CN107317764A - The method and system of flow load balance - Google Patents
The method and system of flow load balance Download PDFInfo
- Publication number
- CN107317764A CN107317764A CN201610264986.2A CN201610264986A CN107317764A CN 107317764 A CN107317764 A CN 107317764A CN 201610264986 A CN201610264986 A CN 201610264986A CN 107317764 A CN107317764 A CN 107317764A
- Authority
- CN
- China
- Prior art keywords
- flow
- node
- client
- difference
- minimum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
Abstract
There is provided a kind of method that flow load balance is carried out in the network including multiple nodes, wherein methods described can include:Collect the flow of each node;Calculate the difference L between the maximum and minimum value in the flow;Determine whether the difference L is more than or equal to predetermined amount of flow trimming threshold M;And in response to determining the difference L more than or equal to predetermined amount of flow trimming threshold M, one or more clients in each client that the node with maximum flow is connected are switched to the node minimum with flow and are connected.
Description
Technical field
The application is related to computer realm, in particular it relates to a kind of method of flow load balance
And system, by the flow load situation of each node of dynamic access cluster, according to real-time dynamic
The actual flow loading condition got carries out dynamic dispatching.
Background technology
With the development of Internet technology, the data of unified log system are more and more huger.Network
System generally requires the forwarding that retransmission center carries out data volume according to the subscription information of user.Should
Retransmission center will carry all data traffics, and separate unit retransmission center can not often meet reception and turn
The data volume for sending out huge, accordingly, it would be desirable to which many retransmission centers are operated.
Although many retransmission centers can solve the above problems, however many retransmission centers face why
The problem of sample ensures flow load balance.If flow load is unbalanced, many retransmission centers
In a certain retransmission center can equally run into and can not meet reception and forward huge data volume
Problem.
Therefore, the various methods of flow load balance are have developed, for example, at random, poll etc..
However, because the flow dynamics of each client transmissions to retransmission center change, each client
Flow it is different, and gap may much times, so according to client between different clients
The quantity at end is polled and is at random not all right certainly.In addition, what is solved the above problems is another
Method is selected according to the flow of retransmission center.This scheme can solve the problem that at some
The distribution flow of moment static state, and the purpose of dynamic implement flow load balance can not be realized.
Under the scene that long connection is transmitted, this method can cause the service of rear end actual treatment flow
Device is very unbalanced, causes part of server very idle, and another part server is very busy,
Possibly even due to handling huge flow and machine of delaying.This equalization scheme is typically only suitable for each
Client traffic all very situations in a balanced way, are not suitable in the field that flow is big and flow often changes
Scape.
Prior art flow equalization scheme does not reach substantially for the scene that flow dynamics change
The efficiency of flow equalization, the quantity that can only realize connection quantity or client is in a balanced way, no
It can guarantee that real flow equalization.
In order to solve the problem of dynamic flow is balanced, present applicant proposes a kind of flow load balance
Method and system, by the flow load situation of each node of dynamic access cluster, according to reality
When the actual flow loading condition that arrives of dynamic access carry out dynamic dispatching.
The content of the invention
It is to solve at least the above and/or shortcoming in terms of the disclosure and provides at least following
Advantage.
One aspect of the present invention is related to a kind of carries out flow in the network including multiple nodes
The method of load balancing, wherein methods described can include:Collect the flow of each node;Meter
Calculate the difference L between the maximum and minimum value in the flow;Whether determine the difference L
More than or equal to predetermined amount of flow trimming threshold M;And in response to determine the difference L be more than or
Equal to predetermined amount of flow trimming threshold M, in each client that the node with maximum flow is connected
One or more clients are switched to the node minimum with flow and are connected.
Preferably, methods described can also include:In response to determining that it is predetermined that the difference L is less than
Flow equalization threshold value M, continues to collect the flow of each node.
Preferably, one in each client that the node by with maximum flow is connected or more
Multiple client is switched to the node minimum with flow and is connected and can include:Collect and maximum flow
The connected each client of node client traffic;And selected according to the client traffic of collection
Select one or more clients and be connected with being switched to the node minimum with flow.
Preferably, it is described according to the client traffic of collection select one or more clients with
The node minimum with flow is switched to be connected and can include:From big to small to the visitor of each client
The flow sequence of family end, and extract the client traffic Y of i-th of clienti, i initial value is 1;
By the client traffic Y of i-th of clientiIt is compared with difference L;In response to client stream
Measure YiLess than or equal to difference L, i-th of client is switched to the node phase minimum with flow
Even;Calculating difference L and client traffic YiTwo times between difference, and result of calculation is made
For new difference L ';Determine whether the new difference L ' is more than or equal to predetermined amount of flow balanced
Threshold value M;And in response to determining that it is balanced that the new difference L ' is more than or equal to predetermined amount of flow
Threshold value M, judges whether to have traveled through the whole clients being connected with the node with peak flow values
Client traffic, wherein in response to determine without the client for having traveled through whole clients
Flow, makes i=i+1, and returns to the extraction operation, until having traveled through all clients flow
Untill.
Preferably, methods described can include:In response to determining that the new difference L ' is less than
Predetermined amount of flow trimming threshold M, terminate one or more clients of the selection be switched to
The connected operation of the minimum node of flow.
Preferably, methods described can include:Carried out when by client traffic Yi and difference L
When comparing, in response to client traffic YiMore than difference L, skip to and described judge whether to have traveled through
The step of client traffic for the whole clients being connected with the node with peak flow values.
Preferably, the flow value with regard to each node in collection network is spaced at predetermined time intervals.
Flowed according to another aspect of the present disclosure there is provided a kind of in the network based on node
The system for measuring load balancing, wherein the system can include:Retransmission center cluster, by two
Or more node constitute;Multiple client, be configured to respectively with described two or more sections
One of point is connected and communicated;Flow load balance scheduler, is configured to:From the forwarding
Center cluster collects the flow value of each node;Calculate the maximum and minimum value in the flow
Between difference L;Determine whether the difference L is more than or equal to predetermined amount of flow trimming threshold M;
And in response to determining that the difference L is more than or equal to predetermined amount of flow trimming threshold M, will be described
One or more clients in each client being connected with the node of maximum flow be switched to
The minimum node of flow is connected.
Preferably, the flow load balance scheduler is also configured as:In response to determining institute
Difference L is stated less than predetermined amount of flow trimming threshold M, continues to collect the flow of each node.
Preferably, the flow load balance scheduler is also configured as:Collect with flow most
The client traffic of the connected each client of big node;And according to the client traffic of collection
Select one or more clients and be connected with being switched to the node minimum with flow.
Brief description of the drawings
Below in conjunction with accompanying drawing, will become apparent from the example embodiment of the disclosure above and other aspect,
In feature and advantage, accompanying drawing:
Fig. 1 shows the signal of the system of the flow load balance according to illustrated embodiments of the invention
Figure;
Fig. 2 shows the flow of the method for the flow load balance according to illustrated embodiments of the invention
Figure;And
Fig. 3 shows the method for the flow load balance according to another example embodiment of the invention
Flow chart.
Embodiment
Fig. 1 shows the signal of the system of the flow load balance according to illustrated embodiments of the invention
Figure.
In Fig. 1, it can be wrapped according to the system of the flow load balance of illustrated embodiments of the invention
Include retransmission center cluster 100, multiple client 200 and flow load balance scheduler 300.Tool
Body, the retransmission center cluster 100 can include one or more retransmission center nodes
101-103, the forwarding of data is carried out for forwarding.Multiple client 200-1,200-2,200-3
Etc. can be with one of the retransmission center node in retransmission center cluster 100 (101,102 or 103)
It is connected and carries out data transmission therewith.Flow load balance scheduler 300 be used for dynamic timing or
The flow of all nodes is obtained with configurable time interval, it is possible to as needed come obtain with
The client traffic of each connected client of a certain node.Flow load balance scheduler 300
According to the flow of each node and/or the situation of the client traffic of individual client, scheduling strategy is formulated,
And the dispatch command that scheduling strategy is formed is handed down to the client for needing to dispatch, and allows client weight
Newly the forwarding interior joint specified is selected to carry out data transmission.
Described hereinafter with reference to Fig. 2 according to the flow load balances of illustrated embodiments of the invention
The schematic diagram of method.Specifically, in step S201, flow load balance scheduler 300 is collected
The flow of each node.These nodes receive the number that the client link being attached thereto is transmitted
According to the wherein connection between these nodes and client can be long connection.For example, in network
Be deployed with it is N number of need the node of flow load balance, flow load balance scheduler 300 is real-time
Collect the flow of N number of node in ground.
Then, in step S202, the difference between the maximum and minimum value in the flow is calculated
Value L.There are a variety of modes for being used to calculate the difference L, for example, flow load balance is adjusted
Degree device can be the N number of flow being collected into according to being ranked up from small to large, then first just
It is the flow corresponding with the node that flow is minimum, last is exactly the node with maximum flow
Corresponding flow, so as to calculate the difference L.It should be noted that according to from small to large
Order be only an example, can also be ranked up according to order flow from big to small.Or
Person can not be ranked up to the N number of flow being collected into, but by seeking maximum and minimum
The algorithm of value calculates difference L therebetween.
Then, in step S203, determine whether the difference L is more than or equal to predetermined amount of flow equal
Weigh threshold value M, and wherein M can be by user preset or system default.When difference L is big
When predetermined amount of flow trimming threshold M, illustrate that the flow load between node is more unbalanced,
Need to carry out flow load balance processing.On the contrary, when difference L is less than predetermined amount of flow trimming threshold
During M, illustrate that the flow load between node is more balanced, it is not necessary to carry out flow load balance
Processing.
In response to determining that the difference L is more than or equal to predetermined amount of flow trimming threshold M, i.e. need
Flow load balance is carried out, in step S204, each visitor that the node with maximum flow is connected
One or more clients in the end of family are switched to the node minimum with flow and are connected.Specifically,
In the disclosure, client is switched to and another node phase from the state being connected with a node
The operation of state even can be realized by following operation:Load balance scheduler is to correspondingly
Client sends dispatch command, and the dispatch command makes the client link to the forwarding specified
Centroid.Client terminates the biography to current data after the dispatch command is received
Defeated and disconnection is linked with original retransmission center node, is then linked to flow load balance scheduling
The retransmission center node that device 300 is specified.If in addition, difference L is less than the balanced threshold of predetermined amount of flow
Value M, i.e. need not now carry out flow load balance, then flow load balance scheduler 300
The flow of each node can be continued to collect.
The foregoing describe the stream of the method for flow load balance according to illustrated embodiments of the invention
Cheng Tu.Described below in conjunction with specific implementation according to another example embodiment of the invention
The flow chart of the method for flow load balance.
Fig. 3 shows the method for the flow load balance according to another example embodiment of the invention
Flow chart.Step S301-S303 in Fig. 3 is similar to the step S201-S203 in Fig. 2,
Therefore will no longer it be repeated.
It is different with Fig. 2, in figure 3, in response to determining that difference L is more than or equal to
Predetermined amount of flow trimming threshold M, i.e. need to carry out flow load balance, in step 304, stream
Amount load balance scheduler 300 can collect each client for being connected with the node of maximum flow
Client traffic, and select one or more clients to cut according to the client traffic of collection
The node minimum with flow is changed to be connected.Specifically, in step 305, from big to small to each
The client traffic sequence of client, and extract the client traffic Y of i-th of clienti.Should
Note, while characterized as by client traffic YiIt is ranked up from big to small, but this
Invention can also be with other sortords (for example, from small to large, or randomly).In addition,
If be ranked up in step S302, sequence can be omitted herein, or based in step
The sequence that S302 is carried out performs subsequent operation.Compared to other sortords, entering from big to small
Row sequence can ensure to switch less client at faster speed, so as to reach stream compared with block
Measure load balancing.In the case of to client traffic being ranked up from big to small, i just
Initial value is preferably equal to 1.
