WO2013185715A1

WO2013185715A1 - Method for implementing virtual network and virtual network

Info

Publication number: WO2013185715A1
Application number: PCT/CN2013/080734
Authority: WO
Inventors: 马苏安; 汪军; 胡永生; 梁亮
Original assignee: 中兴通讯股份有限公司
Priority date: 2012-09-18
Filing date: 2013-08-02
Publication date: 2013-12-19
Also published as: CN102857416B; CN102857416A

Abstract

A method for implementing a virtual network and a virtual network. The method comprises: a controller of the virtual network obtaining a transfer type, a source, and a destination of a data packet sent by a host; the controller formulating a forwarding strategy according to the transfer type, a port identifier of the source, and a destination identifier of the destination; and the controller sending the forwarding strategy to each virtual access switch and each intermediate switch through which the data packet is to be passed; and the virtual access switch and the intermediate switch executing the forwarding strategy on the received data packet.

Description

Method and virtual network for realizing virtual network

Technical field

The present invention relates to data exchange technology, and more particularly to a method and virtual network for implementing a virtual network.

Background technique

The data center provides resource leasing services. For different tenants, the data center needs to provide a virtual network to implement data exchange, so that tenants can use resources in the virtual network without interfering with each other. At the same time, the development of data center services has made data centers increasingly demanding networks, requiring more bandwidth and more reliable data networks.

The number of tenants may be hundreds of thousands or more. In the related art, a virtual local area network (VLAN) is used to isolate tenants. The maximum number of VLANs is 4096, which limits the number of tenants and cannot satisfy the data. The center provides business needs for multi-tenancy. As shown in Figure 1, a Layer 2 (L2, Layer 2) network is superimposed on the Internet Protocol (IP) network. The Ethernet packet is encapsulated in the User Datagram Protocol (UDP) packet. Add one. VXLAN (Virtual Extensible LAN) header, a 24-bit VXLAN Network Identifier (VNI, VXLAN Network Identifier) is added to the packet header to distinguish different virtual networks; the UDP packet forwarding path is determined by a routing protocol; Support for a large number of tenants, multi-path support depends on the underlying IP network. In the IP network 106 shown in FIG. 1, the original data packet 101 is sent from the source host (Host) 104 to the VXLAN Tunnel End Point (VTEP), and the VTEP encapsulates the original data packet 101 in the UDP packet to form a package. The data packet 102 is appended with a VXLAN header (Header), and the VXLan Header internally contains a VNI to identify different virtual networks. The encapsulated data packet 102 is sent to the peer VTEP through the IP network 106, the VTEP is decapsulated, the original data packet 103 is restored, and then forwarded to the destination Hostl07. It can be seen that after the original data packet 101 is encapsulated into the UDP packet, the underlying IP network forwards it as a payload, cannot identify which tenant the original data packet 101 belongs to, and cannot identify the IP address and other information inside the original data packet 101. No optimization or control can be done based on this information. The related technologies have the following problems: The underlying IP network and the virtual network are separated. When the data packets of the virtual network are transmitted in the underlying IP network, the IP network cannot identify the information encapsulated in the data packet, and cannot be optimized for the needs of the tenant. Measures or other strategies; The virtual network does not identify the forwarding mode of the packet in the underlying IP network, and cannot control the forwarding process.

Summary of the invention

Embodiments of the present invention provide a method and a virtual network for implementing a virtual network, which are solved by the bottom layer.

The IP network is separated from the virtual network. When the data packets of the virtual network are transmitted in the underlying IP network, the IP network cannot identify the information encapsulated in the data packet, and cannot implement optimization or other policies for the needs of the tenant; the virtual network does not recognize the data. The forwarding mode of the packet in the underlying IP network cannot control the forwarding process.

An embodiment of the present invention provides a method for implementing a virtual network, including:

The controller of the virtual network acquires the transmission type, source and destination of the data packet sent by the host;

Determining, by the controller, a forwarding policy according to the type of the transmission, the port identifier of the source, and the destination identifier of the destination;

The controller sends the forwarding policy to each virtual access switch and intermediate switch through which the data packet is to be passed, such that the virtual access switch and the intermediate switch perform the forwarding policy on the received data packet.

In the method, the virtual access switch and the intermediate switch perform the forwarding policy on the received data packet, including:

Receiving the received data packet as a first data packet, and packaging the destination identifier with the first data packet to form a second data packet;

Sending the second data packet.

In the method, the transmission type is to forward a data packet in the same subnet; the host that sends the data packet is located on the same subnet as the destination host that receives the data packet;

The destination identifier includes: a destination virtual network identifier of a subnet where the host that sends the data packet is determined according to the port identifier, and a destination route identifier of the destination virtual access switch. In the method, the transmission type is to forward a data packet between subnets; the destination identifier includes: a destination virtual network identifier of the destination subnet, and a destination route identifier of the destination virtual access switch.

