WO2021189821A1

WO2021189821A1 - Network fragmentation control method and network fragmentation system

Info

Publication number: WO2021189821A1
Application number: PCT/CN2020/121096
Authority: WO
Inventors: 汪漪; 李伟超; 黄勇
Original assignee: 鹏城实验室
Priority date: 2020-03-27
Filing date: 2020-10-15
Publication date: 2021-09-30
Also published as: CN111416747A; CN111416747B

Abstract

The present application provides a network fragmentation control method and a network fragmentation system. The method is applied to a network topology structure. The network topology structure comprises a plurality of network sites, each network site comprises at least one link fragmentation switch, the at least one link fragmentation switch is connected to at least one node fragmentation switch, and the link fragmentation switches in the network sites are connected by means of physical links. The method comprises: receiving a configuration request of a network fragment; and configuring the link fragmentation switch and the node fragmentation switch in the network site according to the configuration request. Obviously, the network topology of the present application can be built on the basis of existing commercial networks, and only the node fragmentation switch connected to the link fragmentation switch needs to be added in network sites of the existing commercial networks. Compared with the prior art that an experiment network is separately built, construction and maintenance costs are greatly reduced.

Description

Network fragmentation control method and network fragmentation system

Technical field

This application claims the priority of a domestic application filed with the Chinese Patent Office on March 27, 2020, the application number is 202010230857.8, and the invention title is "a network fragmentation control method and network fragmentation system", the entire content of which is incorporated by reference In this application.

Background technique

With the development of Internet technology, the Internet has penetrated more and more deeply into people's work and life. While providing convenience to users, people have been making in-depth explorations of the Internet.

Due to the large investment in network construction, it is difficult for small research groups to actually put some innovative network ideas into practice due to lack of funds. Establishing an experimental network for future network research can accelerate network innovation. Through the trial network, multiple different network fragments can be established in the network, that is, a physical network is divided into several logical networks. Each logical network can guarantee different business traffic, and can have different network topologies, or Adopt different network protocols/technical composition. Different network fragments can be used to implement different network experiments.

There are also many international projects on the experimental bed, but they have not been very successful. The reason is that the pure experimental network is separated from the existing commercial network, and a large amount of capital and maintenance personnel need to be invested separately, which is too costly and uneconomical.

Summary of the invention

In view of this, the present application provides a network fragmentation control method and a network fragmentation system, which are used to solve the problem of high cost of establishment and maintenance of the existing experimental network. The technical scheme is as follows:

A network slicing control method is applied to a network topological structure, the network topological structure includes a plurality of network sites, each of the network sites includes at least one link slicing switch, and at least one of the link slicing switches is connected There is at least one node fragmentation switch, and the link fragmentation switches in each of the network sites are connected through a physical link;

The method includes:

Receive the configuration request of the network segment;

According to the configuration request, the link fragmentation switch and the node fragmentation switch in the network site are configured.

Preferably, the receiving the configuration request of the network segment includes:

The configuration information of the network segment defined by the user on the network service platform is received, the configuration information includes basic configuration information of virtual nodes, inter-node link configuration information, and link bandwidth parameter configuration information.

Preferably, the configuring the link fragmentation switch and the node fragmentation switch in the network site according to the configuration request includes:

Determining, based on the configuration request, a node fragment switch to be configured corresponding to the virtual node and a target link fragment switch connected to the user terminal and the node fragment switch to be configured;

Separating the virtual link configuration request and the virtual node configuration request from the configuration request;

Determine all the link fragment switches on the virtual link to be configured based on the target link fragment switch;

Configure the link fragment switch to be configured according to the virtual link configuration request, and configure the node fragment switch to be configured according to the virtual node configuration request.

Preferably, the determining all the link fragment switches on the virtual link to be configured based on the target link fragment switch includes:

Sending a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

The link controller calculates the path information of the virtual link to be configured based on the configuration request, and determines all the link fragment switches on the virtual link to be configured on the virtual link to be configured according to the path information;

The configuring the link fragment switch to be configured according to the virtual link configuration request includes:

The link controller controls the to-be-configured link fragment switch to configure the virtual link according to the virtual link configuration request.

Preferably, the configuring the segment switch of the node to be configured according to the virtual node configuration request includes:

Sending the virtual node configuration request to the deep node programming controller;

The in-depth node programming controller configures the node slice switch to be configured.

Preferably, the configuration of the segment switch of the node to be configured by the deep node programming controller includes:

The deep node programming controller obtains the code of the virtual node to be installed based on the virtual node configuration request;

The deep node programming controller queries whether a virtual node has been installed on the slice switch of the node to be configured;

If not, sending the code of the virtual node to be installed to the node fragment switch to be configured, and the node fragment switch to be configured completes the configuration according to the code;

If yes, the deep node programming controller queries whether the idle resources on the shard switch of the node to be configured meet the resource requirements of the virtual node to be installed;

If it is satisfied, the code of the virtual node to be installed is merged with the code of the virtual node that has been installed on the shard switch of the node to be configured, and the merged code is sent to the node shard switch to be configured, and The slicing switch of the node to be configured completes the configuration according to the merged code.

