provide isolation among production lines (PLs) in an industrial private 5G network. To that end, we develop a queuing theory-based model to estimate the end-to-end (E2E) mean packet delay of the infrastructure slices. Then, we use this model to compare the E2E mean delay for two configurations, i.e., dedicated infrastructure slices with segregated resources for each PL against the use of a single shared infrastructure slice to serve the performance-sensitive traffic from PLs. Also we evaluate the use of Time-Sensitive Networking (TSN) against bare Ethernet to provide layer 2 connectivity among the 5G system components. We use a complete and realistic setup based on experimental and simulation data of the scenario considered. Our results support the effectiveness of infrastructure slicing to provide isolation in performance among the different slices. Then, using dedicated slices with segregated resources for each PL might reduce the number of the production downtimes and associated costs as the malfunctioning of a PL will not affect the network performance perceived by the performance-sensitive traffic from other PLs. Last, our results show that, besides the improvement in performance, TSN technology truly provides full isolation in the transport network compared to standard Ethernet thanks to traffic prioritization, traffic regulation, and bandwidth reservation capabilities. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

16 pages, 398 KiB  
Article
Optimizing Traffic Engineering for Resilient Services in NFV-Based Connected Autonomous Vehicles
by Tuan-Minh Pham and Thi-Minh Nguyen
Sensors 2021, 21(24), 8446; https://doi.org/10.3390/s21248446 - 17 Dec 2021
Cited by 1 | Viewed by 2233
Abstract
The massive amount of data generated daily by various sensors equipped with connected autonomous vehicles (CAVs) can lead to a significant performance issue of data processing and transfer. Network Function Virtualization (NFV) is a promising approach to improving the performance of a CAV [...] Read more.
The massive amount of data generated daily by various sensors equipped with connected autonomous vehicles (CAVs) can lead to a significant performance issue of data processing and transfer. Network Function Virtualization (NFV) is a promising approach to improving the performance of a CAV system. In an NFV framework, Virtual Network Function (VNF) instances can be placed in edge and cloud servers and connected together to enable a flexible CAV service with low latency. However, protecting a service function chain composed of several VNFs from a failure is challenging in an NFV-based CAV system (VCAV). We propose an integer linear programming (ILP) model and two approximation algorithms for resilient services to minimize the service disruption cost in a VCAV system when a failure occurs. The ILP model, referred to as TERO, allows us to obtain the optimal solution for traffic engineering, including the VNF placement and routing for resilient services with regard to dynamic routing. Our proposed algorithms based on heuristics (i.e., TERH) and reinforcement learning (i.e., TERA) provide an approximation solution for resilient services in a large-scale VCAV system. Evaluation results with real datasets and generated network topologies show that TERH and TERA can provide a solution close to the optimal result. It also suggests that TERA should be used in a highly dynamic VCAV system. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

23 pages, 3462 KiB  
Article
An SDN-Based Solution for Horizontal Auto-Scaling and Load Balancing of Transparent VNF Clusters
by Alejandro Llorens-Carrodeguas, Irian Leyva-Pupo, Cristina Cervelló-Pastor, Luis Piñeiro and Shuaib Siddiqui
Sensors 2021, 21(24), 8283; https://doi.org/10.3390/s21248283 - 11 Dec 2021
Cited by 6 | Viewed by 3042
Abstract
This paper studies the problem of the dynamic scaling and load balancing of transparent virtualized network functions (VNFs). It analyzes different particularities of this problem, such as loop avoidance when performing scaling-out actions, and bidirectional flow affinity. To address this problem, a software-defined [...] Read more.
This paper studies the problem of the dynamic scaling and load balancing of transparent virtualized network functions (VNFs). It analyzes different particularities of this problem, such as loop avoidance when performing scaling-out actions, and bidirectional flow affinity. To address this problem, a software-defined networking (SDN)-based solution is implemented consisting of two SDN controllers and two OpenFlow switches (OFSs). In this approach, the SDN controllers run the solution logic (i.e., monitoring, scaling, and load-balancing modules). According to the SDN controllers instructions, the OFSs are responsible for redirecting traffic to and from the VNF clusters (i.e., load-balancing strategy). Several experiments were conducted to validate the feasibility of this proposed solution on a real testbed. Through connectivity tests, not only could end-to-end (E2E) traffic be successfully achieved through the VNF cluster, but the bidirectional flow affinity strategy was also found to perform well because it could simultaneously create flow rules in both switches. Moreover, the selected CPU-based load-balancing method guaranteed an average imbalance below 10% while ensuring that new incoming traffic was redirected to the least loaded instance without requiring packet modification. Additionally, the designed monitoring function was able to detect failures in the set of active members in near real-time and active new instances in less than a minute. Likewise, the proposed auto-scaling module had a quick response to traffic changes. Our solution showed that the use of SDN controllers along with OFS provides great flexibility to implement different load-balancing, scaling, and monitoring strategies. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

52 pages, 5086 KiB  
Article
On the Rollout of Network Slicing in Carrier Networks: A Technology Radar
by Jose Ordonez-Lucena, Pablo Ameigeiras, Luis M. Contreras, Jesús Folgueira and Diego R. López
Sensors 2021, 21(23), 8094; https://doi.org/10.3390/s21238094 - 3 Dec 2021
Cited by 14 | Viewed by 9767
Abstract
Network slicing is a powerful paradigm for network operators to support use cases with widely diverse requirements atop a common infrastructure. As 5G standards are completed, and commercial solutions mature, operators need to start thinking about how to integrate network slicing capabilities in [...] Read more.
Network slicing is a powerful paradigm for network operators to support use cases with widely diverse requirements atop a common infrastructure. As 5G standards are completed, and commercial solutions mature, operators need to start thinking about how to integrate network slicing capabilities in their assets, so that customer-facing solutions can be made available in their portfolio. This integration is, however, not an easy task, due to the heterogeneity of assets that typically exist in carrier networks. In this regard, 5G commercial networks may consist of a number of domains, each with a different technological pace, and built out of products from multiple vendors, including legacy network devices and functions. These multi-technology, multi-vendor and brownfield features constitute a challenge for the operator, which is required to deploy and operate slices across all these domains in order to satisfy the end-to-end nature of the services hosted by these slices. In this context, the only realistic option for operators is to introduce slicing capabilities progressively, following a phased approach in their roll-out. The purpose of this paper is to precisely help designing this kind of plan, by means of a technology radar. The radar identifies a set of solutions enabling network slicing on the individual domains, and classifies these solutions into four rings, each corresponding to a different timeline: (i) as-is ring, covering today’s slicing solutions; (ii) deploy ring, corresponding to solutions available in the short term; (iii) test ring, considering medium-term solutions; and (iv) explore ring, with solutions expected in the long run. This classification is done based on the technical availability of the solutions, together with the foreseen market demands. The value of this radar lies in its ability to provide a complete view of the slicing landscape with one single snapshot, by linking solutions to information that operators may use for decision making in their individual go-to-market strategies. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

25 pages, 1040 KiB  
Article
An Energy-Friendly Scheduler for Edge Computing Systems
by Alejandro Llorens-Carrodeguas, Stefanos G. Sagkriotis, Cristina Cervelló-Pastor and Dimitrios P. Pezaros
Sensors 2021, 21(21), 7151; https://doi.org/10.3390/s21217151 - 28 Oct 2021
Cited by 6 | Viewed by 2156
Abstract
The deployment of modern applications, like massive Internet of Things (IoT), poses a combination of challenges that service providers need to overcome: high availability of the offered services, low latency, and low energy consumption. To overcome these challenges, service providers have been placing [...] Read more.
The deployment of modern applications, like massive Internet of Things (IoT), poses a combination of challenges that service providers need to overcome: high availability of the offered services, low latency, and low energy consumption. To overcome these challenges, service providers have been placing computing infrastructure close to the end users, at the edge of the network. In this vein, single board computer (SBC) clusters have gained attention due to their low cost, low energy consumption, and easy programmability. A subset of IoT applications requires the deployment of battery-powered SBCs, or clusters thereof. More recently, the deployment of services on SBC clusters has been automated through the use of containers. The management of these containers is performed by orchestration platforms, like Kubernetes. However, orchestration platforms do not consider remaining energy levels for their placement decisions and therefore are not optimized for energy-constrained environments. In this study, we propose a scheduler that is optimised for energy-constrained SBC clusters and operates within Kubernetes. Through comparison with the available schedulers we achieved 23% fewer event rejections, 83% less deadline violations, and approximately a 59% reduction of the consumed energy throughout the cluster. Full article
(This article belongs to the Special Issue Automation and Learning in MEC and Cloud Systems)
Show Figures

Figure 1

Back to TopTop