The intelligent multi‐domain service function chain deployment: Architecture,challenges and solutions

Network function virtualization (NFV) technology achieves flexible service deployment by replacing the middleboxes with virtual network functions (VNFs). In NFV, a set of VNFs are chained in a given order, called service function chain (SFC), and accordingly, data flow is steered to traverse all the VNFs in order to offer a service. With a large number of network devices and end users being connected into Internet, there is a growing demand for large‐scale multi‐domain networks to dynamically deploy the SFC across multiple network domains, in order to support efficient service provisioning. To this end, in this paper, we first investigate the state of the art of multi‐domain SFC deployment, and then propose an intelligent multi‐domain SFC deployment (IMSD) architecture by leveraging software‐defined networking (SDN), NFV, and deep learning technologies. Furthermore, we discuss the potential challenges to realize the IMSD and provide some promising solutions.     

基于SDN/NFV的未来网络实验平台

依据SDN/NFV集中管控、动态、灵活、高效、可编排等特点,提出了基于SDN/NFV技术的未来网络实验平台的构建方案.该平台主要采用OpenStack和OpenDaylight的开源架构,同时研发设计SDN跨域虚拟网络通信、虚拟网元管理以及网络服务编排三大关键技术,实现了底层异构资源的实时动态管理与开放共享.同时,根据实验用户对网络资源的需求,灵活按需编排各种网络资源与SFC服务,为用户提供端到端的网络实验验证服务.     

Separating VNF and Network Control for Hardware‐Acceleration of SDN/NFV Architecture

A hardware‐acceleration architecture that separates virtual network functions (VNFs) and network control (called HSN) is proposed to solve the mismatch between the simple flow steering requirements and strong packet processing abilities of software‐defined networking (SDN) forwarding elements (FEs) in SDN/network function virtualization (NFV) architecture, while improving the efficiency of NFV infrastructure and the performance of network‐intensive functions. HSN makes full use of FEs and accelerates VNFs through two mechanisms: (1) separation of traffic steering and packet processing in the FEs; (2) separation of SDN and NFV control in the FEs. Our HSN prototype, built on NetFPGA‐10G, demonstrates that the processing performance can be greatly improved with only a small modification of the traditional SDN/NFV architecture.     

基于SDN/NFV的核心网演进关键技术研究

在NFV和SDN的基础上研究核心网架构的演进及应用领域;通过将NFV和SDN技术引入核心网,在核心网虚拟化,分组域网关控制转发分离及基于Service Chaining的即插即用的增值业务及复杂流处理平台进行研究,建议了基于SDN/NFV的核心网长期演进架构。     

Towards SDN/NFV-enabled satellite networks

The telecom community is during the last years witnessing a paradigm shift towards the virtualisation/ “softwarisation” of the network infrastructure, mostly driven by the concepts of software defined networking (SDN) and network functions virtualisation (NFV). At the same time, satellite telecommunication technologies, although rapidly advancing in various fields, have not so far adequately followed this trend. The paper investigates the applicability of SDN and NFV technologies to satcom platforms and determines the benefits and the challenges associated with the integration of satellite infrastructures into future software-based networks. To that end, it identifies specific use cases which clearly benefit from the softwarisation of the satcom network and proposes a functional architecture for federated satellite-terrestrial software-based networks. The architecture is implemented in a lab environment which is used to validate some of the proposed use cases. Furthermore, a techno-economic analysis is conducted, which presents clear economic benefits via the introduction of SDN and NFV in the satcom ecosystem.     

网络功能虚拟化场景下的并行加速

在网络功能虚拟化场景下,每个网络功能都以软件的形式来实现。但传统的串行网络功能组链方式将带来极大的性能损耗。而如今针对网络功能进行加速的工作,主要集中在优化该串行链的每个组成成分上。提出了一个高性能的系统框架,通过让网络功能并行地对数据分组进行处理,提高网络功能虚拟化场景下服务链的整体性能。该系统由 3 个部分组成。首先,该系统为管理员提供了一个策略描述语言来直观地描述串行或并行的组链意图。然后,该系统的编排器智能地鉴别网络功能之间的依赖性,并基于所提供的策略,生成高性能的“服务图”。最后,该系统的底层实现通过执行轻量级的数据分组复制、分布式的并行分组转发和负载均衡后的数据分组合并来支持网络功能的并行处理。基于DPDK技术,在Linux容器中实现了该系统的原型机。通过实验验证可知,该系统能极大地减少真实世界中服务链的处理时延。     

面向随选网络的业务编排系统研发实践

为满足SDN/NFV时代业务快速上线的要求,需要对现有网络运营系统进行重构,以随选网络业务的实践为背景,详细介绍了编排器在运营系统中的定位与作用,然后阐述了编排器的核心设计思路以及核心组件工作流引擎.     

网络切片技术下的5G全新网络架构

移动互联网、物联网的快速发展,对5G移动通信提出了更高的要求。文章先分析了移动通信网络架构IP化、扁平化的发展趋势、4G网络存在问题与5G网络需求;在此基础上分析了全新的5G网络架构,重点剖析了基于SDN/NFV的通用平台、端到端的网络切片和服务化的架构SBA等新特性。     

基于深度确定性策略梯度的虚拟网络功能迁移优化算法

针对NFV/SDN架构下,服务功能链(SFC)的资源需求动态变化引起的虚拟网络功能(VNF)迁移优化问题,该文提出一种基于深度强化学习的VNF迁移优化算法。首先,在底层CPU、带宽资源和SFC端到端时延约束下,建立基于马尔可夫决策过程(MDP)的随机优化模型,该模型通过迁移VNF来联合优化网络能耗和SFC端到端时延。其次,由于状态空间和动作空间是连续值集合,提出一种基于深度确定性策略梯度(DDPG)的VNF智能迁移算法,从而得到近似最优的VNF迁移策略。仿真结果表明,该算法可以实现网络能耗和SFC端到端时延的折中,并提高物理网络的资源利用率。     

Orchestration mechanism for VNF hardware acceleration resources in SDN/NFV architecture

The hardware acceleration mechanism for VNF (virtual network function) is recently a hot research topic in SDN/NFV architecture because of the low processing performance of VNF.Once hardware acceleration resources have been plugged into the network,how to optimally mange and orchestrate these resources under service requirements remains a question to be solved.Firstly,a unified management architecture based on separated control for hardware acceleration resources was proposed.Then,traditional network topology together with hardware acceleration resources were modeled into a unified network model and then the hardware acceleration resource orchestration problem was transferred into a multi-objective linear programming problem.Finally,a hardware-accelerator-card-prior’ heuristic algorithm was designed.Experimental results show that compared with existing methods,the proposed orchestration mechanism can efficiently manage hardware acceleration resources and reduce the processing latency by 30%.     

SDN/NFV——机遇和挑战

SDN/NFV是电信业受云计算、虚拟化技术的深化发展趋势影响,在产业内进行网络IT化的尝试。本文分析了SDN/NFV对于电信领域的商业模式、技术趋势的影响,以及对于运营商、不同层面供应商的影响,并对其带来的机遇和挑战进行了分析和研究。     

IMS核心网网络功能虚拟化方案研究

网络功能虚拟化将传统网络设备的软件功能和硬件功能解耦合,从而降低网络的投资和运营成本,提高新业务的部署效率。SDN为网络功能虚拟化的承载网提供了解决方案,确保网络功能虚拟化的实现。主要对IMS核心网网络功能虚拟化方案进行了论述,并基于SDN架构提出了IMS核心网虚拟化环境下的承载网组网方案。     

Mobile Edge Computing Platform with Container-Based Virtualization Technology for IoT Applications

The next generation of fifth generation (5G) network, implementing mobile edge computing (MEC), network function virtualization (NFV) and software defined networking technologies, establishes a flexible and resilient network in line with various internet of things (IoT) devices. While NFV adds flexibility scale in or out networks by allowing network functions to be dynamically deployed and inter-connected, MEC provide intelligence at the edge of a mobile network; reduces latency, and increases capacity. With the diverse development of networking applications, the proposed MEC with container-based virtualization technology (CVT) as IoT gateway with IoT devices for flow control mechanism in scheduling and analysis methods will effectively enhance the quality of service. In this work, the proposed IoT gateway will be analyzed to elucidate the combined effect of simultaneously deploying virtual network functions and MEC applications on the same network infrastructure. Low latency, high bandwidth and high agility, supporting the connection of large-scale devices, and the efficient combination of resources from network edge and cluster clouds, account for real-time network conditions, reducing the IoT applications and services to indicate that a number is the average of 30% of the latency, that could get more suitable service quality to develop such as both augmented reality and virtual reality application intelligence in coming 5G network.

     

一种5G网络切片的编排算法

网络切片是基于SDN/NFV的5G网络架构实现按需组网的一种重要技术.通过分析5G主要场景,提出了SDN/NFV架构下一种基于GA-PSO优化的网络切片编排算法.该算法利用粒子群算法能够快速收敛于全局最优解的特性,设计网络切片性能的评价函数.并且利用遗传算法快速随机搜索的能力,实现对网络切片的更新和优化,利用粒子群追逐局部最优解与全局最优解得到最优网络切片.仿真实验结果表明,该算法能够实现对多业务场景网络切片的个性化创建,充分发挥SDN的集中控制方式的优点,在降低网络能耗的同时,提高网络资源利用率.     

SDN/NFV‐based handover management approach for ultradense 5G mobile networks

With the great increase of connected devices and new types of applications, mobile networks are witnessing exponential growth of traffic volume. To meet emerging requirements, it is widely agreed that the fifth‐generation mobile network will be ultradense and heterogeneous. However, the deployment of a high number of small cells in such networks poses challenges for the mobility management, including frequent, undesired, and ping‐pong handovers, not to mention issues related to increased delay and failure of the handover process. The adoption of software‐defined networking (SDN) and network function virtualization (NFV) technologies into 5G networks offers a new way to address the above‐mentioned challenges. These technologies offer tools and mechanisms to make networks flexible, programmable, and more manageable. The SDN has global network control ability so that various functions such as the handover control can be implemented in the SDN architecture to manage the handover efficiently. In this article, we propose a Software‐Defined Handover (SDHO) solution to optimize the handover in future 5G networks. In particular, we design a Software‐Defined Handover Management Engine (SDHME) to handle the handover control mechanism in 5G ultradense networks. The SDHME is defined in the application plane of the SDN architecture, executed by the control plane to orchestrate the data plane. Simulation results demonstrate that, compared with the conventional LTE handover strategy, the proposed approach significantly reduces the handover failure ratio and handover delay.     

一种基于Viterbi算法的虚拟网络功能自适应部署方法

为了应对移动数据流量的爆炸性增长,5G移动通信网将引入新型的架构设计。软件定义网络和网络功能虚拟化是网络转型的关键技术,将驱动移动通信网络架构的创新,服务链虚拟网络功能的部署是网络虚拟化研究中亟待解决的问题。该文针对已有部署方法未考虑服务链中虚拟网络功能间顺序约束和移动业务特点的问题,提出一种基于Viterbi算法的虚拟网络功能自适应部署方法。该方法实时感知底层节点的资源变化并动态调整拓扑结构,采用隐马尔科夫模型描述满足资源约束的可用的底层网络节点拓扑信息,基于Viterbi算法在候选节点中选择时延最短的服务路径。实验表明,与其它的虚拟网络功能部署方法相比,该方法降低了服务链的服务处理时间,并提高了服务链的请求接受率和底层资源的成本效率。     

基于改进深度强化学习的虚拟网络功能部署优化算法

针对网络功能虚拟化/软件定义网络 (NFV/SDN)架构下,网络服务请求动态到达引起的服务功能链(SFC)部署优化问题,该文提出一种基于改进深度强化学习的虚拟网络功能(VNF)部署优化算法。首先,建立了马尔科夫决策过程 (MDP)的随机优化模型,完成SFC的在线部署以及资源的动态分配,该模型联合优化SFC部署成本和时延成本,同时受限于SFC的时延以及物理资源约束。其次,在VNF部署和资源分配的过程中,存在状态和动作空间过大,以及状态转移概率未知等问题,该文提出了一种基于深度强化学习的VNF智能部署算法,从而得到近似最优的VNF部署策略和资源分配策略。最后,针对深度强化学习代理通过ε贪婪策略进行动作探索和利用,造成算法收敛速度慢等问题,提出了一种基于值函数差异的动作探索和利用方法,并进一步采用双重经验回放池,解决经验样本利用率低的问题。仿真结果表示,该算法能够加快神经网络收敛速度,并且可以同时优化SFC部署成本和SFC端到端时延。     

智能专线业务编排系统架构设计与实现

在SDN/NFV技术时代,为满足新型网络业务快速自动开通的目标,提出一种智能专线业务编排系统的架构设计和实现方案。按照各功能松耦合、灵活部署的特点,将编排器系统架构设计为业务入口层、网络服务层和适配层三层解耦的架构体系实现。将编排器工作流程拆分为业务准备流程和业务下发流程两部分实现。通过开发实践验证,编排器作为核心系统,基本满足最初对业务编排的业务需求和运维管理需求。     

Intent‐based service management for heterogeneous software‐defined infrastructure domains

One of the main challenges in delivering end‐to‐end service chains across multiple software‐defined networking (SDN) and network function virtualization (NFV) domains is to achieve unified management and orchestration functions. A very critical aspect is the definition of an open, vendor‐agnostic, and interoperable northbound interface (NBI) that should be as abstract as possible and decoupled from domain‐specific data and control plane technologies. In this paper, we propose a reference architecture and an intent‐based NBI for end‐to‐end service management across multiple technological domains. The general approach is tested in a heterogeneous OpenFlow/Internet‐of‐Things (IoT) SDN test bed, where the proposed solution is applied to a rather complex service provisioning scenario spanning three different technological domains: an IoT infrastructure deployment, a cloud‐based data collection, processing, and publishing platform, and a transport domain over a geographic network interconnecting the IoT domain and the data center hosting the cloud services.     

一种基于分层架构的端到端QoS保障系统与方法

随着互联网+、NFV/SDN、5G/MEC、业务编排器的快速发展,业界对端到端质量保障需求旺盛,通过研究一种基于分层架构的端到端QoS保障系统与方法,采用新的QoS质量保障系统,通过省级业务编排器,与全国业务编排器、省分/地(市)核心云、地(市)边缘云、省内核心网、城域网、传输网、宽带网对接为全网提供端到端的全业务QoS保障,并进行云网协同的QoS保障,提升业务体验,为各类应用所服务的用户提供动态带宽加速、QoS保障等服务。