CN113179221A - Internet traffic control method and system - Google Patents

Internet traffic control method and system Download PDF

Info

Publication number
CN113179221A
CN113179221A CN202110737155.3A CN202110737155A CN113179221A CN 113179221 A CN113179221 A CN 113179221A CN 202110737155 A CN202110737155 A CN 202110737155A CN 113179221 A CN113179221 A CN 113179221A
Authority
CN
China
Prior art keywords
flow control
index
flow
value
characteristic
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110737155.3A
Other languages
Chinese (zh)
Other versions
CN113179221B (en
Inventor
窦伊男
宋磊
周力力
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.)
Haohan Data Technology Co ltd
Original Assignee
Haohan Data 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 Haohan Data Technology Co ltd filed Critical Haohan Data Technology Co ltd
Priority to CN202110737155.3A priority Critical patent/CN113179221B/en
Publication of CN113179221A publication Critical patent/CN113179221A/en
Application granted granted Critical
Publication of CN113179221B publication Critical patent/CN113179221B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/20—Traffic policing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
    • H04L41/0636—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis based on a decision tree analysis

Abstract

The invention provides a method and a system for controlling internet flow, wherein the method comprises the following steps: performing characteristic index analysis on the super-reference flow based on a flow control model, and judging whether the super-reference flow needs flow control; and (4) establishing a flow control strategy for the characteristic indexes needing flow control, and implementing the flow control strategy. The invention carries out characteristic index analysis on the over-reference flow through the pre-established flow control model, quickly positions the reason of abnormal flow, and issues a general strategy aiming at the abnormal link group so as to achieve the effect of flow control.

Description

Internet traffic control method and system
Technical Field
The invention relates to the technical field of internet flow control, in particular to a method and a system for controlling internet flow.
Background
The internet service is rapidly developed, more and more network services are provided, and network problems are continuously generated. The operator can relieve link bearing to a certain extent by quickly expanding the capacity of the link, so that user experience is improved, but effective propagation and healthy propagation of network content cannot be guaranteed, so that the operator needs to clearly master the content of network traffic, and link flow control and link expansion are performed on unreasonable network traffic, so that the healthy development of the internet industry is guaranteed. Currently, operators mainly locate abnormal traffic by a "technology for locating abnormal traffic based on traffic proportion" and a "technology for locating abnormal traffic based on flow direction traffic proportion".
The two schemes of the technology for positioning the abnormal flow based on the business flow proportion and the technology for positioning the abnormal flow based on the flow direction flow proportion can reduce the unreasonable flow to achieve a certain flow control effect, but the flow control strategy is implemented for each link, a large amount of workload and working time are consumed in the process of positioning the cause of the problem and issuing the flow control strategy, and the flow control threshold is manually established by experience or exploration, so that the manual processing mode cannot achieve the purposes of monitoring the network quality in real time and issuing the flow control strategy in real time, and lacks the scientificity of data support, so the flow control strategy is relatively low in efficiency, low in reliability and low in yield.
Disclosure of Invention
In view of this, the present invention provides a method and a system for controlling internet traffic, so as to find abnormal traffic in time and perform quick positioning, effectively control traffic with high efficiency and high reliability, and ensure continuous and healthy development of a network.
In order to solve the technical problems, the invention adopts the technical scheme that: a method for controlling internet traffic, the method comprising:
performing characteristic index analysis on the super-reference flow based on a flow control model, and judging whether the super-reference flow needs flow control;
and (4) establishing a flow control strategy for the characteristic indexes needing flow control, and implementing the flow control strategy.
Preferably, the flow control model is a decision tree model.
Preferably, the process of creating a decision tree model comprises:
acquiring sample data, and creating an abnormal judgment reference library;
and constructing a flow control decision tree model based on the data of the abnormal judgment reference library.
Preferably, the process of constructing the flow control decision tree model based on the data of the anomaly decision reference library includes:
determining the influence degree of each characteristic index in the abnormal judgment reference library on determining whether the flow control is carried out or not;
and constructing a decision tree based on the influence degree of the characteristic indexes on the decision of flow control.
Preferably, each characteristic index ultimately determines whether the degree of influence of the flow control is represented by its information gain value.
Preferably, the calculation of the information gain value includes:
obtaining multiple values of each index in sample data, and sequentially arranging;
arbitrarily taking two continuous domain values, and grouping by taking the average value of the two continuous domain values as a temporary dividing point to obtain a plurality of grouping modes;
respectively calculating the statistical information entropy value under each grouping mode;
and comparing the size of each information entropy value under different grouping modes of the same monitoring value index, and taking out the maximum information entropy value and the grouping mode corresponding to the maximum information entropy value.
Preferably, the process of analyzing the characteristic index of the super-reference flow based on the flow control model and judging whether the flow control is required comprises the following steps:
acquiring a characteristic index set in abnormal flow;
and comparing each characteristic index value in the characteristic index set with each node in the decision tree model.
Preferably, the process of formulating the flow control strategy for the characteristic index requiring flow control comprises:
extracting an index and an index value thereof which need flow control, determining a network outlet and a link group generated by the flow which needs flow control, and extracting the index and the index value which correspond to the index and the index value in the last statistical period;
and setting a flow control threshold value.
Preferably, the flow control threshold is calculated by the formula:
Figure 278875DEST_PATH_IMAGE001
a system for controlling internet traffic, the system comprising:
the abnormal positioning module is used for analyzing the characteristic indexes of the super-reference flow based on the flow control model and judging whether the super-reference flow needs to be subjected to flow control; and
and the intelligent flow control module is used for making a flow control strategy for the characteristic indexes needing flow control and implementing the flow control strategy.
The invention has the advantages and positive effects that: the invention carries out characteristic index analysis on the over-reference flow through the pre-established flow control model, quickly positions the reason of abnormal flow, and issues a general strategy aiming at the abnormal link group so as to achieve the effect of flow control.
Drawings
FIG. 1 is a schematic diagram of the structure of a decision tree model of the present invention;
fig. 2 is a schematic configuration diagram of the internet traffic control system of the present invention.
Detailed Description
For a better understanding of the present invention, reference is made to the following detailed description and accompanying drawings that illustrate the invention.
The invention provides a method for controlling internet flow, which comprises the following steps:
performing characteristic index analysis on the super-reference flow based on a flow control model, and judging whether the super-reference flow needs flow control;
and (4) establishing a flow control strategy for the characteristic indexes needing flow control, and implementing the flow control strategy.
For the flow rate exceeding the reference, the detailed details of the flow rate exceeding the reference are generally required to be extracted for one-to-one analysis in the prior art, so as to locate the cause of abnormal flow rate; for example, in the "technology of positioning abnormal traffic based on traffic proportion", detailed extraction of large and small classes of traffic is performed on the super-reference traffic (the assessment index value is out of the specified range), the cause of abnormal traffic is positioned, and a relevant traffic control strategy is formulated according to different causes of abnormal traffic; in the technology for positioning abnormal flow based on flow direction and flow rate ratio, extracting flow direction detail tickets for the over-reference flow direction and flow rate (the assessment index value is out of the group specified range), positioning the reason for generating the abnormal flow rate, and designating a related flow direction and flow control strategy according to different reasons for generating the abnormal flow rate; the two schemes need to consume a large amount of workload and working time in the process of positioning the reason of abnormal flow generation and issuing the flow control strategy.
The invention carries out characteristic index analysis on the over-reference flow through the pre-established flow control model, quickly positions the reason of abnormal flow, and issues a general strategy aiming at the abnormal link group so as to achieve the effect of flow control.
Further, the flow control model is a decision tree model which is established based on an abnormal judgment reference library; the process of creating a decision tree model includes: acquiring sample data, and creating an abnormal judgment reference library; and constructing a flow control decision tree model based on the data of the abnormal judgment reference library.
The method comprises the steps of obtaining characteristic index values and flow control processing modes during past flow super-reference, extracting indexes capable of reflecting flow super-reference phenomenon characteristics as characteristic indexes, enabling the characteristic indexes to form a set DataBase = { value A, value B, value C, value D and value E … }, enabling each characteristic index data value in the DataBase to be collected when the characteristic indexes are checked to be over-reference values, determining whether corresponding cases are subjected to flow control processing or not, forming an abnormity judgment reference base, and calculating information gain values of each characteristic index by means of an information gain algorithm so as to construct a flow control decision tree model.
Specifically, the detailed data format of the anomaly determination reference library is shown in table 1:
TABLE 1 detailed data of anomaly determination reference library
Time In DataBase Index D-A In DataBase Index D-B In DataBase Index D-C In DataBase Indexes D to D In DataBase Indexes D to E In DataBase Index D-F In DataBase Other indicators Flow control whether Y/N
Time1 A1 B1 C1 D1 E1 F1 … Y
Time2 A2 B2 C2 D2 E2 F2 … Y
…. … … … … … … … …
TimeZ AZ BZ CZ DZ EZ FZ … N
(note: the exception criterion library only holds the most recent Z records (Z is more than 1000) and takes the time field as the basis for judgment)
Further, the process of constructing the flow control decision tree model based on the data of the abnormal judgment reference library by means of the decision tree algorithm comprises the following steps:
determining the influence degree of each characteristic index in the abnormal judgment reference library on determining whether the flow control is carried out or not;
and constructing a decision tree based on the influence degree of the characteristic indexes on the decision of flow control.
Wherein, each characteristic index finally determines whether the influence degree of the flow control is expressed by the information gain value.
Specifically, the influence degree of each characteristic index in the index set DataBase in the anomaly determination reference library on the determination of whether the flow control is performed is judged, that is, the information gain of each characteristic index in the index set DataBase is calculated, so that a decision tree is constructed.
The information gain is for each characteristic index, namely, a characteristic index is seen, the information quantity of the index of 'flow control or not' is respectively what the information quantity of the index of 'flow control or not' is when the characteristic index exists, and the difference value of the characteristic index and the index of 'flow control or not' is the information quantity brought by the characteristic index, namely, the gain. Therefore, the information gain of a certain feature index = the information entropy of the entire abnormality determination reference library — the information entropy of a certain feature index.
The information entropy is the chaotic complexity of each characteristic index value in the abnormal judgment reference library. The complexity of each characteristic index in the sample data is judged by calculating the size of the result of the information entropy of each characteristic index, the larger the value of the result of the information entropy is, the higher the disorder degree of the value of the index is, the more scattered the value is, and on the contrary, the smaller the value of the result of the information entropy settlement is, the lower the disorder degree of the value is, and the more concentrated the value is.
The information entropy calculation formula is as follows:
Figure 142926DEST_PATH_IMAGE002
wherein|y|In the calculation of the information entropy of the whole abnormal judgment reference library in the scheme, the classification type of the final result,ythe value of (2) is obtained (the final result comprises two processing schemes of flow control and non-flow control),P k the probability (between 0 and 1) that this decision term is generated corresponding to the resultP k Sum of (1) =).
The information gain calculation formula is as follows:
Gain（A）= Ent – Ent(A)
wherein the content of the first and second substances,Gain（A）to representAThe information gain of the characteristic index is obtained,Entinformation entropy representing the entire abnormality determination reference library,Ent(A)to representAInformation entropy of the characteristic index.Gain（A）The greater the value of (a) is,Athe larger the information gain of the characteristic index is, the more the characteristic index is expressedAThe greater the impact of the characteristic index on whether flow control is required.
According to the algorithm, the information gain of the characteristic indexes in the index set DataBase is calculated in sequence. Further, the calculation process of the information gain value comprises the following steps:
obtaining multiple values of each index in sample data, and sequentially arranging;
arbitrarily taking two continuous domain values, and grouping by taking the average value of the two continuous domain values as a temporary dividing point to obtain a plurality of grouping modes;
respectively calculating the statistical information entropy value under each grouping mode;
and comparing the size of each information entropy value under different grouping modes of the same monitoring value index, and taking out the maximum information entropy value and the grouping mode corresponding to the maximum information entropy value.
Specifically, because the value of each field in the acquired data is a continuous variable and the values cannot be clustered and grouped, the first index field in the acquired data is subjected to clusteringAIs sequentially sorted according to the sequence from small to largeA1，A2，A3…AZAny two consecutive threshold values thereinApAndA（p+1）(wherein0<p<Z-1) The average value is used as a temporary division point for the inclusionZSet of domains We will have (Z-1) Different grouping modes are adopted to respectively calculate the statistical information entropy value under each grouping modeE1，E2，E3…EZComparing the entropy values of the information under different grouping modes of the same monitoring value index, and taking out the maximum valueEnAnd the corresponding grouping mode, performing the operation on each characteristic index in the characteristic index set according to the mode, positioning the grouping mode with the maximum information entropy in each characteristic index, and calculating to obtain each characteristic index in the set DataBaseA，B，C，D，E，F，G…K(assumption set)DataBaseIn which comprisesKIndividual quality index) of the respective information entropy valuesZ1，Z2，Z3，Z4，Z5，Z6…Zk(the values decrease from the left to the right in turn,Z1>Z2>Z3=Z4>Z5>Z6>…> Zk) The corresponding split points are respectivelyAz，Bz，Cz，Dz，Ez，Fz，Gz…Kz。
Further, a decision tree model is constructed according to the information gain value of each characteristic index calculated in the process. Specifically, the calculation formulas of the information entropy and the information gain in the above process can determine the magnitude of the information gain value of each characteristic index, and finally determine whether to use the characteristic index pairThe influence degree of flow control constructs a decision tree, namely: taking the characteristic index with the maximum information gainAThe root of the decision tree is used, and the decision tree is constructed by arranging in sequence, as shown in fig. 1.
The method comprises the steps of monitoring daily check indexes to obtain the super-reference flow, and then performing characteristic index analysis on the super-reference flow through a flow control model to judge whether the super-reference flow needs to be subjected to flow control; specifically, the invention collects the message through the DPI probe, and then synthesizes the xDR ticket (i.e. the user Internet access record, which is the session-level detailed record of the signaling process and the service transmission process generated after the processing based on the internet full data, and contains all the Internet access information of the user, including the big and small service labels identified by the DPI, the domain name information accessed by the user, the resource server attribution IP, the uplink and downlink flow data, etc.) meeting the unified specification; and then, according to the difference of the server where the resource is located and the service classification, respectively marking different labels of the service and the flow direction, and according to the difference of the flow direction, the link group and the network outlet, performing daily monitoring on each assessment index, and if the over-reference flow condition exists, performing alarm.
Further, for the super-reference flow, acquiring a characteristic index set in the super-reference flow; and comparing each characteristic index value in the characteristic index set with each node in the flow control model.
Specifically, if a certain assessment index is detectedXIf the flow exceeds the reference flow, the index will be obtainedXAnd characteristic index set thereofX={Ax，Bx，Cx，Dx，Ex，Fx…KxTaking out the characteristic indexes, and putting the characteristic indexes into a flow control model for characteristic analysis, specifically, a characteristic index setX={Ax，Bx，Cx，Dx，Ex，Fx…KxComparing each node of the flow control model with each node of the flow control model; in the comparison process, if the comparison result points to the root node of other indexes, whether the root node needs flow control needs to be considered, and if the comparison result points to the determination result of flow control or no flow control, the comparison of the indexes is finished.
The following characteristic index setX={Ax，Bx，Cx，Dx，Ex，Fx…KxThe comparative process will be described in detail by taking an example.
Step 1: extracting feature fieldsACorresponding collection data setXMiddle corresponding index calculation resultAxComparison ofAxAnd decision tree rootAzThe size of (2):
1 if the feature fieldAIncremental ratio ofAx>AzAnd if so, carrying out flow control on the current super-reference flow, informing the intelligent flow control module of the abnormal positioning result, simultaneously adding the record and the evaluation result to the abnormal judgment reference library, and carrying out iteration on the abnormal judgment reference library so as to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldAIncremental ratio ofAx<=AzIf the flow control is not needed, it cannot be easily determined whether the current super-reference flow needs to be subjected to flow control, and the comparison operation of the next characteristic index needs to be performed in combination with other characteristic indexes.
Step 2: extracting feature fieldsBCorresponding collection data setXMiddle corresponding index calculation resultBxComparison ofBxAnd decision tree nodesBzThe size of (2):
1 if the feature fieldBIncremental ratio ofBx<=BzIf the flow control is not needed, it cannot be easily determined whether the super-reference flow needs to be controlled, and the next characteristic index needs to be combinedCBreak-up point ofCzA comparison operation is performed.
2 if the feature fieldBIncremental ratio ofBx>BzIt cannot be easily determined whether the current super-reference flow needs to be controlled, and the next characteristic index needs to be combinedDBreak-up point ofDzA comparison operation is performed.
Step 3-1: extracting feature fieldsCCorresponding collection data setXMiddle corresponding index calculation resultCxComparison ofCxAnd decision tree nodesCzThe size of (2):
1 if the feature fieldCIncremental ratio ofCx<=CzAnd if the current super-reference flow is not subjected to flow control, adding a non-flow control label to the record, adding the record and the evaluation result to the abnormal judgment reference library, and iterating the abnormal judgment reference library to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldCIncremental ratio ofCx>CzIf the flow control is not needed, it cannot be easily determined whether the super-reference flow needs to be subjected to flow control, and a comparison operation needs to be performed in combination with the split point Ez of the next characteristic index E.
Step 3-2: extracting feature fieldsDCorresponding collection data setXMiddle corresponding index calculation resultDxComparison ofDxAnd decision tree nodesDzThe size of (2):
1 if the feature fieldDIncremental ratio ofDx<=DzAnd if the current super-reference flow is not subjected to flow control, adding a non-flow control label to the record, adding the record and the evaluation result to the abnormal judgment reference library, and iterating the abnormal judgment reference library to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldDIncremental ratio ofDx>DzAnd if so, carrying out flow control on the current super-reference flow, informing the intelligent flow control module of the abnormal positioning result, simultaneously adding the record and the evaluation result to the abnormal judgment reference library, and carrying out iteration on the abnormal judgment reference library so as to achieve the purpose of updating the abnormal judgment reference library.
Step 4: extracting feature fieldsECorresponding collection data setXMiddle corresponding index calculation resultExComparison ofExAnd decision tree nodesEzThe size of (2):
1 if the feature fieldEIncremental ratio ofEx<=EzAnd if the current super-reference flow is not subjected to flow control, adding a non-flow control label to the record, adding the record and the evaluation result to the abnormal judgment reference library, and iterating the abnormal judgment reference library to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldEIncremental ratio ofEx>EzIf the flow control is not needed, it cannot be easily determined whether the super-reference flow needs to be controlled, and the next characteristic index needs to be combinedFBreak-up point ofFzA comparison operation is performed.
Step 5: extracting feature fieldsFCorresponding collection data setXMiddle corresponding index calculation resultFxComparison ofFxAnd decision tree nodesFzThe size of (2):
1 RuiteSign fieldFIncremental ratio ofFx<=FzAnd if the current super-reference flow is not subjected to flow control, adding a non-flow control label to the record, adding the record and the evaluation result to the abnormal judgment reference library, and iterating the abnormal judgment reference library to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldFIncremental ratio ofFx>FzAnd if so, carrying out flow control on the current super-reference flow, informing the intelligent flow control module of the abnormal positioning result, simultaneously adding the record and the evaluation result to the abnormal judgment reference library, and carrying out iteration on the abnormal judgment reference library so as to achieve the purpose of updating the abnormal judgment reference library.
Step6 repeating the above steps until the characteristic field is extractedKCorresponding collection data setXMiddle corresponding index calculation resultKxComparison ofKxAnd decision tree nodesKzThe size of (2):
1 if the feature fieldKIncremental ratio ofKx<=KzAnd if the current super-reference flow is not subjected to flow control, adding a non-flow control label to the record, adding the record and the evaluation result to the abnormal judgment reference library, and iterating the abnormal judgment reference library to achieve the purpose of updating the abnormal judgment reference library.
2 if the feature fieldKIncremental ratio ofKx>KzAnd if so, carrying out flow control on the current super-reference flow, informing the intelligent flow control module of the abnormal positioning result, simultaneously adding the record and the evaluation result to the abnormal judgment reference library, and carrying out iteration on the abnormal judgment reference library so as to achieve the purpose of updating the abnormal judgment reference library.
And analyzing the super-reference index through a pre-established flow control model to determine whether the super-reference index needs flow control at this time.
Further, a flow control strategy is formulated and implemented for indexes needing flow control, and the method specifically comprises the following steps of
Extracting an index and an index value thereof which need flow control, determining a network outlet and a link group generated by the flow which needs flow control, and extracting the index and the index value which correspond to the index and the index value in the last statistical period;
setting a flow control threshold value;
and issuing a flow control strategy aiming at the link group of the network outlet according to the set flow threshold value.
An index for which flow control is specifically required is, for exampleX={Ax，Bx，Cx，Dx，Ex，Fx..Kx}The index value is N, the network outlet and the link group generated by the super-reference flow are determined, and the index corresponding to the previous statistical period is extractedXAnd an index valueN’(ii) a Then, setting a flow control threshold according to the index value of the index in the current period and the index value of the index in the previous period, wherein the calculation formula of the flow control threshold is as follows:
Figure 957298DEST_PATH_IMAGE003
the flow control strategy can automatically distribute the flow control threshold value to each link according to the flow size and the flow business composition of each link in the link group, and update the strategy of each link every half hour, so the strategy is called as intelligent flow control, and the labor cost is saved and the network safety is ensured while the index X is ensured to meet the requirement of the assessment index.
Further, the present invention provides a system for controlling internet traffic, the system comprising:
the abnormal positioning module is used for analyzing the characteristic indexes of the super-reference flow based on the flow control model and judging whether the super-reference flow needs to be subjected to flow control; and
and the intelligent flow control module is used for making a flow control strategy for the characteristic indexes needing flow control and implementing the flow control strategy.
In a specific embodiment of the present invention, the present invention further includes a resource matching module and a traffic monitoring module, specifically as shown in fig. 2, the packet is collected by a unified DPI probe, and an xDR ticket (i.e., a user internet record) conforming to the unified specification of an operator is synthesized, including data such as a large and small service tag identified by DPI, domain name information and resource server home IP accessed by a user, and uplink and downlink traffic. And then the synthesized xDR call ticket is transmitted to a resource matching module, and the attribution information of the user access resource is identified through matching with the resource matching module. When a user accesses a certain resource of the internet, the resource matching module can mark different labels of service and flow direction for the resource according to the server where the resource is located and different service classifications. The flow monitoring module carries out daily monitoring on each assessment index according to the difference of the flow direction, the link group and the network outlet, and alarms and informs the abnormal positioning module if the over-reference flow condition exists. And the abnormal positioning module calculates indexes in the characteristic index set DataBase, transmits the calculation result to the decision number model and evaluates whether the flow control is performed. If the flow control is needed, the abnormal positioning module informs the intelligent flow control module of the abnormal index and the index value which cause the maximum gain of the abnormal information, and the intelligent flow control module issues the total strategy according to the link group with the abnormal information, so as to achieve the effect of flow control.
The embodiments of the present invention have been described in detail, but the description is only for the preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. All equivalent changes and modifications made within the scope of the present invention should be covered by the present patent.

Claims (10)

1. A control method of internet traffic is characterized in that: the method comprises the following steps:
performing characteristic index analysis on the super-reference flow based on a flow control model, and judging whether the super-reference flow needs flow control;
and (4) establishing a flow control strategy for the characteristic indexes needing flow control, and implementing the flow control strategy.
2. The method of controlling internet traffic as claimed in claim 1, wherein: the flow control model is a decision tree model.
3. The method of controlling internet traffic as claimed in claim 2, wherein: the process of creating a decision tree model includes:
acquiring sample data, and creating an abnormal judgment reference library;
and constructing a flow control decision tree model based on the data of the abnormal judgment reference library.
4. The method of controlling internet traffic as claimed in claim 3, wherein: based on the data of the abnormal judgment reference library, the process of constructing the flow control decision tree model comprises the following steps:
determining the influence degree of each characteristic index in the abnormal judgment reference library on determining whether the flow control is carried out or not;
and constructing a decision tree based on the influence degree of the characteristic indexes on the decision of flow control.
5. The method of controlling internet traffic according to any one of claims 1 to 4, wherein: each characteristic index finally determines whether the influence degree of the flow control is expressed by the information gain value of the characteristic index.
6. The method of controlling internet traffic as claimed in claim 5, wherein: the information gain value calculation process comprises the following steps:
obtaining multiple values of each index in sample data, and sequentially arranging;
arbitrarily taking two continuous domain values, and grouping by taking the average value of the two continuous domain values as a temporary dividing point to obtain a plurality of grouping modes;
respectively calculating the statistical information entropy value under each grouping mode;
and comparing the size of each information entropy value under different grouping modes of the same monitoring value index, and taking out the maximum information entropy value and the grouping mode corresponding to the maximum information entropy value.
7. The method of controlling internet traffic as claimed in claim 1, wherein: the process of analyzing the characteristic indexes of the super-reference flow based on the flow control model and judging whether the flow control is needed or not comprises the following steps:
acquiring a characteristic index set in abnormal flow;
and comparing each characteristic index value in the characteristic index set with each node in the decision tree model.
8. The method of controlling internet traffic as claimed in claim 1, wherein: the process of making the flow control strategy for the characteristic indexes needing flow control comprises the following steps:
extracting an index and an index value thereof which need flow control, determining a network outlet and a link group generated by the flow which needs flow control, and extracting the index and the index value which correspond to the index and the index value in the last statistical period;
and setting a flow control threshold value.
9. The method of controlling internet traffic as claimed in claim 1, wherein: the calculation formula of the flow control threshold value is as follows:
Figure 253746DEST_PATH_IMAGE001
wherein N is "an index value of a sample cycle requiring flow control", and N' is "an index value of a previous cycle requiring flow control".
10. A control system of internet traffic is characterized in that: the system comprises:
the abnormal positioning module is used for analyzing the characteristic indexes of the super-reference flow based on the flow control model and judging whether the super-reference flow needs to be subjected to flow control; and
and the intelligent flow control module is used for making a flow control strategy for the characteristic indexes needing flow control and implementing the flow control strategy.
CN202110737155.3A 2021-06-30 2021-06-30 Internet traffic control method and system Active CN113179221B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110737155.3A CN113179221B (en) 2021-06-30 2021-06-30 Internet traffic control method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110737155.3A CN113179221B (en) 2021-06-30 2021-06-30 Internet traffic control method and system

Publications (2)

Publication Number Publication Date
CN113179221A true CN113179221A (en) 2021-07-27
CN113179221B CN113179221B (en) 2021-09-21

Family

ID=76927962

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110737155.3A Active CN113179221B (en) 2021-06-30 2021-06-30 Internet traffic control method and system

Country Status (1)

Country Link
CN (1) CN113179221B (en)

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105429977A (en) * 2015-11-13 2016-03-23 武汉邮电科学研究院 Method for monitoring abnormal flows of deep packet detection equipment based on information entropy measurement
CN105530138A (en) * 2014-09-28 2016-04-27 腾讯科技（深圳）有限公司 Data monitoring method and data monitoring device
CN106603256A (en) * 2015-10-14 2017-04-26 阿里巴巴集团控股有限公司 Flow control method and apparatus
CN108111433A (en) * 2016-11-25 2018-06-01 阿里巴巴集团控股有限公司 A kind of service dynamic flow control method, device and electronic equipment
CN108632164A (en) * 2018-08-17 2018-10-09 四川新网银行股份有限公司 Open platform gateway intelligence flow control method based on time series forecasting
CN109150845A (en) * 2018-07-26 2019-01-04 曙光信息产业(北京)有限公司 Monitor the method and system of terminal flow
CN109218211A (en) * 2017-07-06 2019-01-15 阿里巴巴集团控股有限公司 The method of adjustment of threshold value, device and equipment in the control strategy of data flow
US20190020670A1 (en) * 2017-07-13 2019-01-17 Cisco Technology, Inc. Bayesian tree aggregation in decision forests to increase detection of rare malware
US20200084153A1 (en) * 2018-09-11 2020-03-12 Cisco Technology, Inc. Packet flow classification in spine-leaf networks using machine learning based overlay distributed decision trees
CN112769796A (en) * 2020-12-30 2021-05-07 华北电力大学 Cloud network side collaborative defense method and system based on end side edge computing
CN112996041A (en) * 2019-12-02 2021-06-18 中国移动通信集团安徽有限公司 Flow control method, device and equipment

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105530138A (en) * 2014-09-28 2016-04-27 腾讯科技（深圳）有限公司 Data monitoring method and data monitoring device
CN106603256A (en) * 2015-10-14 2017-04-26 阿里巴巴集团控股有限公司 Flow control method and apparatus
CN105429977A (en) * 2015-11-13 2016-03-23 武汉邮电科学研究院 Method for monitoring abnormal flows of deep packet detection equipment based on information entropy measurement
CN108111433A (en) * 2016-11-25 2018-06-01 阿里巴巴集团控股有限公司 A kind of service dynamic flow control method, device and electronic equipment
CN109218211A (en) * 2017-07-06 2019-01-15 阿里巴巴集团控股有限公司 The method of adjustment of threshold value, device and equipment in the control strategy of data flow
US20190020670A1 (en) * 2017-07-13 2019-01-17 Cisco Technology, Inc. Bayesian tree aggregation in decision forests to increase detection of rare malware
CN109150845A (en) * 2018-07-26 2019-01-04 曙光信息产业(北京)有限公司 Monitor the method and system of terminal flow
CN108632164A (en) * 2018-08-17 2018-10-09 四川新网银行股份有限公司 Open platform gateway intelligence flow control method based on time series forecasting
US20200084153A1 (en) * 2018-09-11 2020-03-12 Cisco Technology, Inc. Packet flow classification in spine-leaf networks using machine learning based overlay distributed decision trees
CN112996041A (en) * 2019-12-02 2021-06-18 中国移动通信集团安徽有限公司 Flow control method, device and equipment
CN112769796A (en) * 2020-12-30 2021-05-07 华北电力大学 Cloud network side collaborative defense method and system based on end side edge computing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
陈长辉: "对等网络流量信息结构异常的检测技术研究", 《应用激光》 *

Also Published As

Publication number Publication date
CN113179221B (en) 2021-09-21

Similar Documents

Publication Publication Date Title
CN110609759B (en) Fault root cause analysis method and device
CN108415789B (en) Node fault prediction system and method for large-scale hybrid heterogeneous storage system
CN111506478A (en) Method for realizing alarm management control based on artificial intelligence
CN110570012B (en) Storm-based power plant production equipment fault early warning method and system
CN113570200B (en) Power grid running state monitoring method and system based on multidimensional information
CN116797404A (en) Intelligent building operation and maintenance supervision system based on big data and data processing
CN109150845A (en) Monitor the method and system of terminal flow
CN109685140A (en) A kind of DBSCAN algorithm gantry crane state classification method based on principal component analysis
CN112769605A (en) Heterogeneous multi-cloud operation and maintenance management method and hybrid cloud platform
CN109697332A (en) A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method
CN116662282A (en) Service data processing sharing system based on multidimensional data
CN115759640A (en) Public service information processing system and method for smart city
CN113740666B (en) Method for positioning root fault of storm alarm in power system of data center
CN114553671A (en) Diagnosis method for power communication network fault alarm
CN114666686A (en) Optical network alarm processing priority automatic judging method based on dynamic assignment calculation
CN113179221B (en) Internet traffic control method and system
CN108985563A (en) A kind of dynamically labeled method of Mechatronic Systems military service based on self-organizing feature map
CN114205214B (en) Power communication network fault identification method, device, equipment and storage medium
CN116545867A (en) Method and device for monitoring abnormal performance index of network element of communication network
CN114666209A (en) Multi-center-based log monitoring method and device
Wang et al. LSTM-based alarm prediction in the mobile communication network
CN105163343B (en) Risk judgment method and device that the crowd is dense
CN111740856A (en) Network communication equipment alarm acquisition abnormity early warning method based on abnormity detection algorithm
CN115809761B (en) Voltage quality analysis method and system based on low-voltage transformer area
CN116107859B (en) Container fault prediction method and device, electronic equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CP02 Change in the address of a patent holder
CP02 Change in the address of a patent holder

Address after: Room 218, 2nd Floor, Building A, No. 119 West Fourth Ring North Road, Haidian District, Beijing, 100000

Patentee after: HAOHAN DATA TECHNOLOGY CO.,LTD.

Address before: 102, building 14, 45 Beiwa Road, Haidian District, Beijing

Patentee before: HAOHAN DATA TECHNOLOGY CO.,LTD.