CN108984306A - A kind of load balance scheduling system and method according to TCP request quality evaluation - Google Patents

A kind of load balance scheduling system and method according to TCP request quality evaluation Download PDF

Info

Publication number
CN108984306A
CN108984306A CN201810778644.1A CN201810778644A CN108984306A CN 108984306 A CN108984306 A CN 108984306A CN 201810778644 A CN201810778644 A CN 201810778644A CN 108984306 A CN108984306 A CN 108984306A
Authority
CN
China
Prior art keywords
module
business
homophony
tcp
business module
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.)
Pending
Application number
CN201810778644.1A
Other languages
Chinese (zh)
Inventor
罗意
谈加虎
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Gf Securities Co ltd
Original Assignee
Gf Securities Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Gf Securities Co ltd filed Critical Gf Securities Co ltd
Priority to CN201810778644.1A priority Critical patent/CN108984306A/en
Publication of CN108984306A publication Critical patent/CN108984306A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5083Techniques for rebalancing the load in a distributed system
    • G06F9/5088Techniques for rebalancing the load in a distributed system involving task migration
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00Indexing scheme relating to G06F9/00
    • G06F2209/50Indexing scheme relating to G06F9/50
    • G06F2209/508Monitor
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00Indexing scheme relating to G06F9/00
    • G06F2209/50Indexing scheme relating to G06F9/50
    • G06F2209/509Offload

Abstract

A kind of load balance scheduling system and method according to TCP request quality evaluation, it is characterized by: the system comprises: service discovery module, load balancing assembly module, homophony business modules and by tune business module, it include several by tune business process submodule by tune business module, its advantage is that 1) quality evaluation of various dimensions can be carried out to TCP long connection from application layer, to reflect the real-time load condition of business.2) load balancing can be carried out according to business real time load state, can better distribution request, while the dilatation of application server, capacity reducing are all unaware for upper-layer user.3) make load balancing component high-efficiency operation without lockization design, to service impact minimum.4) by the management to TCP connection state, the state of application service can be found in millisecond, so that (or kicking out of) is added in the application that will increase (or deletion) newly.

Description

A kind of load balance scheduling system and method according to TCP request quality evaluation
Technical field
The present invention relates to transmission control procedure, it is related to the distribution for having load balancing, routing disaster tolerance and high performance requirements Background system, espespecially a kind of load balance scheduling system and method according to TCP request quality evaluation.
Background technique
When the performance of a server reaches capacity, server cluster can be used to improve the overall performance of website. In server cluster, the role of a dispatcher is needed, receives all requests of user, then according to every server Loading condition will request assignment to a certain back-end server and go to handle.In this process, dispatcher how appoint by reasonable distribution Business, guarantees that all back-end servers all give full play to performance, thus keep the overall performance of server cluster optimal, here it is Problem of load balancing.There are as below methods for the load balance scheduling of the prior art:
One, polling method
Request is alternately assigned on back-end server in order, it balancedly treats each server of rear end, Without concern for the actual connection number of server and current system load.
Two, randomized
By the random algorithm of system, a service therein is randomly selected according to the list size value of back-end server Device accesses.By Probability Statistics Theory it is known that actual effect is got over as the number at client call service end increases Come each server closer to mean allocation calling amount to rear end, that is, the result of poll.
Two, source address Hash method
The thought of source address Hash is a number being calculated by hash function according to the IP address for obtaining client Value, carries out modulo operation with size of the numerical value to server list, obtained result is that customer side will access server Serial number.Using source address Hash method carry out load balancing, the client of same IP address, when back-end server list is constant, It can be mapped to same back-end server every time and access.
Four, weighted polling method
The configuration of the possible machine of different back-end servers and the load of current system be not identical, therefore their resistance to compression Ability is not also identical.The machine high to configuration, load is low configures higher weight, its processing is allowed more to ask;And matches and set low, bear High machine is carried, lower weight is distributed to it, reduces its system load, weighted polling can handle this problem well, and By request sequence and according to weight distribution to rear end.
Five, Smallest connection number method
Smallest connection number algorithm comparison is flexibly and intelligent, since the configuration of back-end server is not quite similar, for request Faster or slower is handled, it is the connection current according to back-end server, dynamically chooses and wherein currently overstocks connection number most A few server handles current request, improves the utilization efficiency of back-end services as much as possible, will be responsible for reasonably dividing Flow to each server.
These methods of the prior art have the disadvantage in that
It 1) cannot be according to the actual loading ability dynamic change request scheduling rule of back-end services;
2) it cannot really reflect the performance of application layer business module;
3) load balancing cannot be carried out for TCP long connection request.
Summary of the invention
In view of the shortcomings of the prior art, the purpose of the present invention is to provide a kind of loads according to TCP request quality evaluation Balance dispatching system and method.It is intended to propose a kind of assessment business module service for the TCP request quality of different business module The standardized algorithm of performance scores, establish it is a set of can according to the different loads ability of business module carry out dynamic load leveling side Method.
The technical solution adopted by the present invention to solve the technical problems is: providing a kind of according to TCP request quality evaluation Load balance scheduling system and method, which is characterized in that the system comprises: service discovery module, load balancing assembly module, Homophony business module and by tune business module, includes several by tune business process submodule by tune business module, in which:
Service discovery module connects load balancing assembly module and by tune business module, is used for storage and management background system Service module information, version information and authority information including each business module;
Load balancing assembly module connects service discovery module and homophony business module: being used for from service discovery module timing The synchronous service module information for needing to call, and the TCP request data reported within 1 second period according to homophony business module is to clothes Discovery module of being engaged in carries out scoring function;And it is selected in tune business process module according to the performance scores of service discovery module from multiple An optimal TCP connection is taken to send TCP request for homophony business module;It is also responsible for monitoring by the TCP of tune business module process Connect the event whether normal, processing TCP connection is established and closed;
Homophony business module connects load balancing assembly module and by tune business module: when needing to send TCP request packet, By load balancing assembly module afferent module name parameter, call obtain connecting interface get best performance by tune business Process;When receiving TCP request response bag, interface is reported to report request time-consuming and response faultcode by service;It is built in TCP connection TCP connection state is reported when vertical and closing;
By tune business module connection homophony business module and service discovery module comprising several are by tune business process Oneself module information is registered to service discovery module when being started by tune business module, from service discovery module when offline by module The module information of oneself is deleted, provides TCP request service for homophony business module.
The load balance scheduling side for the foundation TCP request quality evaluation based on above system that present invention simultaneously provides a kind of Method, it is characterised in that include the following steps:
S1. the module information of oneself is registered on startup by service discovery module by tune business module;
S2. the timing of load balancing assembly module pulls the module information for needing to call from service discovery module synchronization, newest Module information be compared with original module information, if it is newly-increased IP:PORT to then notify homophony business module into Row TCP connection is established, if it is need the IP:PORT that deletes to then notifying homophony business module to be attached closing and deletion; Homophony business module is also reported to load balancing assembly module when TCP connection is established and is closed, and makes load balancing assembly module Timely update IP:PORT pairs of Algorithms of Selecting.
The step S2 is subdivided into following steps:
S21. external to send TCP request to homophony business module;
S22. homophony business module is incoming is referred to as parameter by tune business module name, calls Get interface, obtains optimal TCP connection;Load balancing assembly module is by carrying out algorithm to according to performance scores to by three IP:PORT of tune business module Selection selects an optimal IP:PORT to returning to homophony business module, while load balancing assembly module records this IP:PORT pairs of transmission state;
S23. homophony business module gets optimal IP:PORT to rear, chooses and arrives corresponding TCP connection, passes through the connection TCP request is sent to by tune business module;
S24. after receiving TCP request by tune business module, the request is handled, while responding to homophony business module;
S25. after homophony business module receives response, statistics responds back packet time-consuming and business return code, then calls and reports Interface feeds back to load balancing assembly module;Load balancing assembly module according to the number, response reported in 1 second period it is time-consuming and After business return code is standardized dimension, request quality score algorithm to IP:PORT to scoring by TCP;
S26. homophony business module carries out responding back packet to external TCP request.
Illustrate that TCP requests quality score algorithm below by concrete instance:
Homophony business module is denoted as Ma, by the entitled Mb of tune business module, by three IP:PORT of tune business module to note Are as follows: B1, B2, B3.It is respectively S1, S2, S3, the response times received that the transmission times of Ma calling Mb, which is (S), within 1 second period (R) be respectively R1, R2, R3, corresponding average time-consuming (C) is respectively C1, C2, C3, system mistake code ratio (E) be respectively E1, E2, E3, for Ma module, scoring (L) is L1, L2, L3 in 1 second period of Mb module,
1) scoring algorithm of transmission times and response times
Score1_i=[(Ri-Si)/Si] * 100, i=1,2,3
2) time-consuming scoring algorithm is responded
Score2_i=[(Ci-min (Ci))/(max (Ci)-min (Ci))] * 100, i=1,2,3
3) business return code scoring algorithm
Score3_i=[(Ei-min (Ei))/(max (Ei)-min (Ei))] * 100, i=1,2,3
4) overall score in period
Li=0.4*Score1_i+0.4*Score2_i+0.2*Score3_i, i=1,2,3
L score is bigger, and expression IP:PORT pairs of the performance is relatively poorer, and the selected probability got also can be lower
5) the selected probability got of B1, B2, B3 node is (100-Li)/100.
The beneficial effects of the present invention are:
1) various dimensions can be carried out to TCP long connection from application layer and (sends response times, response time-consuming and system mistake ratio Example) quality evaluation, to reflect the real-time load condition of business.
2) can according to business real time load state carry out load balancing, can better distribution request, while application clothes The dilatation of business device, capacity reducing are all unaware for upper-layer user.
3) make load balancing component high-efficiency operation without lockization design, to service impact minimum.
4) by the management to TCP connection state, the state of application service can be found in millisecond, to will increase newly (or kicking out of) is added in the application of (or deletion).
Detailed description of the invention
The invention will be further described with reference to the accompanying drawing.
Fig. 1 is configuration diagram of the invention.
Fig. 2 is load balancing assembly module schematic diagram of internal structure of the invention.
In Fig. 1: 1 being service discovery module, 2 be load balanced component module, 3 be homophony business module, 4 be to be adjusted business Module, 41,42,4N be by tune business module 4 by tune business process submodule.
Specific embodiment
Referring to attached drawing, a kind of load balance scheduling system and method according to TCP request quality evaluation of the present invention, feature It is, the system comprises: service discovery module 1, load balancing assembly module 2, homophony business modules 3 and by tune business module 4, by tune business module 4 include several by tune business process submodule 41,42 ... .4N, in which:
Service discovery module 1 connects load balancing assembly module 2 and by tune business module 4, is responsible for storage and management distribution The service discovery module information of formula background system, in an embodiment of the present invention, this information include: a business module title Corresponding multiple processes, each process is by IP:PORT to (such as: 10.2.138.15:10000 indicates that the process operates in 10.2.138.15 on this machine, the external port numbers that provide are 10000) to identify, while the version including each business module This information and authority information etc.;Access layer is disclosed by tune business module 4, two interface services of 0x1000 and 0x1100 are provided, it should For module there are three process, module information example is following (by tune business module 4);
ServiceName:called.gf.com.cn//business module title
Cmd:0x1000 | 0x1100//routing command word
IpPort:127.0.0.1:10000,127.0.0.1:10001,127.0.0.2:10000// IP:PORT couple
IP-PORT pairs of // business process
Version:3.0.0//version number
Privilege:public//whether disclose permission
Load balancing assembly module 2 connects service discovery module 1 and homophony business module 3: being responsible for from service discovery module Timing Synchronization needs the service discovery module information called, and asks according to the TCP that homophony business module 3 reported within 1 second period Data are asked to carry out scoring function to service discovery module 1;The responsible performance scores according to service discovery module 1 are from multiple by tune industry Be engaged in scheduler module 41,42 ... choose an optimal TCP connection in .4N for homophony business module 3 and send TCP request;In addition, It is also responsible for monitoring by whether the TCP connection of 4 process of tune business module is normal, the event that processing TCP connection is established and closed.
Homophony business module 3 connects load balancing assembly module 2 and by tune business module 4: needing to send TCP request packet When, it is called by 2 afferent module title (ServiceName) parameter of load balancing assembly module and obtains connecting interface (Get API) get best performance by tune business process (IP:PORT to);When receiving TCP request response bag, reported by service Interface (Report API) reports request time-consuming with response faultcode, and (time-consuming of request is generally ms, mainly sends out request To the time difference received between response.Response faultcode is the return code of system, for identifying different error categories);In TCP Connection establishes and reports TCP connection state (closed state of TCP connection and establish state) when closing.
By tune business module 4 connect homophony business module 3 and service discovery module 1 comprising several by tune business into Journey submodule 41,42 ... oneself module information 4N: can be registered to by tune business module 4 by service discovery mould on startup Block 1 can delete the module information of oneself from service discovery module 1 when offline, provide TCP request clothes for homophony business module 3 Business.
It is equal in the load that the embodiment of the present invention provides a kind of foundation TCP request quality evaluation based on above system simultaneously Weigh dispatching method, it is characterised in that includes the following steps:
S1. the module information of oneself is registered on startup by service discovery module 1 by homophony business module 3;
S2. 2 timing of load balancing assembly module synchronizes the module information for pulling and needing to call from service discovery module 1, most New module information is compared with original module information, if it is newly-increased IP:PORT to then notifying homophony business module 3 Carry out TCP connection foundation, if it is need the IP:PORT that deletes to then notifying homophony business module 3 to be attached closing and delete It removes;Homophony business module 3 is also reported to load balancing assembly module 2 when TCP connection is established and is closed, and makes load balancing component Module 2 timely updates IP:PORT pairs of Algorithms of Selecting.
The step S2 is subdivided into following steps:
S21. external to send TCP request to homophony business module 3;
S22. homophony business module 3 is incoming is referred to as parameter by tune business module 4, calls Get interface, obtains optimal TCP connection;Load balancing assembly module 2 is by calculating according to performance scores by three IP:PORT of tune business module 4 Method selection, selects an optimal IP:PORT to returning to homophony business module 3, while load balancing assembly module 2 is recorded IP:PORT pairs of the transmission state;
S23. homophony business module 3 gets optimal IP:PORT to rear, chooses and arrives corresponding TCP connection, passes through the connection TCP request is sent to by tune business module 4;
S24. after receiving TCP request by tune business module 4, the request is handled, while to 3 sound of homophony business module It answers;
S25. after homophony business module 3 receives response, statistics responds back packet time-consuming and business return code, then calls and reports Interface feeds back to load balancing assembly module 2;Load balancing assembly module 2 is time-consuming according to the number, response reported in 1 second period After being standardized dimension with business return code, request quality score algorithm to IP:PORT to scoring by TCP;
S26. 3 pairs of homophony business module external TCP requests carry out responding back packet.
Illustrate that TCP requests quality score algorithm below by concrete instance:
Homophony business module 3 is denoted as Ma, by the entitled Mb of tune business module 4, by 4 three IP:PORT pairs of tune business module It is denoted as: B1, B2, B3.It is respectively S1, S2, S3, the response received time that the transmission times of Ma calling Mb, which is (S), within 1 second period Number (R) is respectively R1, R2, R3, and corresponding average time-consuming (C) is respectively C1, C2, C3, and system mistake code ratio (E) is respectively E1, E2, E3, for Ma module, scoring (L) is L1, L2, L3 in 1 second period of Mb module,
1) scoring algorithm of transmission times and response times
Score1_i=[(Ri-Si)/Si] * 100, i=1,2,3
2) time-consuming scoring algorithm is responded
Score2_i=[(Ci-min (Ci))/(max (Ci)-min (Ci))] * 100, i=1,2,3
3) business return code scoring algorithm
Score3_i=[(Ei-min (Ei))/(max (Ei)-min (Ei))] * 100, i=1,2,3
4) overall score in period
Li=0.4*Score1_i+0.4*Score2_i+0.2*Score3_i, i=1,2,3
L score is bigger, and expression IP:PORT pairs of the performance is relatively poorer, and the selected probability got also can be lower.
5) the selected probability got of B1, B2, B3 node is (100-Li)/100.

Claims (3)

1. a kind of load balance scheduling system according to TCP request quality evaluation, it is characterised in that: the system comprises: service Discovery module, load balancing assembly module, homophony business module and by tune business module, include several quilts by tune business module Business process submodule is adjusted, wherein:
Service discovery module connects load balancing assembly module and the clothes by tune business module, for storage and management background system Business module information, version information and authority information including each business module;
Load balancing assembly module connects service discovery module and homophony business module: being used for from service discovery module Timing Synchronization The service module information for needing to call, and the TCP request data reported within 1 second period according to homophony business module sends out service Existing module carries out scoring function;And one is chosen in tune business process module from multiple according to the performance scores of service discovery module A optimal TCP connection sends TCP request for homophony business module;It is also responsible for monitoring by the TCP connection of tune business module process Event whether normal, that processing TCP connection is established and closed;
Homophony business module connects load balancing assembly module and by tune business module: when needing to send TCP request packet, passing through Load balancing assembly module afferent module name parameter, call obtain connecting interface get best performance by tune business into Journey;When receiving TCP request response bag, interface is reported to report request time-consuming and response faultcode by service;It is established in TCP connection TCP connection state is reported with when closing;
By tune business module connection homophony business module and service discovery module comprising several are by tune business process submodule Oneself module information is registered to service discovery module when being started by tune business module, deleted when offline from service discovery module by block Except the module information of oneself, TCP request service is provided for homophony business module.
2. a kind of load balancing based on the load balance scheduling system described in claim 1 according to TCP request quality evaluation Dispatching method, it is characterised in that include the following steps:
S1. the module information of oneself is registered on startup by service discovery module by homophony business module;
S2. load balancing assembly module timing pulls the module information for needing to call, newest mould from service discovery module synchronization Block message is compared with original module information, if it is newly-increased IP:PORT to then notify homophony business module carry out TCP Connection is established, if it is need the IP:PORT that deletes to then notifying homophony business module to be attached closing and deletion;Homophony industry Business module is also reported to load balancing assembly module when TCP connection is established and is closed, and makes load balancing assembly module in time more New IP:PORT pairs of Algorithms of Selecting.
3. the load equilibration scheduling method according to claim 2 according to TCP request quality evaluation, it is characterised in that described Step S2 includes the following steps:
S21. external to send TCP request to homophony business module;
S22. homophony business module is incoming is referred to as parameter by tune business module name, calls Get interface, obtains optimal TCP and connect It connects;Load balancing assembly module by by three IP:PORT of tune business module to according to performance scores carry out algorithms selection, An optimal IP:PORT is selected to returning to homophony business module, while load balancing assembly module records the IP:PORT Pair transmission state;
S23. homophony business module gets optimal IP:PORT to rear, chooses and arrives corresponding TCP connection, is sent by the connection TCP is requested to by tune business module;
S24. after receiving TCP request by tune business module, the request is handled, while responding to homophony business module;
S25. after homophony business module receives response, statistics responds back packet time-consuming and business return code, then calls and reports interface Feed back to load balancing assembly module;Load balancing assembly module is according to number, response time-consuming and the business reported in 1 second period After return code is standardized dimension, request quality score algorithm to IP:PORT to scoring by TCP;
S26. homophony business module carries out responding back packet to external TCP request.
CN201810778644.1A 2018-07-16 2018-07-16 A kind of load balance scheduling system and method according to TCP request quality evaluation Pending CN108984306A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810778644.1A CN108984306A (en) 2018-07-16 2018-07-16 A kind of load balance scheduling system and method according to TCP request quality evaluation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810778644.1A CN108984306A (en) 2018-07-16 2018-07-16 A kind of load balance scheduling system and method according to TCP request quality evaluation

Publications (1)

Publication Number Publication Date
CN108984306A true CN108984306A (en) 2018-12-11

Family

ID=64549287

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810778644.1A Pending CN108984306A (en) 2018-07-16 2018-07-16 A kind of load balance scheduling system and method according to TCP request quality evaluation

Country Status (1)

Country Link
CN (1) CN108984306A (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102057631A (en) * 2008-06-09 2011-05-11 微软公司 Data center interconnect and traffic engineering
US20150040218A1 (en) * 2007-01-24 2015-02-05 Dmitri Alperovitch Detecting image spam
CN104394224A (en) * 2014-11-28 2015-03-04 无锡华云数据技术服务有限公司 Load balance system
CN106375355A (en) * 2015-07-20 2017-02-01 中兴通讯股份有限公司 Load balancing processing method and apparatus
CN107689878A (en) * 2016-08-04 2018-02-13 创盛视联数码科技(北京)有限公司 TCP length connection SiteServer LBSs based on name scheduling
CN108111567A (en) * 2016-11-25 2018-06-01 广东亿迅科技有限公司 Realize the uniform method and system of server load

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150040218A1 (en) * 2007-01-24 2015-02-05 Dmitri Alperovitch Detecting image spam
CN102057631A (en) * 2008-06-09 2011-05-11 微软公司 Data center interconnect and traffic engineering
CN104394224A (en) * 2014-11-28 2015-03-04 无锡华云数据技术服务有限公司 Load balance system
CN106375355A (en) * 2015-07-20 2017-02-01 中兴通讯股份有限公司 Load balancing processing method and apparatus
CN107689878A (en) * 2016-08-04 2018-02-13 创盛视联数码科技(北京)有限公司 TCP length connection SiteServer LBSs based on name scheduling
CN108111567A (en) * 2016-11-25 2018-06-01 广东亿迅科技有限公司 Realize the uniform method and system of server load

Similar Documents

Publication Publication Date Title
CN110266716B (en) Unified service platform system of power grid
WO2020019472A1 (en) Rating method for service node, domain name system (dns) scheduling method and server
EP2563062B1 (en) Long connection management apparatus and link resource management method for long connection communication
CN108712464A (en) A kind of implementation method towards cluster micro services High Availabitity
WO2011000199A1 (en) Method and system for intelligently scheduling cluster servers
CN105933408B (en) A kind of implementation method and device of Redis universal middleware
CN109933431B (en) Intelligent client load balancing method and system
CN103888277B (en) A kind of gateway disaster-tolerant backup method, device and system
CN105337786B (en) A kind of server performance detection method, device and equipment
CN109274707A (en) A kind of load dispatching method and device
CN110177118A (en) A kind of RPC communication method based on RDMA
CN106599711A (en) Database access control method and device
CN110166524B (en) Data center switching method, device, equipment and storage medium
WO2004036344A2 (en) System and method for the optimization of database
CN103607424B (en) Server connection method and server system
CN109547512A (en) A kind of method and device of the distributed Session management based on NoSQL
CN102281190A (en) Networking method for load balancing apparatus, server and client access method
CN112825527A (en) Service node distribution method and device based on MEC and related server
US20080063165A1 (en) System and method for dynamically partitioning context servers
CN107819632A (en) A kind of dynamic load leveling group system based on performance monitoring system and Docker Swarm
CN109960586A (en) A kind of appreciable four-layer load-equalizing device of server state and equalization methods
CN110519354A (en) A kind of distributed objects storage system and its method for processing business and storage medium
CN100484014C (en) Distributed cluster service management system and service management method in intelligent network
CN112543150A (en) Dynamic load balancing method based on server control
CN109600402A (en) Server load balancing method and device, storage medium, server, service system

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20181211

RJ01 Rejection of invention patent application after publication