The significant growth in demand for high-bandwidth network connectivity in recent years has presented pressing network capacity challenges to mobile operators despite the business opportunities it represents for them. Given the minimal return-on-capital investment, increasing the physical infrastructure is not a valuable solution. Therefore, network service providers are continuously searching for innovative and game-chancing solutions that can elevate the network capacity beyond the expected network performance levels.
Virtualization technology has gained significant momentum by offering operators a gateway to reducing the opportunity cost to reducing dependency on proprietary hardware. Despite the vast potential of virtualization technology and the promise it has, there are many questions that remain. For example, questions on the separation of network layers, virtualized network functions (VNFs) and services implementation/placement/dependencies and where all this fits within the telecommunication ecosystem are still to be answered. This has added significance given the recent trend towards adopting distributed computing environments.
This project focuses on developing different innovative optimization as well as intelligent frameworks for the placement, orchestration, and management of different VNFs/5G services while considering the characteristics of the existing workloads (e.g., VNFs, containers, V2X services), components relations, computational constraints, QoS requirements, security requirements, and service’s requests characteristics. Moreover, the goal is to use different data analytics and machine learning (ML) techniques to further automate the placement/orchestration/management process.
Within this project, there are two main activities. The first activity focuses on developing performance-aware, mobility-aware, and latency-aware placement/orchestration/management frameworks for next generation VNFs/5G services using novel optimization models with a broad range of objective functions while satisfying the aforementioned constraints. Additionally, scalable relaxation schemes are being designed to provide computationally efficient lower bound solutions for the proposed optimization models. Using these models and algorithms, decision on where to place and how to manage these VNFs/5G services can be made.
The second activity focuses on developing advanced data analysis and intelligent machine learning models that can learn and emulate optimality in real-time while avoiding the computational complexity of the optimal solutions. Using these models, decisions on the placement, orchestration, and management of VNFs/5G services is done in a dynamic, agile, effective, and intelligent manner.
2019-06-01 till 2023-06-01
Dr. Abdallah Shami, Dr. Abdallah Moubayed, Mr. Dimitrios Manias (Ph.D. Candidate), Mr. Sam Aleyadeh (Ph.D. Candidate), Mr. Ibrahim Shaer (Ph.D. Candidate), and Mr. Ibrahim Tamim (Ph.D. Candidate).