In step S306, by the client traffic Y of i-th of clientiCompared with difference L
Compared with.In response to client traffic YiMore than difference L, then step S311 is skipped to, it is determined whether
The client traffic Y for the whole clients being connected with the node of maximum flow is traveled throughi.If it is,
Then terminate the flow equalization processing.If not traveled through what is be connected with the node of maximum flow
The client traffic Y of whole clientsi, then i=i+1 is made in S312 so that extract next client
The client traffic Y at endi+1.If however, determining client traffic Y in step 306iIt is less than
Or equal to difference L, then flow load balance adjuster 300 can send traffic order, so as to
Make i-th of client be switched to the node minimum with flow to be connected, i.e. by i-th of client
The state being connected from the node with maximum flow is switched to the state that the node minimum with flow is connected.
Then, in step S308, calculating difference L and client traffic YiTwo times between difference,
Flow difference i.e., now between above-mentioned two node, and using result of calculation as new difference L ',
Wherein L '=L-2 × Yi.In step S309, determine the new difference L ' whether be more than or
Equal to predetermined amount of flow trimming threshold M.It is determined that the new difference L ' is more than or equal to pre- constant current
In the case of measuring trimming threshold M, i.e. be switched to and flow minimum by i-th of client
After node is connected, the flow load between node is still unbalanced, still needs to be flowed
Amount is balanced.Now in step S310, difference L is assigned by new difference L ', and in step S311
Judge whether to have traveled through the client for the whole clients being connected with the node with peak flow values
Flow.In response to determining without the client traffic (S311- for having traveled through whole clients
It is no), make i=i+1 in step S312, and return to and described extract operation and (return to step
S305), new client traffic Y is comparedi+1With the difference L of new assignment, until having traveled through
Untill having client traffic.If instead it is determined that the fixed new difference L ' is less than predetermined amount of flow
Trimming threshold M (S311- is), i.e. be switched to and flow minimum by i-th of client
Node reaches flow equalization after being connected, then terminate one or more clients of the selection with
It is switched to the operation that the node minimum with flow is connected.After end operation, it will finally obtain
To the result that maximum flow node and flow minimum node difference are minimum, that is, their two streams
Measure immediate state, i.e. realize flow equalization.It should be noted that can be at predetermined time intervals
Repeat the method according to illustrated embodiments of the invention.
Obtained according to the flow load balance method and system of disclosure example embodiment by dynamic
The flow load situation of each node of cluster is taken, come the actual flow arrived according to real-time dynamic access
Loading condition carries out dynamic dispatching.For example, methods described can apply in unified log system,
The flow of each retransmission center is allowed substantially to be at equilibrium-like at any time by using this scheme
State, improve the server utilization of each retransmission center node, and due to good equal
Weigh effect, it is ensured that the stability of total system.In addition, methods described can also be applied to
During distributed message middleware transmission message, it can be carried out by using this scheme
Dynamic changes the subscribing relationship of message, it is ensured that the flow dynamics of each node of message-oriented middleware are balanced.
In addition to the applications described above, method according to the present disclosure can also be applied to carry out data biography
In other defeated network communicating systems, it is particularly suitable for application to carry out data transmission using long connection
Communication system, so as to solve the unbalanced problem of flow load, carry out dynamic flow in real time
It is balanced.
The program command form that the above method be able to can be held by a variety of computer installations is realized
And record in computer readable recording medium storing program for performing.In this case, computer-readable record is situated between
Matter can include single program command, data file, data structure or its combination.Meanwhile,
The program command of record in the recording medium specially can design or be configured to the disclosure, or
Technical staff's known applications of computer software fields.Computer readable recording medium storing program for performing includes example
Such as hard disk, floppy disk or tape magnetic medium, such as compact disk read-only storage (CD-ROM)
Or optical medium, the magnet-optical medium of such as floptical disk and the example such as digital universal disc (DVD)
Such as ROM, RAM, the flash memory hardware unit of storage and configuration processor order.In addition, journey
The machine language code and computer that sequence order includes compiler formation can by using interpretive program
The high-level language of execution.Hardware unit above can be configured to be used as at least one software module
Operate to perform the operation of the disclosure, and contrary operation is also the same.
