CN109995864A - A kind of SDN load-balancing method based on fuzzy logic - Google Patents
A kind of SDN load-balancing method based on fuzzy logic Download PDFInfo
- Publication number
- CN109995864A CN109995864A CN201910260189.0A CN201910260189A CN109995864A CN 109995864 A CN109995864 A CN 109995864A CN 201910260189 A CN201910260189 A CN 201910260189A CN 109995864 A CN109995864 A CN 109995864A
- Authority
- CN
- China
- Prior art keywords
- load
- server
- fuzzy
- fuzzy logic
- sdn
- 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
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/101—Server selection for load balancing based on network conditions
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention discloses a kind of SDN load-balancing method based on fuzzy logic, key step are as follows: firstly, initialization OpenFlow network；Secondly, obtain server state by improved fuzzy logic equalization algorithm: load balancing module collects server state information by SDN switch, using cpu busy percentage, memory usage, I/O utilization rate and session rate as the input variable of fuzzy logic, and subordinating degree function is defined respectively, corresponding fuzzy class is obtained by improved fuzzy reasoning, and completes the judge of server state by fuzzy matrix；Then, calculation server load value calculates OpenFlow Network Load Balance parameter δ；Finally, setting load balancing adjusts threshold value.Server cluster problem of load balancing is solved using the characteristic that SDN control plane is separated with data plane, solves the problem of conventional load equilibrium hardware price expensive scalability, poor reliability.
Description
Technical field
The invention belongs to the network optimizations, network measure field, and in particular to a kind of SDN load balancing based on fuzzy logic
Method.
Background technique
In recent years, with internet+, the continuous rise and development of the technologies such as e-commerce and big data, network user's number
Proportional increase is measured, the scale and flow of internet are at explosive growth.It virtualizes simultaneously, the quick hair of cloud computing technology
Exhibition, is also a huge challenge to network ability to bear.In order to meet network service demand, many Internet enterprises would generally
The reasonable utilization to resource is realized by load-balancing technique, and high quality, highly reliable service are provided.Certainly due to conventional internet
The defect of body, existing load-balancing device is expensive, while also lacking enough scalabilities and flexibility.
Software defined network (Software Defined Network, SDN) is a kind of optimization and simplified network operation
System structure mode.It will application with the interaction between network service, equipment is closer links together, no matter they are objects
It is reason or virtualization.Core technology one of of the OpenFlow agreement as SDN may be implemented to exchange control layer from tradition
It is independent in equipment, to realize to the more flexible control of network implementation.
Fuzzy logic may be implemented to express fuzzy set with mathematical formulae, solves many complicated and can not establish accurate mathematical mould
The problem of type.In the Task Scheduling Model, each node load situation is in non-linear and unpredictability.Due to collecting nodal information
The limitation of technology, obtains and report information requires to spend extra time.The information of middleware preservation in this way, can only indicate over
The node load information at nearlyr moment, rather than loading condition locating at present, system have certain time delay.In view of receiving
The precision problem for collecting server state assesses the load condition of node, and obtained only estimator is more suitable for blurring
Language is stated.
In conclusion SDN control forwards isolated framework by controller as core, right for load balancing
The monitoring of flow, Real-Time Scheduling in network are highly suitable as data center network and provide load balancing service, and Virtual Service
Device state is also more suitable for the language expression with blurring.There are also scholars to the load-balancing technique progress based on SDN at present
Research, but research concentrates on the balancing link load of OpenFlow network substantially, to VIRTUAL SERVER LOAD Equilibrium Research
It is less.
The invention proposes a load balancing based on SDN, between the parameter by analyzing influence load balancing
Interrelated influence obtains the loading condition of virtual server using fuzzy logic algorithm, and the present load of analysis selection in real time is most
Light virtual server, and server suspend mode/restart strategy is set, to effectively avoid load imbalance, improve Internet resources
Utilization rate.For the validity of authentication policy, the software defined network emulation platform of software and hardware combining has been built, has passed through experiment pair
Traditional load balancing and strategy proposed by the present invention compare, the results showed that, when inventing the response of the system of proposition
Between it is very fast, curvilinear motion is more steady, and system can obtain preferable performance.
Load balancing mainly has hardware realization and software realization two ways.Wherein, hardware load is balanced, is to utilize hardware
Equipment (load balancer) realizes load balancing.By its dedicated asic chip, very high process performance, but its can achieve
Price is generally sufficiently expensive.Software realization is to utilize corresponding load balancing software, and configure related protocol, to reach load
Balanced purpose.The configuration of such method is simple, deployment implementation is more flexible, cost is more cheap relative to hardware realization.But
It is due to the restriction by conventional network architecture, these solutions can not meet both flexible and efficient requirement simultaneously,
And do not have stronger versatility.