In the method, the transmission type is unicast;

The step of the controller formulating the forwarding policy includes: calculating at least one optimized path between the host and the destination of the sending data packet; wherein the forwarding policy includes the at least one optimized path;

The virtual access switch and the intermediate switch perform the forwarding policy on the received data packet, including: the destination virtual access switch of the destination subnet restores the second data packet to a first data packet, and The first data packet is sent to the destination host.

In the method, the transmission type is broadcast;

Before the step of the controller of the virtual network acquiring the transmission type, the source, and the destination of the data packet sent by the host, the method further includes:

The controller establishes a first type of broadcast tree covering the entire subnet in each subnet; or

The controller establishes a second type of broadcast tree in each subnet, and in each of the second type of broadcast trees, selects a virtual access switch connected to the host in the subnet as the first The root node of the second type of broadcast tree.

In the method,

The destination identifier includes: a destination virtual network identifier of the destination subnet corresponding to the first type of broadcast tree of the destination or a destination virtual network identifier of the destination subnet corresponding to the second type of broadcast tree of the destination; the virtual access switch and The performing, by the intermediate switch, the forwarding policy to the received data packet includes: the destination virtual access switch of the destination subnet restores the second data packet to a first data packet, and sends the first data packet Give the destination host.

In the method, the step of the controller sending the forwarding policy to each virtual access switch and the intermediate switch that the data packet is to pass through includes:

The controller sends the forwarding policy by using an OpenFlow protocol, where the forwarding policy is Flow table implementation; or,

The controller sends the forwarding policy by using a network management interface download protocol, where the forwarding policy is implemented by a network management configuration entry; or

The controller delivers the forwarding policy by using a selected private protocol.

The embodiment of the invention further provides a controller, including:

An information obtaining unit, configured to: acquire a transmission type, a source, and a destination of the data packet sent by the host;

a policy making unit, configured to: formulate a forwarding policy according to the type of the transmission, the port identifier of the source, and the destination identifier of the destination;

a policy sending unit, configured to: send the forwarding policy to each virtual access switch and an intermediate switch that the data packet is to pass, so that the virtual access switch and the intermediate switch perform the foregoing on the received data packet Forwarding strategy.

In the controller, the policy making unit includes:

a first setting module, configured to: when the transmission type is to forward a data packet in the same subnet, setting the destination identifier includes: determining, according to the port identifier, a destination virtuality of a subnet where the host that sends the data packet is located The network identifier, and the destination route identifier of the destination virtual access switch.

In the controller, the policy making unit includes:

a second setting module, configured to: when the transmission type is to forward a data packet between subnets, setting the destination identifier includes: a destination virtual network identifier of the destination subnet, and a destination route identifier of the destination virtual access switch .

In the controller, the policy making unit includes:

And a third setting module, configured to: when the transmission type is unicast, calculate at least one optimized path between the host and the destination that sends the data packet; where the forwarding policy includes the at least one optimization path.

The controller further includes:

a broadcast tree unit, configured to: establish a first type of broadcast tree covering the entire subnet in each subnet; or, establish a second type of broadcast tree in each subnet, and in each of the Two In the class broadcast tree, a virtual access switch connected to the host in the subnet is selected as the root node of the second type of broadcast tree. In the controller, the policy making unit further includes:

And a fourth setting module, configured to: when the transmission type is broadcast, setting the destination identifier comprises: a destination virtual network identifier of a destination subnet corresponding to the first type of broadcast tree of the destination or a second category of the destination Destination virtual network ID of the destination subnet corresponding to the broadcast tree.

The embodiment of the present invention further provides a virtual network, including: a virtual access switch, an intermediate switch, and the foregoing controller; wherein each virtual access switch has a route identifier and is connected to the controller; each intermediate switch is The virtual access switch is connected to the controller, and each host is located in a subnet. All hosts in the subnet are in a Layer 2 broadcast domain, and the Layer 2 broadcast domain has a virtual network identifier. ; as well as

A host is connected to a virtual access switch through a port that has a port identifier. In the virtual network,

The virtual access switch includes:

a first policy execution unit, configured to: receive the received data packet as a first data packet, encapsulate the destination identifier with the first data packet to form a second data packet; and send the second data packet;

The intermediate switch includes:

a second policy execution unit, configured to: receive the received data packet as a first data packet, encapsulate the destination identifier with the first data packet to form a second data packet; and send the second data packet;

According to the technical solution of the embodiment of the present invention, in the process of data packet forwarding, the controller can obtain various attributes of the data packet as the destination identifier, formulate a forwarding policy according to the destination identifier, and forward the data packet more finely and flexibly, and support multi-tenancy. , multi-path forwarding and traffic load balancing; and more efficient use of network resources. BRIEF abstract

1 is a schematic diagram showing the working principle of the VXLAN technology in the related art;

2 is a block diagram showing a system of a virtual network according to an embodiment of the present invention;

3 is a schematic structural diagram of a data packet encapsulation according to an embodiment of the present invention;

4 is a schematic diagram showing a routing function between subnets according to an embodiment of the present invention;

FIG. 5 is a schematic diagram showing unicast in a subnet according to an embodiment of the present invention; FIG.