A network slicing system includes: a control system and a number of network sites, each of the network sites includes at least one link slicing switch, and at least one of the link slicing switches is connected to at least one node slicing switch ；

The control system is used to perform configuration management on the network segment requested for configuration;

The link fragment switches in each of the network sites are connected by physical links, and perform link resource configuration according to the virtual link configuration request in the network fragment configuration request under the control of the control system;

The node slicing switch in each network site performs node forwarding processing resource configuration according to the virtual node configuration request in the network slicing configuration request under the control of the control system; the configured virtual link and virtual The network topology formed by the nodes serves as a network fragment.

Preferably, the control system includes an orchestrator, a link controller and a deep node programming controller, wherein:

The orchestrator is configured to determine, based on the network fragment configuration request, the node fragment switch to be configured corresponding to the virtual node and the target link fragment switch connected to the user terminal and the node fragment switch to be configured, and from The virtual link configuration request and the virtual node configuration request are separated from the network fragment configuration request;

The process that the link fragmentation switch performs link resource configuration according to the virtual link configuration request in the network fragmentation configuration request under the control of the control system includes:

The link controller interacts with the orchestrator, determines all the to-be-configured link-slice switches on the virtual link to be configured based on the target link-slice switch, and requests all the links based on the virtual link configuration request. The link fragment switch to be configured is configured;

The process of the node fragmentation switch performing node forwarding processing resource configuration according to the virtual node configuration request in the network fragmentation configuration request under the control of the control system includes:

The deep node programming controller interacts with the orchestrator, and configures the node fragment switch to be configured based on the virtual node configuration request.

Preferably, the link controller interacts with the orchestrator, determines all the link fragment switches to be configured on the virtual link to be configured based on the target link fragment switch, and configures based on the virtual link The process of requesting the configuration of the link fragment switch to be configured includes:

The orchestrator sends a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

The link controller controls the link fragment switch to be configured to configure a virtual link.

Preferably, the deep node programming controller interacts with the orchestrator, and configures the node shard switch to be configured based on the virtual node configuration request, including:

The orchestrator sends the virtual node configuration request to the deep node programming controller;

The deep node programming controller configures the node fragment switch to be configured.

Preferably, the process of the in-depth node programming controller configuring the segment switch of the node to be configured includes:

Preferably, the node fragment switch is a deep programmable white box switch.

The network slicing control method provided in this application is applied to a network topology structure that includes a number of network sites, wherein each network site includes at least one link slicing switch, and at least one link slicing switch is connected with At least one node slicing switch, and the link slicing switches in each network site are connected by physical links. Obviously, the network topology of this case can be built based on the existing commercial network. You only need to add a link to the network site of the existing commercial network. The node slicing switch connected to the channel slicing switch is sufficient. The present application can receive a configuration request for network fragmentation, and then configure link fragmentation switches and node fragmentation switches in a network site according to the request, so as to realize the required network fragmentation configuration in the network topology. In this case, the link fragmentation switch can realize the configuration of virtual link resources, and the node fragmentation switch can realize the configuration of virtual node resources, so that the configured virtual link and virtual node form a network fragment. Users can configure as many as needed. Different network shards can be used for commercial and experimental purposes at the same time. Compared with the existing technology, the construction and maintenance cost of the experimental network is greatly reduced.

Moreover, in this case, in-depth network programming can be supported by setting up node fragment switches. At the same time, at the data link and node layer, it provides full coverage of network fragmentation capabilities and isolation between fragmented networks. Furthermore, since the commercial network and the test network are built on the same physical network, the commercial service traffic can be flexibly guided into the test network, so that the commercial service traffic can be fully utilized to verify the network technology protocol of the test network, making the network experiment more complete.

Description of the drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

Figure 1 illustrates a schematic diagram of the data layer of a network slicing system;

FIG. 2 illustrates a diagram of an embodiment of the data layer of the network slicing system of the present application;

Figure 3 illustrates the schematic diagram of the logical path of VPN+QoS technology;

Figure 4 illustrates the schematic diagram of the logical path of TE technology;

Figure 5 illustrates a schematic diagram of the logical function of a node fragment switch;

Figure 6 illustrates a schematic diagram of the control level control logic of a network slicing system;

Figure 7 illustrates a schematic structural diagram of a deep node programming controller;

Figure 8 illustrates a schematic flow chart of a method for controlling network fragmentation.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

This application provides a network slicing system, the data layer of which includes a network topology structure composed of several network sites, and the control layer of which is composed of a control system.

Among them, the control system is used to perform configuration management on the network segment requested for configuration.

Refer to Figure 1, which illustrates a schematic diagram of a data layer of a network slicing system.

As can be seen from Figure 1, the data plane of the network slicing system is composed of several network sites, which are connected by optical fibers according to a certain topology.

Each network site contains at least one link fragmentation switch, and at least one link fragmentation switch is connected to at least one node fragmentation switch.

The link slicing switches in each network site are connected by physical links to realize the fiber optic interconnection between the sites. Under the control of the control system, the links between different network sites are in accordance with the virtual chain in the network slicing configuration request. Link configuration requests for link resource configuration.

Link fragmentation switches in the same network fragment establish virtual links through physical links between link fragmentation switches and physical links with node fragmentation switches. The node sharding switch performs node forwarding processing resource configuration according to the virtual node configuration request in the network slicing configuration request under the control of the control system, where the node forwarding processing resources include but are not limited to resources such as flow tables and queues.

Finally, the network topology formed by the configured virtual links and virtual nodes is regarded as a network fragment. Users can configure multiple different network fragments according to their needs for different network experimental purposes.

In Figure 1, different filled rectangles represent three different network fragments. At different network sites, virtual nodes of different network fragments can be deployed. The virtual nodes in a network fragment are connected by virtual links. . Different network fragments can eventually realize different network topologies.

The network fragmentation control system provided above in this application includes a network topology composed of several network sites at the data level, wherein each network site contains at least one link fragmentation switch, and at least one link fragmentation switch is connected to at least one Node fragmentation switch, the link fragmentation switches in each network site are connected by physical links. Obviously, the network topology of this case can be built based on the existing commercial network. You only need to add the link division in the network site of the existing commercial network. The node to which the chip switch is connected is just the chip switch. The present application can receive a configuration request for network fragmentation, and then configure link fragmentation switches and node fragmentation switches in a network site according to the request, so as to realize the required network fragmentation configuration in the network topology. In this application, the link slicing switch can realize the configuration of virtual link resources, and the node slicing switch can realize the configuration of virtual node resources, so that the configured virtual link and virtual node form a network slice, and the user can configure it as needed Multiple network shards and different network shards can be used for commercial and experimental purposes at the same time. Compared with the existing technology, the construction and maintenance cost of a separate experimental network is greatly reduced.

In the above embodiment, the link fragmentation switch in the network site can be constructed using mature and stable commercial switches, and mature technologies are used to achieve resource splitting of physical links, such as SR (Segment Routing), MPLS, VxLan, Vlan, etc. technology.

In addition, the node fragmentation switch can adopt a deeply programmable white box switch. The main capability of the white box switch is to virtualize the forwarding plane and realize the flexible resource division of nodes. The white box switch is typically a programmable switch supporting the P4 programming language.

Adopting the above-mentioned link fragmentation switch and node fragmentation switch structure can bring the following benefits: While the network can stably support commercial service bearer, it can also support network fragmentation and innovative network experiments. This is because if any white box switch used for node slicing fails, it will not affect the business traffic carried by the commercial switch in the network site.

Next, in conjunction with FIG. 2, a specific embodiment is used to illustrate how the network fragmentation of the present application is implemented by the link fragmentation switch and the node fragmentation switch at the data level.

As shown in Figure 2, define a network composed of three link fragment switches LN1, LN2, LN3, and a node fragment switch SN100. Among them, SN100 and LN2 are located in the same network site. The physical port 11 and the physical port 22 of the SN100 are respectively connected to the physical ports P5 and P6 of the LN2 through physical links. In addition, the physical port P1 of LN1 and the physical port P2 of LN2 are connected by a physical link. The physical port P3 of LN2 and the physical port P4 of LN3 are connected by a physical link.

LN1 is connected to terminals T1, T2, and T3 through physical ports P10, P11, and P12, respectively; LN3 is connected to terminals T4, T5, and T6 through physical ports P20, P21, and P22, respectively.

Now I hope to use this network to establish three network fragments, namely:

T1 is connected to the T6 terminal through network segment 1. Network segment 1 does not require programmability and only requires stable connection bandwidth;

T2 is connected to a virtual switch A through network segment 2, and then connected to the T4 terminal through this virtual switch for network experiment 1;

T3 is connected to another virtual switch B through network segment 3, and then connected to the T5 terminal through this virtual switch for network experiment 2.

Based on the above requirements, the node slicing switch SN100 can be programmed through the control system of the control plane of the present application to virtualize two switches, namely virtual switch A and virtual switch B. Virtual switch A has two virtual ports 1 and 2, and virtual switch B has two virtual ports 3 and 4.

Further, the configuration can be issued to the link fragmentation switches LN1, LN2, and LN3 through the control system to establish path 1, path 2, path 3, path 4, and path 5, that is, establish virtual links 1-5. The paths 1-5 are as follows:

Path 1: T1-P10-P1-P2-P3-P4-P22-T6;

Path 2: T2-P11-P1-P2-P5-physical port 11-virtual port 1;

Path 3: T3-P12-P1-P2-P5-physical port 11-virtual port 3;

Path 4: virtual port 2-physical port 22-P6-P3-P4-P20-T4;

Path 5: virtual port 4-physical port 22-P6-P3-P4-P21-T5.

It can be seen from the above that multiple paths can share physical ports and physical links, that is, they belong to logical paths. When multipath shares physical link bandwidth, multiple technologies can be used to allocate physical link bandwidth resources to multipath.

Next, several optional implementations of multipath are introduced.

The first way to achieve this is to use VPN+QoS technology.

Use different VPN IDs in the message to distinguish the traffic of different network fragments, and use QoS technology to ensure the bandwidth of the corresponding path. Typically, VxLan technology is used to implement L2VPN. Allows virtual network nodes to be deployed in a flexible location on the network. For physical network sites that have not deployed virtual network nodes in network fragments, VxLan technology can achieve pass-through. In addition, when the network scale is not large, VLAN technology can also be used.

You can refer to the example in Figure 3 to specifically illustrate the solution implemented by the above-mentioned path 2 using the VxLan+QoS technology:

Implement VTEP100 and VTEP101 in LN1 switch and LN2 switch respectively. VTEP (VxLan Tunnel Endpoint) is the endpoint function of VxLan. It can encapsulate MAC encapsulated messages into VxLan encapsulation. VxLan is on top of UDP protocol, so it can realize long-distance L2VPN interconnection through IP network. A VxLan connection is established between VTEP100 and VTEP101 through VNID 100, which is shown as VxLan 100 in the figure. User terminal T2 corresponds to VxLan100 on LN1 through Vlan10, and SN100 is connected to VxLan100 on LN2 through vlan25. To ensure bandwidth, match Vlan10, VxLan100 and Vlan25 on the corresponding physical ports of LN1 and LN2, and implement QoS strategies to ensure the required bandwidth.

The second way of implementation: using TE (Traffic Engineering) technology.

TE technology uses label switching/or label stack technology to flexibly plan the physical path of service traffic. And when there is a multi-hop connection between two virtual nodes in a network fragment, the TE technology has a lower forwarding delay than the VxLan technology. In TE technology, there are two kinds of MPLS-TE and SR-TE. MPLS-TE technology is relatively mature, and SR-TE technology has the potential to become the next-generation network TE technology because of its multi-homing, simple protocol, and automatic discovery and construction capabilities.

Still taking the implementation of the above path 2 as an example, refer to the example of Figure 4 for a specific description of the solution implemented using TE technology:

Two TE tunnels are established between the P11 port of the terminal T2 connected to LN1 and the P5 port of SN100 connected to LN2, and vlan10 is mapped to the tunnel on the P11 port, and vlan25 is mapped to the tunnel on the P5 port. Configure the TE policy to ensure the bandwidth of TE, and then path 2 can be realized.

It should be noted that, regardless of the path implementation scheme 1 or scheme 2, the packets entering the SN100 have a vlan25 TAG, so that the SN100 can recognize that the traffic belongs to the path 2 and inject it into the virtual port 1.

In another embodiment of the present application, an optional implementation manner of a node fragment switch is introduced.

As shown in FIG. 5, FIG. 5 illustrates the implementation logic function diagram of a node fragment switch.

As shown in Figure 5, it illustrates a logical schematic diagram of implementing two virtual switches in an SN100 node slice switch through configuration. After the message enters from the virtual port, it first undergoes message analysis. The message analysis extracts the byte field value of the message header that needs to be used for flow table matching, and composes the key word KEY for subsequent flow table operations. The flow table is an entry constructed by match-action (MATCH-ACTION). The keyword KEY that is parsed for the aforementioned message is compared with the matching flow pre-installed in the flow table. ACTION is the execution action, including discarding and entering Queue, send to the CPU for processing, rewrite a certain field of the message, increase or decrease the field of the message, etc. Messages are sent to one or more queues for buffering, and the scheduler sends the messages in order according to a certain scheduling algorithm. Finally, there is a measurement module, which is used to perform some statistics on the messages, and is used for functions such as network traffic visualization and fault analysis. All the above-mentioned message parsing, flow table, queue, scheduling, measurement and other functions can be completed through programming.

For different virtual switches, their messages undergo different switch processing logic at the entrance based on the tags in the messages. The tag of this message can be a VLAN or an internal identification tag added by the switch to the message.

The above-mentioned node-slicing switch can use the Barefoot Tofino chip supporting the P4 programmable model, and related switch products.

Among them, P4 (Programming Protocol-independent Packet Processors: packet processors independent of programming protocols) is a language used to express how to pass programmable forwarding elements (such as hardware or software switches, network interface cards, routers or networks) The data plane processes data packets.

From the above introduction of the data layer of the network slicing system of the present application, it can be seen that the network data layer can be composed of two types of switches, namely link slicing switches and node slicing switches, both of which coexist at network sites. A network site can be understood as a computer room or a building, etc. Link fragmentation switches are connected to link fragmentation switches in other network sites through physical links. A link fragmentation switch can be directly connected to one or more node fragmentation switches through physical lines. Each node fragment switch can virtualize multiple virtual switches, that is, virtualize multiple virtual nodes. Therefore, such a network site can realize multiple virtual network nodes, so that when such network sites are interconnected into a network, you can Realize multiple network shards.

Logical paths involve multiple paths that are separated on the same physical link and need to be implemented by link fragmentation switches. Specifically, they can be implemented by commercial switches through technologies such as VPN, QoS, and TE.

The node slicing switch can be implemented by a deeply programmable white box switch, such as a P4 switch.

Next, the control system of the control plane of the network slicing system of the present application is introduced.

The control system is mainly used to configure and manage the network fragments requested for configuration. These configurations include, but are not limited to: virtual nodes, paths, and path parameters in the configuration management network segment. Path parameters include path level, bandwidth, and so on.

In order to achieve both flexible network programming to support network experiments and to achieve commercial network requirements, this application can decompose network control into two parts: link programming control and deep node programming control. The link programming can be realized through the link controller. Node programming can be realized by deep node programming controller.

The control system of the present application can receive a user's configuration request for network segmentation. Specifically, a user can initiate a configuration request for network fragments through the network service platform, and the configuration information can include basic configuration information defining virtual nodes, link configuration information between nodes, link bandwidth parameter configuration information, and so on. The control system can perform configuration management on the network segment based on the configuration request.

Specifically, the control system may include an orchestrator, a link controller, and a deep node programming controller, where:

The orchestrator can be used to determine, based on the network fragment configuration request, the node fragment switch to be configured corresponding to the virtual node and the target link fragment switch connected to the user terminal and the node fragment switch to be configured, as well as from the network fragmentation switch. The virtual link configuration request and the virtual node configuration request are separated from the slice configuration request.

Further, the link controller interacts with the orchestrator, determines all the link fragment switches to be configured on the virtual link to be configured based on the target link fragment switch, and based on the virtual link configuration request The slice switch is configured.

In addition, the deep node programming controller interacts with the orchestrator, and configures the shard switch of the node to be configured based on the virtual node configuration request.

The embodiment of the present application introduces a link controller interacting with an orchestrator, based on the target link fragment switch, determining all the link fragment switches to be configured on the virtual link to be configured, and pairing based on the virtual link configuration request The optional implementation process for the configuration of the link fragmentation switch may include:

S1. The orchestrator sends a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

S2. The link controller calculates the path information of the virtual link to be configured based on the configuration request, and determines all the link fragment switches to be configured on the virtual link to be configured according to the path information;

Specifically, the target link fragment switch is located on the path of the virtual link to be configured, but it may not be all the link fragment switches to be configured on the path. For this reason, the link controller needs to calculate based on the configuration request The path information of the virtual link to be configured is further determined according to the path information to determine all the link fragment switches to be configured on the virtual link to be configured.

S3. The link controller controls the to-be-configured link fragment switch to configure a virtual link.

Further, this embodiment of the present application introduces an optional implementation process for a deep node programming controller to interact with an orchestrator and configure the node shard switch to be configured based on the virtual node configuration request, which may specifically include:

S1. The orchestrator sends the virtual node configuration request to the deep node programming controller;

S2. The deep node programming controller configures the segment switch of the node to be configured.

Wherein, step S2, the process in which the deep node programming controller configures the node fragment switch to be configured may include:

S21: The deep node programming controller obtains the code of the virtual node to be installed based on the virtual node configuration request;

S22. The deep node programming controller queries whether a virtual node has been installed on the shard switch of the node to be configured; if not, execute S23, if yes, execute S24;

S23. Send the code of the virtual node to be installed to the node fragment switch to be configured, and the node fragment switch to be configured completes the configuration according to the code;

S24. The deep node programming controller queries whether the idle resources on the shard switch of the node to be configured meet the resource requirements of the virtual node to be installed; if so, execute S25;

S25. Combine the code of the virtual node to be installed with the code of the virtual node that has been installed on the node fragment switch to be configured, and send the combined code to the node fragment switch to be configured. The node fragment switch completes the configuration according to the merged code.

Next, the control logic of the control plane of the network slicing system of the present application will be introduced in conjunction with FIG. 6.

As shown in Figure 6, the user defines the configuration information of the network segment on the network service platform. The network service platform transmits the network segmentation configuration request to the orchestrator of the control system through the network segmentation service API interface A1.

The orchestrator separates the virtual link configuration request and the virtual node configuration request from the network fragment configuration request, and calls the deep node programming controller through the virtual node management interface N2, deploys the virtual node, and calls the link through the virtual link management node C2 The controller completes the deployment and configuration of related virtual links in the network.

The deep node programming controller calls the node shard switch through the node programming interface N3 to deploy virtual nodes. The link controller invokes the link fragment switch through the path programming interface C3 to deploy and configure the virtual link.

Next, several interfaces illustrated in Figure 6 above are introduced.

Interface A1: API interface of network fragmentation service.

This interface provides the addition, deletion, modification, and checking of user-defined network fragments. There are many options for the expression of the network topology, for example, it can be in XML or JSON format.

For example, taking network fragment 2 (T2—virtual switch A—T4 network fragment) in Figure 2 as an example, the network description form of the json sent by interface A1 to the orchestrator is as follows:

Among them, lines 1-4 describe network fragmentation, and lines 5-12 describe path 2. It should be noted that when line 8 describes the location of the T2 terminal, the IP address or MAC address of T2 can be used. Lines 13-20 describe path 4. Lines 22-36 describe the virtual node, and line 24 indicates the download address of the code corresponding to the virtual node of the user. This code is passed to the system by the user in advance. Lines 25-36 describe the two virtual ports of the user virtual node. Line 28 specifies that vlan is used to identify the network traffic to virtual port 1. Line 29 represents the value of vlan id, which is currently empty. This value can be changed from It is specified by the user or allocated by the system.

Of course, for the description of the virtual topology, other forms of description languages such as XML may also be used, which is not limited in this application.

Interface N2: virtual node management interface

The orchestrator uses the interface N2 to install/uninstall the virtual node on the node slice switch. Following the above example, the information sent by the orchestrator to the deep node programming controller through the interface N2 is as follows:

After the deep node programming controller receives the request to establish a virtual node as shown above, it is easy to know that a vnode needs to be installed on the SN100 node slice switch, the name is "myswitchA" and this virtual node has two virtual ports, which are "Vport1" and "vport2", where the connection relationship between the virtual port and the physical port is indicated by the same connectId.

Interface C2: Virtual link management interface

The C2 interface is the interface between the orchestrator and the link controller, and it uses a standard SDN northbound interface. After the orchestrator extracts the path information (ie, virtual link) in the network segment, it identifies the physical port of the corresponding network device. Using the above example, the orchestrator will issue two requests to establish a connection to the link control:

Connection request 1: Establish a pipe connection between port P11 of switch LN1 and port P5 of switch LN2. The guaranteed bandwidth of the connection is 100M. The pipe corresponds to vlan25 on the P5 port. (This connection corresponds to "link2")

Connection request 2: Establish a connection between port P6 of switch LN2 and port P20 of switch LN3, with a guaranteed bandwidth of 100M. The pipe corresponds to vlan28 on port P6 (this connection corresponds to "link4")

Interface N3: Node programming interface

The N3 interface is related to the device, and different node fragment switches may have different interface forms.

The N3 interface provides the following basic programming capabilities:

Send the control program to the node fragmentation switch. The control program is to program the programmable functional components of the node slice switch through a specific programming language or scripting language. Such as P4 program is a kind of control program.

Configure a connection ID (connectId) for each physical port of the SN100 node fragment switch. A typical connection ID is like vlan.

A connection identifier (connectId) used for connecting a virtual port of a virtual node to a physical port is issued. A typical connection identifier is such as vlan.

Interface C3: Path programming interface

The C3 interface is the control interface of the link controller to a specific link fragment switch, and is also called the southbound interface.

Based on the above two pipeline connection requests, the link controller controls the corresponding devices to complete the establishment of related pipeline connections based on different implementation technologies (VPN technology or TE technology).

In an embodiment of the present application, an optional structure of a deep node programming controller is introduced, as shown in Fig. 7:

The deep node programming controller may include: a node programming management module, a virtual node database, a physical node resource database, a synthesis compiler, and a slice switch controller. Further, it may also include a running state entry mapping module.

Among them, after the deep node programming controller receives the virtual node configuration request sent by the orchestrator, the node programming management module obtains the code of the virtual node, and then queries through the database whether there is a virtual node installed on the shard switch of the node to be configured, If not, directly pass the obtained code to the slice switch controller; if it finds that there are virtual nodes installed on the slice switch of the node to be configured, and the free resources are sufficient, then the synthesis compiler is called to generate a merged Run the code and pass it to the shard switch controller.

The fragment switch controller establishes a connection with the fragment switch of the node to be configured, and issues the running code.

When multiple virtual switches are running on the same node shard switch, the user issues the table entry for a certain virtual switch, and the table address translation needs to be realized through the running state entry mapping module, and then the shard switch controller is issued Shard the switch to the nodes on the data plane.

Based on the network fragmentation system introduced in the foregoing embodiment, an embodiment of the present application provides a network fragmentation control method. Next, from the perspective of the control system, the network fragmentation control method is introduced.

The network fragmentation control method of the present application is applied to a network topology structure, the network topology structure includes a number of network sites, each of the network sites includes at least one link fragmentation switch, and at least one of the link fragmentation switches is connected There is at least one node slicing switch, and the link slicing switches in each of the network sites are connected through a physical link.

As shown in combination with FIG. 8, the network fragmentation control method may include:

Step S100: Receive a configuration request for network segmentation.

Specifically, the user can define the configuration information of the network segment on the network service platform, and then the network service platform can send the configuration information of the network segment to the orchestrator of the control system.

The configuration information may include basic configuration information of virtual nodes, link configuration information between nodes, link bandwidth parameter configuration information, and so on.

Step S200: Configure the link fragmentation switch and the node fragmentation switch in the network site according to the configuration request.

Obviously, the network topology of this case can be built based on the existing commercial network, and only the node fragmentation switch connected to the link fragmentation switch can be added to the network site of the existing commercial network. The present application can receive a configuration request for network fragmentation, and then configure link fragmentation switches and node fragmentation switches in a network site according to the request, so as to realize the required network fragmentation configuration in the network topology. In this case, the link fragmentation switch can realize the configuration of virtual link resources, and the node fragmentation switch can realize the configuration of virtual node resources, so that the configured virtual link and virtual node form a network fragment. Users can configure as many as needed. Different network shards can be used for commercial and experimental purposes at the same time. Compared with the existing technology, the construction and maintenance cost of the experimental network is greatly reduced.

Next, in the above step S200, the process of configuring the link fragmentation switch and the node fragmentation switch in the network site according to the configuration request is introduced, which may specifically include:

S1, based on the configuration request, determine the node fragment switch to be configured corresponding to the virtual node and the target link fragment switch connected to the user terminal and the node fragment switch to be configured.

Specifically, the configuration request may specify the node fragment switch and the user terminal corresponding to each virtual node in the network fragment to be configured. Based on this, the node fragment switch to be configured can be determined. At the same time, it can be further determined that the target link fragment switch connected to the node fragment switch to be configured and the user terminal belong to part of the link fragment switch to be configured on the virtual link to be configured.

S2. Separate the virtual link configuration request and the virtual node configuration request from the configuration request.

Specifically, in order to achieve flexible network programming to support network experiments and achieve commercial network requirements, in this step, the configuration request can be divided according to the virtual link configuration request and the virtual node configuration request, so that the link slicing switch and The node fragment switch is configured.

S3: Determine all the link fragment switches on the virtual link to be configured based on the target link fragment switch.

Specifically, as explained above, the target link fragmentation switch belongs to the to-be-configured link fragmentation switch on the to-be-configured virtual link, but it may not be all. Therefore, all the link fragment switches to be configured on the virtual link to be configured can be determined according to the target link fragment switch.

An optional implementation manner, the S3 step may include:

S31: Send a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

S32. The link controller calculates path information of the virtual link to be configured based on the configuration request, and determines all the link fragment switches on the virtual link to be configured based on the path information.

Specifically, the link controller can calculate the path information of the virtual link passing through the target link fragment switch according to the configuration request, and then can determine all the link fragment switches on the virtual link to be configured according to the path information .

S4. Configure the link fragment switch to be configured according to the virtual link configuration request.

Specifically, the link controller controls the link fragment switch to be configured to configure the virtual link according to the virtual link configuration request.

S5. Configure the segment switch of the node to be configured according to the virtual node configuration request.

Optionally, the process of configuring the node fragment switch may include:

S51. Send the virtual node configuration request to the deep node programming controller.

S52. The in-depth node programming controller configures the node slice switch to be configured.

Among them, the implementation process of S52 may include:

S521: The deep node programming controller obtains the code of the virtual node to be installed based on the virtual node configuration request.

S522. The deep node programming controller queries whether a virtual node has been installed on the shard switch of the node to be configured; if not, execute S523, and if yes, execute S524.

S523. Send the code of the virtual node to be installed to the node fragment switch to be configured, and the node fragment switch to be configured completes the configuration according to the code.

S524. The deep node programming controller queries whether the idle resources on the shard switch of the node to be configured meet the resource requirements of the virtual node to be installed, and if so, execute S525.

S525. Combine the code of the virtual node to be installed with the code of the virtual node that has been installed on the node fragment switch to be configured, and send the combined code to the node fragment switch to be configured. The node fragment switch completes the configuration according to the merged code.

Finally, it should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply these entities. Or there is any such actual relationship or sequence between operations. Moreover, the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes those that are not explicitly listed Other elements of, or also include elements inherent to this process, method, article or equipment. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other identical elements in the process, method, article, or equipment that includes the element.

The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined with each other, and the same and similar parts can be referred to each other.

The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown in this document, but should conform to the widest scope consistent with the principles and novel features disclosed in this document.

Claims

A method for controlling network fragmentation, characterized in that it is applied to a network topology structure including a number of network sites, each of the network sites includes at least one link fragmentation switch, and at least one link The fragment switch is connected to at least one node fragment switch, and the link fragment switches in each of the network sites are connected through a physical link;

The method includes:

Receive the configuration request of the network segment;

According to the configuration request, the link fragmentation switch and the node fragmentation switch in the network site are configured.
The method according to claim 1, wherein the receiving a configuration request of a network segment comprises:

The configuration information of the network segment defined by the user on the network service platform is received, the configuration information includes basic configuration information of virtual nodes, inter-node link configuration information, and link bandwidth parameter configuration information.
The method according to claim 1, wherein the configuring the link fragmentation switch and the node fragmentation switch in the network site according to the configuration request comprises:

Determining, based on the configuration request, a node fragment switch to be configured corresponding to the virtual node and a target link fragment switch connected to the user terminal and the node fragment switch to be configured;

Separating the virtual link configuration request and the virtual node configuration request from the configuration request;

Determine all the link fragment switches on the virtual link to be configured based on the target link fragment switch;

Configure the link fragment switch to be configured according to the virtual link configuration request, and configure the node fragment switch to be configured according to the virtual node configuration request.
The method according to claim 3, wherein the determining all the link fragment switches on the virtual link to be configured based on the target link fragment switch comprises:

Sending a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

The link controller calculates the path information of the virtual link to be configured based on the configuration request, and determines all the link fragment switches on the virtual link to be configured on the virtual link to be configured according to the path information;

The configuring the link fragment switch to be configured according to the virtual link configuration request includes:

The link controller controls the link fragment switch to be configured to configure the virtual link according to the virtual link configuration request.
The method according to claim 3, wherein the configuring the shard switch of the node to be configured according to the virtual node configuration request comprises:

Sending the virtual node configuration request to the deep node programming controller;

The in-depth node programming controller configures the node slice switch to be configured.
The method according to claim 5, wherein the configuration of the slice switch of the node to be configured by the deep node programming controller comprises:

The deep node programming controller obtains the code of the virtual node to be installed based on the virtual node configuration request;

The deep node programming controller queries whether a virtual node has been installed on the slice switch of the node to be configured;

If not, sending the code of the virtual node to be installed to the node fragment switch to be configured, and the node fragment switch to be configured completes the configuration according to the code;

If yes, the deep node programming controller queries whether the idle resources on the shard switch of the node to be configured meet the resource requirements of the virtual node to be installed;

If it is satisfied, the code of the virtual node to be installed is merged with the code of the virtual node that has been installed on the shard switch of the node to be configured, and the merged code is sent to the node shard switch to be configured, and The slicing switch of the node to be configured completes the configuration according to the merged code.
A network slicing system, characterized by comprising: a control system and a number of network sites, each of the network sites includes at least one link slicing switch, and at least one of the link slicing switches is connected to at least one Node fragment switch;

The control system is used to perform configuration management on the network segment requested for configuration;

The link fragment switches in each of the network sites are connected by physical links, and perform link resource configuration according to the virtual link configuration request in the network fragment configuration request under the control of the control system;

The node slicing switch in each network site performs node forwarding processing resource configuration according to the virtual node configuration request in the network slicing configuration request under the control of the control system; the configured virtual link and virtual The network topology formed by the nodes serves as a network fragment.
The system according to claim 7, wherein the control system comprises an orchestrator, a link controller, and a deep node programming controller, wherein:

The orchestrator is configured to determine, based on the network fragment configuration request, the node fragment switch to be configured corresponding to the virtual node and the target link fragment switch connected to the user terminal and the node fragment switch to be configured, and from The virtual link configuration request and the virtual node configuration request are separated from the network fragment configuration request;

The process that the link fragmentation switch performs link resource configuration according to the virtual link configuration request in the network fragmentation configuration request under the control of the control system includes:

The link controller interacts with the orchestrator, determines all the to-be-configured link-slice switches on the virtual link to be configured based on the target link-slice switch, and requests all the links based on the virtual link configuration request. The link fragment switch to be configured is configured;

The process of the node fragmentation switch performing node forwarding processing resource configuration according to the virtual node configuration request in the network fragmentation configuration request under the control of the control system includes:

The deep node programming controller interacts with the orchestrator, and configures the node fragment switch to be configured based on the virtual node configuration request.
The system according to claim 8, wherein the link controller interacts with the orchestrator to determine all the link fragments to be configured on the virtual link to be configured based on the target link fragment switch Switch, and the process of configuring the link fragment switch to be configured based on the virtual link configuration request includes:

The orchestrator sends a configuration request of the virtual link to be configured to the link controller based on the target link fragment switch and the virtual link configuration request;

The link controller calculates the path information of the virtual link to be configured based on the configuration request, and determines all the link fragment switches on the virtual link to be configured on the virtual link to be configured according to the path information;

The link controller controls the link fragment switch to be configured to configure a virtual link.
The system according to claim 8, wherein the deep node programming controller interacts with the orchestrator and configures the node shard switch to be configured based on the virtual node configuration request, comprising:

The orchestrator sends the virtual node configuration request to the deep node programming controller;

The deep node programming controller configures the slice switch of the node to be configured.
The system according to claim 10, wherein the process of configuring the slice switch of the node to be configured by the deep node programming controller comprises:

The deep node programming controller obtains the code of the virtual node to be installed based on the virtual node configuration request;

The deep node programming controller queries whether a virtual node has been installed on the slice switch of the node to be configured;

If not, sending the code of the virtual node to be installed to the node fragment switch to be configured, and the node fragment switch to be configured completes the configuration according to the code;

If yes, the deep node programming controller queries whether the idle resources on the shard switch of the node to be configured meet the resource requirements of the virtual node to be installed;

If it is satisfied, the code of the virtual node to be installed is merged with the code of the virtual node that has been installed on the shard switch of the node to be configured, and the merged code is sent to the node shard switch to be configured, and The slicing switch of the node to be configured completes the configuration according to the merged code.
The system according to claim 7, wherein the node fragmentation switch is a deep programmable white box switch.