Domestic and foreign scholars mainly utilize OpenFlow technology at present, load to network link and server respectively
The research of weighing apparatus is such as directed to the LABERIO algorithm of link, and the framework and LOBUS of load balancing are carried out based on server cpu busy percentage
There are words algorithm, and the load-balancing algorithm suitable for WEB server etc..But these algorithms are considering load balancing influence ginseng
Less comprehensive when number, load-balancing algorithm is relatively simple, ununified strategy execution load-balancing algorithm.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, propose a kind of SDN load balancing based on fuzzy logic
Method.
The present invention is the technical problem solved in background technique, and the technical solution of proposition is as follows: one kind being based on fuzzy logic
SDN load-balancing method, include the following steps:
1) OpenFlow network is initialized；
2) obtain server state by improved fuzzy logic equalization algorithm: load balancing module is received by SDN switch
Collect server state information, becomes cpu busy percentage, memory usage, I/O utilization rate and session rate as the input of fuzzy logic
Amount, and subordinating degree function is defined respectively, corresponding fuzzy class is obtained by improved fuzzy reasoning, and complete by fuzzy matrix
At the judge of server state；
3) calculation server load value calculates OpenFlow Network Load Balance parameter δ；
4) setting load balancing adjusts threshold value:
(1) judge whether current network load balance parameters δ exceeds the maximum load upper limit: if so, server load is in
Saturation state opens a server with server migration strategy；If it is not, performing the next step；
(2) judge whether current network load balance parameters δ exceeds the average load upper limit: being calculated if so, enabling fuzzy logic
Method carries out load balancing adjustment, and current Web request is transmitted to the smallest server of load；If it is not, performing the next step；
(3) judge whether current network load balance parameters δ is more than the minimum load upper limit: if so, illustrating that current virtual takes
The load distribution of business device is more balanced, chooses whether to continue to monitor entire server state；If it is not, performing the next step；
(4) when current network load balance parameters δ is less than the minimum load upper limit, server load is in idle state, fortune
With server migration strategy, one server of suspend mode.
The maximum load upper limit is set as 0.8 according to experiment experience in the step 4), minimum load lower limit is 0.2, average
Load threshold is 0.5.
Improved fuzzy logic equalization algorithm uses the SDN network framework for supporting OpenFlow agreement in the step 2),
Mainly available data forwarding control in OpenDayLight controller is extended, according to Virtual Service in OpenFlow network
The load condition of device carries out control forwarding, to realize server load balancing in network.
The system of fuzzy logic is divided into three parts: blurring, fuzzy reasoning, defuzzification in the step 2), to mould
Gelatinization and fuzzy reasoning part are designed and have been improved；
Firstly, characterizing the performance parameter of server load situation in analysis SDN network framework, cpu busy percentage, memory are selected
Utilization rate, I/O utilization rate and session number are patrolled as the load parameter to server node performance evaluation, and using parameter as fuzzy
The input variable of algorithm is collected, and then defines subordinating degree function for each input variable, to the membership between variable and sequence
It is mapped, determines corresponding fuzzy class sequence；
Secondly, most important component part is fuzzy rule base in fuzzy reasoning, classical fuzzy rule is to utilize
Numerical variable constitutes fuzzy rule using linguistic variable；
The present invention improves fuzzy reasoning, realizes that set of factors to the judge of Comment gathers, obtains by building fuzzy matrix
A possibility that virtual server being candidate server size degree.
Beneficial effect
1, server cluster problem of load balancing is solved using the characteristic that SDN control plane is separated with data plane, solved
The problems such as conventional load equilibrium hardware price is expensive, scalability, poor reliability.
2, the load balancing proposed in patent has adjusted VIRTUAL SERVER LOAD, makes the resource utilization total energy of system
It is maintained at a reasonable range, improves system performance.
3, the improved load-balancing algorithm based on fuzzy logic in patent, by problem of load balancing by fuzzy theory into
Row solves, and obtains reasonable, accuracy service device load condition, provides reliable foundation for the implementation of load balancing, drops
The low response time of system, the server load balancing reached.
Detailed description of the invention
Fig. 1 is design framework figure.
Fig. 2 is fuzzy logic system structure chart.
Fig. 3 is load balancing flow chart.
Fig. 4 is system response time figure.
Fig. 5 is system request success rate figure.
Specific embodiment
Below by specific embodiments and the drawings, the present invention is further illustrated.The embodiment of the present invention is in order to more
So that those skilled in the art is more fully understood the present invention well, any limitation is not made to the present invention.
The present invention is a kind of SDN load-balancing method based on fuzzy logic, using the SDN net for supporting OpenFlow agreement
Network framework is mainly extended available data forwarding control in OpenDayLight controller, according in OpenFlow network
The load condition of virtual server carries out control forwarding, to realize server load balancing in network.Since OpenFlow is handed over
It changes planes and unified interface and data forwarding capability is provided controller, controller carries out unified control to the forwarding of OpenFlow flow table
System.Interchanger periodically collects the operating status of virtual server, and result is fed back to OpenDayLight controller, and according to
Load-balancing algorithm calculates the operating status of current working status virtual server, and then according to load balancing, issues control
System instruction, the quantity that the node for keeping processing capacity strong obtains processing task is relatively more, stops when necessary to virtual server realization
It sleeps/restarts, reach load balancing.Design framework proposed by the present invention is as shown in Figure 1.
1, based on the load-balancing algorithm of fuzzy logic
Since the complexity that server calculates makes the resources such as cpu busy percentage, the I/O utilization rate of offer service node
It is a fuzzy concept with load state, system is monitoring load state without accurate mathematical model and control rule
When, consider that fuzzy mathematics is describing the advantage on sudden problem and uncertain problem, fuzzy logic theory is introduced into this
To optimize load balancing strategy in system.Fuzzy logic system can be divided into three parts: blurring, fuzzy reasoning, ambiguity solution
Change.The structure of fuzzy logic system is as shown in Figure 2:
1.1 blurring
Input quantity is carried out change of scale first by blurring, transforms to respective domain range, carries out Fuzzy processing, is used
Fuzzy set indicates.Fuzzification process needs to determine corresponding fuzzy class sequence for every class input variable, such as it is high, in, it is low }.
Each variable corresponding grade sequential element has a degree of membership, it is therefore desirable to selected subordinating degree function, to variable and sequence it
Between membership mapped.
Under Virtual Server Cluster environment, there are many factor that can characterize server load situation.Such as the dominant frequency of CPU
And utilization rate, the size and utilization rate of memory, network speed and network response time, currently performed number of processes etc..Due to carrying out
When task schedule, call request and returns the result and will pass through network transmission；CPU and memory source are occupied when task execution.
For these features, cpu busy percentage μ (C), memory usage μ (M), I/O utilization rate μ (IO) and session number μ (D) conduct are selected
To the load parameter of server node performance evaluation, OpenFlow interchanger timing submits VIRTUAL SERVER LOAD shape to controller
State.
Ambiguity in definition subordinating degree function:
(1) cpu busy percentage
Using utilization rate (C) value of current server as domain, μ is enabledh(C)、μm(C)、μlIt (C) is respectively that cpu busy percentage is subordinate to
In fuzzy subset's " current server cpu busy percentage is high ", " current server cpu busy percentage is medium ", " current server CPU benefit
It is low with rate " degree of membership, then μh(C)、μm(C)、μl(C) it can use:
(2) memory usage
Using utilization rate (M) value of current memory as domain, μ is enabledh(M)、μm(M)、μlIt (M) is respectively that memory usage is subordinate to
In fuzzy subset's " current server memory usage is high ", " current server memory usage is medium ", " current server memory
Utilization rate is low " degree of membership, then μh(M)、μm(M)、μl(M) it can use:
(3) I/O utilization rate
Using utilization rate (I/O) value of current I/O as domain, μ is enabledh(IO)、μm(IO)、μlIt (IO) is respectively that I/O utilization rate is subordinate to
Belong to fuzzy subset's " current server I/O utilization rate is high ", " current server I/O utilization rate is medium ", " current server I/O
Utilization rate is low " degree of membership, then μh(IO)、μm(IO)、μl(IO) it can use:
(4) current sessions number
With Nginx+tomcat configuration virtual server highest load session number for 800, and connected with current virtual server
(D) value for connecing session number utilization rate is domain, enables μh(D)、μm(D)、μlIt (D) is respectively that connection session number is under the jurisdiction of fuzzy subset
" current server connect session number higher ", " current server session number is medium ", " current server session number is lower " person in servitude
Category degree, then μh(D)、μm(D)、μl(D) it can use:
By subordinating degree function after aforementioned four input variable Fuzzy processing, so that it may to VIRTUAL SERVER LOAD shape
Condition makes Comprehensive Evaluation.Unique output of fuzzy logic inference system is the possibility which Virtual Service request should distribute to
Property, it is indicated with R.When carrying out fuzzy evaluation to μ, set of factors can be taken as E={ C, M, IO, D }, and Comment gathers can be taken as F
=it is high, in, it is low }.
1.2 fuzzy reasoning
Fuzzy reasoning is the core of fuzzy controller, the reasoning process based in fuzzy logic implication relation and reasoning rule
Then, it has simulation inferential capability of the people based on fuzzy concept.Most important component part is Fuzzy Control in fuzzy logic inference
Rule base processed, classical fuzzy rule are using numerical variable or to utilize the new method of linguistic variable composition fuzzy rule.This
Invention realizes the judge of set of factors to Comment gathers by building fuzzy matrix.Comprehensive Evaluation is according to following steps:
(1) single factor evaluation is carried out to μ, the fuzzy matrix for indicating fuzzy relation between E and F is constructed using result.The mould
The output of fuzzy logic be overburden in the case of, the size of assignment of traffic to the virtual server possibility, if to cpu busy percentage,
The evaluation result of memory usage, I/O utilization rate and session number utilization rate factor is fuzzy vector:
R1=[uh(C),um(C),ul(C)]
R2=[uh(M),um(M),ul(M)]
R3=[uh(IO),um(IO),ul(IO)]
R4=[uh(D),um(D),ul(D)]
Four vectors constitute the fuzzy matrix R=[R1, R2, R3, R4] of E and F.
(2) weight vector P=[p1, p2, p3, p4] is determined, wherein p1, p2, p3, p4 respectively indicate factor CPU, memory, I/
O, significance level of the session utilization rate in u, p1+p2+p3+p4=1.
(3) make blurring mapping Q=PR, gained fuzzy vector Q is exactly each virtual server commenting on comment set F
Sentence the degree as a result, a possibility that wherein three representation in components virtual servers are candidate servers size.
1.3 defuzzification
Since the output obtained by fuzzy reasoning is still fuzzy variable F=(L, M, H), need to carry out defuzzification processing
To obtain exact output numerical value.Defuzzification can be carried out using classical centre of area method, i.e., with each fuzzy class pair
Definite output of the mass center for the membership function answered as the fuzzy class.The corresponding mass center of fuzzy class L is 0.15, fuzzy class
The corresponding mass center of M is 0.5, and the corresponding mass center of fuzzy class H is 0.85.Finally we can calculate fuzzy patrol using following formula
Collect the exact numerical values recited of output:
M in formulaiIndicate the center of mass values of each output fuzzy class, wiIndicate that corresponding output fuzzy class is MiWeight.
2, the load balancing based on SDN framework
The load balancing based on SDN that the present invention realizes, based on SDN framework, load-balancing algorithm is core,
Service load balancing, specific implementation strategy are completed in conjunction with classical Dynamic Load-balancing Algorithm are as follows:
(1) OpenFlow network is initialized；
(2) load balancing module completes Web request by classical polling algorithm, and load balancing module is exchanged by SDN
Machine collects server state information；
(3) the fuzzy logic algorithm calculation server load value proposed through the invention calculates OpenFlow network load
Balance parameters；
Wherein, i is server number, FavgFor the average load of current time server, FiFor the service value of server i,
N is server node number in OpenFlow network.Load balancing parameter δ is with the form calculus of variance.δ is smaller, network load
It is more balanced.
(4) setting load balancing adjusts threshold value, the maximum load upper limit is set in the present invention as 0.8, minimum load lower limit is
0.2, average load threshold value is 0.5.Fuzzy logic algorithm is enabled when current network load balance parameters δ is greater than 0.5 to be loaded
Current Web request is transmitted to the smallest server of load by equilibrium adjustment；When current network load balance parameters δ is less than 0.2,
Server load is in idle state, with server migration strategy, closes a server；Current network load balance parameters
When δ is greater than 0.8, server load is in a saturated state, with server migration strategy, opens a server；Current network
When load balancing parameter δ is greater than 0.2 less than 0.5, illustrates that the load distribution of current virtual server is more balanced, can continue to monitor
Entire server state.The execution flow chart of strategy is as shown in Figure 3:
3, confirmatory experiment
For the performance for verifying the fuzzy logic load-balancing algorithm proposed by the present invention based on SDN.Emulate the test used
Environment is Ubuntu11.04 operating system, Intel Pentium DualE2180 central processing unit, double-core 4GB memory host.
Experiment selects open source software Mininet2.0 to build OpenFlow network, uses the H3C for supporting OpenFlow1.3 agreement
S6820-4C interchanger realizes the load-balancing method using the OpenDayLight controller based on Java.By
Applied in network performance test tool Iperf software is installed in the running environment of Mininet and generates flow pressure, algorithm performance is surveyed
Examination.The network that wherein Mininet is built is opened up to be interconnected for three OpenFlow interchangers, and initial virtual server is 4, considers clothes
The frequent suspend mode of business device and restart and also will affect system performance, virtual server minimum number is 4.Maximum bandwidth between interchanger
For 10 Mbits/s, the maximum transmission bandwidth of interchanger and server is 100 Mbits/s.
In order to verify the superiority for the fuzzy logic load-balancing algorithm based on SDN that invention proposes, adopted under SDN framework
Compared with the algorithm performance that Weighted Round Robin (WRR), weighting Smallest connection number precedence method (WLC) are realized with the present invention
Compared with three kinds of algorithms use identical hardware environment.In invention Indistinct Input acquire information be service node cpu busy percentage,
Memory usage, I/O utilization rate and session number utilization rate, the information that subsequent fuzzification function acquires these indicate related resource
Fuzziness, and then judge the load state of virtual server, the SDN load balancing of proposition is according to Fig. 3 flow implementation.Through
Experimental test, weight corresponding to four parameters is respectively 0.3,0.3,0.2,0.2.Test request is under big flow network request
The performance of system relates generally to logging in system by user operation, which not only has static page, and there are also database manipulations, belong to
Dynamic requests, the data requested by Iperf pressure test.
Request connection number linearly increasing as unit of minute in test, per minute 2000 connection requests of increase, 10 minutes
Request connection number is linearly reduced as unit of minute afterwards, reduces 2000 per minute.That is:
The test experiments time is 20 minutes, and sampling in every 1 minute is primary, and experiment is repeated 5 times, at that time using average value as system
Average response time, test results are shown in figure 4, the average success response rate of system as shown in Figure 5 shown in.
Experimental data show network request increase sharply and swash subtract in the case where the framework based on SDN WRR algorithm, WLC
The change curve of 3 kinds of policy response time of algorithm and load-balancing algorithm proposed by the present invention and the histogram of request success rate.
From experimental result it can be seen that
(1) when number of requests is less, task management and scheduling work is simpler, and 3 kinds of tactful response times are more close.
But with the increase of task quantity, task management and scheduling becomes increasingly complex, and invention proposes the dispatcher of fuzzy logic profile
The response time of system is less than Weighted Round Robin and weighting Smallest connection figures method, there is very big advantage；With task management with
Scheduling is more and more simpler, and algorithm is obvious with respect to other two kinds of algorithm advantages.
(2) propose that two wave crests occurs in the response time of algorithm in invention.First is: with the increasing of request task quantity
Add, current server bearing capacity reaches peak, and fuzzy logic load balancing proposed by the present invention detects point of load
With preset value is greater than, controller triggers load dispatch mechanism, starts a virtual server, so that load is reached dynamic equilibrium, is
The response time of system occurs gliding and tend towards stability, and second is: with the reduction of request task quantity, current server load
Lower and lower, fuzzy logic load balancing proposed by the present invention detects that the distribution of load is less than preset value, controller touching
Load dispatch mechanism is sent out, one virtual server of suspend mode makes load reach dynamic equilibrium, and gliding simultaneously occurs in the response time of system
It tends towards stability.
(3) it can be seen that the fuzzy logic load proposed by the present invention based on SDN from system request average success rate figure
Balance policy request success rate with higher, can reach 95%.
It should be understood that embodiment and example discussed herein simply to illustrate that, to those skilled in the art
For, it can be improved or converted, and all these modifications and variations all should belong to the protection of appended claims of the present invention
Range.
Claims (4)
1. a kind of SDN load-balancing method based on fuzzy logic, it is characterised in that, include the following steps:
1) OpenFlow network is initialized；
2) obtain server state by improved fuzzy logic equalization algorithm: load balancing module is collected by SDN switch and is taken
It is engaged in device status information, using cpu busy percentage, memory usage, I/O utilization rate and session rate as the input variable of fuzzy logic,
And subordinating degree function is defined respectively, corresponding fuzzy class is obtained by improved fuzzy reasoning, and complete by fuzzy matrix
The judge of server state；
3) calculation server load value calculates OpenFlow Network Load Balance parameter δ；
4) setting load balancing adjusts threshold value:
(1) judge whether current network load balance parameters δ exceeds the maximum load upper limit: if so, server load is in saturation
State opens a server with server migration strategy；If it is not, performing the next step；
(2) judge whether current network load balance parameters δ exceeds the average load upper limit: if so, enable fuzzy logic algorithm into
Current Web request is transmitted to the smallest server of load by the adjustment of row load balancing；If it is not, performing the next step；
(3) judge whether current network load balance parameters δ is more than the minimum load upper limit: if so, illustrating current virtual server
Load distribution it is more balanced, choose whether to continue to monitor entire server state；If it is not, performing the next step；
(4) when current network load balance parameters δ is less than the minimum load upper limit, server load is in idle state, with clothes
Business device migration strategy, one server of suspend mode.
2. a kind of SDN load-balancing method based on fuzzy logic according to claim 1, it is characterised in that, the step
It is rapid 4) in the maximum load upper limit set as 0.8 according to experiment experience, minimum load lower limit is 0.2, and average load threshold value is 0.5.
3. a kind of SDN load-balancing method based on fuzzy logic according to claim 1, it is characterised in that, the step
It is rapid 2) in improved fuzzy logic equalization algorithm using the SDN network framework for supporting OpenFlow agreement, it is main right
Available data forwarding control is extended in OpenDayLight controller, according in OpenFlow network virtual server it is negative
Load state carries out control forwarding, to realize server load balancing in network.
4. a kind of SDN load-balancing method based on fuzzy logic according to claim 1, it is characterised in that, the step
It is rapid 2) in the system of fuzzy logic be divided into three parts: blurring, fuzzy reasoning, defuzzification, to blurring and fuzzy reasoning
It is designed and has been improved in part；
Firstly, characterizing the performance parameter of server load situation in analysis SDN network framework, cpu busy percentage, memory is selected to utilize
Rate, I/O utilization rate and session number are calculated as the load parameter to server node performance evaluation, and using parameter as fuzzy logic
The input variable of method, and then subordinating degree function is defined for each input variable, the membership between variable and sequence is carried out
Mapping, determines corresponding fuzzy class sequence；
Secondly, most important component part is fuzzy rule base in fuzzy reasoning, classical fuzzy rule is to utilize numerical value
Variable constitutes fuzzy rule using linguistic variable；
Fuzzy reasoning is improved, realizes that set of factors to the judge of Comment gathers, obtains Virtual Service by building fuzzy matrix
The degree of a possibility that device is candidate server size.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811550672.4A CN109451052A (en) | 2018-12-18 | 2018-12-18 | A kind of SDN load-balancing method based on fuzzy logic |
| CN2018115506724 | 2018-12-18 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109995864A true CN109995864A (en) | 2019-07-09 |
Family
ID=65560225
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811550672.4A Withdrawn CN109451052A (en) | 2018-12-18 | 2018-12-18 | A kind of SDN load-balancing method based on fuzzy logic |
| CN201910260189.0A Pending CN109995864A (en) | 2018-12-18 | 2019-04-02 | A kind of SDN load-balancing method based on fuzzy logic |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811550672.4A Withdrawn CN109451052A (en) | 2018-12-18 | 2018-12-18 | A kind of SDN load-balancing method based on fuzzy logic |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN109451052A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112187905A (en) * | 2020-09-22 | 2021-01-05 | 烽火通信科技股份有限公司 | Load balancing interface table maintenance method, equipment and system |
| CN113055477A (en) * | 2021-03-15 | 2021-06-29 | 广州市华奕电子科技有限公司 | Method for realizing TCP communication mutual loading balance by middleware software |
| CN113778683A (en) * | 2021-09-14 | 2021-12-10 | 码客工场工业科技(北京)有限公司 | Handle identification system analysis load balancing method based on neural network |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112291161B (en) * | 2020-10-10 | 2022-03-11 | 燕山大学 | Time-sensitive network mixed flow scheduling method |
| CN112346863B (en) * | 2020-10-28 | 2024-06-07 | 河北冀联人力资源服务集团有限公司 | Method and system for processing dynamic adjustment data of computing resources |
| CN112860410A (en) * | 2021-03-08 | 2021-05-28 | 北京睿芯高通量科技有限公司 | Method for enhancing hierarchical control of production system |
| CN113067863B (en) * | 2021-03-17 | 2022-04-19 | 西安电子科技大学 | Dynamic load balancing method based on software defined network |
| CN114422517A (en) * | 2022-01-24 | 2022-04-29 | 广东三合电子实业有限公司 | Server load balancing system and method thereof |
| US11934297B2 (en) | 2022-07-27 | 2024-03-19 | The Toronto-Dominion Bank | System and method for testing applications |
| CN115665160B (en) * | 2022-10-14 | 2024-02-20 | 四川轻化工大学 | Multi-access edge computing system and method for electric power safety tools |
| CN116976681A (en) * | 2023-09-22 | 2023-10-31 | 山东通维信息工程有限公司 | Intelligent operation and maintenance decision system for electromechanical facilities of expressway based on big data |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8311973B1 (en) * | 2011-09-24 | 2012-11-13 | Zadeh Lotfi A | Methods and systems for applications for Z-numbers |
| CN103605418A (en) * | 2013-10-23 | 2014-02-26 | 曙光信息产业(北京)有限公司 | Power consumption adjusting method and device for cluster servers |
| US20140095811A1 (en) * | 2012-09-28 | 2014-04-03 | International Business Machines Corporation | Fuzzy counters for nvs to reduce lock contention |
| US20140129715A1 (en) * | 2012-11-07 | 2014-05-08 | Yahoo! Inc. | Method and system for work load balancing |
| CN103942738A (en) * | 2014-05-16 | 2014-07-23 | 国家电网公司 | Comprehensive evaluation method and system for electric energy meter |
| US20150178365A1 (en) * | 2013-12-19 | 2015-06-25 | Stretchr, Inc. | System And Method For Implementing Nested Relationships Within A Schemaless Database |
| CN105282043A (en) * | 2014-06-20 | 2016-01-27 | 中国电信股份有限公司 | Global network load balancing system, device and method |
| CN105391797A (en) * | 2015-12-03 | 2016-03-09 | 北京航空航天大学 | SDN-based cloud server load balancing method and device |
| CN105743985A (en) * | 2016-03-24 | 2016-07-06 | 国家计算机网络与信息安全管理中心 | Virtual service migration method based on fuzzy logic |
| CN106878420A (en) * | 2017-02-17 | 2017-06-20 | 北京天融信网络安全技术有限公司 | A kind of server load balancing method and device based on SDN |
| CN107689919A (en) * | 2017-09-20 | 2018-02-13 | 北京科技大学 | The dynamic adjustment weight fuzzy routing method of SDN |
| CN108174411A (en) * | 2017-12-12 | 2018-06-15 | 扬州大学 | Wireless sensor network routing method based on fuzzy logic |
| CN108282523A (en) * | 2018-01-18 | 2018-07-13 | 东华大学 | A kind of SiteServer LBS based on SDN |
| CN108712760A (en) * | 2018-03-29 | 2018-10-26 | 北京邮电大学 | High-throughput relay selection method based on random Learning Automata and fuzzy algorithmic approach |
| CN108881110A (en) * | 2017-05-10 | 2018-11-23 | 全球能源互联网研究院 | A kind of safety situation evaluation and defence policies joint decision method and system |
-
2018
- 2018-12-18 CN CN201811550672.4A patent/CN109451052A/en not_active Withdrawn
-
2019
- 2019-04-02 CN CN201910260189.0A patent/CN109995864A/en active Pending
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8311973B1 (en) * | 2011-09-24 | 2012-11-13 | Zadeh Lotfi A | Methods and systems for applications for Z-numbers |
| US20140095811A1 (en) * | 2012-09-28 | 2014-04-03 | International Business Machines Corporation | Fuzzy counters for nvs to reduce lock contention |
| US20140129715A1 (en) * | 2012-11-07 | 2014-05-08 | Yahoo! Inc. | Method and system for work load balancing |
| CN103605418A (en) * | 2013-10-23 | 2014-02-26 | 曙光信息产业(北京)有限公司 | Power consumption adjusting method and device for cluster servers |
| US20150178365A1 (en) * | 2013-12-19 | 2015-06-25 | Stretchr, Inc. | System And Method For Implementing Nested Relationships Within A Schemaless Database |
| CN103942738A (en) * | 2014-05-16 | 2014-07-23 | 国家电网公司 | Comprehensive evaluation method and system for electric energy meter |
| CN105282043A (en) * | 2014-06-20 | 2016-01-27 | 中国电信股份有限公司 | Global network load balancing system, device and method |
| CN105391797A (en) * | 2015-12-03 | 2016-03-09 | 北京航空航天大学 | SDN-based cloud server load balancing method and device |
| CN105743985A (en) * | 2016-03-24 | 2016-07-06 | 国家计算机网络与信息安全管理中心 | Virtual service migration method based on fuzzy logic |
| CN106878420A (en) * | 2017-02-17 | 2017-06-20 | 北京天融信网络安全技术有限公司 | A kind of server load balancing method and device based on SDN |
| CN108881110A (en) * | 2017-05-10 | 2018-11-23 | 全球能源互联网研究院 | A kind of safety situation evaluation and defence policies joint decision method and system |
| CN107689919A (en) * | 2017-09-20 | 2018-02-13 | 北京科技大学 | The dynamic adjustment weight fuzzy routing method of SDN |
| CN108174411A (en) * | 2017-12-12 | 2018-06-15 | 扬州大学 | Wireless sensor network routing method based on fuzzy logic |
| CN108282523A (en) * | 2018-01-18 | 2018-07-13 | 东华大学 | A kind of SiteServer LBS based on SDN |
| CN108712760A (en) * | 2018-03-29 | 2018-10-26 | 北京邮电大学 | High-throughput relay selection method based on random Learning Automata and fuzzy algorithmic approach |
Non-Patent Citations (5)
| Title |
|---|
| AMIRHOSSEIN MORAVEJOSHARIEH等: "A Fuzzy Logic Approach To Increase Quality of Service in Software Defined Networking", 《2018 INTERNATIONAL CONFERENCE ON ADVANCES IN COMPUTING, COMMUNICATION CONTROL AND NETWORKING (ICACCCN)》 * |
| JUN LI，XIANGQING CHANG等: "An Effective Path Load Balancing Mechanism Based on SDN", 《2014 IEEE 13TH INTERNATIONAL CONFERENCE ON TRUST, SECURITY AND PRIVACY IN COMPUTING AND COMMUNICATIONS》 * |
| SHAOJUN ZHANG等: "Online Load Balancing for Distributed Control Plane in Software-Defined Data Center Network", 《IEEE ACCESS》 * |
| 刘小楷: "云计算环境中基于自适应模糊控制的负载均衡算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 梁凯: "基于Fuzzy的负载调度算法在CDN中的应用研究", 《洛阳理工学院学报（自然科学版） 计算机与数理》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112187905A (en) * | 2020-09-22 | 2021-01-05 | 烽火通信科技股份有限公司 | Load balancing interface table maintenance method, equipment and system |
| CN113055477A (en) * | 2021-03-15 | 2021-06-29 | 广州市华奕电子科技有限公司 | Method for realizing TCP communication mutual loading balance by middleware software |
| CN113055477B (en) * | 2021-03-15 | 2022-06-14 | 广州市华奕电子科技有限公司 | Method for realizing TCP communication mutual loading balance by middleware software |
| CN113778683A (en) * | 2021-09-14 | 2021-12-10 | 码客工场工业科技(北京)有限公司 | Handle identification system analysis load balancing method based on neural network |
| CN113778683B (en) * | 2021-09-14 | 2024-05-28 | 码客工场工业科技(北京)有限公司 | Handle identification system analysis load balancing method based on neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109451052A (en) | 2019-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109995864A (en) | A kind of SDN load-balancing method based on fuzzy logic | |
| Liu et al. | Adaptive asynchronous federated learning in resource-constrained edge computing | |
| Chen et al. | Asynchronous online federated learning for edge devices with non-iid data | |
| CN104636197B (en) | A kind of evaluation method of data center's virtual machine (vm) migration scheduling strategy | |
| Vu et al. | A traffic and power-aware algorithm for virtual machine placement in cloud data center | |
| Sun et al. | MARVEL: Enabling controller load balancing in software-defined networks with multi-agent reinforcement learning | |
| CN106385468B (en) | The predictable dynamic load balancing method of Web cluster | |
| Fakhfakh et al. | Simulation tools for cloud computing: A survey and comparative study | |
| CN106534318B (en) | A kind of OpenStack cloud platform resource dynamic scheduling system and method based on flow compatibility | |
| US20120060167A1 (en) | Method and system of simulating a data center | |
| CN105721577B (en) | Software defined network-oriented server load balancing method | |
| CN106649085A (en) | Cloud computing-based software test system | |
| CN110231976A (en) | A kind of edge calculations platform container dispositions method and system based on load estimation | |
| Fallah et al. | NASLA: Novel auto scaling approach based on learning automata for web application in cloud computing environment | |
| CN106790381A (en) | Dynamic feedback of load equalization methods based on weighting Smallest connection | |
| Guo et al. | A container scheduling strategy based on neighborhood division in micro service | |
| CN114726743B (en) | Service function chain deployment method based on federal reinforcement learning | |
| Lan et al. | Task partitioning and orchestration on heterogeneous edge platforms: The case of vision applications | |
| Mahallat | ASTAW: auto-scaling threshold-based approach for web application in cloud computing environment | |
| Zhang | Storage optimization algorithm design of cloud computing edge node based on artificial intelligence technology | |
| Zhu et al. | A multi-resource scheduling scheme of Kubernetes for IIoT | |
| He et al. | Computation offloading and task scheduling based on improved integer particle swarm optimization in fog computing | |
| Liu et al. | Computation offloading strategy for multi user mobile data streaming applications | |
| Zheng et al. | A cloud resource prediction and migration method for container scheduling | |
| Guo et al. | Microservice selection in edge-cloud collaborative environment: A deep reinforcement learning approach |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190709 |
|
| WD01 | Invention patent application deemed withdrawn after publication |