WO2018086569A1 - 一种基于虚拟网络的应用感知的动态sdn配置方法 - Google Patents
一种基于虚拟网络的应用感知的动态sdn配置方法 Download PDFInfo
- Publication number
- WO2018086569A1 WO2018086569A1 PCT/CN2017/110347 CN2017110347W WO2018086569A1 WO 2018086569 A1 WO2018086569 A1 WO 2018086569A1 CN 2017110347 W CN2017110347 W CN 2017110347W WO 2018086569 A1 WO2018086569 A1 WO 2018086569A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- network
- application
- virtual
- virtual machine
- virtual network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
- H04L47/803—Application aware
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/46—Interconnection of networks
- H04L12/4641—Virtual LANs, VLANs, e.g. virtual private networks [VPN]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0888—Throughput
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
- H04L47/805—QOS or priority aware
Definitions
- the present invention relates to the field of computer software technologies, and in particular, to a dynamic network-based application-aware dynamic SDN configuration method.
- SDN Software Defined Network
- OpenFlow realizes the network by separating the control plane of the network device from the data plane. Flexible control of traffic makes the network more intelligent as a conduit.
- Open vSwitch is a network virtual switch developed by OpenStack. It is a high-quality, multi-layer virtual switch (layer of network layer), which is an important technical means to implement software-defined networks. Compared to traditional physical switches, Open vSwitch virtual switches have many advantages:
- the present invention provides a dynamic network-based application-aware dynamic SDN configuration method, and the core idea is to use the OVS to set up a virtual network on a physical server.
- the libvirt monitoring tool is used to obtain the network traffic characteristics of the application, and according to the information such as the frequency of the network access and the data throughput, the application is divided into different types, and corresponding network resource allocation strategies are formulated for different types.
- An application-aware dynamic SDN configuration method based on a virtual network includes the following steps:
- step 2) According to the network traffic of different virtual machines recorded in step 2), the applications in the virtual network are divided into different types, and a network resource allocation policy is set for each type of application;
- step 1) includes installing libvirt and OVS on each server of the physical cluster, setting up a software-defined virtual network, and enabling the virtual machine created by libvirt to use a software-defined virtual network built on OVS by default.
- the software-defined virtual network is a virtual network across physical units; the virtual machine configuration file created by modifying libvirt is used to define a virtual network by default using OVS-based software.
- step 2) the libvirt application programming interface (API) monitors and records the network traffic of different virtual machines in the virtual network architecture.
- API application programming interface
- the resource parameter of the given virtual machine is obtained through the application programming interface of libvirt; before the operation of the virtual machine, the virtual machine hypervisor is first connected, and the application programming interface is periodically called to obtain the number of bytes sent and received by the current virtual machine. According to the number of bytes sent and received, the number of bytes sent and received by the current virtual machine in the unit time, that is, the current virtual machine network read and write rate, is stored in the database.
- the application type in step 3) includes but is not limited to: a high frequency application and a big data application; the high frequency application has the characteristics of frequent network access per unit time and less network traffic consumption; the big data application has characteristics It is: the number of network accesses per unit time is small, and the network traffic is consumed.
- setting the network resource allocation strategy in step 3) includes increasing the network response priority of the high frequency application, and limiting the read and write rates of the high frequency application and the big data application by configuring the software defined virtual network constructed by the OVS.
- step 3) further includes: monitoring virtual machine network access in real time, when the access of the high frequency application increases, exceeding the current allocated network bandwidth, dynamically increasing its priority; when the access of the high frequency application is reduced to the current allocated bandwidth designation
- the priority is dynamically reduced.
- the bandwidth limit is dynamically adjusted, the bandwidth allocation of other big data applications and high-frequency applications is appropriately reduced, and the bandwidth allocation of the big data application is improved.
- a certain big At the end of the data transfer of the data application, its bandwidth allocation is reduced, and the free bandwidth is reallocated to other big data applications and high frequency applications.
- step 4 the network resource configuration of each virtual machine is set and adjusted by the OVS.
- the setting and adjusting the network resource configuration of each virtual machine through the OVS includes configuring a flow table for each port corresponding to the virtual machine through an application programming interface of the OVS, and setting a rule for the network data flow of each port. .
- network resources are prioritized and restricted for each application to ensure the throughput of big data applications and the quality of service (QoS) of high frequency access applications.
- QoS quality of service
- OVS open programming interface is used to dynamically adjust the network resources to improve application efficiency and network resource utilization.
- FIG. 1 is a schematic flowchart of a dynamic network-based application-aware dynamic SDN configuration method according to an embodiment of the present invention.
- FIG. 2 is a schematic diagram showing a software-defined virtual network architecture based on OVS in an embodiment of the present invention.
- Application-aware technology acquires the feature values of an application through some measurement tools or methods, and classifies the applications based on these feature values. Based on these characteristics, the system optimizes the behavior of the application or the performance of other subsystems that the application runs and controls.
- the application-aware system can dynamically allocate and adjust network resources of applications according to relevant information of applications, such as application status and resource requests, to improve application operation efficiency and network resource utilization. Specifically, this is to use libvirt and OVS to build software to define the virtual network environment, and to perform network monitoring and dynamic adjustment to configure network resources.
- a virtual network-based application-aware dynamic SDN configuration method includes the following steps:
- step 2) According to the network traffic of different virtual machines recorded in step 2), the applications in the virtual network are divided into different types, and a network resource allocation policy is set for each type of application;
- Libvirt provides a set of APIs that can get resource parameters for a given virtual machine. Before you can operate on a virtual machine, you first need to get a connection to the virtual machine hypervisor. By periodically calling the API provided by libvirt, you can get the number of bytes sent and received by the current virtual machine. According to the obtained data, the number of bytes sent and received by the current virtual machine in the unit time, that is, the current virtual machine network read/write rate, can be calculated and stored in the database.
- the pseudo code of the algorithm is as follows:
- the input to the algorithm is the identifier Id of all virtual machines and the time interval T for two monitoring.
- the algorithm calls the API provided by libvirt to get the current network read and write bytes Rid and Sid of all virtual machines and record them. After that, the algorithm waits for a given time T. After the waiting time is over, the algorithm calls the libvirt API again. For each virtual machine, the number of network read and write bytes Rid' and Sid' at this time is recorded again, and the two monitoring data corresponding to the id are subtracted, that is, obtained in T. The number of network read/write bytes of the virtual machine, that is, the network read/write rate.
- the high frequency application has the characteristics that the network access is frequent per unit time
- the big data application has the characteristics that the number of network accesses per unit time is small, and the network traffic consumption is large.
- the coping strategy selected by the present invention is to improve the network response priority of the high frequency application, and to respond to the big data application, and to limit the read and write rates of the two types of applications by configuring the virtual network constructed by the OVS. Based on the total bandwidth of the network, it not only ensures the timely response of high-frequency applications, but also satisfies the QoS requirements, and ensures that the rate of data transmission by big data applications is not greatly affected.
- the invention monitors the virtual machine network access situation in real time according to the algorithm in step 2.
- the priority is dynamically increased to ensure timely response of the application.
- the access of the high frequency application is reduced to a threshold, its priority is dynamically lowered to avoid wasting network resources.
- the specific threshold can be adjusted by the system administrator based on experience. The default is 80%.
- the bandwidth limit of the application is dynamically adjusted, the bandwidth allocation of other applications is appropriately reduced, and the bandwidth allocation of big data applications is improved to improve efficiency.
- the bandwidth allocation of the application is reduced, and the free bandwidth is reallocated to other applications to ensure that the network resources are not wasted.
- the pseudo code of the algorithm is as follows:
- the input of the algorithm is the identifier Id of the virtual machine, the type Type of the virtual machine, the time interval T, and the total bandwidth TBW.
- the algorithm in step 2 is invoked to obtain the current virtual machine read/write rate. After the time interval T elapses, the virtual machine read/write rate is calculated again and compared. If the rate is increased, it is processed according to the virtual machine type. If it is a high-frequency application (Web Service) type, the priority of the application is raised. Otherwise, it is a big data application type. According to the total bandwidth TBW, the bandwidth allocation of the application is increased under the condition that the total bandwidth does not exceed TBW, and the other is reduced. Bandwidth allocation for the application. When the read/write rate is reduced, the operation is the same as that described previously, and the operation is reversed from the corresponding operation in the pseudo code.
- the present invention implements this function by using OVS.
- the OVS provides an API, and can configure a flow table for each port corresponding to each virtual machine, and set rules for the network data flow of each port, such as priority, maximum data length, and the like.
- the invention invokes such an API to implement dynamic adjustment of virtual machine network resources, thereby achieving dynamic and rapid configuration of corresponding network resources, and achieving efficient network resource configuration.
- this example builds an OVS-based software-defined virtual network environment based on three physical servers to simulate an enterprise private cloud environment.
- two typical applications namely Web Service application (ie high frequency application) providing Email service and Spark Big Data application (ie Big Data Application) providing data analysis service.
- Web Service application ie high frequency application
- Spark Big Data application ie Big Data Application
- the network resource dynamic adjustment configuration method described in the foregoing embodiment automatically reserves sufficient network resources for the email service during the operation of the big data application, thereby ensuring the service quality of the enterprise web service application.
- the dynamic adjustment preparation method automatically adjusts the priority of the high-frequency application network when the high-frequency application changes, and allocates network resources according to the specific access frequency, thereby ensuring the service quality of the enterprise Web Service application.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Security & Cryptography (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
本发明提供了一种基于虚拟网络的应用感知的动态SDN配置方法,包括以下步骤:1)在数据中心的物理集群中构建虚拟网络;2)监测并记录虚拟网络中不同虚拟机的网络流量;3)根据步骤2)中记录的不同虚拟机的网络流量,将虚拟网络中的应用划分为不同的类型,并为每种类型的应用设定网络资源分配策略;4)根据步骤3)设定的网络资源分配策略,设置并调整每个虚拟机的网络资源配置。为每个应用分配网络资源优先级和限制等,保证大数据应用的吞吐量和高频访问应用的服务质量(QoS)。最后利用OVS开放编程接口实现网络资源动态调整配置,以提高应用运行效率和网络资源利用率。
Description
本发明涉及计算机软件技术领域,特别涉及一种基于虚拟网络的应用感知的动态SDN配置方法。
软件定义网络(Software Defined Network,SDN),是一种新型网络架构,是网络虚拟化的一种实现方式,其核心技术OpenFlow是通过将网络设备控制面与数据面分离开来,从而实现了网络流量的灵活控制,使网络作为管道变得更加智能。
传统IT架构中的网络,根据业务需求部署上线以后,如果业务需求发生变动,重新修改相应网络设备(路由器、交换机、防火墙)上的配置是一件非常繁琐的事情。在互联网/移动互联网瞬息万变的业务环境下,灵活性和敏捷性已经成了IT业务的重要需求。软件定义网络所做的事是将网络设备上的控制权分离出来,由集中的控制器管理,无须依赖底层网络设备(路由器、交换机、防火墙),屏蔽了来自底层网络设备的差异。而控制权是完全开放的,用户可以自定义任何想实现的网络路由和传输规则策略,从而更加灵活和智能。
为了实现软件定义网络,软件定义交换机必不可少。Open vSwitch(OVS)是OpenStack开发的网络虚拟交换机,是一个高质量的,多层虚拟交换机(网络分层的层),是实现软件定义网络的重要技术手段。跟传统的物理交换机相比,Open vSwitch虚拟交换机具备很多优点:
1.配置灵活,因为是软件实现的,一台物理服务器上可以配置数十台或者数百台虚拟交换机,而且端口数目可以灵活选择。
2.成本低廉,通过软件的方式可轻易达到10Gbps的交换速度。
利用Open vSwitch对机房进行软件定义网络改造后,无需对网络中每个节点进行反复配置,网络中的设备天然是自动连通的。用户只需在使用时定义具体的网络规则即可。
然而不同类型的应用具有不同的行为模式,如何定义合适的网络规则对系统管理员提出了新的挑战。
发明内容
针对现有云环境中,不同类型应用的网络资源分配问题,本发明提供了一种基于虚拟网络的应用感知的动态SDN配置方法,其核心思想为:利用OVS在物理服务器上架设虚拟网络。通过libvirt监测工具获取应用的网络流量特征,根据网络访问的频率和数据吞吐量等信息,将应用划分为不同类型,并针对不同类型,制定相应的网络资源分配策略。
本发明采取的具体技术方案是:
一种基于虚拟网络的应用感知的动态SDN配置方法,包括以下步骤:
1)在数据中心的物理集群中构建虚拟网络;
2)监测并记录虚拟网络中不同虚拟机的网络流量;
3)根据步骤2)中记录的不同虚拟机的网络流量,将虚拟网络中的应用划分为不同的类型,并为每种类型的应用设定网络资源分配策略;
4)根据步骤3)设定的网络资源分配策略,设置并调整每个虚拟机的网络资源配置。
进一步地,步骤1)包括在物理集群的各个服务器上安装libvirt和OVS,搭建软件定义虚拟网络,并使libvirt创建的虚拟机默认使用基于OVS构建的软件定义虚拟网络。
进一步地,所述软件定义虚拟网络为跨物理机组虚拟网络;通过修改libvirt创建的虚拟机配置文件使其默认使用基于OVS构建的软件定义虚拟网络。
进一步地,步骤2)中通过libvirt的应用程序编程接口(API)监测并记录虚拟网络构架中不同虚拟机的网络流量。
进一步地,通过libvirt的应用程序编程接口获取给定虚拟机的资源参数;在对虚拟机进行操作之前,首先与虚拟机hypervisor连接,通过定时调用应用程序编程接口,获取当前虚拟机收发字节数,根据收发字节数,计算单位时间内当前虚拟机收发的字节数,即当前虚拟机网络读写速率,并存入数据库中。
进一步地,步骤3)中应用类型包括但不限于:高频应用及大数据应用;所述高频应用具有特征为:单位时间网络访问频繁,消耗网络流量较少;所述大数据应用具有特征为:单位时间网络访问次数较少,消耗网络流量较大。
进一步地,步骤3)中设定网络资源分配策略包括提高高频应用的网络响应优先级,同时通过配置OVS构建的软件定义虚拟网络对高频应用及大数据应用的读写速率进行限制。
进一步地,步骤3)还包括:实时监控虚拟机网络访问情况,当高频应用的访问增加,超过当前分配网络带宽,动态提升其优先级;当高频应用的访问减少至当前分配带宽的指定阈值时,动态降低其优先级,当某一大数据应用的网络读写速率提升时,动态调整其带宽限制,适当降低其他大数据应用和高频应用带宽分配,提高大数据应用的带宽分配;当某一大
数据应用的数据传输结束时,降低其带宽分配,将空闲出的带宽重新分配给其他大数据应用和高频应用。
进一步地,步骤4)中通过OVS设置并调整每个虚拟机的网络资源配置。
进一步地,所述通过OVS设置并调整每个虚拟机的网络资源配置包括通过OVS的应用程序编程接口,对每个虚拟机对应的端口进行流表配置,对每个端口的网络数据流设置规则。
通过采取上述技术方案,为每个应用分配网络资源优先级和限制等,保证大数据应用的吞吐量和高频访问应用的服务质量(QoS)。最后利用OVS开放编程接口实现网络资源动态调整配置,以提高应用运行效率和网络资源利用率。
图1是本发明一实施例中基于虚拟网络的应用感知的动态SDN配置方法的流程示意图。
图2是本发明一实施例中表示基于OVS的软件定义虚拟网络架构示意图。
工作原理介绍,
应用感知技术是通过一些测量工具或方式,获取应用的特征值,并根据这些特征值,将应用进行分类。系统根据这些特征,来优化应用的行为或应用运行、控制的其他子系统的性能。结合软件定义网络技术,应用感知系统可根据应用的有关信息,如应用状态和资源请求等,来动态分配调整应用的网络资源,以提高应用运行效率和网络资源利用率。具体地,本是利用libvirt和OVS构建软件定义虚拟网络环境,并进行网络监测和动态调整配置网络资源。
下面将结合本发明实施例中的附图,对本实用新型实施例中的技术方案进行清楚、完整的描述。
在一实施例中,提供一种基于虚拟网络的应用感知的动态SDN配置方法,如图1所示,包括以下步骤:
1)在数据中心的物理集群中构建虚拟网络;
2)监测并记录虚拟网络中不同虚拟机的网络流量;
3)根据步骤2)中记录的不同虚拟机的网络流量,将虚拟网络中的应用划分为不同的类型,并为每种类型的应用设定网络资源分配策略;
4)根据步骤3)设定的网络资源分配策略,设置并调整每个虚拟机的网络资源配置。
上述步骤的具体技术方案如下:
构建虚拟网络构架。首先在物理服务器中下载并安装libvirt和OVS。利用OVS创建虚拟交换机,并绑定到一个物理网卡上,利用隧道技术跨物理机组建虚拟网络,基于OVS的软件定义虚拟网络架构如图2所示。修改libvirt虚拟机配置文件,使创建的虚拟机网卡默认绑定到基于OVS的软件定义虚拟网络中。
网络流量监控。libvirt提供了一套API,可以获取给定虚拟机的资源参数。在对虚拟机进行操作之前,首先需要得到与虚拟机hypervisor的连接。通过定时调用libvirt提供的API,可以获取当前虚拟机收发字节数。根据获取的数据,可以计算单位时间内当前虚拟机收发的字节数,即当前虚拟机网络读写速率,并存入数据库中。算法伪代码如下:
算法的输入是所有虚拟机的标识符Id,以及两次监控的时间间隔T。首先,算法调用libvirt提供的API,获取所有虚拟机当前的网络读写字节数Rid和Sid,并记录下来。之后,算法等待给定的时间T。等待时间结束后,算法再次调用libvirt的API,对每个虚拟机,再次记录此时的网络读写字节数Rid’和Sid’,将对应id的两次监控数据相减,即得到在T时间内,该虚拟机的网络读写字节数,即网络读写速率。所述高频应用具有特征为:单位时间网络访问频繁,;所述大数据应用具有特征为:单位时间网络访问次数较少,消耗网络流量较大。
应用类型划分及制定策略。根据上一步得到的结果,可以分析得到,部分应用网络访问频繁,但每次读写字节数都较少,即消耗网络流量较少,属于高频应用,如架设在虚拟机中的Apache服务等Web Service应用。另外的应用网络访问较少,但每次读写的数据很大,即消耗网络流量较大,属于大数据交互型应用(大数据应用),如Hadoop、Spark等大数据应用。
具体而言,类似HTTP服务等是典型的高频应用,访问并发度很多,但每次访问传输的包相对较小;而Hadoop分布式计算平台等,则是典型的大数据应用,每次网络访问都会传输很多流量,要求很高的吞吐量,对时延不敏感。可以确定的是,在了解本申请公布的技术内容的基础上,本领域技术人员对于前述“访问频繁”、“读写字节数都较少”、“访问较少”及“读写的数据很大”所限定的程度和范围应有清楚地认识,因此本实施例仅通过举例说明,不再划定具体的范围或设定固定的阀值进行详细说明,本领域技术人员在领会本申请的精神的前提下,是可以实现前后文所描述的技术方案并可解决相应技术问题的。
当网络总带宽一定时,会存在大数据应用因传输数据,造成高频应用延迟响应,此类应用的QoS无法保证。本发明选择的应对策略为,提高高频应用的网络响应优先级,优先于大数据应用进行响应,同时通过配置OVS构建的软件定义虚拟网络对两类应用的读写速率进行限制,在能够充分利用网络总带宽的基础上,既保证高频应用的及时响应,满足QoS需求,又保证大数据应用传输数据的速率不会受到太大影响。
本发明根据步骤2中的算法,实时监控虚拟机网络访问情况。当高频应用的访问增加,超过一定阈值时,动态提升其优先级,以保证此应用的及时响应。当高频应用的访问减少至阈值时,动态降低其优先级,以免浪费网络资源。具体阈值可由系统管理员根据经验调整,默认为80%。当大数据应用的网络读写速率提升时,动态调整应用的带宽限制,适当降低其他应用带宽分配,提高大数据应用的带宽分配,以提高效率。同样,当大数据应用的数据传输结束时,降低此应用的带宽分配,将空闲出的带宽重新分配给其他应用,以保证网络资源不被浪费。算法伪代码如下:
算法的输入是虚拟机的标识符Id,虚拟机的类型Type,时间间隔T,总带宽TBW。首先,调用步骤2中的算法,获取当前虚拟机读写速率,在时间间隔T过后,再次计算虚拟机读写速率,并进行对比。若速率提升,则根据虚拟机类型进行处理。若为高频应用(Web Service)类型,则提升此应用的优先级,否则为大数据应用类型,根据总带宽TBW,在带宽总和不超过TBW的条件下,增加该应用的带宽分配,减少其他应用的带宽分配。当读写速率降低时,与之前描述的处理方式相同,操作与伪代码中的对应操作相反即可。
4.动态调整网络配置。根据步骤3中的算法,在对应条件下,需要提升或降低应用的优先级和带宽分配。本发明通过使用OVS来实现这一功能,OVS提供API,可以对每个虚拟机对应的端口进行流表配置,对每个端口的网络数据流设置规则,如优先级,最大数据长度等。本发明调用此类API,实现对虚拟机网络资源的动态调整,从而达到动态快速配置相应的网络资源,实现高效的网络资源配置的目的。
下面给予一具体的应用实例。该实例模拟了企业私有云内部的实际需求。
首先本例基于3台物理服务器搭建了基于OVS的软件定义虚拟网络环境,用来模拟企业私有云环境。然后选择2种典型应用,分别是提供Email服务的Web Service应用(即高频应用)和提供数据分析服务的Spark大数据应用(即大数据应用)。通过将这2个应用程序部署到上述环境中,在没有应用感知动态调整网络方案的情况下,大数据应用时长会占用过多的网络资源,导致Email的服务质量变差,集中表现在发送邮件超时,无法收到新邮件。而通过前述实施例描述的网络资源动态调整配置方法,在大数据应用运行期间,自动地预留足够的网络资源给Email服务,从而保证了企业Web Service应用的服务质量。另一方面,当高频应用访问发生变化时,若没有应用感知动态调整网络,则无法保证当前访问频率最高的应用的服务质量,集中表现在收发邮件缓慢,而通过签署实施例描述的网络资源动态调整配制方法,在高频应用访问变化时,自动调整高频应用网络优先级,根据具体访问频率,分配网络资源,从而保证企业Web Service应用的服务质量。
最后需要注意的是,公布实施方式的目的在于帮助进一步理解本发明,但是本领域的技术人员可以理解:在不脱离本发明及所附的权利要求的精神和范围内,各种替换和修改都是可能的。因此,本发明不应局限于实施例所公开的内容,本发明要求保护的范围以权利要求书界定的范围为准。
Claims (10)
- 一种基于虚拟网络的应用感知的动态SDN配置方法,包括以下步骤:1)在数据中心的物理集群中构建虚拟网络;2)监测并记录虚拟网络中不同虚拟机的网络流量;3)根据步骤2)中记录的不同虚拟机的网络流量,将虚拟网络中的应用划分为不同的类型,并为每种类型的应用设定网络资源分配策略;4)根据步骤3)设定的网络资源分配策略,设置并调整每个虚拟机的网络资源配置。
- 如权利要求1所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤1)包括在物理集群的各个服务器上安装libvirt和OVS,搭建软件定义虚拟网络,并使libvirt创建的虚拟机默认使用基于OVS构建的软件定义虚拟网络。
- 如权利要求2所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,所述软件定义虚拟网络为跨物理机组虚拟网络;通过修改libvirt创建的虚拟机配置文件使其默认使用基于OVS构建的软件定义虚拟网络。
- 如权利要求2所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤2)中通过libvirt的应用程序编程接口监测并记录虚拟网络构架中不同虚拟机的网络流量。
- 如权利要求4所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,通过libvirt的应用程序编程接口获取给定虚拟机的资源参数;在对虚拟机进行操作之前,首先与虚拟机hypervisor连接,通过定时调用应用程序编程接口,获取当前虚拟机收发字节数,根据收发字节数,计算单位时间内当前虚拟机收发的字节数,即当前虚拟机网络读写速率,并存入数据库中。
- 如权利要求1所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤3)中应用类型包括:高频应用及大数据应用;所述高频应用具有特征为:单位时间网络访问频繁,消耗网络流量较少;所述大数据应用具有特征为:单位时间网络访问次数较少,消耗网络流量较大。
- 如权利要求6所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤3)中设定网络资源分配策略包括提高高频应用的网络响应优先级,同时通过配置OVS构建的软件定义虚拟网络对高频应用及大数据应用的读写速率进行限制。
- 如权利要求6所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤 3)还包括:实时监控虚拟机网络访问情况,当高频应用的访问增加,超过当前分配网络带宽,动态提升其优先级;当高频应用的访问减少至当前分配带宽的指定阈值时,动态降低其优先级;当某一大数据应用的网络读写速率提升时,动态调整其带宽限制,降低其他大数据应用和高频应用带宽分配,提高大数据应用的带宽分配;当某一大数据应用的数据传输结束时,降低其带宽分配,将空闲出的带宽重新分配给其他大数据应用和高频应用。
- 如权利要求1所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,步骤4)中通过OVS设置并调整每个虚拟机的网络资源配置。
- 如权利要求9所述的基于虚拟网络的应用感知的动态SDN配置方法,其特征在于,所述通过OVS设置并调整每个虚拟机的网络资源配置包括通过OVS的应用程序编程接口,对每个虚拟机对应的端口进行流表配置,对每个端口的网络数据流设置规则。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610989862.0 | 2016-11-10 | ||
| CN201610989862.0A CN106656867B (zh) | 2016-11-10 | 2016-11-10 | 一种基于虚拟网络的应用感知的动态sdn配置方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2018086569A1 true WO2018086569A1 (zh) | 2018-05-17 |
Family
ID=58805958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2017/110347 Ceased WO2018086569A1 (zh) | 2016-11-10 | 2017-11-10 | 一种基于虚拟网络的应用感知的动态sdn配置方法 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106656867B (zh) |
| WO (1) | WO2018086569A1 (zh) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110753009A (zh) * | 2019-10-23 | 2020-02-04 | 重庆邮电大学 | 基于多QoS分组的虚拟机和网络带宽联合分配方法 |
| CN114928526A (zh) * | 2022-02-09 | 2022-08-19 | 北京邮电大学 | 基于sdn的网络隔离和资源规划方法及系统 |
| CN115225503A (zh) * | 2022-05-16 | 2022-10-21 | 北京邮电大学 | OpenFlow流表项动态超时分配方法及相关设备 |
| CN115567423A (zh) * | 2021-12-25 | 2023-01-03 | 深圳市奥斯达通信有限公司 | Ovs网络流量加速优化方法及其系统 |
| US11563640B2 (en) | 2018-12-13 | 2023-01-24 | At&T Intellectual Property I, L.P. | Network data extraction parser-model in SDN |
| US11797379B2 (en) | 2022-02-04 | 2023-10-24 | Western Digital Technologies, Inc. | Error detection and data recovery for distributed cache |
| US11899585B2 (en) | 2021-12-24 | 2024-02-13 | Western Digital Technologies, Inc. | In-kernel caching for distributed cache |
| US11934663B2 (en) | 2022-01-10 | 2024-03-19 | Western Digital Technologies, Inc. | Computational acceleration for distributed cache |
| US12182022B2 (en) | 2022-05-10 | 2024-12-31 | Western Digital Tehcnologies, Inc. | In-kernel cache request queuing for distributed cache |
| CN119583454A (zh) * | 2024-12-11 | 2025-03-07 | 深圳市唯仕达通讯科技有限公司 | 一种网络交换机控制方法及系统 |
| US12321602B2 (en) | 2023-05-24 | 2025-06-03 | Western Digital Technologies, Inc. | Disaggregated memory management |
| US12379951B2 (en) | 2022-06-27 | 2025-08-05 | Western Digital Technologies, Inc. | Memory coherence in virtualized environments |
| US12386648B2 (en) | 2022-06-09 | 2025-08-12 | Western Digital Technologies, Inc. | Resource allocation in virtualized environments |
| US12452189B2 (en) | 2022-06-01 | 2025-10-21 | Western Digital Technologies, Inc. | Context-aware NVMe processing in virtualized environments |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106656867B (zh) * | 2016-11-10 | 2019-11-05 | 北京大学(天津滨海)新一代信息技术研究院 | 一种基于虚拟网络的应用感知的动态sdn配置方法 |
| CN106686739B (zh) * | 2016-12-16 | 2020-02-14 | 清华大学 | 面向数据流的基于软件定义网络的无线网络资源管理方法 |
| CN107395621A (zh) * | 2017-08-18 | 2017-11-24 | 国云科技股份有限公司 | 一种虚拟机网卡流量分类监控方法 |
| CN107579850B (zh) * | 2017-09-05 | 2021-05-18 | 郑州云海信息技术有限公司 | 一种云数据中心基于sdn控制的有线无线混合组网方法 |
| CN107579852A (zh) * | 2017-09-15 | 2018-01-12 | 郑州云海信息技术有限公司 | 云服务器中基于历史模型的虚拟网络性能隔离系统和方法 |
| CN107579988B (zh) * | 2017-09-25 | 2020-01-07 | 新华三技术有限公司 | 配置安全策略的方法和装置 |
| CN107634962B (zh) * | 2017-10-11 | 2019-06-18 | Oppo广东移动通信有限公司 | 网络带宽的管理方法及相关产品 |
| CN107861792A (zh) * | 2017-11-06 | 2018-03-30 | 郑州云海信息技术有限公司 | 一种基于Libvirt的虚拟机管理方法及装置 |
| CN108092920B (zh) * | 2017-12-19 | 2020-06-19 | 西安交通大学 | 一种基于软件定义和用户中心网络的应用感知系统及其方法 |
| CN109614229B (zh) * | 2018-11-30 | 2020-11-13 | 深圳先进技术研究院 | 一种基于软件定义的虚拟网络资源分配系统 |
| CN111065118A (zh) * | 2019-12-12 | 2020-04-24 | 西安冉科信息技术有限公司 | 一种移动终端网络速度控制方法 |
| CN113742084A (zh) * | 2021-09-13 | 2021-12-03 | 城云科技(中国)有限公司 | 基于接口特性分配计算资源的方法和装置 |
| CN113923156B (zh) * | 2021-09-29 | 2023-04-11 | 武汉美和易思数字科技有限公司 | 一种基于多路网络的智能网关管理方法及系统 |
| CN114385311B (zh) * | 2021-12-08 | 2025-07-01 | 天翼物联科技有限公司 | 移动物联感知系统的控制方法、装置和存储介质 |
| CN114553785A (zh) * | 2022-02-18 | 2022-05-27 | 上海交通大学 | 一种动态自适应的云平台租户流量监管方法及系统 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156665A (zh) * | 2011-04-13 | 2011-08-17 | 杭州电子科技大学 | 一种虚拟化系统竞争资源差异化服务方法 |
| CN102316483A (zh) * | 2011-09-01 | 2012-01-11 | 中兴通讯股份有限公司 | 一种EVDO系统中保证应用业务QoS的方法及装置 |
| US20140226661A1 (en) * | 2013-02-11 | 2014-08-14 | Cisco Technology, Inc. | Binary compatible extension architecture in an openflow compliant network environment |
| CN104363159A (zh) * | 2014-07-02 | 2015-02-18 | 北京邮电大学 | 一种基于软件定义网络的开放虚拟网络构建系统和方法 |
| CN106656867A (zh) * | 2016-11-10 | 2017-05-10 | 北京大学(天津滨海)新代信息技术研究院 | 一种基于虚拟网络的应用感知的动态sdn配置方法 |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152239A (zh) * | 2013-02-25 | 2013-06-12 | 汉柏科技有限公司 | 一种基于Open VSwitch的虚拟网络实现方法和系统 |
| CN105245365B (zh) * | 2015-09-30 | 2019-04-26 | 联想(北京)有限公司 | 一种虚拟网络的参数配置方法、设备和系统 |
| CN105429995A (zh) * | 2015-12-15 | 2016-03-23 | 国云科技股份有限公司 | 一种基于Openflow的私有云网络动态安全隔离系统及其隔离方法 |
-
2016
- 2016-11-10 CN CN201610989862.0A patent/CN106656867B/zh active Active
-
2017
- 2017-11-10 WO PCT/CN2017/110347 patent/WO2018086569A1/zh not_active Ceased
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156665A (zh) * | 2011-04-13 | 2011-08-17 | 杭州电子科技大学 | 一种虚拟化系统竞争资源差异化服务方法 |
| CN102316483A (zh) * | 2011-09-01 | 2012-01-11 | 中兴通讯股份有限公司 | 一种EVDO系统中保证应用业务QoS的方法及装置 |
| US20140226661A1 (en) * | 2013-02-11 | 2014-08-14 | Cisco Technology, Inc. | Binary compatible extension architecture in an openflow compliant network environment |
| CN104363159A (zh) * | 2014-07-02 | 2015-02-18 | 北京邮电大学 | 一种基于软件定义网络的开放虚拟网络构建系统和方法 |
| CN106656867A (zh) * | 2016-11-10 | 2017-05-10 | 北京大学(天津滨海)新代信息技术研究院 | 一种基于虚拟网络的应用感知的动态sdn配置方法 |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11563640B2 (en) | 2018-12-13 | 2023-01-24 | At&T Intellectual Property I, L.P. | Network data extraction parser-model in SDN |
| CN110753009A (zh) * | 2019-10-23 | 2020-02-04 | 重庆邮电大学 | 基于多QoS分组的虚拟机和网络带宽联合分配方法 |
| CN110753009B (zh) * | 2019-10-23 | 2022-08-26 | 重庆邮电大学 | 基于多QoS分组的虚拟机和网络带宽联合分配方法 |
| US11899585B2 (en) | 2021-12-24 | 2024-02-13 | Western Digital Technologies, Inc. | In-kernel caching for distributed cache |
| CN115567423A (zh) * | 2021-12-25 | 2023-01-03 | 深圳市奥斯达通信有限公司 | Ovs网络流量加速优化方法及其系统 |
| US11934663B2 (en) | 2022-01-10 | 2024-03-19 | Western Digital Technologies, Inc. | Computational acceleration for distributed cache |
| US11797379B2 (en) | 2022-02-04 | 2023-10-24 | Western Digital Technologies, Inc. | Error detection and data recovery for distributed cache |
| CN114928526B (zh) * | 2022-02-09 | 2023-06-16 | 北京邮电大学 | 基于sdn的网络隔离和资源规划方法及系统 |
| CN114928526A (zh) * | 2022-02-09 | 2022-08-19 | 北京邮电大学 | 基于sdn的网络隔离和资源规划方法及系统 |
| US12182022B2 (en) | 2022-05-10 | 2024-12-31 | Western Digital Tehcnologies, Inc. | In-kernel cache request queuing for distributed cache |
| CN115225503B (zh) * | 2022-05-16 | 2023-07-04 | 北京邮电大学 | OpenFlow流表项动态超时分配方法及相关设备 |
| CN115225503A (zh) * | 2022-05-16 | 2022-10-21 | 北京邮电大学 | OpenFlow流表项动态超时分配方法及相关设备 |
| US12452189B2 (en) | 2022-06-01 | 2025-10-21 | Western Digital Technologies, Inc. | Context-aware NVMe processing in virtualized environments |
| US12386648B2 (en) | 2022-06-09 | 2025-08-12 | Western Digital Technologies, Inc. | Resource allocation in virtualized environments |
| US12379951B2 (en) | 2022-06-27 | 2025-08-05 | Western Digital Technologies, Inc. | Memory coherence in virtualized environments |
| US12321602B2 (en) | 2023-05-24 | 2025-06-03 | Western Digital Technologies, Inc. | Disaggregated memory management |
| CN119583454A (zh) * | 2024-12-11 | 2025-03-07 | 深圳市唯仕达通讯科技有限公司 | 一种网络交换机控制方法及系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106656867A (zh) | 2017-05-10 |
| CN106656867B (zh) | 2019-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2018086569A1 (zh) | 一种基于虚拟网络的应用感知的动态sdn配置方法 | |
| Govindarajan et al. | Realizing the quality of service (QoS) in software-defined networking (SDN) based cloud infrastructure | |
| CN109906592B (zh) | 监控切片性能的系统及方法 | |
| US10481939B2 (en) | Bandwidth control in multi-tenant virtual networks | |
| JP6096325B2 (ja) | シンキングDiameterネットワークアーキテクチャを提供するための方法、システム、およびコンピュータ読取可能媒体 | |
| EP3739451A1 (en) | Detect and enforce api slas using cloud access api broker | |
| EP3115897B1 (en) | Intelligent load balancer selection in a multi-load balancer environment | |
| EP4070505B1 (en) | Providing interface between network management and slice management | |
| CN102934403B (zh) | 控制网络上的数据传输 | |
| CN105827523B (zh) | 一种在云存储环境中对多租户的带宽实现动态调整的虚拟网关 | |
| KR102392442B1 (ko) | 분류된 네트워크 스트림의 관리 | |
| WO2016107418A1 (zh) | 云化网络通信路径的分配方法、装置及系统 | |
| CN103428025A (zh) | 一种管理虚拟网络服务的方法、装置和系统 | |
| CN103747059A (zh) | 一种面向多租户的云计算服务器集群网络保障方法和系统 | |
| CN112703774A (zh) | 网络切片管理 | |
| US10382344B2 (en) | Generating and/or receiving at least one packet to facilitate, at least in part, network path establishment | |
| CN116192640A (zh) | 网络切片资源分配方法、装置、sdn控制器和存储介质 | |
| US20230216894A1 (en) | System and methods for real-time data band multi-path routing | |
| US20220214926A1 (en) | Virtual machine monitoring device, virtual machine monitoring method, and program | |
| CN110601897A (zh) | 一种网络资源配置方法及装置 | |
| Chen et al. | Enterprise visor: A Software-Defined enterprise network resource management engine | |
| CN105743999A (zh) | 一种分布式链路接入带宽控制系统 | |
| CN116346727A (zh) | 一种虚拟机的网络限速方法、装置、设备及介质 | |
| US11245630B2 (en) | Network system and network band control management method | |
| US20260005968A1 (en) | Policing Network Traffic on Interfaces Across Multiple Processing Elements |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17869335 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 17869335 Country of ref document: EP Kind code of ref document: A1 |