WO2016101587A1 - 虚拟化服务器的链路聚合方法、系统及其智能网络适配器 - Google Patents

虚拟化服务器的链路聚合方法、系统及其智能网络适配器 Download PDF

Info

Publication number
WO2016101587A1
WO2016101587A1 PCT/CN2015/083369 CN2015083369W WO2016101587A1 WO 2016101587 A1 WO2016101587 A1 WO 2016101587A1 CN 2015083369 W CN2015083369 W CN 2015083369W WO 2016101587 A1 WO2016101587 A1 WO 2016101587A1
Authority
WO
WIPO (PCT)
Prior art keywords
port
aggregation
ports
physical network
network
Prior art date
Application number
PCT/CN2015/083369
Other languages
English (en)
French (fr)
Inventor
李静海
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2016101587A1 publication Critical patent/WO2016101587A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/34Signalling channels for network management communication
    • H04L41/342Signalling channels for network management communication between virtual entities, e.g. orchestrators, SDN or NFV entities

Definitions

  • the present invention relates to the field of link aggregation, and in particular, to a link aggregation method and system for a virtualization server, and an intelligent network adapter thereof.
  • the reliability requirements of the server in the telecom field are very high, and no single point of failure is allowed (for example, a network port failure cannot affect the board and external communication), even in the case of sudden high traffic, the server system must have impact resistance. .
  • one network port is connected to one switch, and the other network port is connected to another switch.
  • Link aggregation is used to improve network link reliability and link load sharing.
  • Method 1 Adopt the method of open source virtual switch (OVS). In this mode, the physical board processing performance is greatly reduced.
  • Method 2 Adopt an open source virtual switch (OVS-DPDK) with a data platform development kit. This method has a corresponding improvement in performance, but the performance gap with the non-virtualized scene is still large.
  • OVS-DPDK open source virtual switch
  • SR-IOV Single Root I/O Virtualization
  • PCIe peripheral component interconnect Express
  • the SR-IOV network ports are divided into two types, Physical Function (PF) ports and Virtual Funciton (VF) ports.
  • the PF port is generally used by the host operating system (HOST OS), and the VF port is generally used by the upper virtual machine (VM).
  • PF is a complete PCIe device that includes comprehensive management and configuration capabilities.
  • a virtualized software platform for example, Hypervisor
  • VF is a simplified PCIe device that only includes I/O functions. It cannot manage physical network cards through VF. All VFs are derived from PF.
  • Each VF is a slice of the physical network card hardware resources.
  • a VF is identical to a normal PCIe network card, and can be used after installing the corresponding driver.
  • the link aggregation method of the existing virtualization server is to associate the logical network port of one VM with one VF port of each SR-IOV network card. As shown in FIG. 1 , the network port of each VM corresponds to two SR-IOVs. The two VF ports of the NIC enable the VM to transmit packets with the external network. The VM aggregates the two links to implement link aggregation. But this approach brings the following problems:
  • the VM deployer needs to allocate a VF port to each SR-IOV NIC for each logical network port of the VM, and bind all VF ports corresponding to the logical network port to the VM.
  • the deployment workload is multiplied. When the number of VMs is large, the deployment workload is even more daunting;
  • the embodiment of the present invention is to provide a link aggregation method and system for a virtualized server, and an intelligent network adapter thereof, which can implement physical network port aggregation binding and SR-IOV virtual VF isolation. , thus eliminating the link binding operation of the VM.
  • An embodiment of the present invention provides a link aggregation method for a virtualization server, where the method includes:
  • the virtual machine VM or the host system Host OS in the virtualization server transmits a message through the VF port;
  • N ⁇ 2 and N is an integer.
  • the method further includes:
  • Each physical network port transmits a packet to the external network through a switching device.
  • the binding of the N physical network ports to one aggregation port includes:
  • the port of the switching device connected to each physical network port is also configured as the preset aggregation binding mode.
  • the preset aggregation binding mode is: a link dynamic aggregation protocol LACP mode.
  • the method further includes:
  • the message is transmitted between the two VF ports.
  • the embodiment of the invention provides an intelligent network adapter, which includes: N physical network ports and an aggregation binding module, an SR-IOV module, a port management module, and a message transmission module;
  • An aggregation binding module is configured to bind N physical network ports to an aggregation interface.
  • the SR-IOV module is configured to use the SR-IOV technology to virtualize the aggregation port as a preset number of VF ports;
  • a port management module configured to allocate an available VF port to the VM or the Host OS, so that the available VF port transmits a message of the VM or the Host OS;
  • the message transmission module is configured to transmit the packet of the VF port to the external network through the aggregation interface;
  • N ⁇ 2 and N is an integer.
  • the aggregation binding module includes:
  • the mode configuration unit is configured to configure the physical network port as the default aggregation binding mode.
  • the intelligent network adapter further includes:
  • An internal transmission module configured to transmit messages between two VF ports.
  • the aggregation binding module, the SR-IOV module, the port management module, the message transmission module, the mode configuration unit, and the internal transmission module may use a central processing unit (CPU) when performing processing. , Central Processing Unit), Digital Signal Processor (DSP), or Field-Programmable Gate Array (FPGA) implementation.
  • CPU central processing unit
  • DSP Digital Signal Processor
  • FPGA Field-Programmable Gate Array
  • An embodiment of the present invention further provides a link aggregation system of a virtualization server, where the system package Including: a virtualization server and the intelligent network adapter described in any of the above scenarios; wherein
  • the virtualized server resource is virtualized by the virtualization software platform into M VMs, and the VM transmits the message through the VF port provided by the intelligent network adapter; the intelligent network adapter enables the VF port to transmit through the at least one physical network port and the external network.
  • the system further includes: N switching devices, configured to be connected to the N physical network ports of the intelligent network adapter, and transmit the packets of the physical network port.
  • the link aggregation method, the system, and the intelligent network adapter of the virtualization server provided by the embodiments of the present invention are configured by performing aggregation of multiple physical network interfaces into one aggregation interface, and then performing SR-IOV processing on the aggregation interface. Multiple physical network ports are virtualized into the same set of VF ports, so that each VF port corresponds to an aggregation port, which actually completes link aggregation, improves the link reliability of the virtualized server, and implements multiple physical network ports.
  • FIG. 1 is a schematic diagram of an implementation of a link aggregation method of an existing virtualization server
  • FIG. 2 is a schematic flowchart of an implementation process of a link aggregation method of a virtualization server according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a correspondence between a VM and a VF port and a corresponding relationship between a Host OS and a VF port according to an embodiment of the present invention
  • FIG. 4 is a schematic diagram of a transmission path of a packet in a virtualized server link aggregation system according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of a transmission path of a packet in a virtualization server according to an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a smart network adapter according to an embodiment of the present invention.
  • a virtualization software platform (hypervisor) is installed in the server, and the hypervisor can virtualize the resources of the server into multiple virtual machines (VMs), and the original system of the server is called a host system (Host OS).
  • hypervisor virtualization software platform
  • VMs virtual machines
  • Host OS host system
  • an intelligent network adapter is installed in the server.
  • the intelligent network adapter includes N physical network ports (N ⁇ 2, N is an integer), and the intelligent network adapter supports SR-IOV technology, intelligent network adapter.
  • the N physical network ports are respectively connected to N switches.
  • FIG. 2 is a schematic flowchart of implementing a link aggregation method of a virtualization server according to an embodiment of the present invention. As shown in FIG. 2, the method includes:
  • Step 101 Bind the N physical network ports into one aggregation port.
  • the physical network port is configured on the virtualized server, and the N physical network ports of the intelligent network adapter are bound to an aggregation port.
  • the configuration mode is the same as that of the prior art. In particular, there are seven types of aggregation binding modes in the LINUX system. You can configure the configuration according to the specific situation. As long as the aggregation binding mode of each physical network port is the same.
  • each network port can be configured as a Link Dynamics Aggregation Protocol (LACP) mode.
  • LACP Link Dynamics Aggregation Protocol
  • the N physical network ports are logically a port.
  • Step 102 The SR-IOV technology is used to virtualize the aggregation port into a preset number of virtual function VF ports.
  • the SR-IOV technology can be used to virtualize the aggregation port to a preset number of virtual function VF ports.
  • the specific process is the same as the existing SR-IOV mechanism.
  • how many VF ports are virtualized can be configured through the virtualization service.
  • Step 103 The VM or Host OS in the virtualization server transmits a packet through the VF port.
  • the Host OS transmits the message through the PF port, but because the VF port in the present invention actually implements link aggregation indirectly through steps 101 and 102, the reliability is better, and therefore, the present invention
  • the Hypervisor also allocates a VF port to the Host OS, and the Host OS transmits the packet through the VF port.
  • the Hypervisor requests the intelligent network adapter to use the VF port when creating the VM, and the hypervisor establishes the correspondence between the VM and the VF port, and notifies the intelligent network adapter of the correspondence, and the intelligent network adapter receives and saves the correspondence. That is, the VF port is assigned to the newly created VM. In this way, the newly created VM can transmit messages through its corresponding VF port.
  • the correspondence between the VM and the VF port and the corresponding relationship between the Host OS and the VF port are as shown in FIG. 3.
  • Step 104 The packet of the VF port is transmitted to the external network through the aggregation interface.
  • the packets of the VF port need to be transmitted with the external network
  • the packets are logically transmitted through the aggregation port, and the aggregation port actually corresponds to the N physical network ports, according to the N physical network ports.
  • packets are finally transmitted to the external network through at least one physical network port.
  • the packets are transmitted through the physical links. When one physical link appears. In the event of a failure, the packets originally transmitted on this physical link are transferred to other normal links, as shown in Figure 4.
  • VM1 is the same as Its corresponding port VF1 transmits the message to the port VFm corresponding to VM2, and is transmitted to VM2 by the port VFm.
  • FIG. 6 is a schematic structural diagram of an intelligent network adapter according to an embodiment of the present invention.
  • the aggregation binding module 51 is configured to bind the N physical network ports 50i to one aggregation port.
  • the SR-IOV module 52 is configured to use the SR-IOV technology to virtualize the aggregation port as a preset number of VF ports;
  • the port management module 53 is configured to allocate an available VF port to the VM or the Host OS, so that the available VF port transmits a message of the VM or the Host OS;
  • the message transmission module 54 is configured to transmit the packet of the VF port to the external network through the aggregation interface.
  • the aggregation binding module includes:
  • the mode configuration unit is configured to configure the physical network port as the default aggregation binding mode.
  • the foregoing intelligent network adapter may further include:
  • An internal transmission module configured to transmit messages between two VF ports.
  • An embodiment of the present invention further provides a link aggregation system of a virtualization server, where the system includes: a virtualization server and any one of the foregoing intelligent network adapters;
  • the virtualized server resource is virtualized by the virtualization software platform into M VMs, and the VM transmits the message through the VF port provided by the intelligent network adapter; the intelligent network adapter enables the VF port to transmit through the at least one physical network port and the external network. Message.
  • M the virtualization software platform into M VMs
  • M the virtualization software platform into M VMs
  • the foregoing link aggregation system may further include: N switching devices, configured to be connected to the N physical network ports of the intelligent network adapter, and transmit the packets of the physical network port.
  • embodiments of the present invention can be provided as a method, system, Or a computer program product. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a network adapter is configured to bind multiple physical network interfaces to an aggregation port, and then perform SR-IOV processing on the aggregation interface to virtualize multiple physical network ports into the same set of VF ports, so that each VF port corresponds to one. The aggregation of the port is actually completed. The link reliability of the virtualized server is improved. The load balancing of multiple physical network ports is also implemented.
  • the VM deployer no longer needs to be in the logical network port of the VM.
  • a VF port is assigned to each SR-IOV NIC. It is no longer necessary to manually complete the link binding operation of each logical network port of the VM.
  • the VM does not need to determine the actual state of the physical network port. Therefore, the communication with the switch is abnormal because the VM misjudges a physical network port.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了一种虚拟化服务器的链路聚合方法,包括:将N个物理网口绑定为一个聚合口;使用SR-IOV技术将所述聚合口虚拟为预设数量的VF端口;所述虚拟化服务器中的VM或Host OS通过所述VF端口传输报文;将VF端口的报文通过所述聚合口与外部网络传输;其中,N≥2,N为整数。本发明还同时公开了一种智能网络适配器和虚拟化服务器的链路聚合系统。

Description

虚拟化服务器的链路聚合方法、系统及其智能网络适配器 技术领域
本发明涉及链路聚合领域,尤其涉及虚拟化服务器的链路聚合方法、系统及其智能网络适配器。
背景技术
随着虚拟化技术的快速发展,其应用范围渗透到了各个行业和领域。鉴于虚拟化对硬件资源的利用率可以有效提升和灵活的动态分配、扩容等优点,在电信领域也进行了硬件设备的虚拟化。
电信领域对服务器的可靠性要求非常高,不允许出现单点故障(比如一个网口故障不能影响单板和外界通信),即使在突发高话务量时,服务器系统也要有抗冲击能力。一般传统非虚拟化场景下都会一个网口连接到一台交换机,另一个网口连接到另外一台交换机,以链路聚合的方式提高网络链路可靠性和实现链路负荷分担。
在虚拟化的场景下,中央处理器(CPU)和内存的虚拟化都得到了高效率的使用,而网络链路逐渐成为新的瓶颈。如下目前比较主流的三种网络链路虚拟化的技术:
方式1:采用开源虚拟交换机(OVS)的方式。这种方式下,物理单板处理性能大大降低。
方式2:采用具备数据平台开发套件的开源虚拟交换机(OVS-DPDK)的方式。这个方式对性能是有相应提升,但和非虚化场景下的性能差距还是较大。
方式3:采用单根I/O虚拟化(Single Root I/O Virtualization,SR-IOV)的方式。这种方式将一个网卡虚拟成多个独立的快速外设部件互连标准 (Peripheral Component Interconnect Express,PCIe)网卡设备,供上层的应用软件使用。SR-IOV的性能可以和非虚拟化的硬件基本接近。所以,在商用的虚拟化环境中SR-IOV成为大家优先考虑的方式。
SR-IOV的网络端口分为两个类型,物理功能(Physical Function,PF)端口和虚拟功能(Virtual Funciton,VF)端口。PF端口一般给宿主操作系统(HOST OS)来使用,VF端口一般给上层虚拟机(VM)来使用。PF是一个完整的PCIe设备,包含了全面的管理、配置功能。当虚拟化软件平台(例如:Hypervisor)识别出一块SR-IOV网卡后,会通过PF来管理和配置网卡的所有I/O资源。VF是一个简化的PCIe设备,仅仅包含了I/O功能,无法通过VF对物理网卡进行管理,所有的VF都是通过PF衍生而来。每一个VF是物理网卡硬件资源的一个切片,对于虚拟化软件平台来说,一个VF同一块普通的PCIe网卡一模一样,安装相应驱动程序后就能够使用。
由于电信服务对可靠性的要求高,在虚拟化的环境下虚拟化服务器也需要实现链路聚合,以提高链路的可靠性,实现链路负荷分担等功能。现有虚拟化服务器的链路聚合方法是将一个VM的逻辑网口与每个SR-IOV网卡的一个VF端口相对应,如图1所示,每个VM的网口对应两个SR-IOV网卡的两个VF端口,使得VM与外部网络传输报文存在两路链路,VM将这两路链路进行聚合绑定,就可实现链路聚合。但这种方式带来了如下问题:
问题1:VM部署者需要为VM的每个逻辑网口在每个SR-IOV网卡里都分配VF端口,并要在VM中绑定逻辑网口对应的所有VF端口,部署工作量成倍增长,当VM数量众多时,部署工作量更是让人生畏;
问题2:VM部署者需要手工完成VM的每个逻辑网口的绑定操作,无疑对VM部署者的能力提出了更高的要求,同时,过多的人工操作也为网络拓扑引入了不确定因素,增加了由于人为疏忽而使网络拓扑出错的可能 性;
问题3:要实现链路聚合(即图1中两个交换机的端口聚合),必须保证VM逻辑网口的聚合绑定模式(例如:LACP模式)与交换机端口的聚合绑定模式相同。然而一个物理网口对应多个VM,每个VM对网口状态的判断可能会产生不一致(例如VM1认为物理网口1可用,而VM2认为物理网口2可用),进而导致无法和对接的交换机正常通信。
发明内容
为解决现有存在的技术问题,本发明实施例期望提供一种虚拟化服务器的链路聚合方法、系统及其智能网络适配器,能实现物理网口聚合绑定和SR-IOV虚拟VF之间隔离,从而免去VM的链路绑定操作。
本发明实施例的技术方案是这样实现的:
本发明实施例提供一种虚拟化服务器的链路聚合方法,该方法包括:
将N个物理网口绑定为一个聚合口;
使用单根I/O虚拟化SR-IOV技术将所述聚合口虚拟为预设数量的虚拟功能VF端口;
所述虚拟化服务器中的虚拟机VM或宿主系统Host OS通过所述VF端口传输报文;
将VF端口的报文通过所述聚合口与外部网络传输;
其中,N≥2,N为整数。
上述方案中,所述方法还包括:
所述每个物理网口分别通过一个交换设备与外部网络传输报文。
上述方案中,所述将N个物理网口绑定为一个聚合口包括:
配置物理网口为预设的聚合绑定模式;
相应的,每个物理网口所连接的交换设备的端口也被配置为所述预设的聚合绑定模式。
上述方案中,所述预设的聚合绑定模式是:链路动态聚合协议LACP模式。
上述方案中,所述方法还包括:
将报文在两个VF端口之间传输。
本发明实施例提供一种智能网络适配器,该智能网络适配器包括:N个物理网口和聚合绑定模块、SR-IOV模块、端口管理模块、报文传输模块;其中,
聚合绑定模块,配置为将N个物理网口绑定为一个聚合口;
SR-IOV模块,配置为使用SR-IOV技术将所述聚合口虚拟为预设数量的VF端口;
端口管理模块,配置为为VM或Host OS分配可用的VF端口,使所述可用的VF端口传输VM或Host OS的报文;
报文传输模块,配置为将VF端口的报文通过所述聚合口与外部网络传输;
其中,N≥2,N为整数。
上述方案中,所述聚合绑定模块包括:
模式配置单元,配置为配置物理网口为预设的聚合绑定模式。
上述方案中,所述智能网络适配器还包括:
内部传输模块,配置为将报文在两个VF端口之间传输。
所述聚合绑定模块、所述SR-IOV模块、所述端口管理模块、所述报文传输模块、所述模式配置单元、所述内部传输模块在执行处理时,可以采用中央处理器(CPU,Central Processing Unit)、数字信号处理器(DSP,Digital Singnal Processor)或可编程逻辑阵列(FPGA,Field-Programmable Gate Array)实现。
本发明实施例还提供一种虚拟化服务器的链路聚合系统,所述系统包 括:虚拟化服务器和上述任一方案中所述的智能网络适配器;其中,
所述虚拟化服务器资源被虚拟化软件平台虚拟为M个VM,VM通过所述智能网络适配器提供的VF端口传输报文;所述智能网络适配器使VF端口通过至少一个物理网口与外部网络传输报文;
其中,M≥2,M为整数。
上述方案中,所述系统还包括:N个交换设备,配置为与所述智能网络适配器的N个物理网口相连,传输物理网口的报文。本发明实施例所提供的虚拟化服务器的链路聚合方法、系统及其智能网络适配器,通过将多个物理网口聚合绑定为一个聚合口,再针对该聚合口进行SR-IOV处理,将多个物理网口虚拟为同一套VF端口,使得每一个VF端口都对应一个聚合口,实际完成了链路聚合,提高了虚拟化服务器的链路可靠性,同时也实现了多个物理网口的负荷分担,同时VM部署者不再需要为VM的每个逻辑网口在每个SR-IOV网卡里都分配VF端口,也不再需要手工完成VM的每个逻辑网口的链路绑定操作,并且由于VM只针对一个聚合口,不需要判断物理网口实际的状态,因而不会因VM对某个物理网口误判而导致与交换机的通信异常。
附图说明
图1为现有的虚拟化服务器链路聚合方法的实现示意图;
图2为本发明实施例提供的虚拟化服务器链路聚合方法的实现流程示意图;
图3为本发明实施例中VM与VF端口的对应关系以及Host OS与VF端口的对应关系示意图;
图4为本发明实施例中报文在虚拟化服务器链路聚合系统中的传输路径示意图;
图5为本发明实施例中报文在虚拟化服务器内部的传输路径示意图;
图6为本发明实施例提供的智能网络适配器的组成结构示意图。
具体实施方式
为了更清楚地说明本发明实施例和技术方案,下面将结合附图及实施例对本发明的技术方案进行更详细的说明,显然,所描述的实施例是本发明的一部分实施例,而不是全部实施例。基于本发明的实施例,本领域普通技术人员在不付出创造性劳动的前提下所获得的所有其他实施例,都属于本发明保护的范围。
在本发明实施例中,服务器中安装有虚拟化软件平台(Hypervisor),Hypervisor可将服务器的资源虚拟化为多个虚拟机(VM),服务器原有的系统称为宿主系统(Host OS),为了使VM能够与网络相连,在服务器中安装有智能网络适配器,智能网络适配器中包含有N个物理网口(N≥2,N为整数),智能网络适配器支持SR-IOV技术,智能网络适配器的N个物理网口分别与N个交换机相连。
图2为本发明实施例提供的虚拟化服务器的链路聚合方法的实现流程示意图,如图2所示,该方法包括:
步骤101,将N个物理网口绑定为一个聚合口;
具体的,通过虚拟化服务器配置物理网口,将智能网络适配器的N个物理网口绑定为一个聚合口。配置方式与现有技术相同,特别的,在LINUX系统下有7种聚合绑定模式,可根据具体情况选择配置,只要保证各个物理网口的聚合绑定模式相同即可。在一个实施例中,可将各个网口配置为链路动态聚合协议(LACP)模式。如此,这N个物理网口在逻辑上就是一个端口。
步骤102,使用SR-IOV技术将所述聚合口虚拟为预设数量的虚拟功能VF端口;
具体的,由于在步骤101中已将N个物理网口绑定为一个聚合口,因 此,可以使用SR-IOV技术将所述聚合口虚拟为预设数量的虚拟功能VF端口,具体过程与现有SR-IOV机制相同,具体需要虚拟出多少个VF端口可通过虚拟化服务配置。
步骤103,所述虚拟化服务器中的VM或Host OS通过所述VF端口传输报文;
一般,在SR-IOV中,Host OS传输报文是通过PF端口完成,但因为本发明中VF端口实际通过步骤101和步骤102间接实现了链路聚合,可靠性更好,因此,在本发明实施例中Hypervisor也为Host OS分配VF端口,Host OS通过该VF端口传输报文。
此外,Hypervisor在创建VM时向智能网络适配器申请可被使用的VF端口,Hypervisor建立VM与该VF端口的对应关系,并将该对应关系通知智能网络适配器,智能网络适配器接收并保存该对应关系,即将该VF端口分配给该新创建的VM。如此,该新创建的VM即可通过它对应的VF端口传输报文。
在一个实施例中,VM与VF端口的对应关系以及Host OS与VF端口的对应关系如图3所示。
步骤104,将VF端口的报文通过所述聚合口与外部网络传输;
具体的,基于SR-IOV技术,VF端口的报文需要与外部网络传输时,逻辑上是通过聚合口传输的,而聚合口又实际对应于N个物理网口,根据这N个物理网口的聚合绑定模式,报文最终通过至少一个物理网口与外部网络传输。例如,N个物理网口被绑定为LACP模式,则当每个物理网口对应的物理链路均正常时,报文通过各条物理链路向外传输;而当其中一条物理链路出现故障时,原本在这条物理链路上传输的报文转由其他正常链路传输,如图4所示。
进一步的,当报文只需要在VM之间传输时,如图5所示,VM1同与 它对应的端口VF1将报文传输至与VM2对应的端口VFm,由端口VFm传输给VM2。
图6是本发明实施例提供的智能网络适配器的组成结构示意图,如图6所示,该智能网络适配器包括:N个物理网口50i(i=1,2……N,N≥2,N为整数)和聚合绑定模块51、SR-IOV模块52、端口管理模块53、报文传输模块54;其中,
聚合绑定模块51,配置为将N个物理网口50i绑定为一个聚合口;
SR-IOV模块52,配置为使用SR-IOV技术将所述聚合口虚拟为预设数量的VF端口;
端口管理模块53,配置为为VM或Host OS分配可用的VF端口,使所述可用的VF端口传输VM或Host OS的报文;
报文传输模块54,配置为将VF端口的报文通过所述聚合口与外部网络传输。
进一步的,上述智能网络适配器中,所述聚合绑定模块包括:
模式配置单元,配置为配置物理网口为预设的聚合绑定模式。
进一步的,上述智能网络适配器中还可包括:
内部传输模块,配置为将报文在两个VF端口之间传输。
本发明实施例还提供一种虚拟化服务器的链路聚合系统,所述系统包括:虚拟化服务器和上述任一种智能网络适配器;其中
所述虚拟化服务器资源被虚拟化软件平台虚拟为M个VM,VM通过所述智能网络适配器提供的VF端口传输报文;所述智能网络适配器使VF端口通过至少一个物理网口与外部网络传输报文。其中,M≥2,M为整数。
进一步的,上述链路聚合系统还可包括:N个交换设备,配置为与所述智能网络适配器的N个物理网口相连,传输物理网口的报文。
本领域内的技术人员应明白,本发明的实施例可提供为方法、系统、 或计算机程序产品。因此,本发明可采用硬件实施例、软件实施例、或结合软件和硬件方面的实施例的形式。而且,本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器和光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
以上所述,仅为本发明的较佳实施例而已,并非用于限定本发明的保护范围。
工业实用性
本发明实施例所提供的虚拟化服务器的链路聚合方法、系统及其智能 网络适配器,通过将多个物理网口聚合绑定为一个聚合口,再针对该聚合口进行SR-IOV处理,将多个物理网口虚拟为同一套VF端口,使得每一个VF端口都对应一个聚合口,实际完成了链路聚合,提高了虚拟化服务器的链路可靠性,同时也实现了多个物理网口的负荷分担,同时VM部署者不再需要为VM的每个逻辑网口在每个SR-IOV网卡里都分配VF端口,也不再需要手工完成VM的每个逻辑网口的链路绑定操作,并且由于VM只针对一个聚合口,不需要判断物理网口实际的状态,因而不会因VM对某个物理网口误判而导致与交换机的通信异常。

Claims (10)

  1. 一种虚拟化服务器的链路聚合方法,所述方法包括:
    将N个物理网口绑定为一个聚合口;
    使用单根I/O虚拟化SR-IOV技术将所述聚合口虚拟为预设数量的虚拟功能VF端口;
    所述虚拟化服务器中的虚拟机VM或宿主系统Host OS通过所述VF端口传输报文;
    将VF端口的报文通过所述聚合口与外部网络传输;
    其中,N≥2,N为整数。
  2. 根据权利要求1所述的链路聚合方法,其中,所述方法还包括:
    所述每个物理网口分别通过一个交换设备与外部网络传输报文。
  3. 根据权利要求2所述的链路聚合方法,其中,所述将N个物理网口绑定为一个聚合口包括:
    配置物理网口为预设的聚合绑定模式;
    相应的,每个物理网口所连接的交换设备的端口也被配置为所述预设的聚合绑定模式。
  4. 根据权利要求3所述的链路聚合方法,其中,所述预设的聚合绑定模式是:链路动态聚合协议LACP模式。
  5. 根据权利要求1所述的链路聚合方法,其中,所述方法还包括:
    将报文在两个VF端口之间传输。
  6. 一种智能网络适配器,所述智能网络适配器包括:N个物理网口和聚合绑定模块、SR-IOV模块、端口管理模块、报文传输模块;其中,
    聚合绑定模块,配置为将N个物理网口绑定为一个聚合口;
    SR-IOV模块,配置为使用SR-IOV技术将所述聚合口虚拟为预设数量的VF端口;
    端口管理模块,配置为为VM或Host OS分配可用的VF端口,使所述可用的VF端口传输VM或Host OS的报文;
    报文传输模块,配置为将VF端口的报文通过所述聚合口与外部网络传输;
    其中,N≥2,N为整数。
  7. 根据权利要求6所述的智能网络适配器,其中,所述聚合绑定模块包括:
    模式配置单元,配置为配置物理网口为预设的聚合绑定模式。
  8. 根据权利要求6所述的智能网络适配器,其中,所述智能网络适配器还包括:
    内部传输模块,配置为将报文在两个VF端口之间传输。
  9. 一种虚拟化服务器的链路聚合系统,所述系统包括:虚拟化服务器和根据权利要求6至8任一项所述的智能网络适配器;其中,
    所述虚拟化服务器资源被虚拟化软件平台虚拟为M个VM,VM通过所述智能网络适配器提供的VF端口传输报文;所述智能网络适配器使VF端口通过至少一个物理网口与外部网络传输报文;
    其中,M≥2,M为整数。
  10. 根据权利要求9所述的链路聚合系统,其中,所述系统还包括:N个交换设备,配置为与所述智能网络适配器的N个物理网口相连,传输物理网口的报文。
PCT/CN2015/083369 2014-12-24 2015-07-06 虚拟化服务器的链路聚合方法、系统及其智能网络适配器 WO2016101587A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410817484.9A CN105790991A (zh) 2014-12-24 2014-12-24 虚拟化服务器的链路聚合方法、系统及其智能网络适配器
CN201410817484.9 2014-12-24

Publications (1)

Publication Number Publication Date
WO2016101587A1 true WO2016101587A1 (zh) 2016-06-30

Family

ID=56149140

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/083369 WO2016101587A1 (zh) 2014-12-24 2015-07-06 虚拟化服务器的链路聚合方法、系统及其智能网络适配器

Country Status (2)

Country Link
CN (1) CN105790991A (zh)
WO (1) WO2016101587A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110688237A (zh) * 2019-06-25 2020-01-14 华为技术有限公司 转发报文的方法、中间设备和计算机设备
CN114189443A (zh) * 2020-09-15 2022-03-15 华为云计算技术有限公司 一种配置信息确定的方法,装置以及设备
CN115484212A (zh) * 2022-07-26 2022-12-16 浪潮通信技术有限公司 VF的QoS调整方法、装置及电子设备
US11575620B2 (en) 2019-03-29 2023-02-07 Intel Corporation Queue-to-port allocation
CN116010130A (zh) * 2023-01-30 2023-04-25 中科驭数(北京)科技有限公司 Dpu虚拟口的跨卡链路聚合方法、装置、设备及介质
CN117692382A (zh) * 2024-02-04 2024-03-12 珠海星云智联科技有限公司 链路聚合方法、网卡、设备以及介质

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110752937B (zh) * 2018-07-23 2022-04-15 中兴通讯股份有限公司 一种物理功能复用方法、装置和计算机存储介质
CN111865626B (zh) * 2019-04-24 2023-05-23 厦门网宿有限公司 一种基于聚合端口的数据收发方法和装置
CN111726251B (zh) * 2020-06-12 2023-04-25 浪潮电子信息产业股份有限公司 一种虚拟化系统中sds存储域的组网方法、系统及装置
CN113472635A (zh) * 2021-05-31 2021-10-01 杭州沃趣科技股份有限公司 一种实现双端口Active-Active的冗余保护方法和系统
CN113630265B (zh) * 2021-06-30 2023-04-07 济南浪潮数据技术有限公司 一种基于智能网卡的虚拟网络冗余备份方法及装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100115174A1 (en) * 2008-11-05 2010-05-06 Aprius Inc. PCI Express Load Sharing Network Interface Controller Cluster
CN102457540A (zh) * 2010-10-25 2012-05-16 成都谛听科技有限公司 低成本nas一体化设备带宽倍增方法
CN103336708A (zh) * 2013-05-31 2013-10-02 华为技术有限公司 聚合网卡资源的方法和相关设备及计算机系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012151392A1 (en) * 2011-05-04 2012-11-08 Citrix Systems, Inc. Systems and methods for sr-iov pass-thru via an intermediary device
US9727386B2 (en) * 2012-10-12 2017-08-08 Futurewei Technologies, Inc. Method and apparatus for network resource virtual partitioning
TWI556174B (zh) * 2014-03-05 2016-11-01 威盛電子股份有限公司 虛擬功能分配系統、方法及其管理主機

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100115174A1 (en) * 2008-11-05 2010-05-06 Aprius Inc. PCI Express Load Sharing Network Interface Controller Cluster
CN102457540A (zh) * 2010-10-25 2012-05-16 成都谛听科技有限公司 低成本nas一体化设备带宽倍增方法
CN103336708A (zh) * 2013-05-31 2013-10-02 华为技术有限公司 聚合网卡资源的方法和相关设备及计算机系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LI, NAN ET AL.: "A VF Allocation Model Based on SR-IOV Extension", COMPUTER APPLICATIONS AND SOFTWARE, vol. 30, no. 11, 30 November 2013 (2013-11-30) *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11575620B2 (en) 2019-03-29 2023-02-07 Intel Corporation Queue-to-port allocation
CN110688237A (zh) * 2019-06-25 2020-01-14 华为技术有限公司 转发报文的方法、中间设备和计算机设备
US11604742B2 (en) 2019-06-25 2023-03-14 Huawei Technologies Co., Ltd. Independent central processing unit (CPU) networking using an intermediate device
CN110688237B (zh) * 2019-06-25 2024-02-09 华为技术有限公司 转发报文的方法、中间设备和计算机设备
CN114189443A (zh) * 2020-09-15 2022-03-15 华为云计算技术有限公司 一种配置信息确定的方法,装置以及设备
CN115484212A (zh) * 2022-07-26 2022-12-16 浪潮通信技术有限公司 VF的QoS调整方法、装置及电子设备
CN115484212B (zh) * 2022-07-26 2023-10-20 浪潮通信技术有限公司 VF的QoS调整方法、装置及电子设备
CN116010130A (zh) * 2023-01-30 2023-04-25 中科驭数(北京)科技有限公司 Dpu虚拟口的跨卡链路聚合方法、装置、设备及介质
CN116010130B (zh) * 2023-01-30 2024-04-19 中科驭数(北京)科技有限公司 Dpu虚拟口的跨卡链路聚合方法、装置、设备及介质
CN117692382A (zh) * 2024-02-04 2024-03-12 珠海星云智联科技有限公司 链路聚合方法、网卡、设备以及介质
CN117692382B (zh) * 2024-02-04 2024-06-07 珠海星云智联科技有限公司 链路聚合方法、网卡、设备以及介质

Also Published As

Publication number Publication date
CN105790991A (zh) 2016-07-20

Similar Documents

Publication Publication Date Title
WO2016101587A1 (zh) 虚拟化服务器的链路聚合方法、系统及其智能网络适配器
US11061712B2 (en) Hot-plugging of virtual functions in a virtualized environment
US9176767B2 (en) Network interface card device pass-through with multiple nested hypervisors
US10230765B2 (en) Techniques to deliver security and network policies to a virtual network function
US10768960B2 (en) Method for affinity binding of interrupt of virtual network interface card, and computer device
US8495217B2 (en) Mechanism for preventing client partition crashes by removing processing resources from the client logical partition when an NPIV server goes down
US9154451B2 (en) Systems and methods for sharing devices in a virtualization environment
US9720712B2 (en) Physical/virtual device failover with a shared backend
CN106557444B (zh) 实现sr-iov网卡的方法和装置、实现动态迁移的方法和装置
US20160342437A1 (en) Data path failover method for sr-iov capable ethernet controller
CN113312143B (zh) 云计算系统、命令处理方法及虚拟化仿真装置
US20190146827A1 (en) Virtualized network function resource management method and device
US8966480B2 (en) System for migrating a virtual machine between computers
US10452419B2 (en) Device virtualization for containers
CN104199716B (zh) 一种虚拟机热迁移的方法、物理主机及系统
US9537797B2 (en) MTU management in a virtualized computer system
US20160259659A1 (en) Method and system for unified application centric connectivity in various virtualization platforms
CN105468429A (zh) 一种高效虚拟化集群管理方法和集群节点
US20160253200A1 (en) Server virtualization method of multi node system and apparatus thereof
US20210006511A1 (en) Software-controlled active-backup mode of link aggregation for rdma and virtual functions
US11025594B2 (en) Secret information distribution method and device
Tasoulas et al. Prototyping Live Migration With SR-IOV Supported InfiniBand HCSs
CN118233414B (zh) 数据处理单元点对点通信方法、计算机设备及介质
US20230325222A1 (en) Lifecycle and recovery for virtualized dpu management operating systems
US10025739B2 (en) Information processing system, information processing method, and recording medium

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: 15871661

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: 15871661

Country of ref document: EP

Kind code of ref document: A1