Smart Virtual Infrastructure

所要求的设备代表了智能数据驱动的网络化虚拟基础设施试验台的所有基本组件。可以对使用的设备进行调整，以模拟不同的连接设备和数字服务的行为，这些设备和数字服务以不同的流量和速度生成不同的流量类型。由此产生的研究将代表真实世界的场景，所取得的结果将提供所提议的协调、管理和安全框架的潜力和功能的真实视图。在接下来的内容中，我们将简要讨论研究计划的一般动机，并描述将从所要求的设备中受益的项目。 近年来，高带宽网络连接需求显著增长。由于网络容量在很大程度上依赖于底层网络基础设施，因此必须进行重大升级才能满足不断增长的需求。然而，从网络服务提供商(NSP)的角度来看，由于回报率较低，进行此类升级所需的资本投资并不理想。因此，世界各地的NSP都转向了网络虚拟化，这是一种可以将基于硬件的静态网络转变为基于软件的动态且强大的网络的技术。这种转换使NSP可以自由地摆脱专有的功能依赖型硬件，并能够将其资本支出减少高达88%。然而，必须解决许多挑战，以充分利用这一颠覆性技术的潜力。首先，必须优化底层虚拟基础设施，以确保其能够满足不断增长的需求和不断变化的数字服务格局。然后，必须考虑在优化的虚拟基础设施上运行的数字服务的优化。除了性能要求外，信息安全仍然是NSP面临的风险最高的挑战，也是客户的主要担忧和采用障碍。例如，随着虚拟基础设施和数字服务的采用，分布式拒绝服务(DDoS)攻击有所增加，2019年第一季度DDoS攻击比前一年增加了84%。 该研究项目旨在通过优化底层虚拟基础设施和覆盖的数字服务，采用双管齐下的基于数据的联网方法。为此，第一个目标是为虚拟基础设施管理开发新的可感知服务质量的智能框架，以维护服务质量要求，遵守网络服务提供商的目标，并提供弹性故障恢复方法。第二个目标是通过使用数学模型和基于ML的模型开发新的数字服务优化、管理和安全框架，为覆盖虚拟基础设施的数字服务设计可感知服务质量的安全框架。

The requested equipment represents all the basic components of an intelligent data-driven networked virtual infrastructure testbed. The equipment used can be adapted to mimic the behavior of different connected devices and digital services that generate varying traffic types at varying volumes and velocities. The resulting research will be representative of real-world scenarios with the achieved results providing a realistic view of the potential and capabilities of the proposed orchestration, management, and security frameworks. In what follows, we briefly discuss the general motivation of the research program and describe the projects that will benefit from the requested equipment. Recent years have seen significant growth in high-bandwidth network connectivity demand. As network capacity greatly depends on the underlying network infrastructure, significant upgrades must be made to meet the growing demand. However, from a Network Service Providers (NSPs) perspective, the capital investment required to make such upgrades is not ideal due to the low rate of return. Therefore, NSPs worldwide have turned towards Network Virtualization, a technology that can transform a static hardware-based network into a dynamic and robust software-based network. This conversion gives NSPs the freedom to move away from proprietary function-dependent hardware and the ability to reduce their capital expenditures by up to 88%. However, many challenges must be addressed to harness the full potential of this disruptive technology. First, the underlying virtual infrastructure must be optimized to ensure its capability to meet both the growing demand and the changing digital service landscape. Then, the optimization of the digital services running on the optimized virtual infrastructure must be considered. In addition to the performance requirements, Information Security remains the challenge with highest risk for NSPs as well as being the clients' main concern and adoption obstacle. For example, Distributed Denial-of-Service (DDoS) attacks have increased with the adoption of virtual infrastructure and digital services as the first quarter of 2019 saw an 84 percent increase in DDoS attacks over the previous year. This research project aims to adopt a two-pronged data-based networking approach through the optimization of both the underlying virtual infrastructure and the overlaying digital services. To that end, the first objective is to develop new QoS-aware and intelligent frameworks for virtual infrastructure management that maintain QoS requirements, adhere to NSPs objectives, and offer resilient failure recovery methods. The second objective is to design QoS-aware and secure frameworks for digital services overlaying the virtual infrastructure by developing novel digital service optimization, management, and security frameworks using mathematical and ML-based models.