Although the operation of context of methods has shown and described with particular order, but it can change every
The order of the operation of individual method so that specific operation can be performed with reverse order or is allowed to
At least in part with other operations while to perform specific operation.In addition, the disclosure is not limited to
State example embodiment, it can on the premise of spirit and scope of the present disclosure are not departed from, including
One or more miscellaneous parts or operation, or omit one or more miscellaneous parts or operation.
Combined preferred embodiment of the present disclosure shows the disclosure above, but this area
Technical staff will be understood that, in the case where not departing from spirit and scope of the present disclosure, can be right
The disclosure carries out various modifications, replaces and change.Therefore, the disclosure should not be by above-described embodiment
To limit, and it should be limited by appended claims and its equivalent.
Claims (10)
1. a kind of method that flow load balance is carried out in the network including multiple nodes, its
Described in method include:
Collect the flow of each node;
Calculate the difference L between the maximum and minimum value in the flow;
Determine whether the difference L is more than or equal to predetermined amount of flow trimming threshold M;And
, will be with stream in response to determining that the difference L is more than or equal to predetermined amount of flow trimming threshold M
One or more clients in the connected each client of the maximum node of amount are switched to and flow
Minimum node is connected.
2. according to the method described in claim 1, wherein also including:It is described in response to determining
Difference L is less than predetermined amount of flow trimming threshold M, continues to collect the flow of each node.
3. according to the method described in claim 1, wherein the node with maximum flow is connected
Each client in one or more clients be switched to the node minimum with flow and be connected bag
Include:
Collect the client traffic for each client being connected with the node of maximum flow;And
One or more clients are selected according to the client traffic of collection to be switched to and flow
Minimum node is connected.
4. method according to claim 3, wherein being selected according to the client traffic of collection
Select one or more clients be switched to the node minimum with flow be connected including:
The client traffic to each client sorts from big to small, and extracts i-th of client
Client traffic Yi, i initial value is 1;
By the client traffic Y of i-th of clientiIt is compared with difference L;
In response to client traffic YiLess than or equal to difference L, i-th of client is switched to
The node minimum with flow is connected;
Calculating difference L and client traffic YiTwo times between difference, and by result of calculation
It is used as new difference L ';
Determine whether the new difference L ' is more than or equal to predetermined amount of flow trimming threshold M;And
In response to determining that the new difference L ' is more than or equal to predetermined amount of flow trimming threshold M, sentence
The disconnected client stream for whether having traveled through the whole clients being connected with the node with peak flow values
Amount, wherein in response to determining without the client traffic for having traveled through whole clients, order
I=i+1, and the extraction operation is returned to, untill having traveled through all clients flow.
5. method according to claim 4, wherein methods described include:In response to true
The fixed new difference L ' is less than predetermined amount of flow trimming threshold M, terminates the selection one or more
Multiple client is to be switched to the operation that the node minimum with flow is connected.
6. method according to claim 4, wherein when by client traffic YiWith difference
When L is compared, in response to client traffic YiMore than difference L, described judge whether is skipped to
The step of the client traffic for the whole clients being connected with the node with peak flow values is traveled through
Suddenly.
7. according to the method described in claim 1, wherein interval is just collected at predetermined time intervals
The flow value of each node in network.
8. a kind of system that flow load balance is carried out in the network based on node, wherein institute
The system of stating includes:
Retransmission center cluster, is made up of two or more nodes;
Multiple client, is configured to one of described two or more nodes be connected respectively and goes forward side by side
Row communication,
Flow load balance scheduler, is configured to:
The flow value of each node is collected from the retransmission center cluster;
Calculate the difference L between the maximum and minimum value in the flow;
Determine whether the difference L is more than or equal to predetermined amount of flow trimming threshold M;With
And
In response to determining that the difference L is more than or equal to predetermined amount of flow trimming threshold M,
By one or more visitors in each client being connected with the node of maximum flow
Family end is switched to the node minimum with flow and is connected.
9. system according to claim 8, wherein the flow load balance scheduler
It is additionally configured to:In response to determining that the difference L is less than predetermined amount of flow trimming threshold M, continue to receive
Collect the flow of each node.
10. system according to claim 8, wherein the flow load balance scheduler
It is additionally configured to:
Collect the client traffic for each client being connected with the node of maximum flow;And
One or more clients are selected according to the client traffic of collection to be switched to and flow
Minimum node is connected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610264986.2A CN107317764B (en) | 2016-04-26 | 2016-04-26 | Traffic load balancing method, system, device and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610264986.2A CN107317764B (en) | 2016-04-26 | 2016-04-26 | Traffic load balancing method, system, device and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107317764A true CN107317764A (en) | 2017-11-03 |
CN107317764B CN107317764B (en) | 2021-12-14 |
Family
ID=60184863
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610264986.2A Active CN107317764B (en) | 2016-04-26 | 2016-04-26 | Traffic load balancing method, system, device and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107317764B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107835131A (en) * | 2017-12-15 | 2018-03-23 | 北京星河星云信息技术有限公司 | Network traffics dispatching method, device and storage medium |
CN108173689A (en) * | 2017-12-27 | 2018-06-15 | 杭州迪普科技股份有限公司 | The output system of load balancing data |
CN109981419A (en) * | 2019-04-11 | 2019-07-05 | 苏州浪潮智能科技有限公司 | Test method, device, system, equipment and the storage medium of load balancing characteristic |
CN110515729A (en) * | 2019-08-19 | 2019-11-29 | 中国人民解放军国防科技大学 | Graph computing node vector load balancing method and device based on graph processor |
CN110532091A (en) * | 2019-08-19 | 2019-12-03 | 中国人民解放军国防科技大学 | Graph computation edge vector load balancing method and device based on graph processor |
WO2019232680A1 (en) * | 2018-06-04 | 2019-12-12 | Beijing Thomson Commerce Co., Ltd. | Method and device for providing load balancing |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1111863A (en) * | 1993-06-02 | 1995-11-15 | 三星电子株式会社 | Alogorithmic controlling method of blind balancing system |
CN101772078A (en) * | 2008-12-31 | 2010-07-07 | 大唐移动通信设备有限公司 | Load control method and device |
CN101965012A (en) * | 2009-07-22 | 2011-02-02 | 中兴通讯股份有限公司 | Load balancing method and device |
CN102111813A (en) * | 2010-03-22 | 2011-06-29 | 电信科学技术研究院 | Method and equipment for balancing loads of carriers |
CN102149136A (en) * | 2010-02-10 | 2011-08-10 | 杭州华三通信技术有限公司 | Method and system for switching channel of access point and access point |
CN103368862A (en) * | 2013-06-21 | 2013-10-23 | 北京京东尚科信息技术有限公司 | Load balance dispatching method and load balance dispatching device |
CN103428100A (en) * | 2013-08-20 | 2013-12-04 | 北京京东尚科信息技术有限公司 | Dynamic flow control method and device based on data classification |
CN104023083A (en) * | 2014-06-23 | 2014-09-03 | 广东睿江科技有限公司 | Method and device for log collection cluster load balance |
CN103118136B (en) * | 2013-02-28 | 2015-08-12 | 杭州电子科技大学 | A kind of method based on block rank cloud memory load balance optimizing |
-
2016
- 2016-04-26 CN CN201610264986.2A patent/CN107317764B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1111863A (en) * | 1993-06-02 | 1995-11-15 | 三星电子株式会社 | Alogorithmic controlling method of blind balancing system |
CN101772078A (en) * | 2008-12-31 | 2010-07-07 | 大唐移动通信设备有限公司 | Load control method and device |
CN101965012A (en) * | 2009-07-22 | 2011-02-02 | 中兴通讯股份有限公司 | Load balancing method and device |
CN102149136A (en) * | 2010-02-10 | 2011-08-10 | 杭州华三通信技术有限公司 | Method and system for switching channel of access point and access point |
CN102111813A (en) * | 2010-03-22 | 2011-06-29 | 电信科学技术研究院 | Method and equipment for balancing loads of carriers |
CN103118136B (en) * | 2013-02-28 | 2015-08-12 | 杭州电子科技大学 | A kind of method based on block rank cloud memory load balance optimizing |
CN103368862A (en) * | 2013-06-21 | 2013-10-23 | 北京京东尚科信息技术有限公司 | Load balance dispatching method and load balance dispatching device |
CN103428100A (en) * | 2013-08-20 | 2013-12-04 | 北京京东尚科信息技术有限公司 | Dynamic flow control method and device based on data classification |
CN104023083A (en) * | 2014-06-23 | 2014-09-03 | 广东睿江科技有限公司 | Method and device for log collection cluster load balance |
Non-Patent Citations (2)
Title |
---|
SHANGZHOU LI 等: "《Load-modifiable content-based Publish/Subscribe Architecture over structured peer-to-peer networks》", 《IEEE》 * |
邬江兴 等: "《新型网络体系结构》", 31 March 2014 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107835131A (en) * | 2017-12-15 | 2018-03-23 | 北京星河星云信息技术有限公司 | Network traffics dispatching method, device and storage medium |
CN108173689A (en) * | 2017-12-27 | 2018-06-15 | 杭州迪普科技股份有限公司 | The output system of load balancing data |
CN108173689B (en) * | 2017-12-27 | 2021-07-23 | 杭州迪普科技股份有限公司 | Output system of load balancing data |
WO2019232680A1 (en) * | 2018-06-04 | 2019-12-12 | Beijing Thomson Commerce Co., Ltd. | Method and device for providing load balancing |
CN109981419A (en) * | 2019-04-11 | 2019-07-05 | 苏州浪潮智能科技有限公司 | Test method, device, system, equipment and the storage medium of load balancing characteristic |
CN110515729A (en) * | 2019-08-19 | 2019-11-29 | 中国人民解放军国防科技大学 | Graph computing node vector load balancing method and device based on graph processor |
CN110532091A (en) * | 2019-08-19 | 2019-12-03 | 中国人民解放军国防科技大学 | Graph computation edge vector load balancing method and device based on graph processor |
CN110532091B (en) * | 2019-08-19 | 2022-02-22 | 中国人民解放军国防科技大学 | Graph computation edge vector load balancing method and device based on graph processor |
CN110515729B (en) * | 2019-08-19 | 2022-05-24 | 中国人民解放军国防科技大学 | Graph computing node vector load balancing method and device based on graph processor |
Also Published As
Publication number | Publication date |
---|---|
CN107317764B (en) | 2021-12-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107317764A (en) | The method and system of flow load balance | |
CN108009028B (en) | Message processing method, device, equipment and computer readable storage medium | |
CN106549878B (en) | Service distribution method and device | |
CN109756559B (en) | Construction and use method for distributed data distribution service of embedded airborne system | |
CN101605092A (en) | A kind of content-based SiteServer LBS | |
CN106453122B (en) | A kind of choosing method and device of Stream Data Transmission node | |
CN105516347A (en) | Method and device for load balance allocation of streaming media server | |
CN109729017B (en) | Load balancing method based on DPI prediction | |
CN109729602B (en) | Link data processing method and device and computer storage medium | |
CN109413202B (en) | System and method for sorting block chain transaction information | |
CN103873523A (en) | Client cluster access method and device | |
CN106357776A (en) | Method and device for selecting transmission resource of streaming data | |
CN110198332A (en) | Dispatching method, device and the storage medium of content delivery network node | |
CN115277598A (en) | Computing resource scheduling method and device and computer readable storage medium | |
CN109495526A (en) | A kind of file transmitting method, device, system, electronic equipment and storage medium | |
WO2010101038A1 (en) | Destination node selection method, node, and recording medium | |
CN105282045B (en) | A kind of distributed computing and storage method based on consistency hash algorithm | |
CN108063814B (en) | Load balancing method and device | |
US9621656B2 (en) | Distributed deployment device and method | |
CN105516271A (en) | Business processing system, business processing method and business processing apparatus | |
CN103096380B (en) | Wireless access point load balancing load balancing | |
CN215298210U (en) | Multistage edge computing system of electric power thing networking | |
CN117459537B (en) | Multi-cloud-based device interaction method, device and readable storage medium | |
CN101482816B (en) | Middleware bridge system and method | |
CN113596079B (en) | Network system, back source disaster recovery method and equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |