Using machine learning for dynamic resource orchestration & task scheduling in a radio access network based edge environment - Doctor of Philosophy thesis

Aug 2, 2026

Upload by Jiven

Mobile wireless generations continue to deliver new capabilities, with 5G enabling cloud-native services through speed, bandwidth, and connectivity. Many such applications—e.g., virtual reality, V2X, AI, and video analytics—require ultra-low latency (≤10 ms) and higher bandwidth, which conventional cloud and radio access network setups often fail to satisfy. The thesis proposes MECRAN (Multi-Access Edge Computing with Cloud Radio Access Networks), leveraging edge computing and machine learning to optimize low-latency round-trip delivery via dynamic task and application scheduling near the user at the radio edge.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages224

Reading Time9 h 24 min

AccessFree to download

Abstract

Background: 5G and edge requirements

Proposed approach: MECRAN

Keywords and terminology

Related published abstracts

Thesis context and acknowledgements

Dedication

Acknowledgements

Using machine learning for dynamic resource orchestration & task scheduling, in a radio access network based edge environment Jude Kojo Amponsah Fletcher St Cross College University of Oxford A thesis submitted to the Department of Engineering Science, in fulﬁlment of the requirements for the degree of Doctor of Philosophy Trinity 2021 Abstract The evolution in mobile wireless communication generations, continues to gift the world with new telecommunication capabilities towards how people live and work. For instance, the roll out of the ﬁfth generation (5G) promises a muchenhanced cloud-native application service, through improved communication speed, higher bandwidth, improved capacity for more connected devices and many more. These exciting new 5G features have given numerous vertical and horizontal industries a reason to explore diﬀerent ways of delivering value, through emerging cloud-native applications that run on access devices (i.e. the Internet of Things (IoTs) and mobile devices) . These cloud-native applications such Virtual Reality, Vehicle to everything communication (V2X), artiﬁcial intelligence, video analytics and so on however have strict performance requirements for extremely low latency (10 milliseconds and below) and higher bandwidth, that 5G alone cannot deliver. Unfortunately, the traditional cloud computing and radio access network setups do not suﬃciently address these key communication needs which are crucial to the performance of these cloud-native applications. There is therefore an urgency to enhance and optimize the traditional mobile telecommunication network architecture, to meet these performance needs. This thesis seeks to demonstrate how we have developed a facility called Multi-Access Edge Computing with Cloud Radio Access Networks (MECRAN) to address this issue. MECRAN is based on the edge computing paradigm and uses machine learning to optimize the round-trip delivery of data at low latency, by dynamically scheduling cloud-native applications, to run in close proximity to a mobile user, at the edge of the radio access network. Keywords: Multi-Access Edge Computing, Mobile Edge Computing, Cloud-Based/ Centralized Radio Access Network (C-RAN), Artiﬁcial Intelligence, Federated Cloud, Security, Containers, Task Scheduling, Ceph, Cloud Computing, 5G, Energy Reduction, MEC Orchestration, eMBB, URLLC, mMTC Related Published Abstracts Whilst it is not the intention to submit this DPhil thesis in an integrated format, it is useful to highlight that this document extends some of the content published and / or submitted to international conferences and journals throughout the course of my study. I conﬁrm all material including the core source codes are my original work and I remain the ﬁrst author on all publications. [1] Jude Fletcher, David Wallom. 2019. Deep Learning based task scheduling in a Cloud RAN enabled edge environment. SEC’19: The Fourth ACM/IEEE Symposium on Edge Computing Proceedings, November 7 {9, 2019, Arlington, VA, USA, 3 pages. [https://doi.org/10.1145/3318216.3363319](https://doi.org/10.1145/3318216.3363319) [2] Jude Fletcher, David Wallom. 2019. Using machine learning to orchestrate cloud resources in a RAN enabled edge environment. SenSys ’19: The 17th ACM Conference on Embedded Networked Sensor Systems Proceedings, November 10 13, 2019, New York, NY, USA, 3 pages. [https://doi.org/10.1145/3356250.3361930](https://doi.org/10.1145/3356250.3361930)[ ](https://doi.org/10.1145/3356250.3361930)Abstract [1] was also accepted in a workshop at the Conference on Neural Information Processing (NeurIPS 2019) . Two additional abstracts were also accepted atthe 12th IEEE/ACM International Conference on Utility and Cloud Computing Conference (UCC, Auckland, New Zealand) and at the 12th IEEE International Conference on Service-Oriented Computing and Applications (SOCA, Kaohsiung, Taiwan) in 2019 . I was unable to attend both UCC and SOCA conferences as they coincided

FAQs

Why do cloud-native applications need edge computing in a radio access network environment?

They require extremely low latency and higher bandwidth, which traditional cloud and conventional radio access network setups do not adequately provide.

What is MECRAN and what role does machine learning play?

MECRAN combines edge computing with Cloud Radio Access Networks and uses machine learning to optimize low-latency round-trip delivery by dynamically scheduling applications at the network edge.

How does the thesis address dynamic resource orchestration and task scheduling?

It schedules cloud-native applications to run close to the mobile user at the edge of the radio access network, optimizing delivery performance under latency constraints.

Aug 2, 2026

Upload by Jiven