WO2015192584A1

WO2015192584A1 - Virtual routing system and method

Info

Publication number: WO2015192584A1
Application number: PCT/CN2014/090190
Authority: WO
Inventors: 张玉军
Original assignee: 中兴通讯股份有限公司
Priority date: 2014-06-18
Filing date: 2014-11-03
Publication date: 2015-12-23
Also published as: CN105227454A; CN105227454B

Abstract

Disclosed are a virtual routing system and method, the system comprises: level-1 routes configured to provide a layer-3 routing function for communication between virtual machines of different virtual networks within the calculation nodes of a data center virtual platform, and providing source address translation/destination address translation to the data center network external to the calculation nodes; level-2 routes configured to provide a layer-3 routing function for communication between virtual machines of different virtual networks between the calculation nodes; border gateway routes configured to provide source address translation/destination address translation to an external network for a data center virtual platform. The present invention solves the problem in the relevant art of a single carrier node for virtual routes, improving network performance and availability of the data center, and reducing the physical network bandwidth consumption by the communication between virtual networks.

Description

Virtual routing system and method

Technical field

The present invention relates to the field of communications, and in particular to a virtual routing system and method.

Background technique

A cloud computing data center that provides infrastructure as a service (IaaS), in order to implement functions such as multi-tenant isolation and virtual machine migration, requires network virtualization. The virtualized network isolates the virtual machines in the private network of the tenant in a Virtual Local Area Network (VLAN)/Virtual Extensible LAN (VXLAN)/Generic Routing Encapsulation (referred to as Generic Routing Encapsulation). For GRE) in a virtual network. With the development of data centers, the requirements for networks are getting higher and higher, requiring high bandwidth, high efficiency, low latency of data center networks, and communication between virtual networks and data center communication. Virtual routing plays an important role in data center network communication. However, the implementation of virtual routing in data centers still has many problems in performance and high availability, such as the east-west flow problem between virtual networks, and the internal virtual network communication between the same computing nodes. The external virtual routing still needs to be performed, and the virtual routing bearer node is single, and there are single point failures and performance bottlenecks.

In view of the single problem of the virtual routing bearer node in the related art, an effective solution has not been proposed yet.

Summary of the invention

The embodiment of the invention provides a virtual routing system and method to solve at least the problem of a single virtual routing bearer node in the related art.

According to an embodiment of the present invention, a virtual routing system is provided, including: a primary routing, configured to provide a three-layer routing function for virtual machine communication of different virtual networks within a computing node of a data center virtualization platform, and provide a source address translation/destination address translation function to a data center network outside the computing node; a secondary route configured to provide a three-layer routing function for virtual machine communication to which different virtual networks belong to the computing node; a border gateway routing, Set to provide source address translation/destination address translation to the external network for the data center virtualization platform.

The primary route is located in the computing node; the secondary route is located in a network element device that can implement a route virtualization function; the border gateway route is located at a gateway function that provides the data center network to an external network. In the routing entity.

The secondary route supports a virtual route dynamic migration protocol VROOM.

There is a unique global identifier for each of the virtual networks.

The border gateway routing supports the Border Gateway Routing Protocol BGP or the Open Shortest Path First Protocol OSPF protocol.

The system also includes a network controller configured to create or delete the primary route, and/or to add or remove an interface of the virtual network for the primary route; and/or set to create, delete, or Migrating the border gateway route, and/or routing an interface for the border gateway to add or remove virtual networks.

According to another embodiment of the present invention, a virtual routing method is provided, which is applied to the virtual routing system, and includes: a source-level route receives a data packet sent by a source virtual machine to a target device, where the source level The route is a primary route corresponding to the source virtual machine; the source primary route forwards the data packet to the target device when the target device is in the node where the source primary route is located a virtual network; or the source-level route forwards the data packet to the secondary route if the target device is located in the data center virtualization platform but not in the node where the source-level routing is located; Or the source-level routing performs source address translation/destination address conversion to the internal network, and sends the data packet to the data center network outside the computing node. Giving the target device; or the source-level routing to the internal network if the target device is not in the data center network Source address translation/destination address translation and forwarding to the border gateway route.

In this embodiment, the source-level route requests a forwarding policy from the network controller; the network controller requests the data center virtualization platform to acquire the location of the target device; and the network controller is configured according to the target device. The location of the routing policy is delivered to the source-level routing, where the forwarding policy includes: when the target device is in the node where the source-level routing is located, the source-level routing The data packet is forwarded to the virtual network where the target device is located; or, in a case where the target device is located in the data center virtualization platform but is not in the node where the source-level routing is located, the source level Route forwarding the data packet to the secondary route; or, if the target device is located in a data center network outside the computing node, the source primary routing performs the data packet to the internal network Source address translation/destination address translation and sent to the target device; or, in the case where the target device is not in the data center network, the source level The routing performs source address translation/destination address translation on the internal network to the internal network, and forwards the packet to the border gateway route; the source primary route is forwarded according to the forwarding policy.

Before the source-level routing requests the forwarding policy from the network controller, the method further includes: the network controller assigning a unique global identifier to each of the virtual networks.

In the case that the secondary route supports the virtual route dynamic migration protocol VROOM, the network controller performs scheduling migration on the secondary route load balancing, and/or recovers the suspended secondary route.

Before the network controller performs the scheduled migration of the secondary routing load balancing, the network controller further determines, according to the collected metric, whether to perform scheduling migration on the secondary routing load balancing, where The metric factor includes at least one of the following: hardware resource utilization, network bearer bandwidth, and power saving requirement of the secondary route.

Before the network controller performs the scheduled migration of the secondary route load balancing, the method further includes: the network controller providing the collected metric to the third party through the northbound interface; the network controller receiving the The third-party scheduling policy is described, and the secondary routing load balancing is scheduled to be migrated according to the scheduling policy.

Before the network controller recovers the hanged secondary route, the network controller further moves all the virtual routes that are responsible for running the secondary route hanged in the secondary routing cluster, and sequentially The hanged secondary route is cloned in other units of the secondary routing cluster.

After the source-level route forwards the data packet to the secondary route, the method further includes: the secondary route forwarding the data packet to a target primary route corresponding to the target device, and by the target The primary route forwards the data packet to the virtual network where the target device is located.

After the source-level routing forwards the data packet to the border gateway route, the method further includes: the border gateway routing performing source address translation/destination address translation on the data packet to the external network, and forwarding to the External network.

With the embodiment of the present invention, a virtual routing system includes: a primary routing, configured to provide a three-layer routing function for virtual machine communication of different virtual networks within a computing node of the data center virtualization platform, and provide the computing to the computing The source address translation/destination address translation function of the data center network outside the node; the secondary route is set to provide a three-layer routing function for the virtual machine communication to which the different virtual networks belong to the computing node; the border gateway routing is set to be data The central virtualization platform provides source address translation/destination address translation to the external network, which solves the problem of single routing of virtual routing bearers in related technologies, improves data center network performance and availability, and reduces virtual network communication to physical network bandwidth. The effect of consumption.

DRAWINGS

The drawings described herein are intended to provide a further understanding of the invention, and are intended to be a part of the invention. In the drawing:

1 is a structural block diagram of a virtual routing system according to an embodiment of the present invention;

2 is a flowchart of a virtual routing method according to an embodiment of the present invention;

3 is a schematic diagram of a method for implementing virtual route of a data center network according to Embodiment 1 of the present invention;

4 is a schematic diagram of a virtual machine routing process between virtual nodes of a computing node according to a first embodiment of the present invention;

5 is a schematic diagram of a virtual machine routing process between virtual nodes of different computing nodes according to the first embodiment of the present invention;

FIG. 6 is a schematic diagram of a process of address translation between a virtual machine and an external network communication network in a virtual network according to Embodiment 1 of the present invention.

detailed description

The invention will be described in detail below with reference to the drawings in conjunction with the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict.

The present embodiment implements a high-performance and high-availability implementation method for virtual routes of a cloud computing data center network in response to the defects of the current data center virtual routing, and effectively solves the single problem of the virtual routing bearer node, and the virtual network between the same computing node. The communication problem realizes a distributed multi-level virtual routing implementation scheme, which can improve the performance and availability of the data center network and reduce the bandwidth consumption of the physical network between the virtual network communication.

In this embodiment, a virtual routing system is provided. FIG. 1 is a structural block diagram of a virtual routing system according to an embodiment of the present invention. As shown in FIG. 1 , the system includes a primary routing 12 and a secondary routing 14 . And border gateway routing 16, as detailed below:

The primary routing 12 is configured to provide a three-layer routing function for virtual machine communication of different virtual networks within the computing node of the data center virtualization platform, and provide source address translation/destination address to the data center network outside the computing node Conversion function

The secondary route 14 is configured to provide a three-layer routing function for virtual machine communication to which the different virtual networks belong to the computing nodes;

Border Gateway Route 16, set to provide source address translation/destination address translation to the external network for the data center virtualization platform.

In this embodiment, the route is separated from the primary route, the secondary route, and the border gateway route in the virtual routing architecture, and the route inside the compute node is processed by the primary route, and the route between the compute nodes is routed by the secondary route. The processing and routing with the external network are handled by the border gateway. The three types of routing work cooperate to solve the problem of single routing of the virtual routing bearer in the related technology, and improve the performance and availability of the data center network and reduce the communication between the virtual networks. The effect of physical network bandwidth consumption.

Preferably, the primary route may be located in the computing node; the secondary route is located in a network element device cluster that implements a route virtualization function; the border gateway route is located to provide the data center network to the external The gateway function of the network is in the routing entity cluster.

Preferably, the secondary route can support Virtual Routers On the Move (VROOM), so that load migration can be implemented between the secondary routes, thereby completing load balancing and preventing a single node from hanging. Affect business implementation.

Preferably, each of the virtual networks may have a unique global identity. This can facilitate the control layer device to manage and control the routing and forwarding of each virtual network.

Preferably, the border gateway route may support a Border Gateway Protocol (BGP) or an Open Shortest Path First (OSPF) protocol.

Preferably, the system may further include a network controller 18 configured to create or delete the primary routing, and/or to add or remove an interface of the virtual network for the primary routing; and/or, set to Creating, deleting, or migrating the border gateway route, and/or routing an interface for the border gateway to add or remove virtual networks.

In this embodiment, a virtual routing method is also provided, which is applied to the virtual routing system. FIG. 2 is a flowchart of a virtual routing method according to an embodiment of the present invention. As shown in FIG. 2, the method includes The following steps:

Step S202: The source-level route receives the data packet sent by the source virtual machine to the target device, where the source-level route is a primary route corresponding to the source virtual machine.

Step S204, the source-level route forwards the data packet to the virtual network where the target device is located if the target device is in the node where the source-level route is located; or the source level Routing, when the target device is located in the data center virtualization platform but not in the node where the source-level routing is located, forwarding the data packet to the secondary route; or the source-level routing is in the Where the target device is located in a data center network outside the computing node, performing source address translation/destination address translation on the internal network to the data packet, and transmitting the data packet to the target device; or The source-level route performs source address translation/destination address conversion to the internal network and forwards the packet to the border gateway route if the target device is not in the data center network.

Through the above steps, the single problem of the virtual routing bearer node in the related technology is solved, thereby improving the performance and availability of the data center network, and reducing the effect of the communication between the virtual networks on the bandwidth consumption of the physical network.

Preferably, the step S204 may be implemented by using a forwarding policy of the network controller, where the source-level route requests a forwarding policy from the network controller, and the network controller requests the data center virtualization platform to acquire the target device. And the network controller sends the forwarding policy to the source-level route according to the location of the target device, where the forwarding policy includes: where the target device is routed at the source level In the case of a node, the source-level route forwards the data packet to a virtual network where the target device is located; or, when the target device is located in the data center virtualization platform but not at the source level In the case that the route is located in the node, the source first-level route forwards the data packet to the secondary route; or, in a case where the target device is located in a data center network outside the computing node, the source one Level routing performs source address translation/destination address translation of the data packet to the internal network and transmits to the target device; or, at the target If the source-level route is not in the data center network, the source-level route performs source address translation/destination address translation on the internal network to the internal network, and forwards the packet to the border gateway route; The route is forwarded according to the forwarding policy.

The data center virtualization platform is a virtual network, and the data center network outside the computing node is a real network in the data center, and the external network of the data center network is a real or virtual network outside the data center.

Preferably, before the source-level routing requests the forwarding policy from the network controller, the method further includes: the network controller assigning a unique global identifier to each of the virtual networks.

Preferably, in a case that the secondary route supports a virtual route dynamic migration protocol (VROOM), the network controller may further perform scheduling migration on the secondary route load balancing, and/or The secondary route is restored.

Preferably, before the network controller performs scheduling migration on the secondary route load balancing, the network controller may determine, according to the collected metric, whether to perform scheduling migration on the secondary routing load balancing. The metric factor may include, but is not limited to, at least one of the following: hardware resource utilization, network bearer bandwidth, and energy saving requirement of the secondary route.

Preferably, the network controller may further provide the collected metrics to the third party through the northbound interface, and then perform scheduling migration on the secondary route load balancing according to the received third-party scheduling policy, where The third party interacts with the northbound interface of the system for other control systems participating in the system.

Preferably, before the network controller recovers the suspended secondary route, the network controller may: move the virtual route responsible for running the secondary route hanged in the secondary routing cluster And cloning the hanged secondary route in the other units of the secondary routing cluster.

Preferably, the network controller may further assign a unique global identifier to each of the virtual networks before the source-level routing requests a forwarding policy from the network controller.

Preferably, after the source-level route forwards the data packet to the secondary route, the secondary route may forward the data packet to a target primary route corresponding to the target device, and The target primary route forwards the data packet to the virtual network where the target device is located.

Preferably, after the source-level routing forwards the data packet to the border gateway route, the border gateway route may perform source address translation/destination address translation on the data packet to the external network, and forward the packet to the Said external network.

The following description is made in conjunction with the preferred embodiments, and the following preferred embodiments incorporate the above-described embodiments and preferred embodiments thereof.

The data center network construction of the data center network virtual route high performance and high availability implementation method provided in the following preferred embodiment includes: the data center network is constructed in a two-dimensional manner of a network control plane and a data forwarding plane, where the network control plane is There is a unique global logic controller; the data center network implements the network virtualization function, and the network controller assigns a unique global identifier to each virtual network; the three-layer routing function in the virtual network consists of multiple levels of distributed virtual routes to implement the data center. High performance of network virtual routing; virtual routing logic entities can implement dynamic migration to achieve high availability of virtual routes in data center networks.

Preferably, the control plane controller allocates a global identifier to the virtualized network, and the virtual network can implement Layer 2 network isolation based on the manner of vlan/vxlan/gre. When the data center provides network-as-a-service services, the tenant needs to be assigned a network namespace to achieve complete isolation of the three- to seven-layer virtual network between tenants.

Preferably, the multi-level distributed virtual route includes: a primary route, a secondary route, and a border gateway route. The first-level routes are distributed in each computing node of the virtualization platform; the secondary routing and border gateway routing can provide high-performance network element device clusters that can perform related functions.

Preferably, the secondary virtual route needs to support a virtual route migration protocol (VROOM), and the controller can dynamically adjust the data center virtual network, including virtual route load balancing scheduling migration and hanging node virtual route recovery.

Preferably, the first-level route implements a three-layer routing function between different virtual networks in the computing node, and the first-level routing is configured according to the control plane controller to create a virtual deletion virtual route, and add and remove a virtual network interface for the virtual route. Implement source address translation/destination address translation.

Preferably, the secondary routing implements a three-layer routing function of different virtual networks between computing nodes, and the secondary routing is carried by a group of network element devices capable of implementing routing virtualization, and each network element device in the cluster can be controlled according to The surface controller schedules the creation and deletion of virtual routes and adds and removes virtual network interfaces for virtual routes.

Preferably, the border gateway route is a set of distributed routing entity clusters that provide gateway functions for the data center network to the external network, and each routing unit can run BGP/OSPF protocol to provide source address translation to the external network for the data center network. Destination address translation.

Preferably, if the virtual machine in the virtual network communicates with the device in the public network of the data center, the primary source address translation/destination address conversion needs to be performed through the first-level route, and the virtual machine in the virtual machine virtual network and the external network of the data center exchange visits. The secondary source address translation/destination address translation needs to be routed through the border gateway.

Preferably, in order to achieve high efficiency of the virtual network, the secondary route needs to migrate the virtual route according to the load balancing controller of the control plane controller, and the controller can collect the measurement factors such as the hardware resource utilization rate, the network bearer bandwidth, and the energy saving requirement of the secondary route. Optimize the use of forwarding plane network element device resources.

Preferably, in order to achieve high availability of the virtual network, when the controller detects that a certain unit in the secondary routing cluster is hanged, all the virtual routes responsible for running are required to be migrated out, and the controller sequentially performs the above virtual according to the relevant scheduling policy. The route clones the virtual route in other units in the cluster to ensure that the virtual network Layer 3 routing function is working properly. After the dead node is restored, it needs to be initialized before re-access, so that it can be scheduled as a new resource.

Preferably, the controller may provide a northbound interface, provide the collected secondary routing metrics to the third party, and formulate a customized secondary virtual routing scheduling policy by the third party.

Embodiment 1

The preferred embodiment solves many problems existing in the virtual routing construction scheme of the cloud computing data center network, and proposes a high-performance and high-availability implementation method for the virtual routing of the data center network, which can effectively improve the performance and availability of the data center network.

3 is a schematic diagram of a method for implementing virtual route of a data center network according to a first embodiment of the present invention. As shown in FIG. 3, the data center network is controlled by a network control plane. The data forwarding plane is constructed in two dimensions. The network control plane has a unique global logic controller (Controller); the data center network implements network virtualization function, and the network controller assigns a unique global identifier to each virtual network; in the virtual network The Layer 3 routing function consists of multi-level distributed virtual routes to achieve high performance of virtual routes in the data center network. The virtual routing logical entities can implement dynamic migration to achieve high availability of virtual routes in the data center network.

In Figure 3, the CONTROLLER represents the network control plane controller, the FVR is the first-level virtual route, the SVR is the secondary virtual route, the VM is the virtual machine, and the SR is the network element device (secondary route) that can implement the route virtualization function. For border gateway routing, BGR CLUSTER is a distributed border gateway routing cluster, and CN is a computing virtualization platform computing node.

The controller allocates a global identifier to each virtualized network in the network in the cloud computing data. The virtualized network may be based on the vlan/vxlan/gre mode to completely isolate the Layer 2 network, and the controller needs to generate one for each virtual network. The mapping between the virtual network identifier and the network isolation identifier. When the data center provides network-as-a-service services, the tenant needs to be assigned a network name space to implement complete isolation of the tenant-to-seven-layer virtual network between tenants. The controller needs to generate a mapping relationship between the tenant ID and the namespace for each tenant.

The multi-level distributed virtual routing provides related three-layer network services such as routing, source address translation, and destination address translation for the virtualized network. The multi-level distributed virtual routes include: primary routing, secondary routing, and border gateway routing. The first-level route is distributed in each computing node of the virtualization platform; the secondary route and the border gateway route are provided by the high-performance network element device cluster that can perform related functions.

The primary routing logical entity resides inside the computing node of the virtualization platform, and provides a three-layer routing function for virtual machine communication between different virtual networks within the computing node, so that virtual machine communication between different networks within the same computing node does not need to occupy the data center network. Bandwidth, when the destination virtual machine does not exist in the same computing node, the primary route forwards the data packet to the secondary routing process, which can effectively reduce the east-west flow data between the virtual machines and improve the performance of the virtual network three-layer communication. The primary route accepts the control plane controller scheduling, implements the function of creating delete and update pseudo virtual routes, and adds and removes virtual network interfaces to virtual routes. At the same time, the virtual route created by the primary route can provide a source address translation/destination address translation service for the virtual machine in the computing node.

The secondary routing implements the three-layer routing communication function of the virtual machine to which the virtual network belongs between the computing nodes. The secondary routing is carried by a group of network element devices that can implement the routing virtualization function. Each network element device in the cluster can be controlled according to the control plane. The controller schedules the creation of delete updates and migration virtual routes, as well as the addition and removal of virtual network interfaces to virtual routes.

A border gateway route is a set of distributed routing entity clusters that provide gateway functions for data center networks to external networks. Each routing unit can run BGP/OSPF protocols and provide source address translation/destination addresses to the external network for the data center network. Conversion function.

Based on the above-mentioned multi-level distributed virtual routing, the virtual machine in the virtual network needs to undergo multiple address translations when accessing the external network. If the virtual machine in the virtual network communicates with the device in the internal public network of the data center, it needs to implement a primary route through the primary routing. Address translation/destination address translation. If the virtual machine in the virtual machine virtual network communicates with the external network of the data center, the secondary source address translation/destination address translation needs to be performed through the border gateway.

To improve data center network performance and availability, the secondary virtual route needs to support the Virtual Route Migration Protocol (VROOM), which can dynamically adjust the secondary virtual route, including virtual route load balancing scheduling migration and hanging node virtual Route recovery.

The controller can collect the metrics such as the hardware resource utilization, the network bearer bandwidth, and the energy saving requirement of the secondary route, and calculate the global resource. The controller migrates the virtual route according to the load balancing scheduling policy, and optimizes the use of the forwarding surface network element device resource. When the controller detects that a unit in the secondary routing cluster is hanged, all the virtual routes that it is responsible for are moved out. The controller clones the virtual route in the other units in the cluster according to the related scheduling policy. In order to ensure the normal operation of the virtual network Layer 3 routing function. After the dead node is restored, it needs to be initialized before re-access, so that it can be scheduled as a new resource.

The controller may provide a northbound interface, and provide the collected secondary routing metric factor data to a third party, and the third party formulates a customized secondary virtual routing scheduling policy.

4 is a schematic diagram of a virtual machine routing process between computing nodes and virtual networks according to the first embodiment of the present invention. As shown in FIG. 4, when the virtual machine first sends data to the target virtual machine, the following steps are included:

Step S402, the source virtual machine sends a data packet to the first virtual route FVR of the computing node;

Step S404, the source FVR is unknown to the destination virtual machine location, and requests the network controller CONTROLLER to install the forwarding policy.

Steps S406-S408, the CONTROLLER requests the virtualization platform Vplat to acquire the target virtual machine location;

Step S410, the CONTROLLER sends the forwarding policy to the source FVR according to the location of the target virtual machine;

Step S412: The target virtual machine is in the computing node, and the source FVR directly forwards the data packet to the destination virtual machine virtual network.

FIG. 5 is a schematic diagram of a virtual machine routing process between virtual nodes of different computing nodes according to the first embodiment of the present invention. As shown in FIG. 5, when the virtual machine first sends data to the target virtual machine, the following steps are included:

Step S502, the source virtual machine sends a data packet to the first virtual route FVR of the computing node;

Step S504, the source FVR is unknown to the destination virtual machine location, and requests the network controller CONTROLLER to install the forwarding policy.

Steps S506-S508, the CONTROLLER requests the virtualization platform Vplat to acquire the target virtual machine location;

Step S510, the CONTROLLER sends the forwarding policy to the source FVR according to the location of the target virtual machine;

Step S512, the target virtual machine does not exist in the computing node, repackage the data packet and forward it to the secondary virtual routing SVR according to the CONTROLLER installation policy;

Step S514, the secondary virtual route forwards the data to the destination gateway FVR;

In step S516, the destination FVR forwards the data packet to the destination host virtual network.

A border gateway route is a set of routing entity clusters that provide gateway functions for data center networks to external networks. Each routing unit can run BGP/OSPF protocols and provide source address translation/destination address translation to the external network for the data center network. .

Based on the above-mentioned multi-level distributed virtual routing, the virtual machine in the virtual network needs to undergo multiple address translations when accessing the external network. FIG. 6 is a virtual network and external network communication network address translation in the virtual network according to the first embodiment of the present invention. Translation, referred to as NAT) process diagram, as shown in Figure 6, if the virtual machine in the virtual network and the device in the internal public network of the data center need to access the primary address translation/destination address conversion through the primary route, if the virtual machine The virtual machine interacts with the external network of the data center in the virtual network, and needs to perform secondary source address translation/destination address translation through the border gateway.

In another embodiment, software is also provided for performing the technical solutions described in the above embodiments and preferred embodiments.

In another embodiment, a storage medium is also provided, in which the above software is stored, including but not limited to an optical disk, a floppy disk, a hard disk, an erasable memory, and the like.

It will be apparent to those skilled in the art that the various modules or steps of the present invention described above can be implemented by a general-purpose computing device that can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein. The steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps thereof are fabricated as a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software.

The above description is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Industrial applicability

As described above, the virtual routing system and method provided by the embodiments of the present invention have the following beneficial effects: solving the single problem of the virtual routing bearer node in the related art, improving the performance and availability of the data center network, and reducing the virtual network. The effect of communication on the bandwidth consumption of the physical network.

Claims

A virtual routing system comprising:

Level 1 routing, configured to provide Layer 3 routing functionality for virtual machine communication of different virtual networks within a compute node of a data center virtualization platform, and provide source address translation/destination address translation to a data center network external to the compute node Features;

The secondary route is configured to provide a three-layer routing function for virtual machine communication to which the different virtual networks belong to the computing node;

Border gateway routing, set to provide source address translation/destination address translation to the external network for the data center virtualization platform.
The virtual routing system according to claim 1, wherein

The primary route is located in the computing node;

The secondary route is located in a network element device that implements a route virtualization function;

The border gateway route is located in a routing entity that provides the data center network with a gateway function to the external network.
The virtual routing system of claim 1, wherein the secondary routing supports a virtual route dynamic migration protocol VROOM.
The virtual routing system of claim 1 wherein each of said virtual networks has a unique global identity.
The virtual routing system of claim 1, wherein the border gateway routing supports a Border Gateway Routing Protocol BGP or an Open Shortest Path First Protocol OSPF protocol.
The virtual routing system of claim 1 wherein the system further comprises:

a network controller configured to create or delete the primary route, and/or to add or remove an interface of the virtual network for the primary route; and/or to set up, delete, or migrate the border gateway route And/or an interface for adding or removing virtual networks for the border gateway.
A virtual routing method, applied to the virtual routing system according to any one of claims 1 to 6, comprising:

The source-level route receives the data packet sent by the source virtual machine to the target device, where the source-level route is a primary route corresponding to the source virtual machine;

The source-level route forwards the data packet to the virtual network where the target device is located if the target device is in the node where the source-level route is located; or

The source-level route forwards the data packet to the secondary route if the target device is located in the data center virtualization platform but is not in the node where the source-level routing is located; or

The source-level routing performs source address translation/destination address conversion to the internal network, and sends the data packet to the internal network in a data center network outside the computing node. Target device; or,

The source-level route performs source address translation/destination address conversion to the internal network and forwards the packet to the border gateway route if the target device is not in the data center network.
The method of claim 7 wherein

The source primary route requests a forwarding policy from the network controller;

The network controller requests the data center virtualization platform to acquire a location of the target device;

The network controller sends the forwarding policy to the source-level route according to the location of the target device, where the forwarding policy includes: in the node where the target device is located in the source-level routing In the case, the source-level route forwards the data packet to the virtual network where the target device is located; or the node where the target device is located in the data center virtualization platform but not at the source level The source-level route forwards the data packet to the secondary route; or, in the case that the target device is located in the data center network outside the computing node, the source-level routing pair Performing source address translation/destination address translation to the internal network and transmitting the data packet to the target device; or, if the target device is not in the data center network, the source level Routing, performing source address translation/destination address translation on the internal network to the data packet, and forwarding to the border gateway route;

The source primary route is forwarded according to the forwarding policy.
The method according to claim 8, wherein before the source-level routing requests the forwarding policy from the network controller, the method further includes:

The network controller assigns a unique global identity to each of the virtual networks.
The method of claim 8 wherein

In the case that the secondary route supports the virtual route dynamic migration protocol VROOM, the network controller performs scheduling migration on the secondary route load balancing, and/or recovers the suspended secondary route.
The method of claim 10, wherein before the network controller performs the scheduled migration of the secondary route load balancing, the method further includes:

The network controller determines, according to the collected metrics, whether to perform scheduling migration on the secondary routing load balancing, where the metric factor includes at least one of: hardware resource utilization of the secondary routing, network Carry bandwidth and energy saving requirements.
The method of claim 11, wherein before the network controller performs the scheduled migration of the secondary route load balancing, the method further includes:

The network controller provides the collected metrics to a third party through a northbound interface;

The network controller receives the scheduling policy of the third party, and performs scheduling migration on the secondary route load balancing according to the scheduling policy.
The method of claim 10, wherein before the network controller recovers the suspended secondary route, the method further includes:

The network controller migrates all the virtual routes that are responsible for running the secondary routes hanged in the secondary routing cluster, and sequentially clones the hanged secondary routes in other units of the secondary routing cluster.
The method according to claim 8, wherein before the source-level routing requests the forwarding policy from the network controller, the method further includes:

The network controller assigns a unique global identity to each of the virtual networks.
The method of claim 7, wherein after the source-level routing forwards the data packet to the secondary route, the method further includes:

The secondary route forwards the data packet to a target primary route corresponding to the target device, and forwards the data packet to the virtual network where the target device is located by the target primary route.
The method of claim 7, wherein after the source-level routing forwards the data packet to the border gateway route, the method further includes:

The border gateway routes the source address translation/destination address translation of the data packet to the external network and forwards to the external network.