CN105282259B - For the load balanced sharing method of backstage cluster service, agency and system - Google Patents
For the load balanced sharing method of backstage cluster service, agency and system Download PDFInfo
- Publication number
- CN105282259B CN105282259B CN201510780814.6A CN201510780814A CN105282259B CN 105282259 B CN105282259 B CN 105282259B CN 201510780814 A CN201510780814 A CN 201510780814A CN 105282259 B CN105282259 B CN 105282259B
- Authority
- CN
- China
- Prior art keywords
- server
- weights
- response time
- time interval
- service
- 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.)
- Active
Links
- 230000004044 response Effects 0.000 claims abstract description 86
- 230000000875 corresponding Effects 0.000 claims abstract description 23
- 230000000052 comparative effect Effects 0.000 claims description 18
- 230000004048 modification Effects 0.000 claims description 5
- 238000006011 modification reaction Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 description 6
- 230000003044 adaptive Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000034 method Methods 0.000 description 3
- 238000011084 recovery Methods 0.000 description 2
- 238000009966 trimming Methods 0.000 description 2
- 230000002950 deficient Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
Abstract
A kind of load balanced sharing method, agency and system for backstage cluster service, method include：S1, in a time interval, the weights that initial weight based on each server or step S2 are reset distribute connection request to corresponding server according to load-balancing algorithm；S2, after a time interval, according to the response time of each server and success rate in the time interval, recalculate and set the weights of each server, start next time interval and enter step S1.The present invention is using time interval as base unit, according to the response time of each server in the time interval and success rate, dynamic corrections are carried out to the weights of each server, response time and success rate have substantially reacted actual performance of each server under the influence of network, these factors of server environment, therefore, the present invention may be implemented to maximize the processing capacity using Servers-all under the variant environment of each server performance, and request is made to obtain optimal distribution.
Description
Technical field
The present invention relates to cluster service field more particularly to a kind of load balancing distribution sides for backstage cluster service
Method, agency and system.
Background technology
In internet mass service development task, the clustering design of background service and clustered deploy(ment) are software developers
The problem that must be solved.The uncertainty of otherness and network between server, to the service invariant of clustered deploy(ment)
Property produces very big influence.How the optimum allocation service between the larger service of server performance diversity ratio, how
It can be rejoined when request being oriented to normal environment and service and normal abnormal restoring when network or abnormal service
Distribution List is asked, becomes background service developer and needs the project solved.
General processing method is to call configuration Target IP and end to each network interface in the configuration file of service
Configuration is loaded into memory by mouth when servicing startup, when network or service exception and normal recovery, manual modification configuration file
Then restart service.Another method is configured by a configuration center, and business service is in real time from configuration center
Newest service profile information is obtained, when Network Abnormal and normal recovery, changes the configuration of configuration center.
The shortcomings that first method is to need modification configuration and the service of restarting when an anomaly occurs, this needs to interrupt service
To prevent operation system from services client, service request can only simply be loaded due to not having dynamic adjustment capability
The forwarding of balanced and service request.Second of method configured by configuration center, although clothes need not be restarted when abnormal
Business, but adaptively actual environment situation, maximization difficult to realize the processing capacity of server cannot be used, it need to when occurring extremely
The correct distribution wanted artificial interference just and can made requests on equally also can only carry out simple load balancing and service to service request
Request forwarding.
Therefore, the prior art is defective, needs to improve.
Invention content
The technical problem to be solved in the present invention is, for the drawbacks described above of the prior art, provides a kind of for collecting from the background
Load balanced sharing method, agency and the system of group's service.
The technical solution adopted by the present invention to solve the technical problems is：Construct a kind of load for backstage cluster service
Equilibrium assignment method, including：
S1, in a time interval, the weights that initial weight based on each server or step S2 are reset, root
Corresponding server is given according to load-balancing algorithm distribution connection request；
S2, after a time interval, according to the response time of each server in the time interval and successfully
Rate recalculates and sets the weights of each server, starts next time interval and enters step S1.
Of the present invention in the load balanced sharing method of backstage cluster service, the step S2 includes：
S21, after a time interval, obtain the sound of each server in the time interval of newest end respectively
Between seasonable and success rate；
S22, the response time based on each server calculate separately the average response time of each server, based on each
The sum of product and average response time of average response time is calculated in the average response time of server；
S23, it is based respectively on the response time factor that formula (1) determines each server：
S24, the weights that formula (2) calculates and sets each server are based respectively on：
W=P* β2*C (2)
If the weights of S25, some server are more than Wmax, then it is W by the weight setting of some servermax；If
The weights of some server are less than 1, then are 1 by the weight setting of some server；
S26, start next time interval and enter step S1；
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum generations
The sum of average response time in table step S22,Represent the average response time of the machine；W represents the weights of the machine, and P is represented
The success rate of the machine, C represent weights radix, are an integer set, WmaxHighest weights are represented, are one preset big
In 1 integer.
Of the present invention in the load balanced sharing method of backstage cluster service, the step S2 further includes：
Configuration information is read before entering each time interval, and according to configuration information that the number of clusters increase institute of cluster service is right
The weight setting for the server answered is 1, and the number of clusters of cluster service is reduced corresponding server in step sl from institute
It is directly deleted in some servers.
Of the present invention in the load balanced sharing method of backstage cluster service, WmaxIt is 100.
Of the present invention in the load balanced sharing method of backstage cluster service, described in the step S1
Include to corresponding server according to load-balancing algorithm distribution connection：
S11, for first time interval, the weights of each server are initial weight, each clothes in other times section
The weights of business device are the weights set in step S2；
S12, the greatest common divisor for determining Servers-all weights, set in Servers-all weights comparative factor to
Maximum weights；
The size of S13, the successively weights and comparative factor of more every server, if the weights of some server are more than
Or be equal to comparative factor, then the server is distributed into next connection request；
S14, when all server relatively after the completion of, by comparative factor using greatest common divisor as increments；If than
It is less than 0 compared with the factor, then comparative factor is re-set as maximum weights, go to step S13.
Of the present invention in the load balanced sharing method of backstage cluster service, initial weight to be set as 1.
The invention also discloses a kind of load balancing assignment agents for backstage cluster service, including
Load distribution module：For in a time interval, the initial weight based on each server or modified weight
The weights that module is reset distribute connection request to corresponding server according to load-balancing algorithm；
Modified weight module：After a time interval, according to the response of each server in the time interval
Time and success rate recalculate and set the weights of each server, start next time interval and trigger load distribution
Module works.
Of the present invention in the load balancing assignment agent of backstage cluster service, the modified weight module packet
It includes：
Response time and success rate acquiring unit：For after a time interval, being obtained respectively in newest end
Time interval in each server response time and success rate；
Response time factor calculating unit：Each server is calculated separately for the response time based on each server
Average response time, when product and the average response of average response time is calculated in the average response time based on each server
The sum of between, then determine based on formula (1) the response time factor of each server：
Weight calculation unit：The weights that each server is calculated and set for being based on formula (2), if some server
Weights be more than Wmax, then it is W by the weight setting of some servermax；It, will if the weights of some server are less than 1
The weight setting of some server is 1；Start next time interval and triggers load distribution module work；
W=P* β2*C (2)
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum generations
The sum of average response time in table step S22,Represent the average response time of the machine；W represents the weights of the machine, and P is represented
The success rate of the machine, C represent weights radix, are an integer set, WmaxHighest weights are represented, are one preset big
In 1 integer.
Of the present invention in the load balancing assignment agent of backstage cluster service, the modified weight module to be also
Including：
Modified weight subelement, for reading configuration information before entering each time interval, and according to configuration information
The weight setting that the number of clusters of cluster service is increased to corresponding server is 1, and the number of clusters of cluster service is reduced
Corresponding server is directly deleted in loading distribution module from all servers.
The invention also discloses a kind of load balancing distribution systems for backstage cluster service, including center configuration service
Module and at least one edge service service module；Wherein, the center configuration service module includes：Configuration database, load
Balanced center cell, configuration management element；The edge service service module includes：Load balancing as described above distributes generation
Reason, business unit, configuration file；
The configuration management element is used to carry out unified modification to configuration information and preserve into configuration database；It is described negative
The management that balanced center cell is used for load balancing data is carried, the configuration read request and interface of load balancing assignment agent are handled
Data statistics；The load balancing assignment agent is used to read configuration and reported data statistics, pipe from load balancing center cell
Simultaneously shared drive is written in reason local profile, is connection request distribution server；Business unit with load balancing by distributing
Agency's interaction obtains optimal destination server information, and service request is distributed to corresponding destination server.
Load balanced sharing method, agency and the system for backstage cluster service for implementing the present invention have with following
Beneficial effect：The present invention is using a time interval as base unit, according to the response time of each server in the time interval
And success rate, a dynamic corrections are carried out to the weights of each server, response time and success rate have substantially reacted each
Actual performance under the influence of these factors of network, server environment and the actual conditions of software service of server, therefore,
The present invention may be implemented to maximize under the variant environment of each server performance to the dynamic trimming of weights uses all clothes
The processing capacity of business device, and can make to ask according to the adaptive adjustment of network, server environment and the actual conditions of software service progress
It asks to obtain optimal distribution.
Description of the drawings
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing：
Fig. 1 is flow chart of the present invention for the load balanced sharing method of backstage cluster service；
Fig. 2 is the flow chart of step S2 in Fig. 1；
Fig. 3 is structural schematic diagram of the present invention for the load balancing assignment agent of backstage cluster service；
Fig. 4 is structural schematic diagram of the present invention for the load balancing distribution system of backstage cluster service.
Specific implementation mode
For a clearer understanding of the technical characteristics, objects and effects of the present invention, now control attached drawing is described in detail
The specific implementation mode of the present invention.
It is flow chart of the present invention for the load balanced sharing method of backstage cluster service with reference to figure 1.The side of the present invention
Method includes mainly：
S1, in a time interval, the weights that initial weight based on each server or step S2 are reset, root
Corresponding server is given according to load-balancing algorithm distribution connection request；
S2, after a time interval, according to the response time of each server in the time interval and successfully
Rate recalculates and sets the weights of each server, starts next time interval and enters step S1.
If it is initial start-up step S1, then load-balancing algorithm is to carry out load balancing distribution according to initial weight, it
Afterwards in executing step S1 from step S2 rebounds, load-balancing algorithm is to carry out load balancing according to the weights set in step S2
Distribution.
Wherein, time interval is arranged as required to, and is not limited to this.
It is understood that existing algorithm may be used in load-balancing algorithm, it is not limited herein.Such as background skill
It is namely based on the weights set in the two schemes of art introduction and carries out equilibrium assignment, only its weights can not be realized dynamic
State adjusts, so simple load balancing can only be carried out.
It is described in detail below in step S2 and specifically how to realize modified weight.
With reference to figure 2, the step S2 is specifically included：
S21, after a time interval, obtain the sound of each server in the time interval of newest end respectively
Between seasonable and success rate；
If the connection quantity of some server was k in last time interval, then the server to be obtained all
The connection corresponding response time：t1、t2…tk；
Success rate be then be calculated according to the number of this k times successful connection, if for example, the connection number of some server
Amount is k, and the number of successful connection is m times (m is less than k), then the success rate of the server is：
S22, the response time based on each server calculate separately the average response time of each serverK in the formula is the connection quantity of a specific server, and certainly, the connection quantity k of each server is
It is determined according to step S1.It is then based on the average response time of each serverThe product of average response time is calculatedThe sum of with average response timeN represents the sum of Servers-all；
S23, it is based respectively on the response time factor that formula (1) determines each server：
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum generations
The sum of average response time in table step S22, T represents the average response time of the machine；
S24, the weights that formula (2) calculates and sets each server are based respectively on：
W=P* β2*C (2)
If the weights of S25, some server are more than Wmax, such as 100, then it is by the weight setting of some server
Wmax；It is 1 by the weight setting of some server if the weights of some server are less than 1；
S26, start next time interval and enter step S1.
Wherein, W represents the weights of the machine, and P represents the success rate of the machine, and C represents weights radix, is set for one whole
Number, WmaxRepresent highest weights, for one it is preset be more than 1 integer.
The effect of weights radix C is so that the weights W that is finally calculated is a larger real number to facilitate follow-up load
The processing of equalization algorithm part.Because the success rate P and response time factor-beta of some server are typically all the number less than 1, institute
In order to enable last W is a larger real number, can uniformly be converted to it using weights radix C.
It is worth noting that, in step s 25, if the weights of some server are more than WmaxThen by some server
Weight setting be WmaxThe advantages of be：When connection request quantity is smaller, avoids and all connection requests are sent to weights most
In this big server, connection is distributed to multiple servers as possible.
In step s 25, it is 1 by the weight setting of some server if the weights of some server are less than 1
Advantage is：If abnormal, to calculate weights W meeting very littles occurs in some server, but the present invention is still still by it
Weights W is set as 1, this ensure that when step S1 carries out load distribution, still has and sounds out the server, so one
It restores normal to denier, then its response time and success rate can be promoted, its weights can be reset by eventually passing through step S2,
The server distribution connection for restoring abnormal can be thus given again.
In view of in the assignment procedure, in fact it could happen that artificial increase number of clusters deletes number of clusters, so excellent
Choosing, the step S2 further includes：Configuration information is read before entering each time interval, and according to configuration information by cluster
The weight setting that the number of clusters of service increases corresponding server is 1, corresponding to the number of clusters reduction by cluster service
Server directly deleted from all servers in step sl.
In addition, the Weighted Round Robin that the load-balancing algorithm in the present invention in step S1 uses, detailed process are as follows：
S11, for first time interval, the weights of each server are initial weight, each clothes in other times section
The weights of business device are the weights set in step S2；
S12, the greatest common divisor for determining Servers-all weights, set in Servers-all weights comparative factor to
Maximum weights；
S13, since First server, the size of the weights and comparative factor of more every server successively, if
The weights of some server are greater than or equal to comparative factor, then the server are distributed to next connection request；
S14, when all server relatively after the completion of, by comparative factor using greatest common divisor as increments；If than
It is less than 0 compared with the factor, then comparative factor is re-set as maximum weights, go to step S12.
If for example, there is 4 servers, cw indicates comparative factor.If after terminating first time interval, by step
The weights of 4 servers are reset to by rapid S2 successively：10,20,30,40, then when it enters second time interval, step
The process of rapid S1 is as follows：
Step S11 obtains 10,20,30,40 weights according to step S2, determines that greatest common divisor is 10.Step S12 settings
Comparative factor is 40.Subsequently into step S13, cw=40, since First server, the weights of First server are
10,10 ＜ cw=40, are unsatisfactory for requiring, then compare the weights of second server, are 20 ＜ cw=40, still not satisfy, then
Compare the weights of third platform server, be 30 ＜ cw=40, still not satisfy, then compare the weights of the 4th server, is 40
4th server is then distributed to next connection request by=cw, and so on, after the completion of all servers compare,
Comparative factor using greatest common divisor as increments, is switched to cw=30 by step S14, then skips back to step S13 again, again
Compare since First server, the weights of First server are 10,10 ＜ cw=30, are unsatisfactory for, second server
Weights be 20,20 ＜ cw=30, be unsatisfactory for, the weights of third platform server are 30,30>=cw=30 then takes third platform
Business device distributes to next connection request, and the weights for continuing the 4th server are 40,40>=cw=30 then takes the 4th
Business device distributes to next connection request, and so on, until the time interval of this bout terminates.
Correspondingly, the present invention also provides a kind of load balancing assignment agent for backstage cluster service, with reference to figure 3, bear
Carrying equilibrium assignment agency includes：
Load distribution module：For in a time interval, the initial weight based on each server or modified weight
The weights that module is reset distribute connection request to corresponding server according to load-balancing algorithm；
Modified weight module：After a time interval, according to the response of each server in the time interval
Time and success rate recalculate and set the weights of each server, start next time interval and trigger load distribution
Module works.
Wherein, the modified weight module includes：
Response time and success rate acquiring unit：For after a time interval, being obtained respectively in newest end
Time interval in each server response time and success rate；
Response time factor calculating unit：Each server is calculated separately for the response time based on each server
Average response time, when product and the average response of average response time is calculated in the average response time based on each server
The sum of between, then determine based on formula (1) the response time factor of each server：
Weight calculation unit：The weights that each server is calculated and set for being based on formula (2), if some server
Weights be more than Wmax, then it is W by the weight setting of some servermax；It, will if the weights of some server are less than 1
The weight setting of some server is 1；Start next time interval and triggers load distribution module work；
W=P* β2*C (2)
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum generations
The sum of average response time in table step S22,Represent the average response time of the machine；W represents the weights of the machine, and P is represented
The success rate of the machine, C represent weights radix, are an integer set, WmaxHighest weights are represented, are one preset big
In 1 integer.
Specifically, the modified weight module further includes：
Modified weight subelement, for reading configuration information before entering each time interval, and according to configuration information
The weight setting that the number of clusters of cluster service is increased to corresponding server is 1, and the number of clusters of cluster service is reduced
Corresponding server is directly deleted in loading distribution module from all servers.
The invention also discloses a kind of load balancing distribution systems for backstage cluster service, are the present invention with reference to figure 4
The structural schematic diagram of load balancing distribution system for backstage cluster service.
System includes center configuration service module and at least one edge service service module；Center configuration service module into
The typing and management of row configuration, the relevant configuration of edge service service module management business and adaptive adjustment operate.
Specifically, the center configuration service module includes：Configuration database, load balancing center cell, configuration management
Unit；The edge service service module includes：Above-mentioned load balancing assignment agent, business unit, configuration file；
The configuration management element is used to carry out unified modification to configuration information, for example, configuration item increasing, delete, look into, changing behaviour
Make；The load balancing center cell is used for the management of load balancing data, and the configuration of processing load balancing assignment agent is read
Request and interface data statistics；In order to realize the flexible management of configuration, configuration file is written by configuration management element and is configured
It is preserved in database.
In order to improve system performance, configuration file is stored using the mode of shared drive.The load balancing distribution
Agency manages local profile and is written in shared for reading configuration and reported data statistics from load balancing center cell
It deposits, optimal distribution and adaptive adjustment is carried out to connection request by the above method；Business unit with load balancing by distributing
Agency's interaction obtains optimal destination server information, and service request is distributed to corresponding destination server.
Load balanced sharing method, agency and the system for backstage cluster service for implementing the present invention have with following
Beneficial effect：The present invention is using a time interval as base unit, according to the response time of each server in the time interval
And success rate, a dynamic corrections are carried out to the weights of each server, response time and success rate have substantially reacted each
Actual performance under the influence of these factors of network, server environment and the actual conditions of software service of server, therefore,
The present invention may be implemented to maximize under the variant environment of each server performance to the dynamic trimming of weights uses all clothes
The processing capacity of business device, and can make to ask according to the adaptive adjustment of network, server environment and the actual conditions of software service progress
It asks to obtain optimal distribution.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited in above-mentioned specific
Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art
Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much
Form, all of these belong to the protection of the present invention.
Claims (8)
1. a kind of load balanced sharing method for backstage cluster service, which is characterized in that including：
S1, in a time interval, the weights that initial weight based on each server or step S2 are reset, according to negative
Equalization algorithm distribution connection request is carried to corresponding server；
S2, after a time interval, according to the response time of each server and success rate in the time interval, weight
The new weights for calculating and setting each server, start next time interval and enter step S1；
Wherein, the step S2 includes：
S21, after a time interval, when obtaining the response of each server in the time interval of newest end respectively
Between and success rate；
S22, the response time based on each server calculate separately the average response time of each server, are based on each service
The sum of product and average response time of average response time is calculated in the average response time of device；
S23, it is based respectively on the response time factor that formula (1) determines each server：
S24, the weights that formula (2) calculates and sets each server are based respectively on：
W=P* β2*C (2)
If the weights of S25, some server are more than Wmax, then it is W by the weight setting of some servermax；If some
The weights of server are less than 1, then are 1 by the weight setting of some server；
S26, start next time interval and enter step S1；
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum ride instead of walk
The sum of average response time in rapid S22,Represent the average response time of the machine；W represents the weights of the machine, and P represents the machine
Success rate, C represents weights radix, is an integer set, WmaxHighest weights are represented, preset are more than 1 for one
Integer；
Wherein, if the connection quantity of some server was k in last time interval, the number of successful connection is m times, m, k
For natural number, m is less than k, then the success rate of the server is：
2. the load balanced sharing method according to claim 1 for backstage cluster service, which is characterized in that the step
Suddenly S2 further includes：Configuration information is read before entering each time interval, and according to configuration information by the cluster of cluster service
The weight setting that quantity increases corresponding server is 1, and the number of clusters of cluster service, which is reduced corresponding server, to exist
It is directly deleted from all servers in step S1.
3. the load balanced sharing method according to claim 1 for backstage cluster service, which is characterized in that WmaxFor
100。
4. the load balanced sharing method according to claim 1 or 3 for backstage cluster service, which is characterized in that institute
It states described in step S1 and includes to corresponding server according to load-balancing algorithm distribution connection：
S11, for first time interval, the weights of each server are initial weight, each server in other times section
Weights be the weights that set in step S2；
S12, the greatest common divisor for determining Servers-all weights set comparative factor to the maximum in Servers-all weights
Weights；
The size of S13, the successively weights and comparative factor of more every server, if the weights of some server are more than or wait
In comparative factor, then the server is distributed into next connection request；
S14, when all server relatively after the completion of, by comparative factor using greatest common divisor as increments；If compare because
Son is less than 0, then comparative factor is re-set as maximum weights, go to step S13.
5. the load balanced sharing method according to claim 1 for backstage cluster service, which is characterized in that initial power
Value is set as 1.
6. a kind of load balancing assignment agent for backstage cluster service, which is characterized in that including
Load distribution module：For in a time interval, the initial weight based on each server or modified weight module
The weights of reset distribute connection request to corresponding server according to load-balancing algorithm；
Modified weight module：After a time interval, according to the response time of each server in the time interval
And success rate, the weights of each server are recalculated and set, starts next time interval and triggers load distribution module
Work；
Wherein, the modified weight module includes：
Response time and success rate acquiring unit：For after a time interval, obtain respectively newest end when
Between the response time of each server and success rate in section；
Response time factor calculating unit：Being averaged for each server is calculated separately for the response time based on each server
Response time, the average response time based on each server be calculated average response time product and average response time it
With, then determine based on formula (1) the response time factor of each server：
Weight calculation unit：The weights that each server is calculated and set for being based on formula (2), if the power of some server
Value is more than Wmax, then it is W by the weight setting of some servermax；If the weights of some server be less than 1, by this certain
The weight setting of a server is 1；Start next time interval and triggers load distribution module work；
W=P* β2*C (2)
Wherein, β represents the response time factor of the machine, and Pro represents the product of the average response time in step S22, Sum ride instead of walk
The sum of average response time in rapid S22,Represent the average response time of the machine；W represents the weights of the machine, and P represents the machine
Success rate, C represents weights radix, is an integer set, WmaxHighest weights are represented, preset are more than 1 for one
Integer；
Wherein, if the connection quantity of some server was k in last time interval, the number of successful connection is m times, m, k
For natural number, m is less than k, then the success rate of the server is：
7. the load balancing assignment agent according to claim 6 for backstage cluster service, which is characterized in that the power
Value correcting module further includes：
Modified weight subelement will collect for reading configuration information before entering each time interval, and according to configuration information
The weight setting that the number of clusters of group's service increases corresponding server is 1, and it is right that the number of clusters of cluster service is reduced institute
The server answered directly is deleted in loading distribution module from all servers.
8. a kind of load balancing distribution system for backstage cluster service, which is characterized in that including center configuration service module
With at least one edge service service module；Wherein, the center configuration service module includes：Configuration database, load balancing
Center cell, configuration management element；The edge service service module includes：As claim 6-7 any one of them loads
Equilibrium assignment agency, business unit, configuration file；
The configuration management element is used to carry out unified modification to configuration information and preserve into configuration database；The load is equal
The center cell that weighs is used for the management of load balancing data, handles the configuration read request and interface data of load balancing assignment agent
Statistics；The load balancing assignment agent is used for from load balancing center cell reading configuration and reported data statistics, and management is originally
Simultaneously shared drive is written in ground configuration file, is connection request distribution server；Business unit by with load balancing assignment agent
Interaction obtains optimal destination server information, and service request is distributed to corresponding destination server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510780814.6A CN105282259B (en) | 2015-11-13 | For the load balanced sharing method of backstage cluster service, agency and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510780814.6A CN105282259B (en) | 2015-11-13 | For the load balanced sharing method of backstage cluster service, agency and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105282259A CN105282259A (en) | 2016-01-27 |
| CN105282259B true CN105282259B (en) | 2018-08-31 |
Family
ID=
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1434393A (en) * | 2003-02-24 | 2003-08-06 | 武汉大学 | Dynamic loading balance method for cluster server |
| US20070143460A1 (en) * | 2005-12-19 | 2007-06-21 | International Business Machines Corporation | Load-balancing metrics for adaptive dispatching of long asynchronous network requests |
| CN101719082A (en) * | 2009-12-24 | 2010-06-02 | 中国科学院计算技术研究所 | Method and system for dispatching application requests in virtual calculation platform |
| CN103401947A (en) * | 2013-08-20 | 2013-11-20 | 曙光信息产业(北京)有限公司 | Method and device for allocating tasks to multiple servers |
| CN105007312A (en) * | 2015-07-03 | 2015-10-28 | 叶秀兰 | Method and system for controlling adaptive load-balancing of cloud computing server |
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1434393A (en) * | 2003-02-24 | 2003-08-06 | 武汉大学 | Dynamic loading balance method for cluster server |
| US20070143460A1 (en) * | 2005-12-19 | 2007-06-21 | International Business Machines Corporation | Load-balancing metrics for adaptive dispatching of long asynchronous network requests |
| CN101719082A (en) * | 2009-12-24 | 2010-06-02 | 中国科学院计算技术研究所 | Method and system for dispatching application requests in virtual calculation platform |
| CN103401947A (en) * | 2013-08-20 | 2013-11-20 | 曙光信息产业(北京)有限公司 | Method and device for allocating tasks to multiple servers |
| CN105007312A (en) * | 2015-07-03 | 2015-10-28 | 叶秀兰 | Method and system for controlling adaptive load-balancing of cloud computing server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107423120B (en) | Task scheduling method and device | |
| US11032210B2 (en) | Software load balancer to maximize utilization | |
| CN105743995B (en) | A kind of system and method for the deployment of portable High Availabitity and management container cluster | |
| US8386512B2 (en) | System for managing data collection processes | |
| US7752262B2 (en) | Slow-dynamic load balancing system and computer-readable medium | |
| CN108712464A (en) | A kind of implementation method towards cluster micro services High Availabitity | |
| US9081624B2 (en) | Automatic load balancing, such as for hosted applications | |
| CN110795203B (en) | Resource scheduling method, device, system and computing equipment | |
| US20210037092A1 (en) | Fault tolerant systems and methods incorporating a minimum checkpoint interval | |
| US7870419B2 (en) | Subscription-based management and distribution of member-specific state data in a distributed computing system | |
| CN113110933B (en) | System with Nginx load balancing technology | |
| WO2023050901A1 (en) | Load balancing method and apparatus, device, computer storage medium and program | |
| EP2710470A1 (en) | Extensible centralized dynamic resource distribution in a clustered data grid | |
| JP2010238051A (en) | Load balancing program and load balancing apparatus | |
| CN111737003A (en) | Pod balanced scheduling method and device, main node and storage medium | |
| US11128697B2 (en) | Update package distribution using load balanced content delivery servers | |
| CN110457138A (en) | Management method, device and the electronic equipment of game server cluster | |
| CN106776008A (en) | A kind of method and system that load balancing is realized based on zookeeper | |
| CN111131486A (en) | Load adjustment method and device of execution node, server and storage medium | |
| CN110515720B (en) | Cloud computing system service cost and reliability driven job scheduling method | |
| WO2024021469A1 (en) | System operation and maintenance management method and apparatus, and electronic device | |
| CN102724104A (en) | Apparatus and method for automatically configuring Java EE application cluster | |
| CN105282259B (en) | For the load balanced sharing method of backstage cluster service, agency and system | |
| US7555544B1 (en) | Implementation of affinities in high availability computer system clusters | |
| Chen et al. | Dynamic replication scheduling for cloud datacenters based on workload statistics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |