Intelligent Management Platform - The Future of Cloud and Networking

分布式数据服务的复杂性正在迅速增加，特别是随着移动计算和物联网(IoT)的发展，数据量呈爆炸式增长。网络功能虚拟化(NFV)、5G网络、软件定义网络(SDN)、机器学习(ML)、边缘计算和人工智能(AI)等新技术将危及新的数据孤岛，并需要使网络和计算更简单的解决方案。因此，业务成功的关键是对用户需求的即时响应和极短的上市时间。我的研究旨在充分利用ML算法和人工智能，为云、边缘计算和NFV/SDN提供复杂的实时和历史分析、全面的实时操作和自动化能力。我的研究将为物联网应用的NFV组件和SDN控制器构建一个分布式智能计算协调框架。这项拟议工作的主要目标是通过预测/检测服务质量(Qos)感知问题(即延迟/安全感知应用)、自动化资源管理以及跨网络基础设施实时预测/平衡工作负载，实现NFV和SDN协调的自动化，并动态增强其性能。它还将专注于使用ML算法管理NFV组件和SDN控制器，同时促进云边缘模型的弹性(向上/向下扩展)，并在保持可扩展性、安全性和弹性的同时处理用户流量的增加。该提案将首先调查网络实体的功能以及微服务体系结构如何映射到NFV和SDN体系结构。然后，它将致力于开发基于云的智能协调方法，以首先促进大规模NFV应用程序和SDN控制器的有效设计和管理，其次将信息技术团队的工作降至最低，以跟上服务级别协议以及快速变化的技术和业务模式的需求。该计划将开发ML模型，以自动在云和边缘部署之间进行选择，特别是针对这些技术的任务关键型应用程序(即，健康感知应用程序)。这项研究将探索针对邻近检测、网络感知、资源自我发现和模式执行的分析，以提高物联网系统的自给自足能力。它还将使用预测分析来根据预测的性能和安全需求/警报管理资源，根据实时流量特征动态更新策略和规则。这项研究提案将基于开源技术，并将整合团队在网络、云计算、分布式系统、ML、数据挖掘和收集、应用和数据建模、可视化技术和管理能力方面的实用和技术专业知识。

The complexity of distributed data services is increasing rapidly, especially with the development of mobile computing and the Internet of Things (IoT) where the data volume is exploding. New technologies; Network Functions Virtualization (NFV), 5G Networks, Software-Defined Networking (SDN), Machine Learning (ML), Edge Computing, and Artificial Intelligence (AI); will jeopardize new data silos and require a solution that makes the network and computing simpler. Therefore, the key to business success is the immediate response to user demands and very low time-to-market. My research will aim at taking full advantage of ML algorithms and AI to empower the Cloud, Edge Computing, and NFV/SDN with sophisticated real-time and historical analytics, comprehensive real-time operations, and automation capabilities. My research will build a distributed intelligent computing orchestration framework for NFV components and SDN controllers for IoT applications. The main objective of this proposed work is automating the NFV and SDN orchestration and dynamically enhancing their performance through predicting/detecting the quality of service (QoS)-aware issues (i.e. latency/security-aware applications), automated resource management, and predicting/balancing workloads in real-time across the network infrastructure. It will also focus on using ML algorithms to manage NFV components and SDN controllers while facilitating elasticity (scaling up/down) across the cloud-edge model and handle the increase in users' traffic while maintaining scalability, security, and resiliency. This proposal will start by investigating the network entities' functionalities and how microservices architecture is mapped to NFV and SDN architectures. It will then aim at developing an intelligent cloud-based orchestration approaches to first, facilitate an effective design and management of large-scale NFV applications and SDN controllers and secondly, minimize the work of the information technology team to keep up with the service level agreements and the needs of the rapidly changing technologies and business model. This proposal will develop ML models to automate the choice between cloud and edge deployments especially for mission-critical applications (i.e. health-aware applications) of these technologies. This research will explore analytics for proximity detection, network awareness, self-discovery of resources, and execution of patterns to increase the self-sufficiency of the IoT systems. It will also use predictive analytics to manage resources based on predicted performance and security needs/alarms, dynamically update policies and rules based on real-time traffic characteristics. This research proposal will be based on open source technologies and will integrate the team's practical and technical expertise in networking, cloud computing, distributed systems, ML, data mining and collection, application and data modeling, visualization techniques, and management capabilities.