CN102098222A - Application service message forwarding method and forwarding node adopting multi-protocol label switching (MPLS) technology - Google Patents

Abstract

The present invention relates to a method and a forwarding node for forwarding an application service message using MPLS technology, the method comprising: the forwarding node receives the message; the forwarding node performs Multi-Protocol Label Switching (MPLS) processing on the received message, wherein, the current When the node is the penultimate hop intermediate node (Transit), the original label in the message is exchanged for the virtual label of the corresponding server group; when the current node is the egress node (Egress PE), the label in the received message is popped up , wherein different server groups correspond to different server group virtual labels; the forwarding node forwards the processed message, wherein, when the current node is an egress node, the message is processed according to the server group information corresponding to the server group virtual label Forward. The invention can realize fast and directed forwarding of content exchange.

Description

Using MPLS technology to forward application service message method and forwarding node

technical field

The present invention relates to MPLS (Multi-Protocol Label Switching, multi-protocol label switching) technology, in particular to a method and a forwarding node for forwarding application service messages by utilizing the MPLS technology.

Background technique

The PTN (Packet Transport Network) bearer network is used to carry multiple services and forward them at high speed. Under the trend of network flattening in the future, the end-to-end path of services may be shortened. With the acceleration of the IPv6 process and the clarity of the prospect, China Telecom announced that it will fully support the V6 network in 2014. The trend of network flattening in the future and increasingly urgent demands such as cloud computing put forward higher requirements on the computing load capacity of network equipment.

With the development of MPLS-TP (Transport Profile for Multi-Protocol Label Switching, transport plane multi-protocol label switching) technology and TMPLS (Transport Multi-Protocol Label Switching, transport plane multi-protocol label switching) and MPLS (Multi-Protocol Label Switching, Multi-protocol label switching) convergence trend, more and more application service technologies are derived in this field, MPLS-TP is independent of the second and third layer protocols, and is the interface of existing routing and switching protocols, while application services belong to At the application layer, MPLS-TP can provide the destination, routing and forwarding and switching capabilities of data forwarding for the application layer technology.

MPLS-TP (Transport Profile for Multi-Protocol Label Switching, Transport Multi-Protocol Label Switching) is a connection-oriented packet switching network technology, which uses MPLS label switching paths, saves MPLS signaling and IP complex functions, and supports multiple services at the same time The bearer is independent of the client layer and the control plane, and can run on various physical layer technologies, with strong transmission capabilities (reflected in QOS (Quality of Service), OAM (Operation Administration and Maintenance) and reliability, etc.). For ITU-T (International Telecommunication Union Telecommunication Standardization Organization) transmission requirements, the IETF MPLS architecture can be extended. The IETF/ITU-T JWT working group is responsible for standard formulation, named MPLS-Transport Profile.

NAT-PT (Network Address Translation-Protocol Translation) technology acts on the devices at the edge of IPV4 and IPV6 networks. All address translation processes are implemented on this device, which is transparent to IPV4 and IPV6 networks That is, the user can realize the communication between the IPV6 network and the IPV4 network without changing the configuration of the host computer in the current IPV4 network.

Content exchange adopts the concept of virtual IP address (Virtual IP, VIP) configured on content, users can connect to this point, and content exchange decides to request to the appropriate server, or handles load balancing. The traditional technology of content exchange and load balancing is shown in Figure 1.

However, the defect of this solution is also obvious. This technology can only be used in a pure IP environment, and the purpose of load balancing can be achieved by replacing the virtual address. IP forwarding has disadvantages such as relying on entire network routing, querying routing tables, and long forwarding prefixes.

Contents of the invention

The invention provides a method and a forwarding node for forwarding an application service message by utilizing the MPLS technology, so as to solve the problem that content exchange cannot be forwarded quickly and directionally.

In order to solve the above technical problems, the present invention provides a method for forwarding application service messages utilizing MPLS technology, the method comprising:

The forwarding node receives the message;

The forwarding node performs multi-protocol label switching (MPLS) processing on the received message, wherein, when the current node is the penultimate hop intermediate node (Transit), the original label in the message is exchanged for the corresponding server group Virtual label; when the current node is an egress PE, pop up the label in the received message, where different server groups correspond to different server group virtual labels;

The forwarding node forwards the processed message, wherein, when the current node is an egress node, the message is forwarded according to the server group information corresponding to the server group virtual label.

Further, before the step of receiving the message, the method also includes assigning a label to the forwarding node and creating a service pseudowire, and different service pseudowires are bound to different service types;

When the forwarding node is an ingress node, after receiving the message and before processing the message, it parses the received message, and judges whether the corresponding service is bound to a service fake according to the service information in the received message. line, and whether the service switch is enabled when the service pseudo-wire is bound; when the service pseudo-wire is not bound or the service pseudo-wire is bound but the service switch is not The service information selects a corresponding service pseudowire and determines a next-hop label.

Further, the egress node is responsible for saving the corresponding relationship between the server group virtual label and the server group information, and the server group information includes the MAC address of the server; when the forwarding node is the egress node, after processing the message and before forwarding the message, it also Perform the following steps:

The egress node queries corresponding server group information according to the server group virtual label;

The egress node balances and schedules the queried server group information according to the load balancing policy, acquires the MAC address of the server, and forwards the information according to the MAC address.

Further, when the forwarding node is an egress node, its processing of the message also includes information replacement, URL replacement or address proxy replacement of the service message.

In order to solve the above technical problems, the present invention also provides another method for forwarding application service messages utilizing MPLS technology, the method comprising:

The ingress node (Ingress PE) receives the message, inserts the label of the corresponding pseudowire, and forwards it;

The intermediate node receives the message and forwards it, wherein the penultimate hop intermediate node (Transit) exchanges the original label in the message with the corresponding server group virtual label, wherein different server groups correspond to different server group virtual labels ;

The egress PE receives the message, pops up the label in the received message, and forwards the message according to the server group information corresponding to the virtual label of the server group.

Furthermore, different pseudowires are bound to different services.

Further, the egress node is responsible for saving the corresponding relationship between the virtual label of the server group and the server group information, and the server group information includes the MAC address of the server; before the egress node forwards the message, the following steps are also performed:

The egress node queries corresponding server group information according to the server group virtual label;

The egress node balances and schedules the queried server group information according to the load balancing policy, acquires the MAC address of the server, and forwards the information according to the MAC address.

In order to solve the above technical problems, the present invention also provides a forwarding node utilizing MPLS technology, the forwarding node comprising:

A message receiving module, configured to receive messages;

A message processing module, configured to perform multi-protocol label switching (MPLS) processing, wherein, when the forwarding node is the penultimate hop intermediate node (Transit), the original label in the message is exchanged for the corresponding server group virtual Label; when the forwarding node is an outgoing node (Egress PE), the label in the received message pops up, wherein different server groups correspond to different server group virtual labels;

A message forwarding module, configured to forward the processed message, wherein, when the forwarding node is an egress node, forward the message according to the server group information corresponding to the server group virtual label.

Further, the forwarding node also includes a service pseudowire creation module, a message parsing and identification module, and a normal processing and forwarding module; wherein:

The service pseudowire creation module is used to assign labels and create service pseudowires, and different service pseudowires are bound to different service types;

When the forwarding node is an ingress node, the message parsing and identification module is configured to parse the received message, judge whether the corresponding service is bound to a service pseudowire according to the service information in the received message, and bind Whether the service switch is enabled when the service pseudo-wire is enabled; and when the service pseudo-wire is bound and the service switch is enabled, select the corresponding service pseudo-wire according to the service information, and determine the next-hop label;

The normal processing and forwarding module is used to perform common MPLS tunnel processing and forwarding on received messages when no service pseudowire is bound or when a service pseudowire is bound but the service switch is not enabled.

Further, when the forwarding node is an egress node, it also includes:

An information table cache module, configured to store the correspondence between the virtual label of the server group and the information of the server group, where the server group information includes the MAC address of the server;

An information table query module, configured to query corresponding server group information according to the server group virtual label in the received message;

A load balancing module, configured to balance and dispatch the queried server group information according to a load balancing policy, and obtain the MAC addresses of the servers.

The present invention utilizes the existing MPLS technology to switch and forward, and the outgoing node (Egress PE) allocates different server group virtual labels for different application service service server groups. During the MPLS switching and forwarding process, the penultimate hop forwarding node will correspond to The virtual label of the server group is used as the label of the last hop forwarding node (that is, the egress node) and exchanged and forwarded. After that, the egress node directly obtains the MAC address of the corresponding server and forwards it, which can improve the switching capability of the application layer service , especially in directional forwarding, it can take advantage of the label forwarding of MPLS-TP technology, and realize fast directional forwarding of application services through the pseudowire technology of MPLS-TP.

Description of drawings

FIG. 1 is a schematic diagram of traditional content exchange and load balancing;

Fig. 2 is a schematic diagram of a method for forwarding application service messages using MPLS technology in the present invention;

Fig. 3 is another schematic diagram of the method for forwarding application service messages using MPLS technology in the present invention;

Fig. 4 is a schematic diagram of an embodiment of the method of the present invention.

Detailed ways

The present invention utilizes existing MPLS technology (comprising MPLS-TP, TMPLS, MPLS etc.) Among them, the penultimate hop forwarding node takes the virtual label of the corresponding server group as the label of the last hop forwarding node (that is, the egress node), and performs exchange and forwarding, and then the egress node directly obtains the MAC address of the corresponding server according to the preset information table Address and forwarding can improve the exchange capability of application layer services, especially in the aspect of directional forwarding, it can take advantage of MPLS-TP technology label forwarding, and realize fast directional forwarding of application services through MPLS-TP pseudowire technology.

For each forwarding node, its method of using MPLS technology to forward the application service message includes:

Step 201: the forwarding node receives the message;

When not specified, the forwarding node mentioned in the present invention can be an ingress node (Ingress PE), an intermediate node (Transit) or an egress node (Egress PE).

Layer 2 VPN (Virtual Private Network, Virtual Private Network) PSN (Packet Switching Network) tunnels between PE devices can be multiplexed into multiple internal tunnels, which are also called pseudowires (Pseudo Wire, PW), Hereinafter, the PW side refers to the VPN network side when MPLS-TP technology is used for VPN networking.

In order to implement packet forwarding, it is necessary to assign labels to each forwarding node and create a service pseudowire. And different service pseudowires are bound to different service types.

A section of MPLS label (Label) space is reserved for service load balancing, called service load balancing label space, and the size of the label space is determined according to the sum of the number of concurrent sessions (sessions) allowed to be accessed by various services. The correspondingly generated VCID space is called a service VCID space.

Preferably, a 1:1 binding scenario between a pseudowire and a service is adopted, that is, only one kind of service is bound to a pseudowire. For different services, such as NAT-PT (IPV6), HTTP, HTTPS, EMAL, FTP and other application services, create an independent pseudowire template. The pseudowire template carries the service type attribute, that is, the above service type.

The difference between a PW template customized for a specific service and an ordinary PW is that the label assigned to an ordinary PW is a private network label generated according to a common VCID (Virtual Connection Identity) distributed through LDP or statically allocated. The label used by the pseudowire is generated according to the service VCID and allocated from the service label control stack. It supports two types of LSP (label switching path), CO-ROUTED LSP and ASSOCIATED LSP, that is, whether to bind labels in the front and back directions . The label allocation of service pseudo-wires also supports dynamic and static methods.

Server group virtual Label and service VCID are dedicated label spaces reserved for different services, and their management involves allocation and release rules. It is mainly allocated and released on demand after the reserved space.

The last hop service virtual label allocated for the service pseudowire, that is, the virtual label of the egress node, is allocated from the service load balancing label space. The pseudowire layer does not perform the penultimate hop pop-up operation during the penultimate hop, but It is exchanged for a server group virtual label, and each server group virtual label corresponds to a specific service server group. The label assigned to a specific link (Session) is obtained from the service load balancing label space, because it has the meaning of service load balancing service and does not participate in ordinary MPLS-TP forwarding and exchange, so it is also called a server group service virtual label. This type of label is generated by the configured service VCID.

The message of an AC received by the Ingress PE will enter the same tunnel, and the inner layer will encapsulate different service pseudowire labels. When the egress PE receives a message carrying a different server group virtual label, it will map the different server group virtual labels to different server group information. AC is the name of the access side of the VPN (Virtual Private Network, virtual private network) on the networking.

Step 202: The forwarding node performs multi-protocol label switching (MPLS) processing on the received message, wherein, when the current node is the intermediate node (Transit) of the penultimate hop, the original label in the message is exchanged as The virtual label of the server group of the corresponding service; when the current node is an egress PE, pop up the label in the received message;

Before MPLS-TP forwards the label exchange, the ingress node analyzes and identifies the message, because in the packet-by-packet forwarding system, analyzing each message one by one will reduce the efficiency, so the analysis and identification function can be controlled by switches. This function mainly performs four-layer identification and protocol type analysis on the type of the forwarded message. Implement corresponding countermeasures for the internal information of the message header, and select different server groups according to the content.

Under the L2VPN instance, multiple customized service switches can be enabled at the same time, and pseudowires of multiple services can be bound to realize pseudowire layer distribution of different types of service packets. When the L2PVN instance is not bound with a service pseudowire or the custom service switch is not enabled, it is forwarded through the common MPLS-TP tunnel.

When processing the MPLS-TP message from the AC (access circuit, Attach Circuit, which is used to connect the user's edge device), the forwarding node (here specifically refers to the ingress node) will analyze the user's MPLS-TP message), such as IPV6 When the message enters the pseudowire channel of IPV6, it can also analyze the 4-7 layer services of the OSI (Open System Interconnect, Open System Interconnection) reference model, and enter different service pseudowire channels according to the requests sent to different URLs , forward the message according to the allocated label channel. Currently supported services are: HTTP, HTTPS, FTP, TELNET, IPV6 (NAT-PT). Ordinary packets that do not belong to the above service types are still forwarded by ordinary MPLS-TP according to the original ordinary pseudowire.

In particular, on the intermediate node of the second-to-last hop, the pop-up operation of the second-to-last hop is not performed, but a server group dummy label is added.

The egress node presets the server group label information table, which stores the corresponding relationship between the server group virtual label and server group information (including the server's MAC address) (the server group virtual label can point to the server group information table in the form of memory index). It is stored in the form of HASH table or binary tree to improve the reading speed. It is written in the way of static allocation and used in the global networking environment.

Step 203: The forwarding node forwards the processed message, wherein, when the current node is an egress node, the message is forwarded according to the server group information corresponding to the server group virtual label.

The egress node presets the server group label information table, which stores the corresponding relationship between the server group virtual label and server group information (including the server's MAC address) (the server group virtual label can point to the server group information table in the form of memory index). It is stored in the form of HASH table or binary tree to improve the reading speed. It is written in the way of static allocation and used in the global networking environment.

After the egress node receives the message, it will extract the business information in the message, query the server group label information table, and obtain the server group information corresponding to the virtual Label. According to the server load balancing strategy (RoundRobin (poll), Weight Round Robin (weighted Polling) or random selection strategy), balance the scheduling of servers in the server group, obtain the MAC information of the corresponding server, and the egress node obtains the business information of the message again or distributes the message to the designated server to achieve the purpose of load balancing. After obtaining the V6 address and corresponding application service information from the server, the processed information is sent back to the PTN device on the AC side.

It is understandable that the forwarding may continue after the outgoing node switches. That is to say, the last hop service equipment may be placed anywhere on the bearer network.

In addition, if the forwarding node is the last hop service device for content exchange, it also has a fast message processing mechanism, that is, quickly replaces the three-layer and above information in the message header, including information replacement of service messages, URL replacement, Address proxy replacement, etc., such as NAT-PT conversion of IPV4 and IPV6 addresses, content exchange of HTTP, HTTPS, FTP, EMAIL and other information.

The IPV4 address in the message is replaced by the fast message processing mechanism, and the IPV6 header address is encapsulated. For the content layer above 3 layers such as HTTP service, the effect achieved is that the request sent to www.aaa.com can be sent to the specified Instead of sending all HTTP traffic to one set of servers, requests to www.zzz.com are sent to a different set of servers. This provides more explicit load balancing than a TCP/UDP (Transmission Control Protocol, Transmission Control Protocol, User Data Protocol, User Datagram Protocol) port and allows virtual hosting. The network exchange can also send the specified Uniform Resource Locator (URL), file type, Uniform Resource Identifier (URI), Cookies, etc. to a specific set of servers. To www.aaa.com? Requests for /products may be sent to product servers, while requests for www.aaa.com/technical may be sent to those designated servers that handle technical content. Increased server throughput speeds up response time and improves user experience quality.

The content-processed message faces two choices on the egress node (such as the foreground PTN (Packet Transport Network) device of the content server): 1. It will be aggregated locally to the AC side. 2. Pop up the service pseudowire label and continue MPLS-TP forwarding after summarizing.

Among them, the method of local aggregation to the AC side: the last hop pops up the virtual label and forwards to the AC; the method of continuing forwarding to the PW side: pops up the server group virtual label, adds the actual forwarding label, and continues MPLS-TP PW forwarding , that is, to forward to another PW. The specific method: bind and forward the related PW or related pseudowire under the instance of the VPN device, and all the packets of the service pseudowire of a certain VPN instance are aggregated under the PW or related pseudowire. .

The advantages are: it supports load balancing in the MPLS-TP protocol environment, uses the fast encapsulation and forwarding of the pseudowire layer, and does not change the forwarding mode of the pseudowire layer, and the PHP (PHP (Penultimate Hop Popping) forwarded by MPLS-TP) , refers to the pop-up function of the penultimate hop label of MPLS, and does not support PHP by default in MPLS-TP, and also needs the penultimate hop not to pop up to support the exchange and identification of server group service virtual labels in the present invention, so MPLS -TP technology has a good supportability for the solution of the present invention.) The default is closed, the label is quickly forwarded, and it supports a variety of dynamic and static label distribution methods, etc., and realizes the combination of business processes and directional forwarding of business messages, and finally One hop realizes the combination of label forwarding and the original load balancing technology. In addition, the PW forwarding method can be continued after the service pseudo-wire label is popped up, so that content exchange can be performed on the PW trunk, and there is no need to terminate the service on the SR or SR+ router. Some networking deployments provide great flexibility.

The above is a description of the method for forwarding application service messages using MPLS technology in the present invention from the perspective of a single forwarding node. Of course, it can also be described from the direction along the forwarding path, as shown in Figure 3, including the following steps:

The ingress node (Ingress PE) receives the message, inserts the label of the corresponding pseudowire, and forwards it;

The intermediate node receives the message and forwards it, wherein the penultimate hop intermediate node (Transit) exchanges the original label in the message with the corresponding server group virtual label, wherein different server groups correspond to different server group virtual labels ;

The egress PE receives the message, pops up the label in the received message, and forwards the message according to the server group information corresponding to the virtual label of the server group.

The specific processing methods of each forwarding node are the same as those described above, and will not be repeated here.

As shown in Figure 4, it is the specific implementation steps of the present invention. It mainly realizes the bearer of the pseudowire to the service in a 1:1 manner, and adopts LDP or static label allocation mode to assign the virtual label of the server group to the pseudowire. From the same AC The incoming packets are based on the same pseudo-wire to realize service-based pseudo-wire traffic separation, corresponding to the corresponding service or content server group to select the target server, and to implement MPLS-TP technology to support content switching and load sharing, so that the bearer network can Better implementation of content switching and load sharing technologies in the field, reducing response time, improving user friendliness and network operability:

Step 401: first create a service pseudowire, assign labels according to the tunnel label allocation type (static, LDP, RSVP), and the last hop is the service virtual label of the server group;

First create a service pseudowire template. Under the PW template, you can specify the service type carried by the pseudowire. Labels are allocated according to the label allocation method (static, LDP (label distribution protocol)), and the VCID range of the pseudowire used to carry the service will be controlled. Within the scope of the service VCID, the PW label generated according to this VC is a virtual service label of the server group. The virtual label needs to be explicitly specified when configuring the tunnel statically or dynamically allocated by the label management module from the service virtual label space, and pushed to the penultimate hop through LDP for label exchange.

Step 402: Enable the switch of the relevant customized service in the L2VPN service instance, and bind the service pseudowire under the L2VPN service instance;

Take the packet analysis and identification service as a customized service, enable the switch of the related customized service in the L2VPN service instance, and bind the service pseudowire under the L2VPN (Layer 2 Virtual Private Network L2VPN) service instance and the outer tunnel, bind After the service is set successfully, the message belonging to this instance will be processed as a message that requires service analysis and directional forwarding. After the information corresponding to the service is identified by the message fast analysis module, the message will enter the MPLS-TP forwarding process.

Step 403: The ingress node parses the service message, and if it is identified as a corresponding enabled service pseudowire, enters the MPLS-TP content exchange and forwarding process, otherwise executes the ordinary message forwarding process;

The ingress node needs to judge whether the corresponding service is bound to the service pseudo-wire according to the service information in the received message, and whether the service switch is enabled when the service pseudo-wire is bound; When the switch is not enabled, the received message is processed and forwarded by the ordinary MPLS tunnel, that is, the ordinary pseudowire in the L2VPN instance is forwarded; otherwise, the corresponding service pseudowire is selected according to the service information, and the next-hop label is determined.

Packets enter the MPLS-TP content switching and forwarding process, including inserting labels at the ingress node and exchanging labels at intermediate nodes.

Step 404: The penultimate hop intermediate node does not pop up the label, and performs virtual label exchange;

Step 405: The last hop (i.e. Egress PE) judges whether the label of the last hop is in the service virtual label space. If the label of the last hop is not in the virtual label space, the server group label information table is not queried, and ordinary message forwarding is performed. process;

Step 406: The outgoing node (Egress PE) queries the server group label information table. If the information corresponding to the label cannot be found, it will also go through the normal forwarding process. If the query obtains the server group information corresponding to the virtual Label, then perform step 407;

When the egress PE receives virtual labels of different server groups, it maps different labels to different server groups.

Step 407: According to the server load balancing strategy, the MAC address information of the server is obtained through balanced scheduling, and the information to be processed related to the message is sent to the target server according to the MAC address;

Step 408: According to whether the content-exchanged message after content processing is locally terminated, the service pseudo-label is popped and then summarized locally to the AC side, or the service pseudo-wire label is popped and summarized, and then forwarded to the PW side of MPLS-TP.

In the above embodiments, the message analysis and identification service is regarded as a customized service, the switch of the related customized service is enabled in the L2VPN service instance, and the service pseudowire is bound under the L2VPN service instance and the outer tunnel, after the binding is successful After the service takes effect, the messages belonging to this instance will be processed as messages that require service analysis and directional forwarding. After the information corresponding to the service is identified by the message fast analysis module, the message will enter the MPLS-TP forwarding process, and the service directional forwarding reflects In the penultimate hop, the label is not popped up, but the last hop label, that is, the virtual label of the corresponding service server group allocated from the service load balancing label space.

For realizing above method, the present invention also provides a kind of forwarding node utilizing MPLS technology, and this forwarding node comprises:

A message receiving module, configured to receive messages;

A message processing module, configured to perform multi-protocol label switching (MPLS) processing, wherein, when the forwarding node is the penultimate hop intermediate node (Transit), the original label in the message is exchanged for the server group of the corresponding service Virtual label; when the forwarding node is an outgoing node (Egress PE), the label in the received message pops up, wherein different server groups correspond to different server group virtual labels;

A message forwarding module, configured to forward the processed message, wherein, when the forwarding node is an egress node, forward the message according to the server group information corresponding to the server group virtual label.

Further, the forwarding node further includes a service pseudowire creation module, configured to assign labels, create a service pseudowire, and bind different service types to different service pseudowires.

The forwarding node also includes a message parsing and identification module and a common processing and forwarding module; wherein, when the forwarding node is an ingress node, the message parsing and identification module is used to parse the received message, and according to the received message The service information in the text determines whether the corresponding service is bound to a service pseudowire, and whether the service switch is enabled when the service pseudowire is bound; and when the service pseudowire is bound and the service switch is enabled, select the corresponding service according to the service information Service Pseudowire, to determine the next hop label;

The normal processing and forwarding module is used to perform common MPLS tunnel processing and forwarding on received messages when no service pseudowire is bound or when a service pseudowire is bound but the service switch is not enabled.

When the forwarding node is an egress node, it also includes:

An information table cache module, configured to store the correspondence between the virtual label of the server group and the information of the server group, where the server group information includes the MAC address of the server;

An information table query module, configured to query corresponding server group information according to the server group virtual label in the received message;

A load balancing module, configured to balance and dispatch the queried server group information according to a load balancing policy, and obtain the MAC addresses of the servers.

The present invention proposes a method for forwarding application service messages by using MPLS technology. By combining content application services into the PTN bearer network, services are introduced into the PTN bearer network, thereby constructing an operable and manageable PTN bearer network. Improve the competitiveness of bearer network products in an all-round way, contribute to the competition in the next network era, and deepen the interaction with the needs of high-end customers without changing the existing telecommunication equipment pattern.

The present invention can support the load sharing of the server group through the reasonable transformation of the pseudo-wire technology, uses the technology of combining label distribution and pseudo-wire technology, does not use VIP address replacement schemes, improves the forwarding efficiency, and makes the process of content forwarding and load sharing organic combined.

The present invention realizes the bearer of the pseudowire to the service in a 1:1 manner, realizes fast and directional forwarding of the application service, adopts the LDP or static label allocation mode to assign virtual labels to the server group for the pseudowire, and realizes separation of traffic based on the pseudowire. Corresponding to NAT-PT, HTTP, HTTPS, EMAIL, FTP and other content server groups, select servers to provide content exchange services.

This patent uses load balancing technology to provide NAT-PT services. On the layer 2 switch, it will communicate with the NAT-PT server group through a dedicated encrypted channel, provide address translation services and customizable layer 4 and above content exchange, and customize HTTP , HTTPS, EMAIL, FTP and other application services for content exchange, and use MPLS-TP technology to enhance application layer forwarding capabilities.

The suitable networking environment of the present invention is the typical or atypical networking of all PTNs using MPLS-TP technology, and the server group (server group) can be deployed after the RNC of the convergence ring, or after the convergence node on the access ring , the server group can also be deployed after the access node of the access ring

The present invention selects different server groups according to the internal information of the message header to implement the load sharing strategy. In the penultimate hop, it is not popped up, but replaced with a virtual label of the server group, and the virtual label is popped up in the last hop to query the corresponding server group information, and then according to the server group load sharing strategy: Round Robin, random selection, Weight Round Robin Select the target server. The egress node directs the message or information to the dedicated connection port of the server group, performs load balancing countermeasures according to the internal information of the message header, and optimizes the WEB service for special functions such as image service, SSL dialogue and database transaction service.

Claims (10)

1. forwarding application service message method of utilizing the MPLS technology is characterized in that this method comprises:

Forward node receives message;

Described forward node carries out multiprotocol label switching (mpls) to the message that receives to be handled, and wherein, when present node is the intermediate node (Transit) of the last two jump, the former label in the described message is exchanged for the empty label of respective server group; When present node is egress (Egress PE), eject the label in the message that receives, wherein, the empty label of the server group that different server groups is corresponding different;

Described forward node is transmitted the message after handling, and wherein, when present node is egress, carries out message according to the empty label corresponding server of described server group group information and transmits.

2. the method for claim 1 is characterized in that:

Before receiving the step of message, this method also comprises, be described forward node distributing labels, and create professional pseudo-line, and the different different types of service of professional pseudo-line binding;

When described forward node is ingress, after it receives described message, handle described message before, the message that receives is resolved, judge according to the business information in the message that receives whether corresponding service binds professional pseudo-line, and whether the business switch enables when binding professional pseudo-line; Do not bind professional pseudo-line or bind professional pseudo-line but professional switch when not enabling, the message that receives is carried out common MPLS tunnel handle and transmit; Otherwise select the pseudo-line of corresponding business according to described business information, determine next jumping label.

3. the method for claim 1 is characterized in that: egress is responsible for preserving the corresponding relation of empty label of server group and server group information, and described server group information comprises the MAC Address of server; When described forward node is egress, behind the processing message, before E-Packeting, also carry out following steps:

Described egress is according to the empty searching label corresponding server of described server group group information;

Described egress is according to the load balancing strategy, and the server group information that balance dispatching inquires is obtained the MAC Address of described server, and transmits according to described MAC Address.

4. the method for claim 1 is characterized in that: when described forward node was egress, its processing to message comprised that also the information of service message is replaced, URL replaces or the address agency replaces.

5. forwarding application service message method of utilizing the MPLS technology is characterized in that this method comprises:

Ingress (Ingress PE) receives message, inserts the label of corresponding pseudo-line, and transmits;

Intermediate node receives message, and transmits, and wherein the intermediate node of the last two jump (Transit) is exchanged for the empty label of respective server group with the former label in the described message, wherein, and the empty label of the server group that different server groups is corresponding different;

Egress (Egress PE) receives message, and ejects the label in the message that receives, and carries out message according to the empty label corresponding server of described server group group information and transmits.

6. method as claimed in claim 5 is characterized in that: the different different business of pseudo-line binding.

7. method as claimed in claim 5 is characterized in that: described egress is responsible for preserving the corresponding relation of empty label of server group and server group information, and described server group information comprises the MAC Address of server; Before described egress E-Packets, also carry out following steps:

Described egress is according to the empty searching label corresponding server of described server group group information;

Described egress is according to the load balancing strategy, and the server group information that balance dispatching inquires is obtained the MAC Address of described server, and transmits according to described MAC Address.

8. a forward node that utilizes the MPLS technology is characterized in that, this forward node comprises:

The message receiver module is used to receive message;

Message processing module (MPM) is used to carry out multiprotocol label switching (mpls) and handles, and wherein, when described forward node is the intermediate node (Transit) of the last two jump, the former label in the message is exchanged for the empty label of corresponding server group; When described forward node is egress (Egress PE), eject the label in the message that receives, wherein, the empty label of the server group that different server groups is corresponding different;

Packet forwarding module is used to transmit the message after the processing, wherein, when described forward node is egress, carries out message according to the empty label corresponding server of described server group group information and transmits.

9. forward node as claimed in claim 8 is characterized in that, described forward node also comprises professional pseudo-line creation module, packet parsing identification module and common process forwarding module; Wherein:

The pseudo-line creation module of described business is used for distributing labels, creates professional pseudo-line, and the different different types of service of professional pseudo-line binding;

When described forward node is ingress, described packet parsing identification module, be used for the message that receives is resolved, judge according to the business information in the message that receives whether corresponding service binds professional pseudo-line, and whether the business switch enables when binding professional pseudo-line; And when binding professional pseudo-line and professional switch enable, select the pseudo-line of corresponding business according to described business information, determine that next jumps label;

Described common process forwarding module is used for the message that receives being carried out common MPLS tunnel handling and transmit not binding professional pseudo-line or binding professional pseudo-line but professional switch when not enabling.

10. forward node as claimed in claim 8 is characterized in that, when described forward node is egress, also comprises:

The information table cache module is used to preserve the corresponding relation of empty label of server group and server group information, and described server group information comprises the MAC Address of server;

The information table enquiry module is used for the empty searching label corresponding server of the server group group information according to the message that receives;

Load balancing module is used for according to the load balancing strategy, and the server group information that balance dispatching inquires is obtained the MAC Address of described server.

Images