Joint service function chain deploying and path selection for bandwidth saving and VNF reuse

Service function chain (SFC) is managed as a set of chained virtual network functions (VNFs) in a particular order. The efficiency of a network closely depends on how these VNFs are deployed in the network. In this paper, we study the joint problem of service function chain deploying and path selection for bandwidth saving and VNF reuse. We first describes the problem as a multiobjective and multirestriction problem, and then a heuristic service function chain deployment algorithm based on longest function assignment sequence is proposed. Via simulations, we show that our design can reduce the number of deployed VNFs and link bandwidth requirement jointly and serve more requests than other algorithms.     

基于联合备份的服务功能链可靠性保障的部署方法

在网络功能虚拟化(NFV)环境中,针对服务功能链(SFC)部署时的可靠性问题,该文提出对备份虚拟网络功能选择、备份实例放置和服务功能链部署的联合优化方法。首先,定义一个单位开销可靠性提高值的虚拟网络功能衡量标准,改进备份虚拟网络功能选择方法;其次,采用联合备份的方式调整相邻备份实例之间的放置策略,以降低带宽资源开销;最后,将整个服务功能链可靠性保障的部署问题构建成整数线性规划模型,并提出一种基于最短路径的启发式算法,克服整数线性规划求解的复杂性。仿真结果表明,该方法在优先满足网络服务可靠性需求的同时,优化资源配置,提高了请求接受率。     

基于强化学习的5G网络切片虚拟网络功能迁移算法

针对5G网络切片架构下业务请求动态性引起的虚拟网络功能(VNF)迁移优化问题,该文首先建立基于受限马尔可夫决策过程(CMDP)的随机优化模型以实现多类型服务功能链(SFC)的动态部署,该模型以最小化通用服务器平均运行能耗为目标,同时受限于各切片平均时延约束以及平均缓存、带宽资源消耗约束。其次,为了克服优化模型中难以准确掌握系统状态转移概率及状态空间过大的问题,该文提出了一种基于强化学习框架的VNF智能迁移学习算法,该算法通过卷积神经网络(CNN)来近似行为值函数,从而在每个离散的时隙内根据当前系统状态为每个网络切片制定合适的VNF迁移策略及CPU资源分配方案。仿真结果表明,所提算法在有效地满足各切片QoS需求的同时,降低了基础设施的平均能耗。     

一种面向运营成本优化的虚拟网络功能部署和路由分配策略

随着网络功能虚拟化(NFV)技术的发展,虚拟网络功能(VNF)可以通过服务功能链(SFC)的形式部署在如虚拟机的通用平台中,为管理带来灵活性。但是对于服务提供商来说,由于网络基础设施的复杂性和日益增长的服务需求,给VNF的部署带来了高昂的运营成本(OPEX)。针对此问题,该文提出一种面向OPEX优化的策略,旨在最小化OPEX中的激活、能耗和传输成本,得到VNF部署和路由分配优化方案。为此建立一种全新的混合整数线性规划(MILP)模型,并设计包括遗传算法(GA)在内的3种OPEX优化算法。仿真实验评估在不同资源配给下MILP和3种算法的OPEX及其性能,其中GA算法在节点资源配比60%以上时可以得到近似于MILP模型的解决方案。     

基于流量演化感知的服务功能链在线弹性编排策略

随着网络功能虚拟化（NFV）的引入，运营商能够提供更为灵活的网络服务.然而现有服务功能链（SFC）编排局限于静态反应式策略，业务流量发生变化时网络资源供应量难以匹配负载需求，虚拟网络功能（VNF）频繁部署与迁移，运营开销增大.针对上述问题，该文提出一种基于流量演化感知的服务功能链在线弹性编排策略（OEOP），该策略将在线学习引入到SFC流量演化感知的过程，预先确定细粒度的VNF弹性需求.此外，以实时更新的SFC路径与节点负载两因子为导向，完成新增VNF的在线弹性部署，代替VNF迁移应对系统负载变化.仿真表明，该策略明显增强了虚拟资源供应量与负载需求的匹配特性，VNF吞吐量利用率提高10.2%~24.8%，运营开销平均降低26.7%.     

基于深度强化学习的服务功能链可靠部署算法

针对当前关于服务功能链(SFC)的部署问题都未考虑到虚拟网络功能(VNF)的失效重要度,该文提出了基于深度强化学习的SFC可靠部署算法。首先建立VNF和虚拟链路可靠映射模型,为重要的VNF设置高可靠性需求,并通过链路部署长度限制尽可能保证虚拟链路可靠性需求。其次,以负载均衡为资源协调原则,与VNF可靠性联合优化,最终使用深度强化学习得到服务功能链部署策略。另外,提出了基于重要度的节点备份和链路备份策略,用于应对部署过程中VNF/链路可靠性难以满足的情况。仿真结果表明,该文的可靠部署算法在保证可靠性需求的基础上能够有效减少SFC失效损失,同时使虚拟网络更加稳定可靠。     

5G网络切片场景中基于预测的虚拟网络功能动态部署算法

针对无线虚拟化网络在时间域上业务请求的动态变化和信息反馈时延导致虚拟资源分配的不合理,该文提出一种基于长短时记忆(LSTM)网络的流量感知算法,该算法通过服务功能链(SFC)的历史队列信息来预测未来负载状态。基于预测的结果,联合考虑虚拟网络功能(VNF)的调度问题和相应的计算资源分配问题,提出一种基于最大最小蚁群算法(MMACA)的虚拟网络功能动态部署方法,在满足未来队列不溢出的最低资源需求的前提下,采用按需分配的方式最大化计算资源利用率。仿真结果表明,该文提出的基于LSTM神经网络预测模型能够获得很好的预测效果,实现了网络的在线监测;基于MMACA的VNF部署方法有效降低了比特丢失率的同时也降低了整体VNF调度产生的平均端到端时延。     

一种基于主-备份节点同步映射的VNF热备份机制

网络功能虚拟化实现了网络功能与专属硬件平台的分离,虚拟网络功能（Virtual Network Function,VNF）通过动态组合形成服务功能链,实现同一硬件平台部署不同的功能服务。针对软件定义网络功能服务可靠性和物理网络资源效率问题,提出了一种基于主备份节点同步映射的VNF节点热备份机制。首先,选择满足节点资源约束的物理节点为主备份节点候选映射对;然后,选择满足带宽约束的最短物理链路为服务路径,实现节点与链路资源总开销的优化。实验结果表明,与传统的备份机制相比,该方法能够满足电信级高可靠性的服务需求,降低了VNF节点备份资源总开销,提高了服务请求的接受率和底层物理网络资源的利用率。     

Method of network slicing deployment based on performance-aware

In order to deal with the performance degradation caused by resource contention due to the sharing of physical resources between VNF in the network slicing,a network slicing deployment method based on performance-awareness was proposed.When deploying network slice instances,first two-phase deployment that mapping virtual nodes was adopted,and then virtual links were mapped.In the virtual nodes mapping phase,the VNF performance influences factor was defined from the perspective of resource supply and demand to quantify the degree of VNF performance impact.Then the sum of the performance influence factors of all VNF on the deployable physical server was found in the network slice instance,the physical server with the smallest sum of performance influence factors was used as the mapping location,and the simulated annealing-discrete particle swarm algorithm was used to find the nodes mapping result.In the virtual links mapping phase,the shortest path algorithm was used to obtain the link mapping result.The simulation results show that the proposed method reduces the impact of other network slices on its own service performance.     

An optimal bandwidth allocation and data droppage scheme for differentiated services in a wireless network

This paper presents an optimal proportional bandwidth allocation and data droppage scheme to provide differentiated services (DiffServ) for downlink pre‐orchestrated multimedia data in a single‐hop wireless network. The proposed resource allocation scheme finds the optimal bandwidth allocation and data drop rates under minimum quality‐of‐service (QoS) constraints. It combines the desirable attributes of relative DiffServ and absolute DiffServ approaches. In contrast to relative DiffServ approach, the proposed scheme guarantees the minimum amount of bandwidth provided to each user without dropping any data at the base‐station, when the network has sufficient resources. If the network does not have sufficient resources to provide minimum bandwidth guarantees to all users without dropping data, the proportional data dropper finds the optimal data drop rates within acceptable levels of QoS and thus avoids the inflexibility of absolute DiffServ approach. The optimal bandwidth allocation and data droppage problems are formulated as constrained nonlinear optimization problems and solved using efficient techniques. Simulations are performed to show that the proposed scheme exhibits the desirable features of absolute and relative DiffServ. Copyright © 2009 John Wiley & Sons, Ltd.     

特征关联的服务功能链可靠性优化算法

针对优化服务功能链(Service Function Chain,SFC)可靠性时资源利用率低的问题,提出了特征关联的服务功能链可靠性优化算法。首先,从底层节点的资源可用率、周边资源密集度、全网中心性评价其作为虚拟网络功能（Virtual Network Function,VNF）实例备选节点的优势,从VNF实例的节点度数与带宽流量评价其作为备选对象的优势。其次,逐个对优越性最高的VNF实例选择备份底层节点,直到所有SFC的可靠性满足客户要求。在为VNF实例选择备份底层节点时,提出了基于资源特征关联感知的备份节点选择算法,将当前VNF实例所映射的底层节点与其他待备份VNF实例的底层节点的邻接节点进行关联,为当前VNF实例选择最优备份底层节点的同时,减少为其他待备份VNF实例选择备份底层节点时链路的开销。仿真结果表明,相比于已有算法,该SFC可靠性优化算法下备份实例数降低了约29%,备份带宽资源消耗降低了约42%,SFC接受率提升了约12%。     

基于微服务架构的粒度可变服务功能链映射算法

针对5G环境下服务功能链（SFC）端到端时延无法满足时延敏感型应用需求的问题,将传统虚拟网络功能（VNF）拆分成粒度不一的映射单元,提出了基于微服务架构的粒度可变服务功能链映射（VG-SFCM）算法。首先将传统粗粒度的VNF解耦成细粒度的微服务单元,随后通过SFC内冗余微服务单元的合并及SFC间微服务单元的复用,减少微服务单元的实例化,降低SFC的处理时间。仿真结果表明,所提算法在降低平均部署网络成本的同时,其SFC端到端时延相较于传统的映射算法降低了14.81%。     

基于迁移演员-评论家学习的服务功能链部署算法

针对5G网络切片环境下由于业务请求的随机性和未知性导致的资源分配不合理从而引起的系统高时延问题,该文提出了一种基于迁移演员-评论家(A-C)学习的服务功能链(SFC)部署算法(TACA)。首先,该算法建立基于虚拟网络功能放置、计算资源、链路带宽资源和前传网络资源联合分配的端到端时延最小化模型,并将其转化为离散时间马尔可夫决策过程(MDP)。而后,在该MDP中采用A-C学习算法与环境进行不断交互动态调整SFC部署策略,优化端到端时延。进一步,为了实现并加速该A-C算法在其他相似目标任务中(如业务请求到达率普遍更高)的收敛过程,采用迁移A-C学习算法实现利用源任务学习的SFC部署知识快速寻找目标任务中的部署策略。仿真结果表明,该文所提算法能够减小且稳定SFC业务数据包的队列积压,优化系统端到端时延,并提高资源利用率。     

NFV业务链资源分配技术

SDN/NFV为运营商实现网络重构和架构演进提供了支撑技术,业务链及其组网技术是NFV的一个重要应用场景,相关的业务链资源分配问题是业界的研究热点。首先对业务链现状进行了梳理,重点分析了已有的资源分配机制和存在的问题,提出了一种业务链资源分配方案,充分考虑了对网络资源的优化利用,实现了VNF实例放置和业务链路径优化选择。实验结果表明,本方案具有更大的系统用户容量和更短的业务路径。     

Resource optimization algorithm of combination of NFV and SDN for application of multiple services

Various services of internet of things (IoT) require flexible network deployment to guarantee different quality of service (QoS).Aiming at the problem of IoT service function chain deployment,network function virtualization (NFV) and software defined networking (SDN) were combined to optimize resources.Considering forwarding cost and traffic load balance,a joint optimization model of virtual network function placement and service function chain routing was given and was proved to be NP-Hard.In order to solve this model,two heuristic algorithms were proposed.One was the service chain deployment algorithm of first routing then placing (FRTP) and the other was the placing followed by routing (PFBR) based on node priority.Simulation results demonstrate that FRTP and PFBR algorithm can significantly balance network traffic load while alleviating congestion and improving the acceptance ratio of the chain requests compared with other algorithms.     

基于改进灰狼优化算法的服务功能链映射算法

随着工业互联网、车联网、元宇宙等新型互联网应用的兴起，网络的低时延、可靠性、安全性、确定性等方面的需求正面临严峻挑战。采用网络功能虚拟化技术在虚拟网络部署过程中，存在服务功能链映射效率低与部署资源开销大等问题，联合考虑节点激活成本、实例化开销，以最小化平均部署网络成本为优化目标建立了整数线性规划模型，提出基于改进灰狼优化算法的服务功能链映射（improved grey wolf optimization based service function chain mapping，IMGWO-SFCM）算法。该算法在标准灰狼优化算法基础上添加了基于无环K最短路径（K shortest path，KSP）问题算法的映射方案搜索、映射方案编码以及基于反向学习与非线性收敛改进三大策略，较好地平衡了其全局搜索及局部搜索能力，实现服务功能链映射方案的快速确定。仿真结果显示，该算法在保证更高的服务功能链请求接受率下，相较于对比算法降低了11.86%的平均部署网络成本。     

一种支持硬件加速的虚拟网络功能部署模型

为解决以软件实现的虚拟网络功能(VNF)性能受限问题,软件定义网络和网络功能虚拟化(SDN/NFV)等新型网络架构引入了硬件加速资源。硬件加速资源的部署,使得VNF能够为日益增长的数据流量提供服务保障。该文针对已有研究未考虑具有高性能数据处理需求的服务链VNF部署问题,提出一种支持硬件加速的VNF部署模型。该模型基于硬件加速资源的承载特性,在保证未加速VNF到商用服务器的优化部署下,优先实现交换机中加速资源的复用,并根据网络业务的性能需求,灵活调整加速资源与VNF的映射约束。仿真实验表明,与其他典型部署方法相比,在引入相同硬件加速资源的情况下,该模型可以承载更多的业务流量,满足服务链高性能数据处理需求,有效提高了部署在网络中加速硬件的资源利用率。     

边缘网络中一种VNF需求预测方法

针对当前虚拟网络功能（Virtualization Network Functions,VNF）需求预测方法准确率较低且不适用于边缘网络的问题,提出了一种在边缘网络中基于支持向量回归（Support Vector Regression,SVR）与门控循环单元（Gated Recurrent Unit,GRU）神经网络模型结合的VNF需求预测方法。考虑到网络边缘流量具有突发性、自相似性及长相关性等特点,结合SVR和GRU两种模型的优点,利用计算复杂度较低的SVR和GRU模型分别提取网络服务历史时序数据的短期特征和长期特征,以提高VNF需求预测准确率,实现边缘网络中VNF的提前放置。实验表明,所提出的预测方法在边缘网络中针对不同网络服务的预测较于传统方法、循环神经网络（Recurrent Neural Networks,RNN）、长短期记忆网络（Long Short-Term Memory,LSTM）模型能够降低20%~30%的误差,有更佳的预测效果。     

A resource‐tuned QoS multicast routing protocol

This paper presents a novel framework for quality‐of‐service (QoS) multicast routing with resource allocation that represents QoS parameters, jitter delay, and reliability, as functions of adjustable network resources, bandwidth, and buffer, rather than static metrics. The particular functional form of QoS parameters depends on rate‐based service disciplines used in the routers. This allows intelligent tuning of QoS parameters as functions of allocated resources during the multicast tree search process, rather than decoupling the tree search from resource allocation. The proposed framework minimizes the network resource utilization while keeping jitter delay, reliability, and bandwidth bounded. This definition makes the proposed QoS multicast routing with resource allocation problem more general than the classical minimum Steiner tree problem. As an application of our general framework, we formulate the QoS multicast routing with resource allocation problem for a network consisting of generalized processor sharing nodes as a mixed‐integer quadratic program and find the optimal multicast tree with allocated resources to satisfy the QoS constraints. We then present a polynomial‐time greedy heuristic for the QoS multicast routing with resource allocation problem and compare its performance with the optimal solution of the mixed‐integer quadratic program. The simulation results reveal that the proposed heuristic finds near‐optimal QoS multicast trees along with important insights into the interdependency of QoS parameters and resources.     

Power control for delay constrained multi‐channel communications using outdated CSI

In this paper, we study the power allocation scheme for a single user, multi‐channel system, e.g., orthogonal frequency‐division multiplexing (OFDM) systems, under time‐variant wireless fading channels. We assume the receiver feeds back perfectly estimated channel state information (CSI) to the transmitter after a processing delay. The objective of the power allocation is to maximize throughput subject to quality‐of‐service (QoS) constraint. The QoS measure of our consideration is a triplet of data rate, delay, and delay bound violation probability. A two‐step sub‐optimal power allocation scheme is proposed to address the impact of outdated CSI. In the first step, the total transmission power that can be used within one block is determined according to the summation of the channel gains of all the channels. In the second step, the total transmission power is allocated among all the channels. The proposed power control scheme is less sensitive to the feedback delay. Compared to the optimal power allocation scheme designed for the perfect CSI scenario, it has lower computational complexity while achieving comparable capacity. Copyright © 2010 John Wiley & Sons, Ltd.