WO2007131441A1

WO2007131441A1 - A method and a means for load balancing based on sip

Info

Publication number: WO2007131441A1
Application number: PCT/CN2007/001540
Authority: WO
Inventors: Ming Xu
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2006-05-16
Filing date: 2007-05-11
Publication date: 2007-11-22
Also published as: CN1984142A; CN100558109C

Abstract

A method and a means for load balancing on the basis of SIP is provided. The method includes the following steps: the device at side of the network receives the SIP request message and sets the dialog identifier for the message and then sends the SIP request message and the dialog identifier to the scheduler; the scheduler determines the application server to process the SIP request message according to the predefined relation between the dialog identifier and the application server. The means includes: storage unit for storing the relation between the dialog identifier and the application server; and scheduler for distributing the application server. The invention realizes sending the different SIP request to the same node to process by setting the request of dialog and the dialog identifier.

Description

Load balancing implementation method and device based on initial protocol of the tongue This application claims to be submitted to the Chinese Patent Office on May 16, 2006, the application number is 200610080814.6, and the invention name is "based on the initial protocol based load balancing implementation method and system" Priority of the patent application, the entire contents of which is incorporated herein by reference. Technical field

The present invention relates to the field of information technology, and in particular, to a load balancing implementation method and apparatus, and in particular, to a load balancing implementation method and apparatus based on a session initial protocol. Background technique

With the promotion of applications such as VoIP and IP Multimedia Subsystem (SIP), Session Initiation Protocol (SIP) has also been widely used. SIP is the basis for establishing a session, and the session is from the user. From the beginning to the end of the business, the complete interaction process between the user and the server, all application server sessions need to use the SIP protocol.

To reduce costs, clustering technology is a priority technology. Broadly speaking, a group of systems that are independent of each other is a cluster. When a client interacts with a cluster, the cluster behaves like a separate system. In a narrow sense, a cluster is a number of independent computers that provide services as a whole. For the customer, these computers behave like a computer, but their processing power, reliability, and input/output (I/O) capabilities have been greatly improved. Cluster technology is a general-purpose technology. Its purpose is to solve the problem of single-computer computing power and I/O capability, improve service reliability, achieve scale scalability, and strive for lower overall operation and maintenance costs (operation and upgrade). , maintenance costs). At the core of the clustering technology is load balancing. The so-called load balancing refers to the equal distribution of external requests to multiple computers with the same functions.

The application is a specific service, or it can be a set of software. Considering different applications, in actual load balancing, it is also necessary to consider the maintenance of the session, that is, all messages in a session need to be sent in the same node in the cluster. .

The existing SIP-based load balancer implementation method is based on CALLID or SourcelP in the SIP protocol, that is, the same CALLID or the same SourcelP's SIP request can be processed to the same node.

However, in practical applications, this solution cannot solve the need for different CALLID or SourcelP SIP requests to be sent to the same node for processing. For example: The status of the subscription conference, because the subscriber's CalllD will not be the same as the conference creator's CalllD, and the IP is different. The existing load balancing implementation method cannot satisfy this application scenario. Summary of the invention

In the embodiment of the present invention, a load balancing implementation method and apparatus based on a session initial protocol are provided. By configuring a request and a tongue mark, different SIP requests can be sent to the same node for processing.

The embodiment of the invention provides a load balancing implementation method based on a session initial protocol, which includes:

Receiving, by the network side, a session initiation protocol SIP request message sent by the terminal, setting a session identifier for the session initiation protocol SIP request message, and sending the session initiation protocol SIP request message and its session identifier to the SIP request for the session initiation protocol a scheduling unit for the message distribution application server;

The scheduling unit reads the session flag, and determines an application server that processes the session initiation protocol SIP request message according to a correspondence between a predetermined session identifier and an application server.

The embodiment of the present invention further provides a load balancing device based on a session initial protocol, including: a storage unit, configured to save a correspondence between a session identifier and an application server;

a scheduling unit, configured to allocate an application server for a session initiation protocol SIP request message; after receiving the SIP request message, the scheduling unit queries the storage unit according to the session identifier carried in the received SIP request message to obtain the session identifier and the application server Correspondence relationship to determine an application server that processes the SIP request message.

In the technical solution provided by the embodiment of the present invention, different SIP requests having the same session flag are scheduled to be processed by the same application server (node) by setting the request message type of the session and the session flag, thereby implementing different CALLID or SourcelP. SIP requests are sent to the same node for processing purposes. DRAWINGS

FIG. 1 is a flowchart of a load balancing implementation method based on an initial protocol according to an embodiment of the present invention; Figure

2 is a structural diagram of a network topology in an embodiment of the present invention;

3 is a flowchart of a load balancing implementation method based on an initial protocol according to another embodiment of the present invention;

Figure 4 is a flow chart in the embodiment based on Figure 3;

FIG. 5 is a schematic structural diagram of a load balancing apparatus based on a session initial protocol according to an embodiment of the present invention; FIG.

FIG. 6 is a schematic structural diagram of a load balancing apparatus based on a live initial protocol according to another embodiment of the present invention. detailed description

In order to make the principles, the features and the advantages of the present invention more clear, the technical solutions of the present invention are further described in detail below with reference to the accompanying drawings and embodiments.

The device used to allocate the application server for the session initial request message may be a load balancer, a server, or the like.

FIG. 1 is a flowchart of a method for implementing a load balancing based on a session initial protocol according to an embodiment of the present invention. The present invention uses a Push to Talk over Cellular (PoC) service as a preferred embodiment, and is assumed to be in this embodiment. The network topology involved is shown in Figure 2. There are four servers, PoCl, PoC2, PoC3, and PoC4, which are used to process PoC services, load balancer Α2 (scheduling device or unit), and two participating in the PoC service. User terminals UE1 and UE2.

When the load balancer initializes, it creates two independent storage containers in the memory, one of which saves all the nodes in the cluster. For the convenience of description, it is called the container one; the other saves the correspondence between the session and the node, which is called For the container two.

Step 101 specifically includes the following two steps:

Step 1011: The terminal UE1 sends a session initial request message SipRequest() to the network side to request a PoC service session.

Step 1012: The network side first session initial message SipRequest() is set to an Invite class request, and the session flag of the PoC service session requested by the UE1 is set as the contact header field of the response, that is, the contact header in the 200 response of the Invite request. Domain dynamic configuration is PoCSessionl.

Step 102: The network side sends a PoC service initial request message whose session flag is PoCSession1 to the load balancer SIP Balance _0.

Step 103: The load balancer SIP Balance reads the session identifier PoCSessionl of the session initial request message SipRequest(), and obtains the application server that processes the PoCSession1 according to the correspondence between the session identifier recorded in the container 2 and the application server, and assumes that the container II records The application server that handles PoCSessionl is PoC3.

If the PoCSession1 is not necessarily recorded in the container 2, that is, the PoCSession1 does not necessarily exist, the load balancer SEP Balance first determines whether the session flag PoCSession1 exists, and if so, performs step 104; otherwise, the load balancer SIP Balance finds the idle from the container one. The server PoC2, and the PoC service session initial request message with the session flag PoCSession1 is sent to the server PoC2, and the PoC service session PoCSessionl is processed by the server PoC2, and the relationship between the PoC2 and the PoCSessionl server PoCessionl is stored in the container 2 In step 105, go to step 105.

Step 104, the load balancer SEP Balance sends the session initial request message

The SipRequestO is redirected to the application server PoC3 that processes the session initial request message, and is processed by the server PoC3.

Step 105, end.

The application server that is configured to process the PoC service session request sent by the terminal UE1 is the PoC2. At this time, the other terminal UE2 also sends an initial request to the network side, and when it is required to join the PoC service session, as shown in FIG. 3, the specific execution is performed. The following steps:

Step 201 specifically includes the following two steps:

Step 2011: The terminal UE2 sends a session initial request message SipRequest() to the network side to request to join the PoC service session.

Step 2012: The network side sets the session initial request message SipRequest() to the Invite class SIP message whose session flag is to the header field, and the to header field is dynamically configured as PoCSessionl.

Step 202: The network side sends the SIP message to the load balancer SIP Balance;

Step 203: The load balancer obtains the application server PoC2 that processes the PoC service session PoCSession1 according to the corresponding relationship between the recorded tongue mark and the application server; Step 204: The load balancer SIP Balance redirects the SIP message to the application server PoC2, and the application server PoC2 processes the SIP request sent by the terminal UE2.

Step 205, the end.

In this way, SIP requests from different terminals are sent to the unified node (ie, application server) for processing. Since the CALLID and SourcelP of different terminals are different, the present invention solves the request for the SIP request of different CALLID or SourcelP to be sent to the unified node for processing.

In the above embodiment, the load balancer pre-configures a SIP message that all services can create a session. If the load balancer only pre-configures a part of the service, the SIP message of the session can be created. At this time, as shown in FIG. 4, the steps in FIG. Between step 202 and step 203, the method further includes the following steps: Step 301: The load balancer SIP Balance determines whether the Invite class SIP message exists, and if yes, step 302 is performed; otherwise, step 304 is performed;

Step 302: The load balancer SIP Balance determines whether the session identifier PoCSession1 in the Invite SIP message exists. If yes, the corresponding application server that processes the PoC service session PoCSession1 is obtained according to the corresponding relationship between the recorded session identifier and the application server; otherwise, the step is performed. 303;

Step 303: The load balancer SIP Balance finds an idle application server, such as PoC4, from the container one, and records the relationship between the PoC4 and the PoCSessionl. In order to enable the load balancer to know whether the PoC4 receives the PoCSession1, the PoC4 may first load balance the load. The SIP Balance returns a session initial request response message, and then ends the load balancing process. Step 304: The load balancer SIP Balance finds an idle application server from the container one, for example, PoC4, and hands the PoCSession1 service session to the application server PoC4. End the load balancing process.

FIG. 5 is a schematic structural diagram of a load balancing apparatus based on a session initiation protocol according to an embodiment of the present invention. The embodiment of the present invention further provides a corresponding apparatus to implement load balancing based on a session initial protocol. As shown in FIG. 5, the apparatus includes: a load balancing module S1 for distributing an application server for a session initial request message, also called a scheduling unit; a container S2 for storing a list of all application servers in the cluster, and a container for storing a correspondence between the tongue mark and the application server Two S3. The container one S2 and the second container S3 are respectively connected to the load balancing module SI for the negative The load balancing module performs the query.

The load balancing module allocates a corresponding application server for the session initial request message by reading the session flag and the query container in the session initial request message, thereby implementing load balancing based on the session initial protocol.

A load balancing device based on a session initiation protocol according to an embodiment of the present invention includes: a storage unit, configured to save a correspondence between a session identifier and an application server;

The load balancing device further includes:

a determining unit, configured to determine whether a session identifier of the session initial request message exists; and when the determining unit determines that the session identifier of the session initial request message exists, forwarding the session initial request message to the scheduling unit; And the scheduling unit redirects the session initial request message to an idle application server for processing, and records a session identifier in the session initial message and a correspondence between the session identifier and the idle application server.

FIG. 6 is a schematic structural diagram of a device for implementing load balancing based on an initial protocol according to another embodiment of the present invention. In this embodiment, a load balancing unit S1 is provided with a receiving module S11, an allocating module S12, and a sending module S13;

The receiving module S11 is connected to the allocating module S12, and is configured to receive a session initial request message and provide it to the allocating module S12;

The distribution module S12 is connected to the container S2 and the container S3, and is configured to query the application server list saved by the container S2 and the correspondence between the session identifier saved by the container S3 and the application server, and allocate the session initial request message. application server;

The sending module S13 is connected to the allocating module S12, and is configured to send a session initial message to the application server, where the application server is an application server allocated by the allocating module S12 for the session initial request message.

Those skilled in the art can understand that all or part of the functional modules in the above embodiments and The steps may be implemented by a program instructing related hardware, and the program may be stored in a computer readable storage medium, such as a ROM RAM, a magnetic disk, a optical disk, or the like. Alternatively, they may be fabricated into individual integrated circuit modules, or a plurality of modules or steps may be fabricated into a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software.

It can be seen from the foregoing embodiment that the present invention schedules different SIP requests having the same tongue mark to the same application server (node) for processing by setting the request message type of the session and the session flag, thereby implementing SIP of different CALLID or SourcelP. The request is sent to the same node for processing purposes.

The SIP message processing in the IMS service is complex, different services, and different application scenarios. For cluster applications, the distribution policy is also flexible. The configurable method of the distribution policy in the present invention can meet the service requirements of various SIP applications, so that the cluster technology can be widely used in the IMS application server (AS), and the cost of the AS is reduced.

The above embodiments are only used to illustrate the principles of the present invention, and are not to be construed as limiting the scope of the embodiments of the present invention. Make modifications or equivalent replacements. Modifications or equivalent substitutions are intended to fall within the scope of the invention.

Claims

A method for implementing load balancing based on a session initial protocol, comprising: a network side receiving a session initial protocol SIP request message sent by a terminal, setting a session flag for the tongue initial protocol SIP request message, and The Session Initiation Protocol SIP Request message and its session flag are sent to a scheduling unit for allocating an application server for the Session Initiation Protocol SIP Request message;

The scheduling unit reads the rights session flag, and determines an application server that processes the tongue initial protocol SIP request message according to a correspondence between a predetermined session flag and an application server.

The load-to-U balance implementation method according to claim 1, further comprising: pre-establishing and saving a correspondence between the session identifier and the application server.

The load balancing and balancing implementation method according to claim 1 or 2, further comprising:

Determining, by the scheduling unit, whether a session identifier of the SIP request message exists, and if yes, the scheduling unit reads a session identifier in the session initiation protocol SIP request message, and according to a correspondence between the predetermined session identifier and the application server The relationship determines an application server that processes the SIP request message; otherwise, the SIP message is redirected to an idle application server for processing.

The load balancing implementation method according to claim 3, further comprising: recording a session identifier in the SIP message and a correspondence between the session identifier and the idle application server.

The load balancing implementation method according to claim 4, comprising: after the scheduling unit redirects the SIP request message, the idle application server returns a session initial protocol response message to the scheduling unit. .

A load balancing device based on a session initial protocol, comprising: a storage unit, configured to save a correspondence between a session identifier and an application server;

a scheduling unit, configured to allocate an application server for a session initiation protocol SIP request message; after receiving the SIP request message, the scheduling unit queries the storage unit according to the session identifier carried in the received SIP request message to obtain the session identifier server Correspondence to determine the application server that handles the SIP request message.

The load balancing device according to claim 6, further comprising: a determining unit, configured to determine whether a session flag of the SIP request message exists; and when the determining unit determines a session of the SIP request message a flag exists to forward the SIP message to the scheduling unit; otherwise, the scheduling unit redirects the SIP request message to an idle application server for processing, and records a session flag in the SIP request message and Correspondence relationship between the session identifier and the idle application server.

The load balancing device according to claim 6 or 7, wherein the scheduling unit is provided with a receiving module, an allocating module and a transmitting module;

The receiving module is configured to receive a SIP message and provide the information to the distribution module, where the allocation module is configured to query a correspondence between a session identifier saved in the storage unit and an application server, and according to the session identifier and the application The corresponding relationship of the server is

SIP request message distribution application server;

The sending module is connected to the allocating module, and configured to send a SIP request message to an application server that is allocated by the allocating module for the session initial request message.