6 is a flowchart showing an implementation of an optimal path according to an embodiment of the present invention;

FIG. 7 is a flowchart showing a broadcast in a subnet according to an embodiment of the present invention;

FIG. 8 is a schematic structural diagram of a controller according to an embodiment of the present invention.

Preferred embodiment of the invention

The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, in the case of non-conflict, the features in the embodiments and the embodiments in the present application may be arbitrarily combined with each other.

Software Defined Network (SDN) technology is a communication network implementation method whose principle is control, forwarding and separation. The control plane of the SDN is composed of a centralized controller (ie, Forwarding Policy Server (FPS)), and controls the behavior of the forwarding plane through a development flow (OpenFlow) protocol. The control method is a flow table, and the flow table includes Match rules and behaviors that indicate behaviors that should be taken for packets that match the matching rules. After receiving the data packet, the forwarding plane of the SDN matches the flow entry in sequence, and executes the matching command after the match is found. If no match is found, the data packet can be forwarded to the controller and processed by the controller.

A tenant can have multiple subnets. Each host of the tenant belongs to a subnet. The subnet serves as a Layer 2 broadcast domain. Each Layer 2 broadcast domain is assigned a virtual network identifier (also known as a broadcast domain identifier (Broadcast ID). )). Each host is connected to a port on a virtual access switch. Each port has a unique port identifier. Each Virtual Access Switch is assigned a route ID by the controller.

An embodiment of the present invention provides a method for implementing a virtual network, including the following steps: Step 1: The controller of the virtual network acquires the transmission type, source and destination of the data packet sent by the host.

Step 2: The controller formulates a forwarding policy according to the transmission type, the port identifier of the source, and the destination identifier of the destination.

Step 3: The controller sends the forwarding policy to each virtual access switch and the intermediate switch that the data packet is to pass, so that the virtual access switch and the intermediate switch perform the forwarding policy on the received data packet.

Applying the technology provided by the embodiment of the present invention, in the process of data packet forwarding, the controller can acquire various attributes of the data packet, such as a Broadcast ID, a Route ID, an Ethernet packet header, an IP packet header, and a TCP (Transmission Control Protocol). The protocol/UDP port number is used as the destination identifier. The forwarding policy is formulated according to the destination identifier. The forwarding plane forwards the data packet more finely and flexibly. It supports multi-tenancy, multi-path forwarding, and traffic load balancing, making more efficient use of the network. Resources.

The structure of the virtual network is shown in Figure 2. The virtual network includes a forwarding plane and a control plane.

The forwarding plane includes a virtual access switch 202 and an intermediate switch 206; the intermediate switch 206 is not directly connected to the host 203 and is responsible for forwarding packets in the middle. The host 203 accesses the virtual network through the virtual access switch 202, and the host 203 includes a physical machine and a virtual machine.

The control plane includes a controller 201 that interconnects the virtual access switch 202 and the intermediate switch 206, and delivers a forwarding policy to the virtual access switch 202 and the intermediate switch 206 to control the forwarding behavior of the forwarding plane; the forwarding policy includes controlling which port the data packet is from. Forward, modify the properties of the packet, etc.

The following describes the working principle of each device in the virtual network.

The controller 201 is a device that constitutes a control plane, and the control logic of the entire virtual network is set on the controller 201. The controller 201 responds to the network event from the forwarding plane, and delivers the forwarding policy to each virtual access switch according to the virtual network requirement. 202 and intermediate switch 206 to implement control of the entire virtual network.

The virtual access switch 202 is a device that constitutes a forwarding plane and is responsible for connecting the host 203 to the virtual network.

The intermediate switch 206 is a device that constitutes a forwarding plane, and is an intermediate node of the virtual network, and is not directly Access to host 203. The virtual access switch 202 and the intermediate switch 206 forward the data packet according to the forwarding policy delivered by the controller 201.

Each host 203 is connected to a port of a virtual access switch 202.

The controller 201 is connected to the interface 204 of the forwarding plane, and the forwarding policy can be delivered by using the OpenFlow protocol, the network management protocol, or the private protocol.

The virtual machine manager (VMM) 205 is responsible for managing the host 203, and transmits information such as the Broadcast ID, the port identifier of the port where the host 203 is located, and the route ID to the controller 201.

The identifier of the host 203, the port identifier of the port where the host 203 is located, the virtual network identifier of the subnet to which the host 203 belongs, and the correspondence between them are stored in the controller 201.

Among the various attributes of the data packet, the destination identifier includes at least the destination Broadcast ID, and may also include the destination Route ID.

The source is the subnet where the source host that sent the packet is located;

The destination is the subnet where the destination host that receives the packet is located.

In a preferred embodiment, this information is provided by other systems, such as a virtualization management platform (VMP,

The Virtualization Management Platform) is transmitted to the controller 201 through the interface 204.

When the data packet enters the virtual network, the virtual network identifier and the route identifier are marked. When the data packet is transmitted in the virtual network, the forwarding plane forwards according to the routing identifier of the data packet; the forwarding plane distinguishes different virtual networks according to the virtual network identifier of the data packet. Packets can only be sent to hosts on the same virtual network.

The packet encapsulation mode is as shown in FIG. 3. The data packet 301 sent by the source host 304 includes a layer 2 packet header and data, and is encapsulated into a data packet 302 by the virtual access switch 202. In this process, the actual network label is used. For example, a 20-bit MPLS (Multi Protocol Label Switching) label is used to implement the Broadcast ID and the Route ID, and the Broadcast ID and the Route ID are encapsulated with the data packet 301 to form a data packet 302, or only the Broadcast ID and the Broadcast ID are The data packet 301 is encapsulated to form a data packet 303. The Broadcast ID indicates that the virtual network identifier and the Route ID are used to identify the destination virtual access switch. In this case, the MPLS label serves only as a carrier of the Broadcast ID and the Route ID, and does not have the original meaning of the MPLS label. 20bit Broadcast ID can Support for more than 1 million independent subnets, so it can support a large number of tenants.

After receiving the data packet 302, the destination virtual access switch removes the data packet 304 of the Broadcast ID and the Route ID, restores the state of the initial data packet 301, and sends the data packet 304 to the destination host 305.

The broadcast packet 303 is sent to all the broadcast domains because it is not forwarded to a specific Host.

Host, so there is no need to encapsulate the Route ID, but only the Broadcast ID.

There are various options for the interface protocol between the controller 201 and the forwarding plane. For example, the OpenFlow protocol can be used, and the forwarding policy can be implemented through a flow table. The network management interface can also be used to download the protocol. The forwarding policy is implemented by the network management configuration entry.

In a preferred embodiment, the controller 201 delivers the forwarding policy by using an OpenFlow protocol; the forwarding policy is implemented by a flow table;

Or,

The controller sends the forwarding policy by using a network management interface download protocol; the forwarding policy is implemented by a network management configuration entry;

Or,

The controller delivers the forwarding policy by using a proprietary protocol.

In a preferred embodiment, the steps of the virtual access switch 202 and the intermediate switch 206 performing the forwarding policy on the received data packet to perform the forwarding policy include:

The received data packet is used as a first data packet, and the destination identifier is encapsulated with the first data packet to form a second data packet;

Sending the second data packet.

The provided technical solution can forward data packets in the same subnet.

In a preferred embodiment, the transmission type is to forward a data packet in the same subnet; the destination identifier includes: a destination broadcast ID of a subnet where the host that sends the data packet is found according to the port identifier, and a destination. The purpose of the virtual access switch is the Route ID.

In an application scenario, when forwarding a data packet in a virtual network, the controller 201 applies two destination identifiers to the data packet by using a forwarding policy: Broadcast ID and Route ID, and the process package Including the following steps:

Step 01: When the host sends a data packet, the controller 201 queries the database according to the port where the host is located, finds the Broadcast ID corresponding to the subnet where the host is located, and the destination route ID of the destination virtual access switch, thereby forming a forwarding policy.

Step 02: The controller 201 sends a forwarding policy to the virtual access switch connected to the host, and the intermediate switch, and the intermediate switch is connected to the virtual access switch or other intermediate switch.

Step 03: When the forwarding policy is executed, the Broadcast ID and the Route ID are sealed in the first data packet to form a second data packet. To forward a packet in the same subnet, you need to type the Broadcast ID. Otherwise, you cannot distinguish which virtual network the packet is. When the second data packet with the broadcast ID and the route ID is transmitted in the subnet, the intermediate switch matches the destination identifier to ensure that the data packet is only sent to the destination host in the same subnet, thereby implementing the same subnet. The isolation of transport packets and the transmission of packets between subnets.

The provided technical solution can forward data packets between different subnets. The virtual network ID of the virtual network where the destination host is located is encapsulated in the data packet to ensure that only hosts in the destination virtual network can receive the data packet.

In a preferred embodiment, the transmission type is to forward data packets between subnets;

The destination identifier includes: the destination virtual network identifier of the destination subnet and the destination route identifier of the destination virtual access switch. The destination virtual access switch is located in the destination subnet.

The inter-subnet routing process is similar to the unicast process in the subnet. The difference is that the inter-subnet routing replaces the MAC (Medium Access Control) address of the packet because the source MAC address is sent when the packet is sent. The MAC address of the source host, the destination MAC address is the MAC address of the router; where the MAC address of the router is the MAC address of the router configured by the host (not the MAC address of the virtual access switch), and the router configured by the host is virtual, each The host configures the IP address of the router. When the host wants to send the data packet between the subnets, the host obtains the MAC address corresponding to the IP address of the router through the ARP (Address Resolution Protocol). The controller 201 replaces the MAC address. The virtual router answers this ARP request and returns the MAC address of the virtual router to the host. The packet passes through the virtual access switch, the source MAC address is replaced with the router's MAC address, and the destination MAC address is replaced with the source host's MAC address. In this process, you need to set the address of a virtual router, but you do not need to set up a physical router, and only complete the routing function by replacing the MAC address.

In an application scenario, as shown in FIG. 4, the OpenFlow protocol is used as an interface protocol between the controller 201 and the forwarding plane. The forwarding policy corresponds to the flow table defined in the OpenFlow protocol, and includes the following steps:

Step 401: The source host sends the first data packet to the source virtual access switch 202a connected thereto. Step 402: The source virtual access switch 202a forwards the first data packet to the controller 201 through the Packetln message because there is no matching flow table.

Step 403: The controller 201 sends a flow table to the source virtual access switch 202a. The flow table is configured to encapsulate the broadcast ID, the route ID, and the replacement MAC address. The replacement MAC address includes: replacing the source MAC address with the MAC address of the router, and the destination MAC address. Replace with the MAC address of the source host.

Step 404: The controller 201 sends a flow table to the destination virtual access switch 202b, where the flow table is responsible for popping the broadcast ID and the Route ID from the received first data packet to restore the data packet, and forwarding the restored data packet to the appropriate one. The destination host.

Step 405: The controller 201 sends the first data packet back to the source virtual access switch 202a through the PacketOut message.

Step 406: The source virtual access switch 202a matches the flow table mentioned in step 403, encapsulates the broadcast ID and the route ID, replaces the source MAC address with the MAC address of the router, and replaces the destination MAC address with the MAC address of the source host. Form a second data packet.

Step 407: The source virtual access switch 202a matches the flow table mentioned in step 403, and forwards the second data packet according to the optimal path.

Step 408: The intermediate switch 206 matches the flow table, and forwards the second data packet according to the optimal path. Step 409: The destination virtual access switch 202b matches the flow table mentioned in step 404, and pops up.

Broadcast ID and Route ID, recovering the first data packet from the second data packet;

Step 410: The destination virtual access switch 202b matches the flow table and forwards the first data packet to the destination host. Unicast packets can also be forwarded in the provided technical solution.

In a preferred embodiment, the transmission type is unicast;

The second step further includes: calculating at least one optimized path between the host and the destination of the sending data packet, where the forwarding policy includes the at least one optimized path;

The virtual access switch 202 and the intermediate switch 206 perform the forwarding policy on the received data packet, including: the destination virtual access switch of the destination subnet restores the second data packet to the first data packet, and The first data packet is sent to the destination host.

When the host sends a unicast data packet, the controller 201 queries the route ID of the virtual access switch where the destination host is located, and sends a forwarding policy. When the virtual access switch executes the forwarding policy, the route ID is encapsulated in the data packet.

The controller 201 calculates an optimal path between each two virtual access switches, and delivers a forwarding policy. The source virtual access switch 202a forwards the Route ID in the data packet matching the policy, and forwards it along a specific path until it arrives. The destination virtual access switch 202b forwards the data packet to the destination host by the destination virtual access switch 202b.

The controller 201 can also calculate multiple paths at the same time, so that the current data packet is forwarded according to a predefined forwarding policy, thereby implementing multipath load balancing. Different forwarding methods are determined by the controller 201, and different options can be determined at any time as needed.

The unicast packets in the subnet need to be labeled with the broadcast ID and the route ID. The two IDs are used to identify the destination virtual access switch 202b. The broadcast ID is used to uniquely identify a Layer 2 broadcast domain. Network), to ensure that hosts in the same subnet can receive this packet, and good isolation between subnets.

