Intelligent Strategies for Optimal Virtual Network Function Placement in 5G Core

拟议的项目旨在为5G核心网部署开发新算法，以满足关键性能指标（kpi）。最终目标是在满足延迟要求的同时，最大限度地减少5G分组核心中的虚拟网络功能数量。因此，该项目旨在解决5G网络中延迟和运营成本之间的权衡，预计到2025年5G网络将覆盖82%的人口。更高的速度、更低的延迟和更大的5G网络容量将使直播和增强现实得到改善。这些改进将使联网汽车、高级游戏、触觉互联网、远程医疗等各种服务成为可能。拟议的项目将遵循围绕三个主要研究主题的整体方法：1)开发实时可行的算法，用于在5G分组核心中放置用户和控制平面虚拟网络功能（VNF）， 2) VNF放置的优化和人工智能支持模型，以及3)测试平台实现和验证所开发的算法。实现这些方案将能够满足高效和低延迟5G网络的核心逻辑组件。提出的研究主题新颖，响应了5G分组核心管理和控制需求日益增长的新需求。因此，该项目依赖于启发式、优化和基于机器学习的规划方案，为每个研究主题开发智能且具有成本效益的策略、技术、算法和方法。这将节省因无法满足延迟需求而导致的SLA违规恢复费用，并节省在网络核心上放置许多VNFs的运营支出，从而带来数百万美元的收益。此外，在解决合作伙伴公司Ciena的延迟和高效管理需求的同时，加拿大各地的许多企业将受益于该成果，因为它们将满足加拿大信息和通信技术（ICT）行业的需求，从而在5G支持的人工智能5G网络和生态系统中处于领先地位。

The proposed project aims at developing novel algorithms for the 5G Core Network Placement to meet Key Performance Indicators (KPIs). The ultimate goal is to minimize the number of virtual network functions in 5G packet core while meeting the latency requirements. Thus, this project aims to address the trade-off between latency and operational costs in a 5G network, which is expected to cover 82% of the population by 2025. Higher speed, lower latency and greater capacity of 5G networks will enable improved livestreaming and improved augmented reality. These improvements will enable various services such as connected cars, advanced gaming, tactile Internet, remote healthcare and many others. The proposed project will follow a holistic approach anchored around three main research themes: 1) Development of real time and feasible algorithms for user and control plane virtual network function (VNF) placement in the 5G packet core, 2) Optimization and AI-enabled models for VNF placement , and 3) Test bed implementation and verification of the developed algorithms. Achieving each of these schemes will enable meeting the core logical components of efficient and low latency 5G networks. The proposed research themes are novel and respond to the emerging needs of the increasing demands in the management and control of 5G packet core. Accordingly, this project relies on heuristic, optimization and machine learning-based planning schemes to develop intelligent and cost-efficient strategies, techniques, algorithms, and methods for each of the studied themes. This will consequently result in million dollars of gains by saving from what is spent for recovering SLA violations due to not being able to meet latency requirements, as well as savings from operational expenditures for placing many VNFs across the network core. Furthermore, while addressing the latency and efficient management needs of the partner company, Ciena, many businesses across the country will benefit from the outcomes as they will serve for the needs of the Canadian Information and Communication Technology (ICT) sector to take the lead in the AI-enabled 5G networks and ecosystems enabled by 5G.