CN1691796A - Communication load sharing system and method thereof - Google Patents

Abstract

The invention relates to mobile communication technology, discloses a system and method for contributing the communication load, so that a plurality of GGSN can contribute the load effectively and supply redundancy protection. The system and method for contributing the communication load sets a plurality of GGSN network in a master/slave mode, in which there is at least one GGSN serves as host node, each host node maintenances the source occupation of all the host nodes, when the SGSN enquires the DNS, the DNS comes back to the address of the host node, then SGSN sends creating PDP test solicited message to the host node, the host node judges that whether local GGSN is in minimum source occupation, if it is minimum, processing the solicited message direct, if not, it repeats the solicited message to the GGSN which source occupation is minimum, not allowed to repeat more then one time; after creating PDP test solicited message successfully, data flow interact between the SGSN and the active GGSN direct.

Description

Communication load sharing system and method thereof

technical field

The invention relates to mobile communication technology, in particular to load sharing technology among multiple communication nodes.

Background technique

General Packet Radio Service (General Packet Radio Service, referred to as "GPRS") is one of the contents of the implementation of the Global System for Mobile (Global System for Mobile, referred to as "GSM") Phase 2.1 specification. It introduces GPRS service support in the GSM system Node (Serving GPRSSupporting Node, referred to as "SGSN"), GPRS gateway support node (Gateway GPRSSuporting Node, referred to as "GGSN") and packet control unit, so that the GPRS system can provide higher data than the existing GSM network 9.6kbit/s rate. The advantages of GPRS are high resource utilization, high transmission rate, short access time, and support for IP protocol and X.25 protocol.

The GPRS mobile station communicates with the GSM base station and sends GPRS packets to the base station. But different from the circuit-switched call, GPRS groups are sent from the base station to the SGSN, rather than connected to the voice network through a mobile switching center. SGSN communicates with GGSN and sends GPRS packets to GGSN. The GGSN performs corresponding processing on the packet data, and then sends it to the destination network, such as the Internet or X.25 network. The IP packet marked with the address of the mobile station from the Internet is received by the GGSN, and then forwarded to the SGSN, and then transmitted to the mobile station. The SGSN is connected with the transceiver subsystem of the base station through frame relay, and is the interface between the GSM network and the mobile station. The main role of the SGSN is to record the current location information of the mobile station, and complete the sending and receiving of mobile packet data between the mobile station and the GGSN. GGSN is connected to SGSN through GPRS backbone network based on IP protocol, and is the gateway connecting GSM network and external packet switching network. GGSN mainly acts as a gateway, so GGSN is also called a GPRS router. GGSN can convert the protocol of GPRS packet in GSM network, so that these packets can be sent to the remote TCP/IP or X.25 network. SGSN and GGSN use GPRS tunnel protocol to encapsulate IP or X.25 packets to realize data transmission between them.

When users need to use the wireless packet service provided by GPRS, they first need to set the access point name (Access Point Name, "APN" for short) on the mobile station, and then send an activation request message to the SGSN to initiate the activation process. APN represents the name of the packet network accessed by GPRS. After the SGSN receives the activation request message sent by the mobile station, it extracts the APN information, and then queries the domain name server (Domain Name Server, referred to as "DNS") for the address of the GGSN gateway that needs to be accessed. After obtaining the address of the GGSN, the SGSN initiates a packet data protocol (Packet Data Protocol, referred to as "PDP") context request message to the selected GGSN, and after receiving the response sent back by the GGSN, the logical link is activated, and the user can The grouping services provided by GPRS such as Internet browsing and multimedia short messages have been implemented through the selected GGSN.

Generally, only one GGSN needs to be set up in a city or site; however, with the development of mobile data services, the demand for GPRS packet services of mobile users is increasing, so it is possible that multiple GGSNs need to be set up in a certain area. At this time, there are problems of networking and service load sharing between multiple GGSN nodes. A solution to this problem is to divide different GGSNs according to different access point names. When DNS resolves the domain name, the addresses of different GGSNs will be given according to the access point names, so that different GGSNs can share the load of different networks.

In practical application, the above solution has the following problems: Since the traffic volume of each access network is unequal in most cases, this method cannot realize uniform load sharing of each GGSN. In addition, if each access network uses one GGSN gateway, the resources of each GGSN cannot be used for mutual redundant backup, and the reliability cannot be guaranteed.

One of the main reasons for this situation is that this solution simply assigns a certain access network to a certain GGSN, resulting in uneven loads on each GGSN, and each GGSN cannot be a backup for each other.

Another prior art is to configure an address round-robin function on the SGSN and the DNS. Configure multiple GGSN addresses for each domain name in the DNS domain name resolution table. Assume that there are 3 GGSNs with addresses IP1, IP2, and IP3. DNS rotates the address list given by it each time it performs domain name resolution. Arrangement sequence, such as IP1, IP2, IP3 given for the first time (DNS returns multiple GGSN addresses for mobile stations to choose at a time), IP2, IP3, IP1 given for the second time, IP3, IP1, IP3 given for the third time IP2, and so on. Similarly, one of multiple addresses may be selected in turn on the SGSN as the address of the selected GGSN. In this way, load sharing among the GGSNs can also be achieved to a certain extent.

In practical application, the above solution has the following problems: the objectives of evenly load sharing and redundant protection cannot be achieved among multiple GGSNs.

One of the main reasons for this situation is that there is no resource information exchange means between the SGSN and multiple GGSNs in this scheme, and traffic allocation cannot be intelligently performed according to the load status of each GGSN.

Contents of the invention

In view of this, the main purpose of the present invention is to provide a communication load sharing system and method thereof, so that multiple GGSNs can effectively share the load and provide redundant protection.

To achieve the above object, the present invention provides a communication load sharing system, comprising at least two GPRS gateway support nodes, at least one of which is a master node, and at least one of which is a slave node;

The master node is used to maintain the resource occupancy of all GPRS gateway support nodes in the system, and when receiving a packet data protocol context creation request message from a GPRS service support node, judges whether the master node is The GPRS GW support node with the least current resource occupation, if so, processes the request message, or forwards the request message to the GPRS GW support node with the minimum current resource occupation;

The slave node is configured to process the request message from the master node.

Wherein, the system includes at least two master nodes.

The master node obtains resource occupation information of other GPRS gateway support nodes through the private information unit of the GPRS tunneling protocol message.

The resource occupation situation may be the number of activated users or the number of occupied address pools.

The GPRS gateway support node belongs to a GPRS system or a broadband code division multiple access system.

The present invention also provides a communication load sharing method. There are at least two GPRS GW support nodes, at least one of which is a master node, and each master node maintains all GPRS GW support nodes resource occupancy; the method includes the following steps:

B. The GPRS service support node sends a packet data protocol context creation request message to the master node;

C. The master node judges whether the master node is the GPRSG support node with the least current resource occupation, and if so, processes the request message, or forwards the request message to the GPRS with the least current resource occupation The wireless service gateway supports nodes to process;

Among them, the following steps are also included:

D After the packet data protocol context is created successfully, the data flow directly interacts between the GPRS gateway support node processing the request message and the GPRS service support node.

Before the step B, the following steps are also included:

A GPRS service support node queries the domain name service system and obtains the address of the master node;

Wherein, the following sub-steps are also included in the step C:

If the master node has processed the request message, the master node sends a packet data protocol context creation response message to the GPRS service support node;

If other GPRS GW support nodes have processed the request message, then the GPRS GW support node sends a packet data protocol context response message to the master node, and the master node sends a packet data protocol context response message to the GPRS The business service support node forwards the packet data protocol context creation response message.

In the step C, when the master node receives a packet data protocol context creation request message from another master node, the master node directly processes the request message.

Described step A also further comprises following sub-steps:

The GPRS service support node in A1 queries the domain name service system for the address of the GPRS gateway support node;

A2 The domain name service system returns the addresses of multiple master nodes in random order to the GPRS service support node;

A3 The GPRS service support node arbitrarily selects one of the multiple master node addresses returned to send the packet data protocol context creation request message.

The GPRS gateway support node belongs to a GPRS system or a broadband code division multiple access system.

By comparison, it can be found that the difference between the technical solution of the present invention and the prior art is that a master-slave GGSN networking mode is introduced, there can be multiple master GGSNs, function enhancements are performed on the master GGSN, and GGSN is implemented using the standard GTP protocol. Information exchange and signaling forwarding between nodes realize link detection, and maintain the load status information of each GGSN on the main GGSN. The maintenance of the load state information mentioned in the present invention refers to continuously inquiring the load state information of each GGSN, and constantly updating the load state information of each GGSN saved by itself with the inquired information.

The difference in this technical solution has brought obvious beneficial effects, that is, because the load status information of each GGSN is maintained on the primary GGSN, the primary GGSN can find the GGSN with the lightest load to forward when receiving the signaling , to achieve even load sharing of multiple GGSNs; and because multiple primary GGSNs are used, each primary GGSN has the same function and is connected to other GGSNs, so as long as there is one primary GGSN that can still work normally, the entire system can be guaranteed It will not completely fail, and the reliability of the system is guaranteed; this master-slave GGSN networking mode is very flexible, and can continue to expand with the increase of network traffic, meeting the needs of GPRS and WCDMA networks with large traffic volumes .

Description of drawings

Fig. 1 is a plurality of GGSNs according to an embodiment of the present invention network structure diagram in master-slave (single master GGSN) mode;

Fig. 2 is a plurality of GGSNs according to an embodiment of the present invention in a master-slave (multi-master GGSN) mode networking structure diagram;

Fig. 3 is a flow chart of service messages of multiple GGSNs networking in a master-slave manner according to an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings.

As shown in Fig. 1, in one embodiment of the present invention, the multi-GGSN system of networking with single master GGSN mode is made up of SGSN 11, master GGSN 21, No. 1 slave GGSN 31, No. 2 slave GGSN 32 and DNS 41. Wherein SGSN 11 is connected with DNS 41 and main GGSN 21 respectively. The master GGSN 21 is connected to the slave GGSN 31 and the slave GGSN 32 respectively. From the perspective of SGSN 11, only the main GGSN 21 exists. In another embodiment of the present invention, there may be more than two slave GGSNs, and all slave GGSNs are connected to the master GGSN.

SGSN 11 is responsible for receiving GPRS service requests from users, converting the requests into corresponding PDP context creation request messages and sending them to the main GGSN 21 according to the GGSN address returned by DNS. In addition to having all the functions of a common GGSN, the master GGSN 21 also maintains the resource occupation information of the slave GGSN 31 and the slave GGSN 32. The master GGSN 21 constantly learns about the resource occupation of each slave GGSN (including the number of activated users, the number of address pools occupied, etc.) through the link detection message and updates the information. It should be noted that the link detection uses the standard GPRS Tunnel Protocol (GPRS Tunnel Protocol, "GTP") handshake message, and uses the private information unit of the message to expand the information that needs to be exchanged between GGSN nodes. After the main GGSN 21 receives the request message to create a PDP context, it selects the most idle GGSN to forward according to the information maintained by itself, thereby effectively realizing load sharing. The forwarding message also adopts the standard GTP message, and the purpose is to ensure that the secondary GGSN can process it normally.

The slave GGSN 31 and the slave GGSN 32 only need to inform the master GGSN 21 of its own resource occupation information and process the creation PDP context request message transferred to the master GGSN 21, and they themselves do not need to maintain the resource occupation information.

DNS 41 is mainly responsible for providing a suitable GGSN address to SGSN 11, adopting the system structure shown in Figure 1, only need to fix the address as the address of the main GGSN 21.

The system structure shown in Fig. 1 can well realize the load sharing between multiple GGSNs, but once the link between SGSN 11 and master GGSN 21 or the link between master GGSN 21 and any slave GGSN fails, it will cause large-scale impact. For security considerations, the present invention makes the following improvements to the system structure shown in Fig. 1 again:

In a preferred embodiment of the present invention, another main GGSN 32 has been added, as shown in Figure 2, the multi-GGSN system with a plurality of main GGSN mode networking consists of SGSN 11, main GGSN 21, main GGSN 22, from GGSN 31 , formed from GGSN 32 and DNS 41. Wherein SGSN 11 is not only connected with DNS 41, but also connected with main GGSN 21 and main GGSN 22 respectively. The master GGSN 21 and the master GGSN 22 are respectively connected to the slave GGSN 31 and the slave GGSN 32, and the two master GGSNs are also connected to each other. From the perspective of SGSN11, there are only master GGSN 21 and master GGSN 22. For the master GGSN 21, the master GGSN 22, the slave GGSN 31 and the slave GGSN 32 are all its slave GGSNs; similarly, the master GGSN 22 also regards the master GGSN 21, the slave GGSN 31 and the slave GGSN 32 as its slaves GGSN.

The functions of each entity (including DNS, SGSN, primary GGSN, and secondary GGSN) in Figure 2 and the mutual messages are the same as those in Figure 1, except that the SGSN 11 can arbitrarily select the primary entity according to the resolution results of the DNS 41 during the user's Internet access process. GGSN 21 or master GGSN 22 accesses, and these two master GGSNs can choose to process or forward to the 3 slave GGSNs connected to themselves according to the busy or idle information of the resources they maintain (the other master GGSN is also regarded as the current GGSN) from any one of GGSN). However, it should be noted that each PDP context request message can only be forwarded once at most, otherwise the PDP context request message may be forwarded unlimitedly. Compared with the system shown in Figure 1, the improved system has a certain degree of redundancy in addition to effective load sharing, thereby improving the reliability of the system.

It is worth mentioning that with the expansion of the network scale, when the GGSN load of the original system is too heavy, it can be easily expanded according to the structure shown in Figure 2, and the master GGSN and slave GGSN can be added to establish a super-large GGSN node. For example, the master GGSN can be expanded to 3, and the slave GGSN can be expanded to 6. Those skilled in the art can know that both WCDMA network and GPRS network have SSNN and GGSN nodes, and the functions are also consistent, so the structure proposed by the present invention is applicable to both GPRS network and WCDMA network.

The communication load sharing system of the present invention has been described above, and the communication load sharing method of the present invention will be described below in conjunction with the system.

As shown in Figure 3, in step 101, after SGSN 11 receives the user's Internet access request, it initiates a DNS query request message to DNS41 to inquire about the address of the main GGSN that needs to be accessed.

Then enter step 102, DNS 41 returns DNS query response message to SGSN 11, with the address of all or part of the main GGSN that can be accessed in the message. These address information are usually pre-configured in DNS 41. For example, for the system shown in Figure 2, the addresses of each main GGSN (GGSN 21 and GGSN 22) are stored in DNS 41.

Then enter step 103, the SGSN 11 selects the main GGSN 21 to access according to the address returned by the DNS 41 and initiates a PDP context request message to it, usually the message includes information such as the protocol type adopted, the name of the access point, and the quality of service . The selection of the main GGSN is random, and GGSN 22 can also be selected. Here, GGSN 21 is taken as an example for illustration.

Then enter step 104, master GGSN 21 forwards the establishment PDP context request message sent from SGSN 11 to comparatively idle from GGSN 31 according to the resource usage of each GGSN node maintained by itself. It should be noted that the master GGSN 21 includes the resource occupation information of itself and all slave GGSN nodes connected to it, and if the multi-master GGSN mode is adopted, it also includes resource occupation information of other master GGSN nodes connected to it. In addition to its own information, the information of other nodes is obtained through the link detection mechanism. Except the situation shown in the figure, if the main GGSN 21 itself is the most idle, the session can be processed by itself without forwarding the message, and the PDP context response message is directly replied to the SGSN 11, and step 106 is entered.

Then enter step 105, after receiving the PDP context request message forwarded by the main GGSN 21 from the GGSN 31, send a PDP context response message to the main GGSN 21, which includes the address of the GGSN 31.

Then enter step 106, the main GGSN 21 forwards the PDP context response message sent from the GGSN 31 to the SGSN 11, and the SGSN 11 feeds back to the user. For the session that the master GGSN 21 handles by itself, it is not forwarding but actively sends a PDP context response message to the SGSN 11.

From step 103 to step 106, if the master GGSN21 needs to forward the create PDP context request message, it can adopt the technology proposed in the Chinese patent application number 02129375.9. In this patent, it is assumed that there are SGSN code-named A, GGSN code-named B and GGSN code-named C in the system. If the connection between A and C is to be realized through the forwarding of B, the following steps can be taken:

First: A sends a PDP context activation request message to B. The source address in the IP header of the message is A, and the destination address is B. The message contains two information elements: "SGSN Address for signaling" and "SGSN Address for user traffic ", respectively representing the SGSN signaling address and data address used by the PDP context. The key information of the message is as follows:

IP header: A->B

SGSN Address for signaling: A

SGSN Address for user traffic: A

Second: B forwards the message to C. The source address in the IP header is B, the destination address is C, and the content of the GTP packet is not modified. The key information of the message is as follows:

IP header: B->C

SGSN Address for signaling: A

SGSN Address for user traffic: A

Again: C will establish a PDP context after receiving the activation request message. According to the GTP protocol, the SGSN signaling address should be the cell SGSN Address for signaling carried in the activation request message: A, and the SGSN data address is the signal carried in the message. Set SGSN Address for usertraffic: A, and specify the GGSN signaling and data address as its own address C, and then return an activation response message. According to the GTP protocol, the source address in the IP header of the message should be C, and the destination address should be B. The activation response message contains two information elements: "GGSN Address for signaling" and "GGSNAddress for user traffic", respectively representing the GGSN signaling address and data address used by the PDP context, and its value is C. The key information of the message is as follows:

IP header: C->B

GGSN Address for signaling: C

GGSN Address for user traffic: C

Then: B forwards to A after receiving the activation response message, the source address in the message IP header is B, the destination address is A, B does not care about the content of the activation response message and does not make any modification. The key information of the message is as follows:

IP header: B->A

GGSN Address for signaling: C

GGSN Address for user traffic: C

Finally: A receives the activation response message and establishes a PDP context. According to the GTP protocol, the SGSN address in the PDP context should be the address A specified by itself, and the GGSN signaling address should be the information element GGSN Address for carried in the activation response message. signaling: C, the GGSN data address is the cell carried in the message GGSN Address for user traffic: C; in this way, a pair of PDP contexts is established between A and C, where the SGSN address is A and the GGSN address is C, which is equivalent to That is, a logical channel for GTP message transmission is established between A and C. Subsequent GTP signaling messages and GTP data packets are directly interacted between A and C, and B is no longer involved.

Then enter step 107, after the user obtains the address from the GGSN 31, just use the GGSN 31 as a gateway to start surfing the Internet and carry out the data transmission process. It should be noted that the data flow no longer needs to be forwarded by the master GGSN 21, but directly interacts between the SGSN 11 and the actually activated slave GGSN 31.

When the user finishes surfing the Internet, enter step 108 . The SGSN 11 directly initiates a delete PDP context request message to the slave GGSN 31 to inform the end of the data transmission.

Then enter step 109, send back the deletion PDP context response message from GGSN 31 to SGSN 11, so far a complete GPRS online process ends.

Although the present invention has been illustrated and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein, and without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (12)

1. A communication load sharing system, characterized in that it comprises at least two GPRS gateway support nodes, wherein at least one is a master node, and at least one is a slave node; The master node is used to maintain the resource occupancy of all GPRS gateway support nodes in the system, and when receiving a packet data protocol context creation request message from a GPRS service support node, judges whether the master node is The GPRS GW support node with the least current resource occupation, if so, processes the request message, or forwards the request message to the GPRS GW support node with the minimum current resource occupation; The slave node is configured to process the request message from the master node. 2. The communication load sharing system according to claim 1, wherein the system comprises at least two master nodes. 3. The communication load sharing system according to claim 2, wherein the master node obtains resource occupancy information of other GPRS gateway support nodes through the private information unit of the GPRS tunneling protocol message. 4 . The communication load sharing system according to claim 2 , wherein the resource occupancy status can be the number of activated users or the number of occupied address pools. 5. The communication load sharing system according to any one of claims 1 to 3, wherein the GPRS gateway support node belongs to a GPRS system or a broadband code division multiple access system. 6. A communication load sharing method, characterized in that there are at least two GPRS GW support nodes, at least one of which is a master node, and each master node maintains all GPRS GW support nodes The resource occupation situation of; Described method comprises the following steps: B. The GPRS service support node sends a packet data protocol context creation request message to the master node; C. The master node judges whether the master node is the GPRSG support node with the least current resource occupation, and if so, processes the request message, or forwards the request message to the GPRS with the least current resource occupation The wireless service gateway supports nodes for processing. 7. communication load sharing method according to claim 6, is characterized in that, also comprises the following steps: D After the packet data protocol context is created successfully, the data flow directly interacts between the GPRS gateway support node processing the request message and the GPRS service support node. 8. communication load sharing method according to claim 7, is characterized in that, also comprises the following steps before described step B: The GPRS service support node in A queries the domain name service system and obtains the address of the master node; 9. The communication load sharing method according to claim 6, characterized in that, the step C also includes the following sub-steps: If the master node has processed the request message, the master node sends a packet data protocol context creation response message to the GPRS service support node; If other GPRS GW support nodes have processed the request message, then the GPRS GW support node sends a packet data protocol context response message to the master node, and the master node sends a packet data protocol context response message to the GPRS The business service support node forwards the packet data protocol context creation response message. 10. The communication load sharing method according to claim 6, characterized in that, in the step C, when the master node receives the packet data protocol context request message from other master nodes, the master node The request message is processed directly. 11. The communication load sharing method according to claim 8, wherein said step A further comprises the following sub-steps: The GPRS service support node in A1 queries the domain name service system for the address of the GPRS gateway support node; A2 The domain name service system returns the addresses of multiple master nodes in random order to the GPRS service support node; A3 The GPRS service support node arbitrarily selects one of the multiple master node addresses returned to send the packet data protocol context creation request message. 12. The communication load sharing method according to any one of claims 6 to 11, wherein the GPRS gateway support node belongs to a GPRS system or a broadband code division multiple access system.

Images