In an application scenario, as shown in FIG. 5, the OpenFlow protocol is used as an interface protocol between the controller 201 and the forwarding plane, and the forwarding policy corresponds to the flow table defined by the OpenFlow protocol, including the following steps:

Step 501: The source host sends the first data packet to the source virtual access switch 202a connected thereto. Step 502: The first data packet has no matching flow table on the source virtual access switch 202a, and the source virtual access switch 202a sends a Packetln message to the controller 201. The Packetln message includes the content and input port of the first data packet. Step 503: The controller 201 determines, according to the content of the Packetln message, a flow table that needs to be matched, and sends a flow table to the source virtual access switch 202a, where the flow table is responsible for encapsulating the broadcast ID and the Route ID.

Step 504: The controller 201 sends a flow table to the destination virtual access switch 202b, where the flow table is responsible for popping the Broadcast ID and the Route ID from the first data packet, and forwarding the first data packet to the appropriate host;

The flow table can be downloaded when the Packetln message is received, or it can be downloaded before the host sends the packet.

Step 505, the controller 201 sends the first data packet back to the source virtual access switch 202a through the PacketOut message;

Step 506: The source virtual access switch 202a matches the data packet to the flow table, performs the behavior specified by the flow table, and encapsulates the first data packet with the Broadcast ID and the route ID to form a second data packet.

Step 507: The source virtual access switch 202a matches the flow table, and forwards the second data packet according to the optimal path.

Step 508, the intermediate switch 206 matches the flow table, and forwards the second data packet according to the optimal path. Step 509: The destination virtual access switch 202b matches the second data packet with the flow table, and pops the broadcast ID and the Route ID from the second data packet. Restore the first data packet;

Step 510: The destination virtual access switch 202b matches the flow table, and forwards the restored first data packet to the destination host.

Each packet is forwarded by the optimal path or by a specific policy corresponding to the tenant. For example, different routing paths can be selected according to the level to which the tenant belongs. In this process, each node in the routing path needs to be set. Flow table.

In various embodiments, there are multiple ways to implement the optimal path for forwarding the forwarding policy. The forwarding policy may be sent immediately when the network topology changes, or the optimal path may be calculated first. Then, the forwarding policy is delivered. As shown in Figure 6, the process of implementing the optimal path includes the following steps:

Step 601: The controller 201 assigns a Route ID to each virtual access switch 202 and detects Network topology.

When the OpenFlow protocol is used as the protocol between the control plane and the forwarding plane, the controller 201 detects the network topology. The controller 201 instructs each port to send a Link Layer Discovery Protocol (LLDP) packet. If the LLDP packet is received on another port on the forwarding plane, then there is a link between the two ports.

Step 602: Calculate an optimal path between each virtual access switch 202 and other virtual access switches 202.

Step 603: The controller 201 sends a forwarding policy to the virtual access switch 202 and the intermediate switch 206. The forwarding policy indicates that the data packet from the virtual access switch 202 to the destination virtual access switch 202b should be output, and the matching field is Route ID.

The broadcast data packet can also be forwarded in the provided technical solution.

In a preferred embodiment, the transmission type is broadcast;

Before step 1, the method further includes:

The controller establishes a first type of broadcast tree covering the entire subnet in each subnet; or, the controller establishes a second type of broadcast tree in each subnet, and in each of the In the second type of broadcast tree, a virtual access switch connected to the host in the subnet is selected as the root node of the second type of broadcast tree.

There are several ways to establish a broadcast tree. You can set up a broadcast tree when sending broadcast packets, or you can set up a broadcast tree in advance. The broadcast tree can be set up for each subnet, or you can connect Virtual Access to each host in each subnet. The Switch creates multiple ports. Since the node of the broadcast tree is related to the location of the host, when the host location changes, the broadcast tree is re-established and the flow table is updated.

In a preferred embodiment, the destination identifier includes: a destination virtual network identifier of a destination subnet corresponding to the first type of broadcast tree of the destination or a destination virtual network identifier of the destination subnet corresponding to the second type of broadcast tree of the destination;

In an application scenario, as shown in Figure 7, the broadcast packets in the subnet need to be encapsulated by the MPLS standard. Signed the implementation of the Broadcast ID. The OpenFlow protocol is used as an example of the interface protocol between the controller 201 and the forwarding plane. The forwarding policy corresponds to the flow table defined by the OpenFlow protocol. The controller 201 establishes a broadcast tree by sending the flow table to propagate the data packet according to the broadcast tree. The method includes the following steps: Step 701: The source host sends the first data packet to the source virtual access switch 202a connected to the source host.

Step 702: The source virtual access switch 202a does not have a matching flow table, and forwards the first data packet to the controller 201 through the Packetln message.

Step 703: The controller 201 sends a flow table to the source virtual access switch 202a, where the flow table is responsible for encapsulating the Broadcast ID for the first data packet.

Step 704: The controller 201 establishes a broadcast tree according to the network topology, and sends a flow table to the intermediate switch 206 corresponding to the broadcast tree node according to the result of establishing the broadcast tree, where the flow table is responsible for forwarding the second data packet in the broadcast tree.

Step 705: The controller 201 sends a flow table to the destination virtual access switch 202b. The flow table is responsible for popping the Broadcast ID from the second data packet and forwarding the second data packet to the appropriate destination host.

Step 706, the controller 201 sends the first data packet back to the source virtual access switch 202a through the PacketOut message.

Step 707: The source virtual access switch 202a matches the flow table, and encapsulates the first data packet with the Broadcast ID to form a second data packet.

Step 708, the source virtual access switch 202a matches the flow table, and forwards the second data packet according to the broadcast tree. Step 709, the intermediate switch 206 allocates a flow table, and forwards the data packet according to the broadcast tree; the data packet arrives at the destination virtual access switch 202b.

Step 710: The destination virtual access switch 202b matches the flow table, and pops the Broadcast ID from the second data packet to restore the first data packet.

Step 711: The destination virtual access switch 202b matches the flow table, and forwards the first data packet to the destination host. An embodiment of the present invention provides a controller, as shown in FIG. 8, including:

The information obtaining unit 801 is configured to: acquire a transmission type, a source, and a destination of the data packet sent by the host;

a policy making unit 802, configured to: formulate a forwarding policy according to the type of the transmission, the port identifier of the source, and the destination identifier of the destination;

The policy issuance unit 803 is configured to: send the forwarding policy to each of the virtual access switches and the intermediate switch that the data packet is to pass, so that the virtual access switch and the intermediate switch pair the received data packet The forwarding policy is executed.

In a preferred embodiment, policy formulation unit 802 includes:

The first setting module 8021 is configured to: when the transmission type is to forward a data packet in the same subnet, the setting the destination identifier comprises: determining, according to the port identifier, a destination virtuality of a subnet where the host that sends the data packet is located The network identifier and the destination route identifier of the destination virtual access switch.

In a preferred embodiment, policy formulation unit 802 includes:

The second setting module 8022 is configured to: when the transmission type is to forward a data packet between subnets, setting the destination identifier includes: a destination virtual network identifier of the destination subnet, and a destination route of the destination virtual access switch Logo.

In a preferred embodiment, policy formulation unit 802 includes:

The third setting module 8023 is configured to: when the transmission type is unicast, calculate at least one optimized path between the host and the destination that sends the data packet; wherein the forwarding policy includes the at least one Optimize the path.

In a preferred embodiment, as shown in FIG. 8, the controller further includes:

a broadcast tree unit 804, configured to: establish a first type of broadcast tree covering the entire subnet in each subnet; or, establish a second type of broadcast tree in each subnet, and in each of the In the second type of broadcast tree, a virtual access switch connected to the host in the subnet is selected as the root node of the second type of broadcast tree.

In a preferred embodiment, policy formulation unit 802 includes:

a fourth formulation module 8024, configured to: when the transmission type is broadcast, set the target The identifier includes: a destination virtual network identifier of the destination subnet corresponding to the first type of the broadcast tree of the destination or a destination virtual network identifier of the destination subnet corresponding to the second type of the broadcast tree of the destination.

The embodiment of the present invention provides a virtual network, as shown in FIG. 2, including: a virtual access switch 202, an intermediate switch 206, and a controller 201;

Each virtual access switch 202 has a routing identifier and is coupled to the controller 201;

Each intermediate switch is connected to the virtual access switch 202 and connected to the controller 201;

Each host 203 is located in a subnet, and all hosts 203 in the subnet are in a Layer 2 broadcast domain, and the Layer 2 broadcast domain has a virtual network identifier;

One host 203 is connected to the virtual access switch 202 through a port having a port identifier.

In a preferred embodiment, as shown in FIG. 8, the controller includes:

The policy making unit 802 is configured to: formulate a forwarding policy according to the transmission type, the port identifier of the source, and the destination identifier of the destination;

The policy issuance unit 803 is configured to: send the forwarding policy to each of the virtual access switches and the intermediate switch that the data packet is to pass, so that the virtual access switch and the intermediate switch pair the received data packet Performing the forwarding policy;

In a preferred embodiment, the virtual access switch includes:

a first policy execution unit, configured to: use the received data packet as the first data packet, The destination identifier is encapsulated with the first data packet to form a second data packet; and the second data packet is sent;

as well as,

Intermediate switches include:

And a second policy execution unit, configured to: use the received data packet as the first data packet, and encapsulate the destination identifier with the first data packet to form a second data packet; and send the second data packet.

By using the solution of the embodiment of the present invention, multi-tenant support, multi-path, and traffic load balancing capability are provided by applying virtual network labels and routing labels to data packets; and the data network can be more comprehensively managed and implemented more flexibly. Data forwarding strategies, even for the optimization and control of traffic for specific tenant needs.

One of ordinary skill in the art will appreciate that all or a portion of the above steps may be accomplished by a program instructing the associated hardware, such as a read-only memory, a magnetic disk, or an optical disk. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiment may be implemented in the form of hardware or in the form of a software function module. Embodiments of the invention are not limited to any particular form of combination of hardware and software.

The above is a preferred embodiment of the embodiments of the present invention. It should be noted that those skilled in the art can make some improvements and refinements without departing from the principles of the embodiments of the present invention. Improvements and retouching should also be considered as protection of the present invention.

Industrial applicability

According to the technical solution of the embodiment of the present invention, in the process of data packet forwarding, the controller can obtain various attributes of the data packet as the destination identifier, formulate a forwarding policy according to the destination identifier, and forward the data packet more finely and flexibly, and support multi-tenancy. , multi-path forwarding and traffic load balancing; and more efficient use of network resources.

Claims

Claim

1. A method for implementing a virtual network, comprising:

The controller of the virtual network acquires a transmission type, a source location, and a destination of a data packet sent by the host;

2. The method according to claim 1, wherein the virtual access switch and the intermediate switch perform the forwarding policy on the received data packet, including:

Sending the second data packet.

3. The method according to claim 2, wherein the transmission type is to forward a data packet in the same subnet; the host that sends the data packet is located on the same subnet as the destination host that receives the data packet; and the destination identifier includes: The destination virtual network identifier of the subnet where the host that sends the data packet is determined according to the port identifier, and the destination route identifier of the destination virtual access switch.

4. The method according to claim 2, wherein the transmission type is to forward a data packet between subnets;

The destination identifier includes: a destination virtual network identifier of the destination subnet, and a destination route identifier of the destination virtual access switch.

5. The method according to claim 1, wherein the transmission type is unicast; the step of the controller formulating a forwarding policy comprises: calculating at least one optimization between a host and a destination of the sending data packet a path, where the forwarding policy includes the at least one optimized path;

The virtual access switch and the intermediate switch perform the forwarding on the received data packet The policy includes: the destination virtual access switch of the destination subnet restores the second data packet to a first data packet, and sends the first data packet to a destination host.

6. The method according to claim 1, wherein the transmission type is broadcast; before the step of the controller of the virtual network acquiring a transmission type, a source and a destination of a data packet sent by the host, the method Also includes:

7. The method according to claim 6, wherein

8. The method according to claim 1, wherein the step of the controller transmitting the forwarding policy to each virtual access switch and the intermediate switch to which the data packet is to pass includes: the controller adopting an open flow The forwarding policy is delivered by the protocol; where the forwarding policy is implemented by a flow table; or

9. A controller comprising:

a policy making unit, configured to: according to the type of transmission, the port identifier of the source, the destination The destination identifier of the place to formulate a forwarding strategy;

The controller according to claim 9, wherein the policy making unit comprises: a first formulating module, configured to: set the purpose when the transfer type is to forward a data packet in the same subnet The identifier includes: a destination virtual network identifier of a subnet where the host that sends the data packet is determined according to the port identifier, and a destination route identifier of the destination virtual access switch.

The controller according to claim 9, wherein the policy making unit comprises: a second formulating module, configured to: set the purpose when the transfer type is to forward a data packet between subnets The identifier includes: a destination virtual network identifier of the destination subnet, and a destination route identifier of the destination virtual access switch.

The controller according to claim 9, wherein the policy making unit comprises: a third formulating module, configured to: when the transfer type is unicast, calculate a host and a destination for sending a data packet At least one optimized path between the at least one optimized path;

13. The controller of claim 9, further comprising:

a broadcast tree unit, configured to: establish a first type of broadcast tree covering the entire subnet in each subnet; or, establish a second type of broadcast tree in each subnet, and in each of the In the second type of broadcast tree, one of the subnets is selected to be a virtual access switch connected to the host as a root node of the second type of broadcast tree.

The controller according to claim 12, wherein the policy making unit comprises: a fourth formulating module, configured to: when the transmission type is broadcast, setting the destination identifier comprises: The destination virtual network identifier of the destination subnet corresponding to the broadcast tree or the destination virtual network identifier of the destination subnet corresponding to the second type of broadcast tree of the destination.

A virtual network, comprising: a virtual access switch, an intermediate switch, and the controller according to any one of claims 9-14; Each virtual access switch has a routing identifier and is connected to the controller; each intermediate switch is connected to the virtual access switch and connected to the controller; each host is located in a subnet, All hosts in the subnet are in a Layer 2 broadcast domain, and the Layer 2 broadcast domain has a virtual network identifier;

A host is connected to a virtual access switch through a port that has a port identifier.

16. The virtual network according to claim 15, wherein

The virtual access switch includes:

The intermediate switch includes: