CN106452829B - A kind of cloud computing center intelligence O&M method and system based on BCC-KNN - Google Patents
A kind of cloud computing center intelligence O&M method and system based on BCC-KNN Download PDFInfo
- Publication number
- CN106452829B CN106452829B CN201610686202.5A CN201610686202A CN106452829B CN 106452829 B CN106452829 B CN 106452829B CN 201610686202 A CN201610686202 A CN 201610686202A CN 106452829 B CN106452829 B CN 106452829B
- Authority
- CN
- China
- Prior art keywords
- abnormal
- bcc
- running state
- state data
- knn
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- 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
- H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Whether the cloud computing center intelligence O&M method and system based on BCC-KNN that the invention discloses a kind of, method include: to detect the operating statuses of all services to be abnormal, and acquire its running state data to the service being abnormal；Processing is carried out to the running state data collected and it is subjected to classification processing using BCC-KNN algorithm；Corresponding solution is called from scheme base according to the abnormal class obtained after classification and is executed.System includes: abnormality acquisition module, abnormality detection module and exception processing module.The present invention is by successively detecting, determining and automating the process for solving to execute to abnormal conditions, intelligent independent processing can be carried out to the abnormal conditions that cloud service occurs, effectively reduce artificial intervention, improve the speed of abnormality processing, it realizes the availability of the service of raising and reduces the purpose of the workload of staff, greatly reduce operation cost.The present invention can be widely applied in cloud service as a kind of cloud computing center intelligence O&M method and system based on BCC-KNN.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of cloud computing center intelligence O&Ms based on BCC-KNN
Method and system.
Background technique
It is a model that NIST, which defines cloud computing, can be matched for quickly accessing one by network on demand anywhere or anytime
Computing resource (network, server, storage, the application program and service) shared pool set.Computing resource only needs the management of very little
Cost and the interaction less with service provider just quickly can be obtained and be discharged.Cloud computing model is by 5 essential characteristics, 3
Kind service model, 4 kinds of deployment models compositions.
Essential characteristic has on demand service certainly, resource pool, extensive internetwork connection mode, quick resilient expansion, can count
The service of amount.
Refer to that user automatically uniaxially can prepare computing resource as needed from service on demand, does not need to mention with service
It is interacted for quotient.
Resource pool refers to that computing resource is pooled in resource pool that (type includes storage, processing, interior by ISP
Deposit, bandwidth and virtual machine etc.), multiple consumers are shared to by multi-tenant mode, according to consumer demand to different objects
Reason resource and virtual resource are dynamically distributed or are reassigned.The interface of customer-centric can be realized cloud infrastructure to answer
With the transparency to user, user does not have to consider service is where to provide based on demand access service self.
Extensive internetwork connection mode refers to that the resource that cloud computing provides can be by the thin end of various isomeries by standardization
End or specific client, such as mobile phone, tablet computer, work station are accessed.
Quick resilient expansion is an important feature of cloud computing, because of resource poolization and quickly elasticity, user
The resource that can be obtained is seemingly inexhaustible, and can obtain any amount of resource at any time.The weight of elasticity
The property wanted is the resource that can distribute according to need, and application program, which resides in, quickly to be carried out in the data center of horizontal extension, i.e. cloud
The scale of service can quick-expansion, to adapt to the dynamic change of business load automatically.Demand of the resource that user uses with business
It is consistent, it avoids because of service quality decline or the wasting of resources caused by server performance overload or redundancy.It is worth noting
, the extension of resource must be fine granularity and quickly just computing resource can be made to match well with computational load enough.Cause
It is 3-10 times higher than average value for the peak value workload of server, so the server utilization of traditional data center is probably only
10% to 30% or so.In order to guarantee service quality, server will be disposed by peak value workload demands, this just necessarily causes
The waste of non-peak time resource, and the resource that the fluctuation more high seas loaded are taken is more.
Measurable service refer to cloud system using a kind of function of measuring come auto-control and optimization the utilization of resources, according to not
With service type measured, monitored and report resource service condition according to suitable Measure Indexes, promote ISP
With the transparency of service consumer.
However, the scale of cloud computing center is excessive now, low-frequency anomaly in traditional environment will be because in cloud computing center
Scale and become high-frequency anomaly, and many exceptions can repeat.Therefore, if cannot in the way of effective but
Very important person is to be responsible for a large amount of abnormal processing, this will lead to higher exception of the same race and has manually reprocessed cost, i.e., abnormal hair
After giving birth to and handling, process repeats are manually also wanted when occurring again.
In addition, cloud application itself can generate relevant status data in the process of running, these data can instruct O&M people
Member determines abnormal, but the running state data of the magnanimity of a large amount of cloud application generation will lead to operation maintenance personnel and spend greatly
The time and efforts of amount goes to analyze these data to extend abnormal repair time to position exception and determine its type.
Finally, due to which the design of the means of the maintenance management under traditional environment, system, model is not towards in cloud computing
The heart, so many maintenance tasks cannot once be completed with simple step, this just forces cloud management person to put into a large amount of people
Power resource leads to higher running cost.
Summary of the invention
In order to solve the above-mentioned technical problem, abnormal conditions can be quickly handled the object of the present invention is to provide a kind of, and can dropped
A kind of cloud computing center intelligence O&M method and system based on BCC-KNN of low running cost.
The technical scheme adopted by the invention is that:
A kind of cloud computing center intelligence O&M method based on BCC-KNN, comprising the following steps:
A, whether the operating status for detecting all services is abnormal, and acquires its operating status to the service being abnormal
Data；
B, processing is carried out to the running state data collected and it is subjected to classification processing using BCC-KNN algorithm；
C, corresponding solution is called from scheme base according to the abnormal class obtained after classification and executed.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly A includes:
A1, cycle detection is carried out to the running state data of all services, judges whether it is abnormal, if so, holding
Row step A2；Otherwise it carries out executing step A1；
A2, the acquisition that running state data is carried out to the service being abnormal.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly B includes:
B1, Screening Treatment is carried out to the running state data collected；
B2, by treated, running state data is converted into meeting the data format of BCC-KNN algorithm；
B3, real-time grading processing, the service of obtaining are carried out using BCC-KNN algorithm to the running state data after format transformation
Corresponding abnormal class.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly B3 includes:
B31, by the example set of storage according to abnormal class carry out divide cluster；
B32, calculating respectively cluster interior all examples data mean value, obtain the abnormal center that clusters；
B33, the center that clusters to non-classified running state data and each exception carry out similarity calculation, sorted out to
The highest exception of its similarity clusters clustering where center, and using the abnormal class to cluster as its corresponding exception class
Not, to obtain the corresponding abnormal class of the service.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly C includes:
C1, corresponding solution is called from scheme base according to the abnormal class obtained after classification；
C2, the solution called as needed generate its automated execution script, and execute.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
After rapid C2 further include:
C3, detection script execution after the exception whether handle success, if unsuccessful, send alert notification staff into
Pedestrian is processing；
C4, according to the running state data to cluster after classification, the exception center of clustering to cluster is updated.
It is of the present invention another solution is that
A kind of cloud computing center intelligence operational system based on BCC-KNN, comprising:
Whether abnormality acquisition module, the operating status for detecting all services are abnormal, and to being abnormal
Service acquire its running state data；
Abnormality detection module, for carrying out processing to the running state data collected and calculating it using BCC-KNN
Method carries out classification processing；
Exception processing module, for calling corresponding solution simultaneously from scheme base according to the abnormal class obtained after classification
It executes.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often state acquisition module includes:
Whether condition monitoring module judges it for carrying out cycle detection to the running state data of all services
It is abnormal, if so, executing data acquisition module；Otherwise it carries out executing condition monitoring module；
Data acquisition module, for carrying out the acquisition of running state data to the service being abnormal.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often detection module includes:
Data processing module, for carrying out Screening Treatment to the running state data collected；
Data conversion module, for running state data to be converted into meeting the data lattice of BCC-KNN algorithm by treated
Formula；
Data categorization module, for being divided in real time using BCC-KNN algorithm the running state data after format transformation
Class processing, the corresponding abnormal class of the service of obtaining.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often processing module includes:
Processing scheme generation module, for calling corresponding solution party from scheme base according to the abnormal class obtained after classification
Case；
Automated execution module, the solution for calling as needed generate its automated execution script, and hold
Row.
The beneficial effects of the present invention are:
A kind of cloud computing center intelligence O&M method based on BCC-KNN of the present invention is by successively examining abnormal conditions
The process for solving to execute is surveyed, determined and automated, intelligent independent processing can be carried out to the abnormal conditions that cloud service occurs, effectively subtracted
Few artificial intervention, improves the speed of abnormality processing, realizes the availability of the service of raising and reduces the workload of staff
Purpose greatly reduces operation cost.
Another beneficial effect of the invention is:
A kind of cloud computing center intelligence operational system based on BCC-KNN of the present invention passes through abnormality acquisition module, different
Normal detection module and exception processing module are successively detected, determined and are automated the process for solving to execute, energy to abnormal conditions
Intelligent independent processing is carried out to the abnormal conditions that cloud service occurs, artificial intervention is effectively reduced, improves the speed of abnormality processing,
It realizes the availability of the service of raising and reduces the purpose of the workload of staff, greatly reduce operation cost.
Detailed description of the invention
Specific embodiments of the present invention will be further explained with reference to the accompanying drawing:
Fig. 1 is a kind of step flow chart of the cloud computing center intelligence O&M method based on BCC-KNN of the present invention；
Fig. 2 is a kind of block diagram of the cloud computing center intelligence operational system based on BCC-KNN of the present invention.
Specific embodiment
With reference to Fig. 1, a kind of cloud computing center intelligence O&M method based on BCC-KNN of the present invention, comprising the following steps:
A, whether the operating status for detecting all services is abnormal, and acquires its operating status to the service being abnormal
Data；
B, processing is carried out to the running state data collected and it is subjected to classification processing using BCC-KNN algorithm；
C, corresponding solution is called from scheme base according to the abnormal class obtained after classification and executed.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly A includes:
A1, cycle detection is carried out to the running state data of all services, judges whether it is abnormal, if so, holding
Row step A2；Otherwise it carries out executing step A1；
A2, the acquisition that running state data is carried out to the service being abnormal.
Wherein, machine or service are detected, running state data when acquisition abnormity occurs if being abnormal.This
In need to carry out setting range to normal situation data in advance, more than being then determined as exception, for example be pair to a certain service request
The response time of one request was no more than 3 seconds, therefore the detection instrument write can carry out timing to access time, surpass
It spends three seconds and thinks that the service is abnormal, cycle detection then is carried out to the service, (the corresponding time surpasses 3 when an exception occurs
Second) just acquire the running state data of the service.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly B includes:
B1, Screening Treatment is carried out to the running state data collected；
B2, by treated, running state data is converted into meeting the data format of BCC-KNN algorithm；
B3, real-time grading processing, the service of obtaining are carried out using BCC-KNN algorithm to the running state data after format transformation
Corresponding abnormal class.
For BCC-KNN algorithm mainly there are three advantage, first point is still have in the case where there is inclination class distribution
Preferable classification performance, second point are that improving traditional KNN needs the shortcomings that storing entire training set, need to only store by training
Collect the data structure of conversion, first mate reduces storage cost and improves the real-time of classification, is thirdly that BCC-KNN supports increment
Study can evolve as the continuous generation of abnormal data is autonomous, the new environment of lasting adaptation.
Further, it includes processing invalid value, missing values that the described pair of running state data collected, which carries out Screening Treatment,
And remove incomplete data, the data of mistake, duplicate data and the data volume for injecting processing system is reduced, accelerate data
Processing speed so that promote treatment effeciency.
Further, by treated, running state data is converted into meeting the data format of BCC-KNN algorithm comprising that will count
According to being converted into other suitable numerical representations, or data are normalized, so that the value model of each attribute
Enclosing will not have big difference, and each attribute of data is computable.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly B3 includes:
B31, by the example set of storage according to abnormal class carry out divide cluster；
B32, calculating respectively cluster interior all examples data mean value, obtain the abnormal center that clusters；
B33, the center that clusters to non-classified running state data and each exception carry out similarity calculation, sorted out to
The highest exception of its similarity clusters clustering where center, and using the abnormal class to cluster as its corresponding exception class
Not, to obtain the corresponding abnormal class of the service.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
Suddenly C includes:
C1, corresponding solution is called from scheme base according to the abnormal class obtained after classification；
C2, the solution called as needed generate its automated execution script, and execute.
Further, solution generation calling be composed using relatively simple abnormality processing scheme, that is,
Say that an abnormal solution is can to increase the flexibility and reusability of scheme in this way by many sub- forecast scheme configurations.It solves
Scheme needs are write according to the format pre-defined, and when execution can automatically generate corresponding according to the environment of machine
Script.
As a kind of cloud computing center intelligence O&M further improvements in methods based on BCC-KNN, the step
After rapid C2 further include:
C3, detection script execution after the exception whether handle success, if unsuccessful, send alert notification staff into
Pedestrian is processing；
C4, according to the running state data to cluster after classification, the exception center of clustering to cluster is updated.
Further, when a new exception occurs, the scheme deposited cannot carry out processing well leads to processing failure, this
When just need to notify staff artificially write solution and be put into scheme base, the correspondence of update abnormal classification and scheme
Relationship, while predictablity rate of the training set to improve such is added in new abnormality, and then work as the abnormality again
It is secondary can be from main process task without with artificially participating in again when occurring.
With reference to Fig. 2, a kind of cloud computing center intelligence operational system based on BCC-KNN of the present invention, comprising:
Whether abnormality acquisition module, the operating status for detecting all services are abnormal, and to being abnormal
Service acquire its running state data；
Abnormality detection module, for carrying out processing to the running state data collected and calculating it using BCC-KNN
Method carries out classification processing；
Exception processing module, for calling corresponding solution simultaneously from scheme base according to the abnormal class obtained after classification
It executes.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often state acquisition module includes:
Whether condition monitoring module judges it for carrying out cycle detection to the running state data of all services
It is abnormal, if so, executing data acquisition module；Otherwise it carries out executing condition monitoring module；
Data acquisition module, for carrying out the acquisition of running state data to the service being abnormal.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often detection module includes:
Data processing module, for carrying out Screening Treatment to the running state data collected；
Data conversion module, for running state data to be converted into meeting the data lattice of BCC-KNN algorithm by treated
Formula；
Data categorization module, for being divided in real time using BCC-KNN algorithm the running state data after format transformation
Class processing, the corresponding abnormal class of the service of obtaining.
It is described different as a kind of further improvement of cloud computing center intelligence operational system based on BCC-KNN
Often processing module includes:
Processing scheme generation module, for calling corresponding solution party from scheme base according to the abnormal class obtained after classification
Case；
Automated execution module, the solution for calling as needed generate its automated execution script, and hold
Row.
Wherein, the BCC-KNN algorithm core concept in the present invention is that example set will be trained to be converted into the abnormal center collection that clusters.
One abnormal class includes many examples, and each example has specific object data self, we will be real according to abnormal class
Example is divided, i.e., all examples with identical abnormal class belong to same cluster.
After HTTP data set is divided by abnormal class, the example classification in the same division is identical, therefore each classification
It can be regarded as one to cluster.Because tradition KNN algorithm will carry out similarity operation with each whole training set examples, this just has
Very big computing cost and entire training set is stored, if so we can be by clustering with a model after division
It indicates, then model can be converted to training dataset, the model that we use is exactly to calculate the interior all realities that cluster
The mean value of example, and using this mean value as the center to cluster, it each clusters just only need a record that can indicate in this way, store
Amount substantially reduces.There is exception to cluster after the collection of center it is necessary to classify to non-classified example, algorithm is by non-classified example
With the exception after conversion cluster center carry out similarity calculation, rather than with entire training set calculate similarity, then will be similar
Spend classification of the abnormal class at the highest center that clusters as example.Specifically it is exactly the thought using KNN algorithm, will divide
The example of class and the center that each clusters carry out similarity calculation.If similarity uses Euclidean distance, then clustering with all
After center successively calculates distance, using the abnormal class where the center that clusters nearest with its distance as the classification knot of the example
Fruit.Because the classification of algorithm uses similarity calculation, and chooses highest classification knot of the center as example that cluster of similarity
Fruit, the i.e. value of K are 1, this just reduces the influence of deflection class distribution.
In addition, training set is converted into the abnormal center that clusters by algorithm so that algorithm only need to it is abnormal cluster center rather than
Similarity calculation is carried out with entire training set, not only makes the speed of classification greatly speed up reduction time complexity in this way, simultaneously
Space complexity is also reduced, because same only need to store the center that clusters rather than entire training set, so that all exceptions
The center of clustering can all be put into memory, be conducive to the quick update at the center of clustering, and importance herein is time and space
The reduction of complexity enables the algorithm to greatly improve the real-time of classification.
Because data set is converted to the abnormal center collection that clusters by BCC-KNN, in order to realize that incremental learning algorithm is wanted
It can generate exception example according to new and adjust the abnormal center collection that clusters, by new knowledge in conjunction with old knowledge, so that during exception clusters
Heart collection can adapt to new environment with the addition constantly adjustment of new knowledge.Simultaneously in order to prevent expired old knowledge to new environment
Processing impacts, and also to abandon expired part and cluster central value.When generating exception example, abnormal classification only has new
Extremely and abnormal two types are had occurred and that.It is as follows for the incremental learning strategy of the anomaly algorithm of both types:
It was had occurred and that if abnormal, then in the exception example is just divided into that the abnormal class is corresponding clustering,
It is then based on the example and the center of clustering is recalculated at the center of clustering.
New exception if abnormal, then at this time just by the exception example be divided into one it is new cluster, then by the reality
Example is as the center that clusters.
After storing the abnormal center collection that clusters, we will be updated the abnormal center of clustering after exception example generates, I
Newly generated HTTP data are described with (Anew, Bnew, Cnew, Nnew ..., class) vector and abnormal classification is correctly right
It answers, specially (0.75,0.78,0.79,0.29,0,0.77, load too high), is denoted as (VCTnew, class).We sentence first
The value for determining class just illustrates that the exception example generated is because the value " load too high " of class can be found clustering to arrange in not
Through what is occurred, then the exception example center that clusters corresponding with its abnormal class is directly carried out mean value calculation.Load
The excessively high corresponding center vector CLUcenter that clusters carries out mean value calculation, and new center vector is denoted as NEWCLUcenter, therefore
Calculation formula is as follows:
If abnormal class belonging to exception example is that occur for the first time, we equally use (VCTnew, class) to describe,
Specific data and abnormal class correspond to (0.8,0.1,0.1,0,0,0, configuration error), i.e., the value of its class is that configuration is wrong
Accidentally.Since the abnormal center that clusters is concentrated just without the corresponding entry that clusters of configuration error classification, it may thus be appreciated that the abnormal class is
Occur for the first time, thus we by the exception example be divided into one it is new cluster, the corresponding classification that clusters is exactly to configure
Mistake, and the center NEWCLUcenter that clusters of example is just initialized as (VCTnew, class).
Pass through the description of front, it is known that algorithm can constantly adjust the abnormal center that clusters according to newly generated exception example
Collection, but legacy data is classified the influence of this process to using new knowledge in order to prevent, algorithm can abandon expired data, algorithm
A fixed time period is defined, this is also the validity period of data, then as time goes by, the data more than validity period will
It is dropped, allows for the newest environment of adaptation that the abnormal center collection that clusters can always continue in this way.
Since our center of clustering is the average value for calculating all examples under the abnormal class, in order to delete
The data of phase, one timestamp of each data entry that we concentrate to the abnormal center that clusters, then by timestamp in validity period
Interior son clusters center combination jointly to calculate average value.It is briefly exactly that the center of clustering is made of several subcenters, every height
There is the timestamp of oneself at center, passs at any time, and the son of the expired timestamp mark center that clusters just is deleted, cannot again by
It clusters center as generation.
For example expired time is limited to three months by we, and legacy data is deleted as unit of the moon, then we can will be every
As soon as a month is used as a sub- cluster centre, therefore has 3 sons to cluster center, it is assumed that the current moon is 0, last month 1, the The Month Before Last
It is 2.The actual center of clustering is calculated by three son centers of clustering, we indicate that timestamp is with CLU0sub-center
0 entry, CLU1sub-center come indicate timestamp be 1 entry, CLU2sub-center come indicate timestamp be 2 item
Mesh, and us are assumed with (hub attribute Areal, hub attribute Breal, hub attribute Creal ...) vector to indicate to load
The center that clusters of high abnormal class is denoted as CLUreal-center, and this center of clustering is really to be used to calculate similarity
It uses, then CLUreal-center calculation formula is as follows；
After the time entering new month, all newly generated examples are all identified as the timestamp in new month, during son clusters
The initialization procedure of the heart is similar with the process that new abnormal class occurs, i.e., by first abnormal class generated in new month
Example clusters center as son, the example of generation thereafter and recalculates mean value with the son center of clustering again.Next will
Time be more than trimestral son cluster center deletion, by taking more timestamps cluster center collection as an example, when enter new month when, each
The value of timestamp adds 1, and entry of all timestamps mark greater than 2 for the center concentration that clusters extremely should be deleted.We assume that into
Enter the load too high that new month generates is (0.73,0.782,0.80,0.22,0,0.76, load too high) extremely.
It is to be illustrated to preferable implementation of the invention, but the invention is not limited to the implementation above
Example, those skilled in the art can also make various equivalent variations on the premise of without prejudice to spirit of the invention or replace
It changes, these equivalent deformations or replacement are all included in the scope defined by the claims of the present application.
Claims (4)
1. a kind of cloud computing center intelligence O&M method based on BCC-KNN, it is characterised in that: the following steps are included:
A, whether the operating status for detecting all services is abnormal, and acquires its operating status number to the service being abnormal
According to；
B, processing is carried out to the running state data collected and it is subjected to classification processing using BCC-KNN algorithm；
C, corresponding solution is called from scheme base according to the abnormal class obtained after classification and executed；The step A includes:
A1, cycle detection is carried out to the running state data of all services, judges whether it is abnormal, if so, executing step
Rapid A2；Otherwise it carries out executing step A1；
A2, the acquisition that running state data is carried out to the service being abnormal；
The step B includes:
B1, Screening Treatment is carried out to the running state data collected；
B2, by treated, running state data is converted into meeting the data format of BCC-KNN algorithm；
B3, real-time grading processing is carried out using BCC-KNN algorithm to the running state data after format transformation, the service of obtaining corresponds to
Abnormal class；
The step C includes:
C1, corresponding solution is called from scheme base according to the abnormal class obtained after classification；
C2, the solution called as needed generate its automated execution script, and execute.
2. a kind of cloud computing center intelligence O&M method based on BCC-KNN according to claim 1, it is characterised in that:
The step B3 includes:
B31, by the example set of storage according to abnormal class carry out divide cluster；
B32, calculating respectively cluster interior all examples data mean value, obtain the abnormal center that clusters；
B33, the center that clusters to non-classified running state data and each exception carry out similarity calculation, sorted out to its phase
It clusters clustering where center like highest exception is spent, and using the abnormal class to cluster as its corresponding abnormal class, from
And obtain the corresponding abnormal class of the service.
3. a kind of cloud computing center intelligence O&M method based on BCC-KNN according to claim 1, it is characterised in that:
After the step C2 further include:
Whether the exception handles success after C3, detection script execution, if unsuccessful, send alert notification staff and carries out people
For processing；
C4, according to the running state data to cluster after classification, the exception center of clustering to cluster is updated.
4. a kind of cloud computing center intelligence operational system based on BCC-KNN, it is characterised in that: include:
Whether abnormality acquisition module, the operating status for detecting all services are abnormal, and to the clothes being abnormal
Business acquires its running state data；
Abnormality detection module, for the running state data collected carry out processing and by its using BCC-KNN algorithm into
Row classification processing；
Exception processing module, for calling corresponding solution from scheme base according to the abnormal class obtained after classification and holding
Row；
The abnormality acquisition module includes:
Condition monitoring module judges whether it occurs for carrying out cycle detection to the running state data of all services
It is abnormal, if so, executing data acquisition module；Otherwise it carries out executing condition monitoring module；
Data acquisition module, for carrying out the acquisition of running state data to the service being abnormal；The abnormality detection module
Include:
Data processing module, for carrying out Screening Treatment to the running state data collected；
Data conversion module, for running state data to be converted into meeting the data format of BCC-KNN algorithm by treated；
Data categorization module, for being carried out at real-time grading to the running state data after format transformation using BCC-KNN algorithm
Reason, the corresponding abnormal class of the service of obtaining；
The exception processing module includes:
Processing scheme generation module, for calling corresponding solution from scheme base according to the abnormal class obtained after classification；
Automated execution module, the solution for calling as needed generate its automated execution script, and execute.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610040862 | 2016-01-21 | ||
| CN2016100408626 | 2016-01-21 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106452829A CN106452829A (en) | 2017-02-22 |
| CN106452829B true CN106452829B (en) | 2019-07-19 |
Family
ID=58181284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610686202.5A Active CN106452829B (en) | 2016-01-21 | 2016-08-18 | A kind of cloud computing center intelligence O&M method and system based on BCC-KNN |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106452829B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108540298B (en) * | 2017-03-01 | 2022-06-17 | 中兴通讯股份有限公司 | Method and device for automatically processing garbage service |
| CN108847953A (en) * | 2018-04-25 | 2018-11-20 | 合肥智圣新创信息技术有限公司 | A kind of O&M service system and method |
| CN110659173B (en) | 2018-06-28 | 2023-05-26 | 中兴通讯股份有限公司 | Operation and maintenance system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102457525A (en) * | 2011-12-19 | 2012-05-16 | 河海大学 | Load-based anomaly intrusion detection method and system |
| CN103227734A (en) * | 2013-04-27 | 2013-07-31 | 华南理工大学 | Method for detecting abnormity of OpenStack cloud platform |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9912529B2 (en) * | 2014-08-20 | 2018-03-06 | International Business Machines Corporation | Tenant-specific log for events related to a cloud-based service |
-
2016
- 2016-08-18 CN CN201610686202.5A patent/CN106452829B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102457525A (en) * | 2011-12-19 | 2012-05-16 | 河海大学 | Load-based anomaly intrusion detection method and system |
| CN103227734A (en) * | 2013-04-27 | 2013-07-31 | 华南理工大学 | Method for detecting abnormity of OpenStack cloud platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106452829A (en) | 2017-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109961204A (en) | Quality of service analysis method and system under a kind of micro services framework | |
| CN105046327B (en) | A kind of intelligent grid information system and method based on machine learning techniques | |
| CN109725899B (en) | Data stream processing method and device | |
| CN107733986A (en) | Support the protection of integrated deployment and monitoring operation big data support platform | |
| US10331156B2 (en) | System and method for big data geographic information system discovery | |
| AU2017258970A1 (en) | Testing and improving performance of mobile application portfolios | |
| CN107103064B (en) | Data statistical method and device | |
| Zhang et al. | A data set accuracy weighted random forest algorithm for IoT fault detection based on edge computing and blockchain | |
| CN106452829B (en) | A kind of cloud computing center intelligence O&M method and system based on BCC-KNN | |
| CN106940677A (en) | One kind application daily record data alarm method and device | |
| CN113448812A (en) | Monitoring alarm method and device under micro-service scene | |
| CN109034580B (en) | Information system overall health degree evaluation method based on big data analysis | |
| CN109117421A (en) | Data are handled to improve the quality of data | |
| KR20150038905A (en) | Apparatus and method for preprocessinig data | |
| US9600523B2 (en) | Efficient data collection mechanism in middleware runtime environment | |
| Natu et al. | Holistic performance monitoring of hybrid clouds: Complexities and future directions | |
| Sîrbu et al. | Towards data-driven autonomics in data centers | |
| CN114090378A (en) | Custom monitoring and alarming method based on Kapacitor | |
| Wladdimiro et al. | Disaster management platform to support real-time analytics | |
| CN118132326A (en) | Multisource heterogeneous data analysis method and system based on intelligent enhanced data analysis | |
| Ahmed et al. | Automated diagnostic of virtualized service performance degradation | |
| CN103326880B (en) | Genesys calling system high availability cloud computing monitoring system and method | |
| Iuhasz et al. | Monitoring of exascale data processing | |
| CN114500229B (en) | Network alarm positioning and analyzing method based on space-time information | |
| CN106874215B (en) | Serialized storage optimization method based on Spark operator |
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 |