CN104135525B - The resource expansion method and apparatus of cloud platform ELB components - Google Patents

The resource expansion method and apparatus of cloud platform ELB components Download PDF

Info

Publication number
CN104135525B
CN104135525B CN201410375345.5A CN201410375345A CN104135525B CN 104135525 B CN104135525 B CN 104135525B CN 201410375345 A CN201410375345 A CN 201410375345A CN 104135525 B CN104135525 B CN 104135525B
Authority
CN
China
Prior art keywords
service
elb
cloud platform
service instance
components
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
Application number
CN201410375345.5A
Other languages
Chinese (zh)
Other versions
CN104135525A (en
Inventor
王旭
周冠宇
温云龙
宋吉鹏
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.)
GCI Science and Technology Co Ltd
Original Assignee
GCI Science and Technology 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 GCI Science and Technology Co Ltd filed Critical GCI Science and Technology Co Ltd
Priority to CN201410375345.5A priority Critical patent/CN104135525B/en
Publication of CN104135525A publication Critical patent/CN104135525A/en
Application granted granted Critical
Publication of CN104135525B publication Critical patent/CN104135525B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Debugging And Monitoring (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a kind of resource expansion method and apparatus of cloud platform ELB components, first according to the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components of reading, maximum service number of requests maxRpt and the average increment m per second of single Service Instance that single Service Instance is accepted are calculated;Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;By real time service number of requests compared with Service Instance maximum processing capability;When real time service number of requests is more than Service Instance maximum processing capability, polymerizing value growth rate k and the average increment m per second of single Service Instance ratio are calculated, resource expansion is carried out according to the ratio.The efficiently quickly request response of response high load condition of the invention, improves the responding ability of cloud platform, improves resource utilization, cost-effective, application value is high.

Description

The resource expansion method and apparatus of cloud platform ELB components
Technical field
The present invention relates to cloud computing technology applied technical field, expands more particularly to a kind of resource of cloud platform ELB components Open up method and apparatus.
Background technology
The characteristics of cloud platform one is natural is exactly the access towards high concurrent, and solves response speed in high concurrent access Problem, it would be desirable to have the method for a load balancing to solve.The core concept of load balancing is exactly by the way that request is distributed to Different service ends are handled, and will so lift the throughput efficiency of whole platform.Similar ELB (Elastic Load Balancer, elastic load balanced device) in CloudFoundry (a PaaS cloud platform), Google App Engine (Googles Web application) it can all play critically important effect.In Stratos (a kind of PaaS service platforms), ELB is mainly pair The load balancing that Cartridge is accessed, when excessive developer needs same Cartridge, it will first pass through ELB A request shunting is carried out, then different request flow points are dealt into different Cartridge, so as to obtain Cartridge mapping.
Because elasticity is one of mostly important characteristic of cloud computing platform, and the important mark of evaluation cloud computing platform ability Standard, so the resilient expansion decision-making of cloud computing platform is particularly important.The quality of resilient expansion decision-making technique will be directly connected to cloud Whether calculating platform can be extended according to the requirement of user in reasonable time, so being determined in cloud computing platform resilient expansion Substantial amounts of research is generated in terms of plan, also generates the product of maturation.Auto Scaling (resilient expansion) are Amazon EC2 Automatic expansion service on cloud computing platform, its function are to be automatically created or terminate EC2 void according to user-defined trigger Plan machine example.The fundamental mode of Auto Scaling services is as shown in Figure 1.Trigger (trigger) in Fig. 1 is to use The trigger that family defines.According to the rule of trigger definition in figure, when average CPU (Central Processing Unit, in Central processor) utilization rate be more than 80% when, by EC2 examples increase by 10%, and when average CPU utilization be less than 40% when, by EC2 Example reduces 10%.Launch Configurations (starting configuration) in figure, which are used to specify, to be created needed for new EC2 examples The parameter wanted.Another service CloudWatch (cloud monitoring) on Amazon EC2 is responsible for monitoring EC2, according to CloudWatch Obtained extension defined in monitoring data and trigger and rule is shunk, Auto Scaling will expand EC2 automatically Exhibition is shunk.
Response type extension decision-making technique is entered according to the load of current cloud computing platform and the extension rule being manually set A kind of method of row automatic elastic extension.Auto Scaling resilient expansion method can range response type extension decision-making party Method.The monitored results of cloud computing platform and the extension being manually set are advised because this one kind extension decision-making technique places one's entire reliance upon Then, realize it is relatively simple, so commercialized cloud computing platform more in this way carry out resilient expansion decision-making.But this The defects of class method be it is fairly obvious, response type extension decision-making technique do not consider required for cloud computing platform resilient expansion when Between, simply simply start resilient expansion when monitored results meet extension rule, so causing user to need one section of experience to prolong The slow time can just obtain required resource.
Prediction type extension decision-making technique is to be predicted by the load capacity following to cloud computing platform so as to carry out elasticity Extend the method for decision-making.This kind of method regards the history monitoring value of cloud computing platform as a predictable sequence, and to the sequence Row carry out mathematical modeling, the methods of being analyzed using recurrence, time series similarity, then the monitoring value at next time point Can is predicted using value of the mathematical modeling at next time point.Carried out in advance to cloud computing platform future load amount After survey, it is possible to determine when carry out resilient expansion using prediction result, make resilient expansion decision-making.Based on pattern match Extension decision-making technique, by improving future load of KMP (linear session string matching algorithm) algorithms to cloud computing platform Amount is predicted, so as to be extended decision-making.Prediction type extension decision-making technique copes with prolonging needed for cloud computing platform extension The slow time, but simultaneously because need to predict future load amount according to historic load amount, this method needs accurate history in detail Monitoring data is as support.The existing research on quality of service monitor employs different frameworks, agreement or algorithm and reached The purpose of quality of service monitor is carried out in the case where producing a small amount of traffic load, but with the proviso that in quality of service monitor Accurate monitoring value need not be obtained in scene, or does not require to obtain monitored results in a short time, so existing service Quality control method can not obtain detailed accurate monitoring data in time, so extension decision-making can not be supported well.
In the practical application of cloud platform ELB components, in general way is directly to use response type expanding policy, response type Extension decision-making technique does not consider the time required for cloud computing platform resilient expansion, simply simply meets to extend in monitored results Start resilient expansion when regular, so causing user to need one section of time delay of experience just to obtain required resource.Prediction type Extension decision-making technique copes with the time delay needed for cloud computing platform extension, but simultaneously because needs according to historic load amount Future load amount is predicted, this method needs history monitoring data accurate in detail, and workload is past in real work as support It is past larger, it is impossible to meet the needs of real work.
The content of the invention
Based on the above situation, the present invention proposes a kind of resource expansion method of cloud platform ELB components, efficiently quickly rings Answer the request of high load condition to respond, improve the responding ability of cloud platform, improve resource utilization, it is cost-effective, have very High application value.
To achieve these goals, the embodiment of technical solution of the present invention is:
A kind of resource expansion method of cloud platform ELB components, comprises the following steps:
According to the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components of reading, single service is calculated The average increment m per second of maximum service number of requests maxRpt and single Service Instance that example is accepted;
Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;
By real time service number of requests compared with Service Instance maximum processing capability, the maximum processing of the Service Instance Ability passes through the maximum service number of requests maxRpt that the single Service Instance is accepted and real time execution Service Instance number nIn real time Product determine;The polymerizing value growth rate k passes through formula:It is determined that wherein vi (t) represents history The service request quantity of i-th of node of t in data, n represent node total number, and △ t were represented between the time of tasks carrying twice Every;
When the real time service number of requests is more than the Service Instance maximum processing capability, calculates the polymerizing value and increase Long rate k and the average increment m per second of the single Service Instance ratio, the Service Instance for needing to start according to ratio prediction Number, carry out resource expansion.
For prior art problem, the invention also provides a kind of resource expansion device of cloud platform ELB components, effectively solution Certainly existing response type extension causes user to need one section of time delay of experience, prediction type extension workload to meet actual work greatly The problem of making demand, it is adapted to application.
The embodiment of technical solution of the present invention is:
A kind of resource expansion device of cloud platform ELB components, including:
Processing module, for the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components according to reading, Calculate maximum service number of requests maxRpt and the average increment m per second of single Service Instance that single Service Instance is accepted, and root Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;
Comparison module, for by real time service number of requests compared with Service Instance maximum processing capability, the clothes The maximum service number of requests maxRpt and real time execution that pragmatic example maximum processing capability is accepted by the single Service Instance Service Instance number nIn real timeProduct determine;The polymerizing value growth rate k passes through formula:It is determined that its Middle vi (t) represents the service request quantity of i-th of node of t in historical data, and n represents node total number, and △ t represent to appoint twice The time interval that business performs;
Resource expansion module, for being more than the Service Instance maximum processing capability when the real time service number of requests When, the polymerizing value growth rate k and the average increment m per second of the single Service Instance ratio are calculated, is entered according to the ratio Row resource expansion.
Compared with prior art, beneficial effects of the present invention are:The resource expansion method of cloud platform ELB components of the present invention and Device, by introducing the average increment per second of maximum service number of requests, single Service Instance that single Service Instance accepts and poly- The parameters such as conjunction value growth rate, polymerizing value growth rate and the ratio of the average increment per second of single Service Instance are calculated, according to the ratio Prediction needs the number of the Service Instance started, pre-cooling Service Instance, before peak value arrival, reaches peak performance demand Stock number extension, effectively overcome response type extension decision-making technique and do not consider time required for resilient expansion, to cause to use Family needs to undergo the problem of just obtaining required resource one section of time delay, while reduces prediction type extension decision-making and need to count The a large amount of historic load amount prediction future load amounts of Fitting Analysis are calculated, efficiently the quickly request response of response high load condition, is carried The responding ability of high cloud platform, resource utilization is improved, it is cost-effective, there is very high application value.
Brief description of the drawings
Fig. 1 is that existing Auto Scaling service fundamental mode schematic diagram;
Fig. 2 is the resource expansion method flow diagram of cloud platform ELB components in one embodiment;
Fig. 3 is the resource expansion method flow diagram based on cloud platform ELB components in method one shown in Fig. 2 specific example;
Fig. 4 is the resource expansion apparatus structure schematic diagram of cloud platform ELB components in one embodiment.
Embodiment
For the objects, technical solutions and advantages of the present invention are more clearly understood, below in conjunction with drawings and Examples, to this Invention is described in further detail.It should be appreciated that embodiment described herein is only to explain the present invention, Do not limit protection scope of the present invention.
The resource expansion method of cloud platform ELB components in one embodiment, as shown in Fig. 2 comprising the following steps:
Step S201:According to the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components of reading, calculate The average increment m per second of maximum service number of requests maxRpt and single Service Instance that single Service Instance is accepted;
Step S202:Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;
Step S203:By real time service number of requests compared with Service Instance maximum processing capability, the service is real Example maximum processing capability passes through the maximum service number of requests maxRpt that the single Service Instance is accepted and real time execution service Example number nIn real timeProduct determine;
Step S204:When the real time service number of requests is more than the Service Instance maximum processing capability, institute is calculated Polymerizing value growth rate k and the average increment m per second of the single Service Instance ratio are stated, resource expansion is carried out according to the ratio Exhibition.
It is evidenced from the above discussion that the resource expansion method of cloud platform ELB components of the present invention, efficiently quickly response high capacity The request response of situation, improves the responding ability of cloud platform, improves resource utilization, be adapted to practical application.
As one embodiment, the maximum service number of requests maxRpt that the single Service Instance is accepted passes through formula: MaxRpt=Rps × △ t × AUR determine that the average increment m per second of the single Service Instance passes through formula:Really Fixed, wherein Rps represents the single Service Instance maximum service number of requests per second accepted, △ t represent tasks carrying twice when Between be spaced, AUR represent high alarm setting, t represent Service Instance start-up study;
By the parameter that is configured in cloud platform ELB components calculate maximum service number of requests that single Service Instance accepts, The parameter value of the average increment per second of single Service Instance, predicts following peaking capacity, pre-cooling Service Instance, has very high Application value.
As one embodiment, the polymerizing value growth rate k passes through formula:Really Fixed, wherein vi (t) represents the service request quantity of i-th of node of t in historical data, and n represents node total number, and △ t are represented The time interval of tasks carrying twice, ensure that subsequent treatment is normally carried out, be adapted to practical application.
As one embodiment, automatic elastic spreading parameter and service in the cloud platform ELB components according to reading Before collection swarm parameter is calculated, in addition to step:
The automatic elastic spreading parameter and service cluster parameter of cloud platform ELB components are configured, is configured and joined according to actual conditions Number, fully meets actual demand, is adapted to application.
As one embodiment, in the automatic elastic spreading parameter and service cluster ginseng of the configuration cloud platform ELB components After number, the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components according to reading carry out calculating it Before, in addition to step:
The log information of cloud platform ELB components is configured, journal function is realized, tracks detailed algorithm design cycle and calculating Method, it is adapted to practical application.
In order to more fully understand this method, a present invention detailed below is applied to the resource of cloud platform ELB components The application example of extended method, the application example are realized in WSO2-StratOS PaaS cloud platform ELB components.
WSO2Stratos is a PaaS platform increased income, and it is mainly that enterprise-level application development provides the solution based on cloud Certainly scheme, at present WSO2Stratos externally provide ApplicationServer (application server), Elastic Load 16 kinds of serviced components such as Balancer (elastic load balanced device), every kind of serviced component can be all responsible for providing specific function, And all serviced components are in communication with each other further through WebService (web application program), finally by it is all this A little serviced components constitute WSO2Stratos platforms, WSO2Stratos overall architectures, and WSO2Stratos is put down as a PaaS The solution that platform externally provides, realized by combining different serviced components.
As shown in figure 3, the application example may comprise steps of:
Step S301:Configure automatic elastic spreading parameter, the services set of WSO2-StratOS PaaS cloud platform ELB components Swarm parameter:
Automatic elastic spreading parameter configures:(1) resilient expansion is started, the automatic extension under default situations in ELB is Close, it is necessary to arrive<ELB_HOME>In/repository/conf/loadbalancer.conf files, to ' Loadbalancer' changes attribute, ' enable_autoscaler'='true';(2) autoscaler_task_delay tables Show the time interval between two tasks carryings;(3) server_startup_delay represents automatic extension task in startup In the maximum duration cycle of preceding wait, refer to because network problem Service Instance adds ELB time delay;(4)instances Represent minimum service example number in ELB;(5) enable_autoscaler indicates whether to start automatic elastic extension;(6) Autoscaler_task represents that automatic elastic extension determines generation task, is configured to org.wso2.carbon.mediator.au toscale.lbautoscale.task.ServiceRequestsIn FlightAutoscaler;(7)use_embedded_ Autoscaler indicates whether to use built-in automatic elastic expander;(8) size_of_cache represents cache size;(9) Autoscaler_service_epr represents that automatic elastic expander service terminal point is quoted, and is configured to https://host_ address:https_port/services/AutoscalerService;(10) autoscaler_task_interval tables Show the time interval (△ t, millisecond) of tasks carrying twice;(11) server_startup_interval represents that Service Instance opens Dynamic delay (t);(12) session_timeout represents session timeout;(13) whether fail_over standbies start;
Service cluster parameter configuration:(1) min_app_instances represents minimum service example number, automatic expander Extension is not less than the value downwards;(2) max_app_instances represents maximum Service Instance number, automatic expander to Upper extension is not over the value;(3) max_requests_per_second (Rps) single Service Instance maximum per second accepted Service request quantity;(4) iterations that rounds_to_average (r) extends automatically;(5)alarming_upper_rate (AUR) high alarm setting is represented, span is 0~1, and acquiescence is 0.7;(6) message_expiry_time represents that message is expired Time;
Step S302:The log information of WSO2-StratOS PaaS cloud platform ELB components is configured,<ELB_HOME>/ Log4j.logger.org.wso2.carbon.mediat is added in repository/conf/log4j.properties files Or.autoscale.lbautoscale.task.ServiceRequestsInFlightAut oscaler=DEBUG restart clothes Business device realizes journal function, tracks detailed algorithm design cycle and computational methods;
Step S303:Read automatic elastic spreading parameter, the services set of WSO2-StratOS PaaS cloud platform ELB components Swarm parameter, the maximum service number of requests maxRpt and single service accepted according to the single Service Instance of the parameter of reading calculating are real The average increment m per second of example, wherein the maximum service number of requests maxRpt that single Service Instance is accepted passes through formula:MaxRpt= Rps × △ t × AUR determine that the average increment m per second of single Service Instance passes through formula:It is determined that wherein Rps tables Show the single Service Instance maximum service number of requests per second accepted, △ t represent the time interval of tasks carrying twice, AUR tables Show high alarm setting, t represents Service Instance start-up study;
Step S304:Polymerizing value Vtotal is calculated according to service request quantity in the historical data of acquisition, average nodal is gathered Conjunction value Vavg and polymerizing value growth rate k, wherein polymerizing value Vtotal pass through formula:It is determined that average nodal is gathered Conjunction value Vavg passes through formula:It is determined that polymerizing value growth rate k passes through formula: It is determined that wherein vi (t) represents the service request quantity of i-th of node of t in historical data, n represents node total number, △ t tables Show the time interval of tasks carrying twice, in this application example interior joint sum n=1;
Step S305:By real time service number of requests compared with Service Instance maximum processing capability, the Service Instance Maximum processing capability is serviced real by the maximum service number of requests maxRpt that above-mentioned single Service Instance is accepted with real time execution Example number nIn real timeProduct determine;
Step S306:When real time service number of requests is more than above-mentioned Service Instance maximum processing capability, calculate above-mentioned poly- Conjunction value growth rate k and the average increment m per second of single Service Instance ratio, resource expansion is carried out according to the ratio;
Configuration parameter explanation:Time interval △ t 60000ms, the single Service Instance of tasks carrying are per second twice accepts Maximum service number of requests Rps 5, high alarm setting AUR 0.7, minimum service example number 1, Service Instance start-up study t 180000ms, i.e. autoscaler_task_interval (△ t) 60000ms, max_requests_per_second (Rps) 5、alarming_upper_rate(AUR)0.7、min_app_instances 1、server_startup_interval(t) 180000ms;
The maximum service number of requests that single Service Instance is accepted is calculated:
The average increment per second of single Service Instance:
The process one that resource expansion is carried out in WSO2-StratOS PaaS cloud platform ELB components can be such as table 1 below institute Show;
The process one of table 1
3rd iteration real time service number of requests as shown in table 1 is more than Service Instance maximum processing capability, that is, is more than described The maximum service number of requests maxRpt that single Service Instance is accepted and real time execution Service Instance number nIn real timeProduct, i.e., 250 >210, polymerizing value growth rate k and the average increment m per second of single Service Instance ratio are calculated, resource expansion is carried out according to this ratio Exhibition, the 3rd iteration polymerizing value growth rate k value is 4.15, then polymerizing value growth rate k and the average increment per second of single Service Instance M ratioRound, obtain needing when the 2nd iteration, 4 Service Instances should be started and carry out resource Extension, so when the 3rd iteration, in the case where considering startup of server delay, it can meet to ask performance requirement, Service Instance operating overload effectively is prevented, the problem of causing request delay or interrupt, in WSO2-StratOS PaaS cloud platforms As shown in Table 2 below the process two that resource expansion is carried out in ELB components can be;
The process two of table 2
Server_startup_interval (t) 180000ms in this application example, i.e. Service Instance start-up study t =180000ms, the time interval of each iteration is 1min, so 4 Service Instances started when the 2nd iteration, by Step starts, and has had been started up 1 when the 3rd iteration, real time execution Service Instance number nIn real time=2, the 4th iteration when Time has had been started up 2, real time execution Service Instance number nIn real time=3,3 have been had been started up when the 5th iteration, in real time Operation service example number nIn real time4 are had been started up when=4, the 6th iteration, real time execution Service Instance number nIn real time= 5。
Polymerizing value, average nodal polymerizing value and polymerizing value growth rate are introduced in ELB components in WSO2-StratOS Deng parameter value, by calculating the ratio of the average increment per second of polymerizing value growth rate and single Service Instance, prediction needs what is started The number of Service Instance, therefore, it is possible to which before request next time peak value arrives, the quantity of Service Instance can just meet performance Demand, overcome the service pressure overload that startup of server delay is brought, it is impossible to meet the pressure requirements of peak value request, also overcome The intrinsic algorithm errors of ELB components, reduce the extensive work that history matching is brought in paas frameworks, improve resource use Efficiency, it is cost-effective.
The resource expansion device of cloud platform ELB components in one embodiment, as shown in figure 4, including:
Processing module, for the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components according to reading, Calculate maximum service number of requests maxRpt and the average increment m per second of single Service Instance that single Service Instance is accepted, and root Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;
Comparison module, for by real time service number of requests compared with Service Instance maximum processing capability, the clothes The maximum service number of requests maxRpt and real time execution that pragmatic example maximum processing capability is accepted by the single Service Instance Service Instance number nIn real timeProduct determine;
Resource expansion module, for being more than the Service Instance maximum processing capability when the real time service number of requests When, the polymerizing value growth rate k and the average increment m per second of the single Service Instance ratio are calculated, is entered according to the ratio Row resource expansion.
As shown in figure 4, a preferred embodiment of each module annexation of the present apparatus is:Processing module, comparison module It is linked in sequence successively with resource expansion module.
Processing module first according to the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components of reading, Maximum service number of requests maxRpt and the average increment m per second of single Service Instance that single Service Instance is accepted are calculated, according to Service request quantity calculates polymerizing value growth rate k in the historical data of acquisition;Then comparison module is by real time service number requests Compared with Service Instance maximum processing capability;When real time service number requests are more than Service Instance maximum processing capability, Resource expansion module calculates polymerizing value growth rate k and the average increment m per second of single Service Instance ratio, is carried out according to the ratio Resource expansion.
It is evidenced from the above discussion that the resource expansion device of cloud platform ELB components of the present invention, effectively solves existing response type and expands The problem of exhibition causes user to need one section of time delay of experience, prediction type extension workload to meet real work demand greatly, It is adapted to application.
As one embodiment, the maximum service number of requests maxRpt that the single Service Instance is accepted passes through formula: MaxRpt=Rps × △ t × AUR determine that the average increment m per second of the single Service Instance passes through formula:Really Fixed, wherein Rps represents the single Service Instance maximum service number of requests per second accepted, △ t represent tasks carrying twice when Between be spaced, AUR represent high alarm setting, t represent Service Instance start-up study;
By the parameter that is configured in cloud platform ELB components calculate maximum service number of requests that single Service Instance accepts, The parameter value of the average increment per second of single Service Instance, predicts following peaking capacity, pre-cooling Service Instance, has very high Application value.
As one embodiment, the polymerizing value growth rate k passes through formula:Really Fixed, wherein vi (t) represents the service request quantity of i-th of node of t in historical data, and n represents node total number, and △ t are represented The time interval of tasks carrying twice, ensure that subsequent treatment is normally carried out, be adapted to practical application.
As one embodiment, in addition to configuration module, for configuring the automatic elastic spreading parameter of cloud platform ELB components With service cluster parameter;
The processing module according to the automatic elastic spreading parameters of the cloud platform ELB components read from the configuration module and Service cluster parameter, calculates maximum service number of requests maxRpt that single Service Instance accepts and single Service Instance is averagely every Second increment m, according to actual conditions configuration parameter, fully meets actual demand, is adapted to application.
As one embodiment, the configuration module is additionally operable to configure the log information of cloud platform ELB components, realizes daily record Function, detailed algorithm design cycle and computational methods are tracked, be adapted to practical application.
Embodiment described above only expresses the several embodiments of the present invention, and its description is more specific and detailed, but simultaneously Therefore the limitation to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that for one of ordinary skill in the art For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the guarantor of the present invention Protect scope.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.

Claims (8)

  1. A kind of 1. resource expansion method of cloud platform ELB components, it is characterised in that comprise the following steps:
    According to the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components of reading, single Service Instance is calculated The average increment m per second of maximum service number of requests maxRpt and single Service Instance accepted;
    Polymerizing value growth rate k is calculated according to service request quantity in the historical data of acquisition;The polymerizing value growth rate k passes through public affairs Formula:It is determined that wherein vi (t) represents the service request number of i-th of node of t in historical data Amount, n represent node total number, and △ t represent the time interval of tasks carrying twice;
    By real time service number of requests compared with Service Instance maximum processing capability, the Service Instance maximum processing capability Pass through the maximum service number of requests maxRpt that the single Service Instance is accepted and real time execution Service Instance number nIn real timeMultiply Product determines;
    When the real time service number of requests is more than the Service Instance maximum processing capability, the polymerizing value growth rate is calculated K and the average increment m per second of the single Service Instance ratio, resource expansion is carried out according to the ratio.
  2. 2. the resource expansion method of cloud platform ELB components according to claim 1, it is characterised in that the single service The maximum service number of requests maxRpt that example is accepted passes through formula:MaxRpt=Rps × △ t × AUR determinations, the single clothes The average increment m per second of pragmatic example passes through formula:Accepted it is determined that wherein Rps represents that single Service Instance is per second Maximum service number of requests, △ t represent the time interval of tasks carrying twice, and AUR represents high alarm setting, and t represents Service Instance Start-up study.
  3. 3. the resource expansion method of cloud platform ELB components according to claim 1, it is characterised in that described according to reading Before the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components taken are calculated, in addition to step:
    Configure the automatic elastic spreading parameter and service cluster parameter of cloud platform ELB components.
  4. 4. the resource expansion method of cloud platform ELB components according to claim 3, it is characterised in that in the configuration cloud After the automatic elastic spreading parameter and service cluster parameter of platform ELB components, the cloud platform ELB components according to reading Before automatic elastic spreading parameter and service cluster parameter are calculated, in addition to step:
    Configure the log information of cloud platform ELB components.
  5. A kind of 5. resource expansion device of cloud platform ELB components, it is characterised in that including:
    Processing module, for the automatic elastic spreading parameter and service cluster parameter of the cloud platform ELB components according to reading, calculate The average increment m per second of maximum service number of requests maxRpt and single Service Instance that single Service Instance is accepted, and according to obtaining Service request quantity calculates polymerizing value growth rate k in the historical data taken;
    Comparison module, for compared with Service Instance maximum processing capability, the service to be real by real time service number of requests Example maximum processing capability passes through the maximum service number of requests maxRpt that the single Service Instance is accepted and real time execution service Example number nIn real timeProduct determine;The polymerizing value growth rate k passes through formula:It is determined that wherein vi (t) the service request quantity of i-th of node of t in historical data is represented, n represents node total number, and △ t represent that two subtasks are held Capable time interval;
    Resource expansion module, for when the real time service number of requests is more than the Service Instance maximum processing capability, counting The polymerizing value growth rate k and the average increment m per second of the single Service Instance ratio are calculated, resource is carried out according to the ratio Extension.
  6. 6. the resource expansion device of cloud platform ELB components according to claim 5, it is characterised in that the single service The maximum service number of requests maxRpt that example is accepted passes through formula:MaxRpt=Rps × △ t × AUR determinations, the single clothes The average increment m per second of pragmatic example passes through formula:Accepted it is determined that wherein Rps represents that single Service Instance is per second Maximum service number of requests, △ t represent the time interval of tasks carrying twice, and AUR represents high alarm setting, and t represents Service Instance Start-up study.
  7. 7. the resource expansion device of cloud platform ELB components according to claim 6, it is characterised in that also include configuration mould Block, for configuring the automatic elastic spreading parameter and service cluster parameter of cloud platform ELB components;
    Automatic elastic spreading parameter and service of the processing module according to the cloud platform ELB components read from the configuration module Collect swarm parameter, calculate maximum service number of requests maxRpt and the average increasing per second of single Service Instance that single Service Instance is accepted Measure m.
  8. 8. the resource expansion device of cloud platform ELB components according to claim 7, it is characterised in that the configuration module It is additionally operable to configure the log information of cloud platform ELB components.
CN201410375345.5A 2014-07-31 2014-07-31 The resource expansion method and apparatus of cloud platform ELB components Active CN104135525B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410375345.5A CN104135525B (en) 2014-07-31 2014-07-31 The resource expansion method and apparatus of cloud platform ELB components

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410375345.5A CN104135525B (en) 2014-07-31 2014-07-31 The resource expansion method and apparatus of cloud platform ELB components

Publications (2)

Publication Number Publication Date
CN104135525A CN104135525A (en) 2014-11-05
CN104135525B true CN104135525B (en) 2017-12-01

Family

ID=51808047

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410375345.5A Active CN104135525B (en) 2014-07-31 2014-07-31 The resource expansion method and apparatus of cloud platform ELB components

Country Status (1)

Country Link
CN (1) CN104135525B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105068919B (en) * 2015-07-15 2018-04-13 华南理工大学 A kind of WS BPEL flow load testing instruments based on WSO2 Stratos
CN107643975A (en) * 2017-09-25 2018-01-30 郑州云海信息技术有限公司 A kind of method, apparatus and computer-readable recording medium for counting pressure information
CN113783799A (en) * 2020-06-09 2021-12-10 马上消费金融股份有限公司 Flow control method and device
CN113254224A (en) * 2021-07-15 2021-08-13 中电金信软件有限公司 Computing resource capacity expansion method and device, electronic equipment and readable storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103220337A (en) * 2013-03-22 2013-07-24 合肥工业大学 Cloud computing resource optimizing collocation method based on self-adaptation elastic control
CN103425535A (en) * 2013-06-05 2013-12-04 浙江大学 Agile elastic telescoping method in cloud environment
CN103778474A (en) * 2012-10-18 2014-05-07 华为技术有限公司 Resource load capacity prediction method, analysis prediction system and service operation monitoring system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778474A (en) * 2012-10-18 2014-05-07 华为技术有限公司 Resource load capacity prediction method, analysis prediction system and service operation monitoring system
CN103220337A (en) * 2013-03-22 2013-07-24 合肥工业大学 Cloud computing resource optimizing collocation method based on self-adaptation elastic control
CN103425535A (en) * 2013-06-05 2013-12-04 浙江大学 Agile elastic telescoping method in cloud environment

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"TeraPELB:云计算中基于预测的弹性负载均衡算法";吴和生等;《系统仿真学报》;20130917;第25卷(第8期);1751-1760 *

Also Published As

Publication number Publication date
CN104135525A (en) 2014-11-05

Similar Documents

Publication Publication Date Title
Liu et al. Aggressive resource provisioning for ensuring QoS in virtualized environments
Moreno et al. Improved energy-efficiency in cloud datacenters with interference-aware virtual machine placement
US8782215B2 (en) Performance testing in a cloud environment
US10783002B1 (en) Cost determination of a service call
Inomata et al. Proposal and evaluation of a dynamic resource allocation method based on the load of VMs on IaaS
CN112003763B (en) Network link monitoring method, monitoring device, monitoring equipment and storage medium
CN104135525B (en) The resource expansion method and apparatus of cloud platform ELB components
CN111966289B (en) Partition optimization method and system based on Kafka cluster
CN107566535B (en) Self-adaptive load balancing method based on concurrent access timing sequence rule of Web map service
CN103713935A (en) Method and device for managing Hadoop cluster resources in online manner
Ali et al. Cedule: A scheduling framework for burstable performance in cloud computing
Yin et al. System resource utilization analysis and prediction for cloud based applications under bursty workloads
Singh Study of response time in cloud computing
CN114780233A (en) Scheduling method and device based on microservice link analysis and reinforcement learning
US20160225043A1 (en) Determining a cost of an application
Tseng et al. An energy efficient vm management scheme with power-law characteristic in video streaming data centers
Liu et al. ScaleFlux: Efficient stateful scaling in NFV
EP3787231B1 (en) Server configuration method and apparatus
CN112817719A (en) Method, device and equipment for adjusting parameters of thread pool and readable storage medium
US11374869B2 (en) Managing bandwidth based on user behavior
CN111124669A (en) Operation method, system, terminal and storage medium of distributed SaaS software
Oral et al. Supporting performance isolation in software as a service systems with rich clients
Ibrahim PRESENCE: a framework for monitoring, modelling and evaluating the performance of cloud SaaS web services
CN111598390B (en) Method, device, equipment and readable storage medium for evaluating high availability of server
Hanczewski et al. A Multiparameter Analytical Model of the Physical Infrastructure of a Cloud-Based System

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant