WO2014047902A1

WO2014047902A1 - Load balancing method, device, system and computer readable medium

Info

Publication number: WO2014047902A1
Application number: PCT/CN2012/082382
Authority: WO
Inventors: 宁德刚
Original assignee: 华为技术有限公司
Priority date: 2012-09-28
Filing date: 2012-09-28
Publication date: 2014-04-03
Also published as: CN103918239A

Abstract

Embodiments of the present invention provide a load balancing method, device, system and computer readable medium. The load balancing method comprises: obtaining a first load of a first processing node in a distributed processing system and an average load of all processing nodes in the distributed processing system; if the first load is greater than the average load, migrating at least one virtual node distributed on the first processing node to at least one processing node whose load is less than the average load in the distributed processing system. By means of the embodiment of the present invention, at least one virtual node corresponding to a processing node whose load is greater than an average load in a distributed system can be migrated to at least one processing node whose load is less than the average load in the distributed processing system, so that the load distribution on processing nodes in the distributed processing system becomes more balanced, thereby enhancing the parallel processing capability of the distributed processing system and improving the efficiency of the distributed processing system.

Description

The present invention relates to communication technologies, and in particular, to a load balancing method, apparatus, system, and computer readable medium. Background technique

With the rapid development of the Internet, enterprises are increasingly demanding distributed processing capabilities. The distributed processing system clusters a large number of inexpensive servers, each of which is a processing node, which provides an overall high-performance cluster processing capability by processing node parallel processing.

At present, in a distributed processing system, a processing task is usually decomposed into multiple processing requests and distributed to multiple processing nodes for parallel processing, so that high parallel processing capability can be obtained, which requires balancing multiple processing requests. Assigned to each processing node.

In the process of implementing the embodiments of the present invention, the inventor finds that the processing method in the prior art often has a problem of unbalanced processing request allocation, which in turn causes the heavily loaded processing node to become the bottleneck of the entire distributed processing system, and the load is light. The processing node wastes its processing power, which reduces the processing efficiency of the distributed processing system. SUMMARY OF THE INVENTION Embodiments of the present invention provide a load balancing method, apparatus, system, and computer readable medium, so that processing nodes in a distributed processing system are more balanced in load distribution.

In a first aspect, a load balancing processing method is provided, including:

Obtaining a first load of a first processing node in the distributed processing system and an average load of all processing nodes in the distributed processing system;

And if the first load is greater than the average load, migrating at least one virtual node distributed on the first processing node to at least one processing node in the distributed processing system with a load less than the average load.

In a first possible implementation manner of the first aspect, the at least one virtual node that is distributed on the first processing node is migrated to the distributed processing system, and the load is less than At least one processing node of the average load, including:

Migrating at least one virtual node distributed on the first processing node with a total load less than or equal to an overload amount to at least one processing node in the distributed processing system, wherein the total amount of underload is greater than or equal to the total amount of the load, The overload amount is a load amount of the first load exceeding the average load, and the total amount of underload is a sum of load of each of the at least one processing node being lower than a load of the average load .

With reference to the first aspect or the first possible implementation manner of the first aspect, in a second possible implementation, the at least one virtual node that is distributed on the first processing node is migrated to the Before the load in the distributed processing system is less than the average load on at least one processing node, the method further includes:

Obtaining load information of each processing node in the distributed processing system, where the load information includes load information of each virtual node distributed on each processing node;

Determining, according to load information of each processing node, the at least one processing node; sending, to the at least one processing node, a migration request message, where the migration request message carries information of the at least one virtual node;

Receiving a migration request response message of the at least one processing node;

The at least one virtual node distributed on the first processing node is migrated to the at least one processing node in the distributed processing system where the load is less than the average load, specifically:

And if the migration request response message is a success response message, migrating at least one virtual node distributed on the first processing node to the at least one processing node.

With reference to the first aspect or the first possible implementation manner of the first aspect, in a third possible implementation, the at least one virtual node that is distributed on the first processing node is migrated to the The at least one processing node in the distributed processing system whose load is less than the average load includes:

Updating a mapping relationship table between the first processing node and the at least one virtual node;

And transmitting, to the other processing nodes in the distributed processing system, update information, so that other processing nodes in the distributed processing system update the mapping table between the locally stored processing node and the virtual node according to the update information. With reference to the first aspect or the first possible implementation manner of the first aspect, in a fourth possible implementation, the at least one virtual node that is distributed on the first processing node is migrated to the After the load in the distributed processing system is less than the processing load on the at least one processing node, the method further includes:

Synchronizing data corresponding to the at least one virtual node to the at least one processing node.

In a second aspect, a load balancing processing apparatus is provided, including:

An obtaining module, configured to acquire a first load of a first processing node in the distributed processing system and an average load of all processing nodes in the distributed processing system;

a processing module, if the first load is greater than the average load, migrating at least one virtual node distributed on the first processing node to at least one processing in the distributed processing system that has a load less than the average load On the node.

In a first possible implementation manner of the second aspect, the processing module is specifically configured to migrate at least one virtual node that is distributed on the first processing node to a total amount of load less than or equal to an overload amount to the distributed The at least one processing node in the processing system is greater than or equal to the total amount of the load, the overload amount is a load amount of the first load exceeding the average load, and the total underload is the at least The load of each processing node in a processing node is lower than the sum of the load of the average load.

With reference to the second aspect or the first possible implementation manner of the second aspect, in a second possible implementation, the determining module, the sending module, and the receiving module are further included;

The acquiring module is further configured to acquire load information of each processing node in the distributed processing system, where the load information includes load information of each virtual node distributed on each processing node;

The determining module is configured to determine, according to load information of each processing node, the at least one processing node;

The sending module is configured to send a migration request message to the at least one processing node, where the migration request message carries information of the at least one virtual node;

The receiving module is configured to receive a migration request response message that is sent by the at least one processing node according to the migration request message;

The processing module is specifically configured to: if the migration request response message is a success response message, And then migrating at least one virtual node distributed on the first processing node to the at least one processing node.

With reference to the second aspect or the first possible implementation manner of the second aspect, in a third possible implementation manner, the method further includes:

And an update module, configured to update a mapping relationship table between the first processing node and the at least one virtual node;

The sending module is further configured to send update information to other processing nodes in the distributed processing system, so that other processing nodes in the distributed processing system update the locally stored processing node and the virtual according to the update information. A mapping table between nodes.

With reference to the second aspect or the first possible implementation manner of the second aspect, in a fourth possible implementation, the method further includes:

And a synchronization module, configured to synchronize data corresponding to the at least one virtual node to the at least one processing node.

In conjunction with the second aspect or the first possible implementation of the second aspect, in a fifth possible implementation, the device is a processing node or a master control center.

In a third aspect, a distributed processing system is provided, comprising: at least two processing nodes described in the above fifth possible implementation manner.

In a fourth aspect, a distributed processing system is provided, including: at least two processing nodes and a master control center as described in the fifth possible implementation manner.

In a fifth aspect, a processing node is provided, including a processor and a memory, the memory storing execution instructions, when the device is running, the processor is in communication with the memory, the processor executing the Executing the instructions causes the processing node to perform the method of the first aspect.

In a sixth aspect, a master control center is provided, including a processor and a memory, the memory storing execution instructions, when the device is running, the processor is in communication with the memory, and the processor executes Executing the instructions causes the master control center to perform the method of the first aspect.

In a seventh aspect, a distributed processing system is provided, comprising at least two processing nodes described in the fifth aspect.

In an eighth aspect, a distributed processing system is provided, including at least two processing nodes And the main control center of the sixth aspect.

In a ninth aspect, a computer readable medium is provided, comprising computer executed instructions for causing a processing node to perform the method of the first aspect.

In a tenth aspect, there is provided a computer readable medium comprising computer executed instructions for causing a master control center to perform the method of the first aspect.

The load balancing processing method provided by the embodiment, the device, the system, and the computer readable medium, wherein at least one virtual node corresponding to the processing node with a load greater than the average load is migrated to the distributed processing system in the distributed system, and the load is less than the average load. At least one processing node can make the processing nodes in the distributed processing system more balanced in load distribution, thereby enhancing the parallel processing capability of the distributed processing system and improving the efficiency of the distributed processing system. DRAWINGS

The drawings used in the embodiments or the description of the prior art are briefly described. It is obvious that the drawings in the following description are some embodiments of the present invention, and are not creative to those skilled in the art. Other drawings can also be obtained from these drawings on the premise of labor.

1 is a flowchart of Embodiment 1 of a load balancing processing method provided by the present invention;

2 is a flowchart of Embodiment 2 of a load balancing processing method provided by the present invention;

3 is a flowchart of Embodiment 3 of a load balancing processing method provided by the present invention;

4 is a schematic structural diagram of Embodiment 1 of a load balancing processing apparatus according to the present invention; FIG. 5 is a schematic structural diagram of Embodiment 2 of a load balancing processing apparatus according to the present invention; FIG. 6 is a third embodiment of a load balancing processing apparatus according to the present invention. FIG. 7 is a schematic structural diagram of Embodiment 4 of a load balancing processing apparatus according to the present invention; FIG. 8 is a schematic structural diagram of Embodiment 1 of a processing node according to the present invention;

FIG. 9 is a schematic structural diagram of Embodiment 1 of a main control center according to the present invention;

FIG. 10 is a schematic structural diagram of Embodiment 1 of a distributed processing system according to the present invention; FIG. 11 is a schematic structural diagram of Embodiment 2 of a distributed processing system according to the present invention. detailed description The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is a partial embodiment of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

FIG. 1 is a flowchart of Embodiment 1 of a load balancing processing method according to the present invention. As shown in FIG. 1, the load balancing processing method in this embodiment includes:

Step 101: Acquire a first load of a first processing node in a distributed processing system and an average load of all processing nodes in the distributed processing system;

For example, the manner of obtaining the first load of the first processing node in the distributed processing system may be: setting a load counter on the first processing node, and receiving, by the first processing node, a processing request, In addition to the operation, the processing request amount recorded by the load counter is the load of the first processing node; another case may be: setting a load counter on each virtual node corresponding to the first processing node, then each virtual node The sum of the loads is the load of the first processing node described above.

The average load of all processing nodes of the distributed processing system may be obtained by: Each processing node in the distributed processing system elects a main processing node by using a consistency algorithm such as pasox, and a software timer is set on the main processing node, and the main processing is performed. The node timing sends its load information to other processing nodes in the distributed processing system. After receiving the load information of the main processing node, the other processing nodes send their own load information to the main processing node, and the main processing node according to the distributed processing. The load information of all processing nodes in the system calculates the average load of all processing nodes of the distributed processing system, and then broadcasts the average load to other processing nodes.

The average load of all processing nodes of the distributed processing system may also be obtained by: The master control center in the distributed processing system obtains all the processing of the distributed processing system by acquiring load information of all processing nodes in the distributed processing system. The average load of the nodes, and then broadcast the above average load to other processing nodes.

In a distributed processing system, a distributed hash called a virtual node is employed. Suppose there are N processing nodes, and the spatial ring of the keyword is randomly divided into Q intervals (Q»N) by an algorithm. One interval is called a virtual node. Then randomly divide the Q virtual nodes It is allocated to N processing nodes, and each processing node stores about Q/N virtual nodes. Where N and Q are both positive integers. The distributed processing system in the embodiments of the present invention is also applicable to the distributed processing system described in the following specific embodiments.

Step 102: If the first load is greater than the average load, migrating the at least one virtual node distributed on the first processing node to at least one processing node in the distributed processing system whose load is less than the average load.

For example, if the first load of the first processing node is greater than the average load, and the load amount of the first load exceeding the average load is called an overload amount, an overload amount threshold may be set according to the situation, that is, the overload amount at the first processing node. The above migration process is performed when the threshold is exceeded. After the virtual node is migrated, the at least one processing node that receives the load of the migrated virtual node that is less than the average load may process a part of the processing request corresponding to the migrated virtual node, thereby reducing the load of the first processing node.

The load balancing processing method provided in this embodiment is configured to migrate, by using a distributed system, at least one virtual node distributed on a processing node with a load greater than an average load to at least one processing node in a distributed processing system with a load less than an average load, The processing nodes in the distributed processing system are more balanced in load distribution, thereby enhancing the parallel processing capability of the distributed processing system and improving the efficiency of the distributed processing system.

Optionally, on the basis of the corresponding embodiment of FIG. 1, the at least one virtual node that is distributed on the first processing node is migrated to the at least one processing node in the distributed processing system where the load is less than the average load, including:

At least one virtual node distributed on the first processing node with a total load less than or equal to the overload amount is migrated to at least one processing node in the distributed processing system with the total underload being greater than or equal to the total load, and the overload is greater than the first load. The load of the average load, the total amount of underload is the sum of the load of each processing node in at least one processing node being lower than the load of the average load.

Specifically, if the at least one virtual node is a virtual node, the total load is the load of the one virtual node, and if the at least one virtual node is a plurality of virtual nodes, the total load is multiple virtual The sum of the load of the node. Similarly, if the at least one processing node is a processing node, the total amount of the underload is a load of the one processing node that is lower than the average load, if the at least one processing node is multiple Processing the node, the total amount of the underload is lower than the load of each of the at least one processing node The sum of the load of the average load.

For example, if the overload of processing node A in the distributed processing system is 50, the load of one virtual node A1 distributed on processing node A is 50, and the underload of another processing node B in the distributed processing system is 50, the virtual node A1 is migrated to the processing node B, and the other case may be: if the overload of the processing node A in the distributed processing system is 50, the two virtual nodes A1 and A2 distributed on the processing node A The load is 20 and 30 respectively, and the other node B in the distributed processing system has an underload of 60, and the virtual nodes A1 and A2 are migrated to the processing node B.

If the overload of processing node A in the distributed processing system is 50, the loads of the two virtual nodes A1 and A2 distributed on the processing node A are 20 and 30, respectively, and the other two processing nodes B and C in the distributed processing system. The underloads are 20 and 30 respectively, then the virtual node A1 is migrated to the processing node B, and the virtual node A2 is migrated to the processing node C. The other case may be: if the distributed processing system processes the node A The overload is 50, the load of the three virtual nodes A1, A2 and A3 distributed on the processing node A are 15, 20 and 10 respectively, and the underload of the other processing nodes B and C in the distributed processing system are respectively 25 and 45, the virtual node A1 is migrated to the processing node B, and the virtual nodes A2 and A3 are migrated to the processing node C. The load balancing processing method provided in this embodiment is configured to migrate at least one virtual node in the processing node with a total load less than or equal to the overload amount to at least one processing node in the distributed processing system with the total amount of underload being greater than or equal to the total load. The processing nodes in the distributed processing system can be more balanced in load distribution, enhance the parallel processing capability of the distributed processing system, and improve the efficiency of the distributed processing system.

2 is a flowchart of a second embodiment of a load balancing processing method according to the present invention. As shown in FIG. 2, the load balancing processing method in this embodiment includes:

Step 201: Acquire load information of each processing node in the distributed processing system, where the load information includes load information of each virtual node distributed on each processing node.

For example, the method for obtaining load information of each processing node in the distributed processing system may be: Each processing node in the distributed processing system elects a main processing node by using a consistency algorithm such as pasox, and sets a software timing on the main processing node. The main processing node periodically sends its load information to other processing nodes in the distributed processing system. After receiving the load information of the main processing node, the other processing nodes send their own load information to the main processing section. Point, the primary processing node broadcasts load information of all processing nodes to other processing nodes in the distributed processing system, or the primary processing node periodically sends its load information to other processing nodes in the distributed processing system, and the other processing nodes receive After the main processing node loads the information, it sends its own load information to other processing nodes in the distributed processing system.

The method for obtaining load information of each processing node in the distributed processing system may also be: The master control center in the distributed processing system acquires load information of all processing nodes in the distributed processing system, and then performs other processing in the distributed processing system. The node broadcasts the above load information.

The load information of the processing node includes the virtual nodes distributed on the processing node and the load on each virtual node.

Step 202: Determine at least one processing node according to load information of each processing node. Since each processing node in the distributed processing system can learn the load information of other processing nodes, if the first load of the first processing node is greater than the average load of the distributed processing system, the processing node or the control center is according to the foregoing The overload of the processing node and the load of the virtual node distributed on the first processing node determine at least one processing node in the distributed processing system that has a load less than the average load.

Step 203: Send a migration request message to at least one processing node, where the migration request message carries information of at least one virtual node.

After the at least one processing node is determined, the first processing node or the control center sends a migration request message to the determined at least one processing node, where the migration request message carries the virtual node information to be migrated.

Step 204: Receive a migration request response message of the determined at least one processing node.

After receiving the foregoing migration request message, the at least one determined processing node feeds back a migration request response message to the first processing node.

Step 205: If the migration request response message is a success response message, at least one virtual node distributed on the first processing node is migrated to at least one processing node in the distributed processing system whose load is less than the average load.

For example, the migration request response message is a failure response cancellation, and the foregoing may determine that the total underload of the at least one processing node determined by the foregoing is smaller than the total load of the at least one virtual node to be migrated.

The load balancing processing method provided in this embodiment firstly is based on the load signal of each processing node. And determining, by the at least one processing node, sending a migration request message to the at least one processing node, and if the migration request response message is a success response message, migrating the at least one virtual node distributed on the first processing node to the distributed processing system The at least one processing node with a medium load less than the average load can make the processing nodes in the distributed processing system more balanced in load distribution, enhance the parallel processing capability of the distributed processing system, and improve the efficiency of the distributed processing system.

On the basis of the corresponding embodiments of FIG. 1 and FIG. 2, further, the at least one virtual node distributed on the first processing node is migrated to the at least one processing node in the distributed processing system where the load is less than the average load, including:

Updating a mapping relationship table between the first processing node and the at least one virtual node.

The update information is sent to other processing nodes in the distributed processing system to cause other processing nodes in the distributed processing system to update the mapping table between the locally stored processing nodes and the virtual nodes based on the update information.

Specifically, one implementation manner of migrating at least one virtual node distributed on the first processing node to at least one processing node whose load is less than the average load in the distributed processing system is: the foregoing first processing node or The master control center updates the mapping relationship table between the first processing node and the at least one virtual node according to the information of the at least one virtual node, and the mapping relationship table can be, for example, as shown in Table 1:

Table 1

For example, if the overload of processing node A is 50, the load of one virtual node A1 distributed on processing node A is 50, and the load of another processing node B in the distributed processing system is 50, it will be virtual. Node A1 is migrated to processing node B, and the updated mapping table is as shown in Table 2:

Table 2

Processing node A virtual node A2 (range), processing node B virtual node A1 (range), virtual node B1 (range), virtual node B2 (range), ......

Processing node C virtual node C1 (range), virtual node C2 (range), ... the first processing node or the master control center updates the first processing node and the at least one virtual node according to the information of the at least one virtual node After the mapping table between the nodes, the update information is sent to other processing nodes in the distributed processing system, and the updated information may include an updated mapping relationship table, and other processing nodes in the distributed processing system are updated according to the The information updates the mapping table between the locally stored processing node and the virtual node.

The load balancing processing method provided in this embodiment updates the mapping relationship table between the first processing node and the at least one virtual node, and sends update information to other processing nodes in the distributed processing system, so as to enable other in the distributed processing system. The processing node updates the mapping table between the locally stored processing node and the virtual node according to the update information, so that at least one virtual node on the processing node with a load greater than the average load is migrated to the distributed processing system in the distributed system, and the load is less than At least one processing node of the average load realizes that the processing nodes in the distributed processing system are more balanced in load distribution, enhance the parallel processing capability of the distributed processing system, and improve the efficiency of the distributed processing system.

On the basis of the embodiment corresponding to FIG. 1, after the at least one virtual node that is distributed on the first processing node is migrated to the at least one processing node whose load is less than the average load in the distributed processing system, the method further includes:

Synchronizing data corresponding to the at least one virtual node to at least one processing node.

Specifically, if the overload amount of the processing node A is 50, the load of one virtual node A1 distributed on the processing node A is 50, and the load of another processing node B in the distributed processing system is 50, the virtual node is After A1 is migrated to the processing node B, the data corresponding to A1 distributed in the processing node A is synchronized to the processing node B. If the processing request received by the virtual node A1 is a read operation, the data corresponding to A1 is synchronized to the processing node. After B, the processing node B can also process the above read operation processing request; if the processing request received by the virtual node A1 is a write operation, the processing node B can also process the write operation for the virtual node distributed on the processing node A. If A1 stores hot resources and causes the processing node A to be overloaded, the virtual node A1 distributed on the processing node A is migrated to the processing node B, and the processing node B processes the request. The load balancing processing method provided in this embodiment synchronizes the data corresponding to the virtual node on the first processing node to the at least one processing node after the virtual node is migrated, and in the case that some processing nodes store the hot resources, The processing nodes in the distributed processing system can be more balanced in load distribution, enhance the parallel processing capability of the distributed processing system, and improve the efficiency of the distributed processing system.

FIG. 3 is a flowchart of a third embodiment of a load balancing processing method according to the present invention. As shown in FIG. 3, the load balancing processing method in this embodiment includes:

Step 301: The timer expires.

The above-mentioned timers can be set on the main processing node or the main control center. The main processing nodes are elected by the processing nodes in the distributed processing system through a consistency algorithm such as pasox. For details, see Embodiment 1.

Step 302: The main processing node sends its load information to other processing nodes in the distributed processing system.

Step 303: The other processing nodes in the distributed processing system send their own load information to the primary processing node.

Step 304: The main processing node broadcasts load information of all processing nodes to other processing nodes in the distributed processing system.

Step 305: Each processing node calculates an average load of the distributed processing system according to the received load information, compares the self load and the average load, and determines whether the load needs to be migrated.

In this embodiment, the load of the processing node A is greater than the average load, and the load of the processing node A exceeds the average load is called the overload amount of the processing node A, and the threshold is set in advance. When the overload amount is greater than the threshold, the load migration is performed. .

Step 306: The processing node A determines, according to the load information of each processing node, at least one virtual node and at least one processing node whose total underload is greater than or equal to the total load of the virtual node to be migrated.

Specifically, if the at least one virtual node is a virtual node, the total load is the load of the one virtual node, and if the at least one virtual node is a plurality of virtual nodes, the total load is multiple virtual The sum of the loads of the nodes. Similarly, if the at least one processing node is a processing node, the total amount of the underload is the one processing node. The load is lower than the load of the average load. If the at least one processing node is a plurality of processing nodes, the total amount of underload is the sum of the load of each processing node of the at least one processing node being lower than the average load. .

In this embodiment, the load of the processing node B is less than the average load, and the underload amount is greater than or equal to the total load of the virtual node to be migrated, and the load of the virtual node A1 distributed on the processing node A is smaller than the overload of the processing node A. .

Step 307: The processing node A sends a migration request message to the processing node B.

The above migration message carries the load information of the virtual node A1.

Step 308: The processing node B sends a migration request response message to the processing node A.

If the underload of the processing node B is greater than or equal to the load of the virtual node A1, the processing node B sends a migration request response message to the processing node A.

Step 309: The processing node A updates the mapping relationship table between the virtual node A1 and the processing node A.

Processing node A updates the pre-migration A1-A to the migrated Al-B.

Step 310: The processing node B updates the mapping relationship table between the virtual node A1 and the processing node B.

Step 311: The processing node A sends update information to other processing nodes in the distributed processing system.

Step 312: The other processing nodes in the distributed processing system update the mapping table between the locally stored processing node and the virtual node according to the update information.

The load balancing processing method provided by the embodiment, by updating a mapping relationship table between the first processing node and the virtual node distributed on the first processing node, and the determined at least one processing node and the at least one processing node distributed in the determining The mapping manner of the virtual node on the virtual node is implemented in the distributed system to migrate at least one virtual node distributed on the processing node with a load greater than the average load to at least one processing node in the distributed processing system with a load less than the average load, The processing nodes in the distributed processing system are more balanced in load distribution, enhance the parallel processing capability of the distributed processing system, and improve the efficiency of the distributed processing system.

4 is a schematic structural diagram of Embodiment 1 of a load balancing processing apparatus according to the present invention. As shown in FIG. 4, the load balancing processing apparatus provided in this embodiment includes an obtaining module 41 and a processing module. 42 . The obtaining module 41 is configured to acquire a first load of the first processing node in the distributed processing system and an average load of the distributed processing system, and the processing module 42 is configured to: if the first load is greater than the average load, the first The at least one virtual node corresponding to the processing node is migrated to at least one processing node in the distributed processing system whose load is less than the average load.

The device in this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 1. The implementation principle and technical effects are similar, and details are not described herein again.

Optionally, the processing module 42 is configured to migrate at least one virtual node that is distributed on the first processing node with a total load less than or equal to the overload amount to at least one processing in the distributed processing system, where the total amount of underload is greater than or equal to the total amount of the load. On the node, the overload is the load of the first load exceeding the average load, and the total amount of the underload is the sum of the load of each processing node of the at least one processing node being lower than the load of the average load.

Specifically, if the at least one virtual node is a virtual node, the total load is the load of the one virtual node, and if the at least one virtual node is a plurality of virtual nodes, the total load is multiple virtual The sum of the load of the node. Similarly, if the at least one processing node is a processing node, the total amount of the underload is a load of the one processing node that is lower than the average load, if the at least one processing node is multiple Processing the node, the total amount of underload is the sum of the load of each processing node in the at least one processing node being lower than the load of the average load.

FIG. 5 is a schematic structural diagram of a second embodiment of a load balancing processing apparatus according to the present invention. As shown in FIG. 5, the load balancing processing apparatus provided in this embodiment, based on the embodiment shown in FIG. 4, further includes a determining module 43, a sending module 44, and a receiving module 45, wherein the obtaining module 41 is further configured to acquire load information of each processing node in the distributed processing system, where the load information includes each distributed on each processing node The load information of the virtual node, the determining module 43 is configured to determine at least one processing node according to the load information of each processing node, and the sending module 44 is configured to send a migration request message to the at least one processing node, where the migration request message carries the The information of the at least one virtual node, the receiving module 45 is configured to receive the migration request response message that is sent by the processing node according to the migration request message, and the processing module 42 is specifically configured to: if the migration request response message is a success response message, At least one virtual node on the first processing node To at least one processing node.

The device of this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 2, The implementation principle is similar to the technical effect, and will not be described here.

FIG. 6 is a schematic structural diagram of a third embodiment of a load balancing processing apparatus according to the present invention. As shown in FIG. 6, the load balancing processing apparatus provided in this embodiment, based on the embodiment shown in FIG. 5, further includes The update module 46 is configured to update a mapping relationship table between the first processing node and the at least one virtual node, where the sending module 44 is further configured to send to other processing nodes in the distributed processing system. The information is updated such that other processing nodes in the distributed processing system update the mapping table between the locally stored processing node and the virtual node according to the update information.

The device in this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 3, and the implementation principle and the technical effect are similar, and details are not described herein again.

FIG. 7 is a schematic structural diagram of Embodiment 4 of a load balancing processing apparatus according to the present invention. As shown in FIG. 7, the load balancing processing apparatus provided in this embodiment, based on the embodiment shown in FIG. 6, further includes The synchronization module 47, the synchronization unit 47 is configured to synchronize data corresponding to the at least one virtual node to the at least one processing node.

Optionally, the load balancing processing apparatus provided by the embodiment of the present invention may be a processing node or a master control center.

Specifically, the case where the load balancing processing device is a processing node may include: any processing node in the distributed processing system may perform load balancing processing on the distributed processing system or may be performed by a primary processing node in the distributed processing system. The distributed processing system performs load balancing processing.

For the case where the load balancing processing device provided by the embodiment of the present invention is the main control center, a main control center may be set in the distributed processing system for performing load balancing processing on the distributed processing system.

FIG. 8 is a schematic structural diagram of Embodiment 1 of a processing node according to the present invention. As shown in FIG. 8, the processing node provided in this embodiment includes: at least one processor 801, at least one network interface 804 or other user interface 803, and a memory 805. And at least one communication bus 802. The load balancing processing device optionally includes a user interface 803 including a display, a keyboard or a pointing device. The memory 805 may include a high speed RAM memory and may also include a non-volatile memory such as at least one disk memory. The memory 805 can optionally include at least one storage device located remotely from the load balancing processing device. Save The memory 805 stores execution instructions. When the load balancing processing device is in operation, the processor 801 communicates with the memory 805, and the processor 801 executes the execution instructions so that the load balancing processing device can execute the above-described method embodiments. Operating system 806, which includes various programs for implementing various basic services and processing hardware-based tasks.

The processing node provided by the embodiment of the present invention may perform the technical solution of the embodiment of the load balancing processing method, and the implementation principle and the technical effect thereof are similar, and details are not described herein again. FIG. 9 is a schematic structural diagram of Embodiment 1 of a main control center according to the present invention. As shown in FIG. 9, the main control center provided in this embodiment includes: at least one processor 901, at least one network interface 904, or another user interface 903. A memory 905, and at least one communication bus 902. The load balancing processing device optionally includes a user interface 903 including a display, a keyboard or a pointing device. Memory 905 may contain high speed RAM memory and may also include non-volatile memory, such as at least one disk memory. The memory 905 can optionally include at least one storage device located remotely from the load balancing processing device. The memory 905 stores execution instructions. When the load balancing processing device operates, the processor 901 communicates with the memory 905, and the processor 901 executes the execution instructions so that the load balancing processing device can execute the above-described method embodiments. Operating system 906, which contains various programs for implementing various basic services and handling hardware-based tasks.

The main control center provided by the embodiment of the present invention can perform the technical solution of the embodiment of the load balancing processing method, and the implementation principle and the technical effect are similar, and details are not described herein again.

FIG. 10 is a schematic structural diagram of Embodiment 1 of a distributed processing system according to the present invention, as shown in FIG.

10, the distributed processing system of this embodiment includes at least two processing nodes 1000, and the processing node 1000 can adopt the structure of any device embodiment of FIG. 4 to FIG. 8, correspondingly, any one of the processing nodes 1000 can execute The technical solution of the embodiment of the load balancing processing method is similar to the technical solution, and details are not described herein again.

FIG. 11 is a schematic structural diagram of Embodiment 2 of a distributed processing system according to the present invention, as shown in FIG.

As shown in FIG. 1 , the distributed processing system of the present embodiment includes at least two processing nodes 1010 and a main control center 1020. The processing node 1010 can adopt the structure of the processing node in the prior art, and the main control center 1020 can adopt FIG. The configuration of the device embodiment of FIG. 8 is correspondingly configured to implement the technical solution of the embodiment of the load balancing processing method. The implementation principle and technical effects are similar, and details are not described herein again. The embodiment of the present invention further provides a distributed processing system, including at least two processing nodes in the embodiment shown in FIG.

The embodiment of the present invention further provides a distributed processing system, which includes two processing nodes and a main control center in the embodiment shown in FIG.

The embodiment of the invention further provides a computer readable medium, comprising computer execution instructions, wherein the computer execution instructions are used to enable any one of the processing nodes in the above embodiments to execute the load balancing processing method.

The embodiment of the present invention further provides a computer readable medium, comprising computer execution instructions, wherein the computer execution instructions may be used to enable any one of the above-mentioned embodiments to execute the load balancing processing method.

It will be understood by those skilled in the art that all or part of the steps of implementing the above method embodiments may be performed by hardware related to program instructions. The foregoing program may be stored in a computer readable medium, and when executed, the program includes the above The steps of the method embodiments; and the aforementioned computer readable medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

It should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

Claim

A load balancing processing method, comprising:

2. The method according to claim 1, wherein the migrating at least one virtual node distributed on the first processing node to the distributed processing system is less than at least one load of the average load On the processing node, including:

The method according to any one of claims 1 or 2, wherein the migration of at least one virtual node distributed on the first processing node to the distributed processing system is less than Before the at least one processing node of the average load, the method further includes: acquiring load information of each processing node in the distributed processing system, where the load information includes load information of each virtual node distributed on each processing node;

And if the migration request response message is a success response message, migrating the at least one virtual node distributed on the first processing node to the at least one processing node.

The method according to any one of claims 1 or 2, wherein The at least one virtual node distributed on the first processing node is migrated to the at least one processing node in the distributed processing system where the load is less than the average load, including:

The update information is sent to other processing nodes in the distributed processing system such that other processing nodes in the distributed processing system update the mapping table between the locally stored processing node and the virtual node based on the update information.

The method according to any one of claims 1 or 2, wherein the migration of at least one virtual node distributed on the first processing node to the distributed processing system is less than After the at least one processing node of the average load, the method further includes: synchronizing data corresponding to the at least one virtual node to the at least one processing node.

6. A load balancing processing device, comprising:

The device according to claim 6, wherein the processing module is configured to migrate at least one virtual node distributed on the first processing node to a total amount of load less than or equal to an overload amount to the distribution In the processing system, the total amount of underload is greater than or equal to the total load of the at least one processing node, the overload amount is the load of the first load exceeding the average load, and the total amount of the underload is the The load of each of the at least one processing node is lower than the sum of the load of the average load.

The device according to any one of claims 6 or 7, further comprising a determining module, a sending module and a receiving module;

The determining module is configured to determine, according to load information of each processing node, the One less processing node;

The processing module is specifically configured to: if the migration request response message is a success response message, migrate at least one virtual node distributed on the first processing node to the at least one processing node.

The apparatus according to any one of claims 6 or 7, further comprising:

10. Apparatus according to any one of claims 6 or 7 further comprising:

And a synchronization module, configured to synchronize the at least one virtual node corresponding data to the at least one processing node.

The device according to any one of claims 6 or 7, wherein the device is a processing node or a master control center.

12. A processing node, comprising: a processor and a memory, the memory storing execution instructions, when the device is in operation, the processor is in communication with the memory, the processor executing the Executing the instructions causes the apparatus to perform the method of claims 1 to 5.

13. A master control center, comprising: a processor and a memory, wherein the memory stores execution instructions, when the device is in operation, the processor communicates with the memory, and the processor executes Executing the instructions causes the apparatus to perform the method of claims 1 to 5.

14. A distributed processing system, comprising: at least two Find the processing node described in 11.

A distributed processing system, comprising: at least two processing nodes and a master control center according to claim 11.

16. A distributed processing system, comprising at least two processing nodes as claimed in claim 12.

A distributed processing system, comprising: at least two processing nodes and a master control center according to claim 13.

18. A computer readable medium, comprising computer executed instructions for causing a processing node to perform the method of any one of claims 1 to 5.

A computer readable medium, comprising: computer-executable instructions for causing a master control center to perform the method of any one of claims 1 to 5.