CN110247827B - NFV network element full-surrounding test method and device based on digital twin technology - Google Patents
NFV network element full-surrounding test method and device based on digital twin technology Download PDFInfo
- Publication number
- CN110247827B CN110247827B CN201910674326.5A CN201910674326A CN110247827B CN 110247827 B CN110247827 B CN 110247827B CN 201910674326 A CN201910674326 A CN 201910674326A CN 110247827 B CN110247827 B CN 110247827B
- Authority
- CN
- China
- Prior art keywords
- network element
- service
- index
- nfv
- simulation
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/50—Testing arrangements
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to a digital twin technology-based NFV network element full-enclosure testing method and device. The invention includes NFV business system module, NFV network element surround test module and intelligent scoring module, the invention constructs the network element surround test module at first, establish the virtual and real mapping relation between business system module and network element surround test module; secondly, in a network element surrounding test module, performing network element simulation, service flow customization, message simulation, dynamic telephone traffic model construction, index statistical analysis and transmitting the counted indexes to an intelligent scoring module; then, the intelligent scoring module is used for carrying out comprehensive evaluation on the network element to obtain a network element evaluation result, and the result is fed back to the NFV network element surrounding test module; and finally, applying the grading data to the operation and maintenance of the network element of the current network to help complete the operation and maintenance work of the actual service. The invention can complete the simulation of the operation environment of the current network service in the test stage and verify the tested service by using real data.
Description
Technical Field
The invention belongs to the technical field of new generation information, particularly relates to the technical fields of information communication, NFV/SDN, digital twinning, machine learning and the like, and discloses a NFV network element full-surrounding testing method and device based on a digital twinning technology.
Background
NFV, Network Function Virtualization. By using general purpose hardware such as x86 and virtualization technology, very versatile software processing is carried. Thereby reducing the cost of expensive equipment for the network. The functions of the network equipment can be independent of special hardware through software and hardware decoupling and function abstraction, resources can be shared fully and flexibly, rapid development and deployment of new services are achieved, and automatic deployment, elastic expansion, fault isolation, self-healing and the like are carried out based on actual service requirements.
The SDN is a novel network innovation architecture and is an implementation mode of network virtualization. The core technology OpenFlow separates the control plane and the data plane of the network equipment, thereby realizing the flexible control of network flow, enabling the network to be more intelligent as a pipeline, and providing a good platform for the innovation of a core network and application.
With the emergence of NFV/SDN technology and the evolution of 5G networks, the mode that the traditional relies on proprietary hardware gradually appears as software, the NFV network architecture is relatively complex compared with the traditional network, and mutual cooperation and influence exist among independent unit modules while hierarchical decoupling exists between the architectures, and further instability of the whole network element is caused. Therefore, effective means and methods are needed to measure and evaluate the structural design rationality, the resource planning and allocation accuracy, the stability of the current network operation, the high-pressure and high-traffic impact bearing capacity of the service and other series of problems.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a digital twin technology-based NFV network element full-enclosure testing method and device.
The technical scheme adopted by the invention for solving the technical problem is as follows:
the invention relates to a NFV network element full-enclosure testing method, which comprises the following steps:
s1, constructing an NFV network element enclosure test module, and establishing a virtual-real mapping relationship between the NFV service system module and the NFV network element enclosure test module.
S2, in the NFV network element surrounding test module, performing network element simulation, service flow customization, message simulation, dynamic traffic model construction, index statistical analysis, and transmitting the statistical index to the intelligent scoring module.
S3, the network element is comprehensively evaluated through the intelligent evaluation module, the network element evaluation result is obtained, and the result is fed back to the NFV network element surrounding test module.
S4, the network element scoring data and the network element testing effect are applied to the operation and maintenance of the current network element, and the operation and maintenance work of the actual service is completed.
Furthermore, in the network element simulation, the tested network element is in butt joint with the simulation network element, and the simulation network element has partial capacity of a real network element, so that the aim of butt joint of the real network element is fulfilled; the EPC core network simulates eNodeB, HSS, SAEGW and MME network elements, has the link capacity of establishing S1-MME, S1-U, S6a, S11-C, S11-U, S5/S8-C, S5/S8-U, and realizes the single-node networking architecture which is the same as that of the existing network; in the VoLTE domain, simulating P-CSCF/SBC, I/S-CSCF, VoLTE AS, SCC AS and HSS network elements, and having the communication capability of establishing Mw, ISC, Cx and Sh links.
Furthermore, the service flow customization is to analyze the current network service scene and service flow, determine and customize the service message flow, wherein the service message flow is supposed to cover all service interfaces of the service networking; in the EPC core network, the attachment scene of 4G users is simulated, and the service flow covers S1, S6a, S11/S11-U, S5/S8 and S5/S8-U interfaces.
Furthermore, in the message simulation, the simulation network element initiates a simulation service message, the simulation message conforms to the service protocol specification, and the tested network element actively identifies and triggers the service function; in EPC core network, the service simulation message follows S1AP/NAS, GTP and Diameter protocol specification, and the IMS domain has SIP and Diameter protocol specification.
Furthermore, in the construction of a dynamic telephone traffic model, a telephone traffic model of a service is determined according to different service scenes, and model parameters are dynamically adjusted, so that a service flow is dynamically injected in the surrounding test process, and the test purpose that the service flow is the same as the operation flow of the current network service is achieved.
Furthermore, the traffic model is calculated according to the system design capacity and the busy hour traffic, and finally the average traffic initiated per second and the concurrency per second are obtained.
Furthermore, the index statistical analysis comprises a service index and a network element overall performance index; the service index makes the detection points of the service index in the specific simulation message flow different according to different service scenes and service flows, and the service index analysis is statistical analysis based on index detection of the service flows and is the final output result of the test, which is used as a data basis for evaluating the capability of the tested network element. The network element overall performance index comprises a physical layer, a virtual layer, a network and a stored all-round monitoring index;
further, if the total number of the service indexes and the overall performance index of the network element is 53 (examples), the following procedures are executed in the intelligent scoring module:
and S31, performing dimensionality reduction on the service index and the network element overall performance index to obtain a feature set with 4 dimensionalities, wherein the dimensionality reduction adopts a principal component analysis method.
S32, for sample set D ═ x
1,x
2,x
3…x
mConsists of m samples, each x consisting of 53 features, x ∈ R
53×1(ii) a All samples are centered, i.e., each element is subtracted by the mean of all samples of its same feature.
And S33, calculating a covariance matrix A according to the samples, wherein the size of the covariance matrix A is 53 multiplied by 53.
S34, calculating the characteristic value lambda of the image through the covariance matrix A
1,λ
2…λ
53And taking the maximum four eigenvalues and obtaining the corresponding four eigenvectors to form a projection matrix W, wherein the size of the matrix is 4 multiplied by 53.
S35 by performing x on each sample
*An operation of W.x associates x with R
53×1Is mapped to x
*∈R
4×1On the feature vector space.
And S36, utilizing K-means clustering to obtain 10 clustering points of each feature dimension in the converted feature set, respectively corresponding to 10 values, and sequencing the 10 clustering points.
And S37, finding the nearest clustering point in the converted characteristic dimension, and subtracting the value of the point to obtain a value which is the distance T from the actual characteristic value to the clustering point.
And S38, converting the distance into a score S in a certain range by using a sigmoid function.
And S39, calculating to obtain corresponding scores in each dimension through the converted characteristics, obtaining 4 scores in 4 dimensions, and obtaining the final corresponding network element performance comprehensive score by taking the average value of the 4 scores, wherein the score becomes the final result of network element evaluation.
A digital twin technology-based NFV network element full-surrounding testing device comprises an NFV service system module, an NFV network element surrounding testing module and an intelligent scoring module.
And the NFV service system module and the NFV network element surrounding test module form a virtual-real mapping relation.
And the NFV network element surrounding test module obtains network element test data and then sends the network element test data to the intelligent scoring module.
And the intelligent scoring module comprehensively evaluates the network element, obtains a network element evaluation result and feeds the result back to the NFV network element surrounding test module.
Furthermore, the NFV network element surrounding test module comprises a network element simulation unit, a service flow customization unit, a message simulation unit, a dynamic traffic model building unit and an index statistical analysis unit.
The network element simulation unit is used for being in butt joint with a network element to be tested; the business process customizing unit is used for analyzing the current network business scene and business process, and determining and customizing the business message process; the message simulation unit is used for simulating a simulation network element in the unit to initiate a simulation service message; the dynamic traffic model building unit determines a traffic model of a service according to different service scenes, dynamically adjusts model parameters, and dynamically injects a service flow into the surrounding test process to achieve the test purpose of the same operation flow as the current network service; and the index statistical analysis unit is used for acquiring the service index and the overall performance index of the network element as a data basis for evaluating the capability of the tested network element.
The invention has the beneficial effects that:
1. the general test method is difficult to simulate the complex service scene and network environment of the existing network, so that the test points are not completely covered, and the problem is not thoroughly found. The invention adopts the surrounding test method to complete the simulation of the operation environment of the current network service in the test stage and verify the tested service by using real data.
2. The surrounding test can be adopted to simulate the operation and maintenance environment of the current network of the service in the test stage, so that the indexes measured by the entity of the tested service system are more comprehensive, and meanwhile, the digital twin technology is applied.
3. The application of the digital twin technology in the service testing stage can find the problem of the tested service in advance, the service is not required to be checked for leakage and gap filling after the service operation and maintenance stage, and meanwhile, an accurate testing and predicting model is established, so that the commercial risk of the service is greatly reduced, and the service operation and maintenance capability is improved.
4: and constructing a service index and system index system, constructing an intelligent scoring algorithm by fusing Principal Component Analysis (PCA), k-means clustering and sigmoid functions, extracting key evaluation indexes, mining the internal relation between the indexes and evaluation results, and realizing comprehensive evaluation on the network element performance. Compared with the conventional evaluation method using a single index and an independent index, the method has the advantages of comprehensiveness and objectivity.
Drawings
FIG. 1 is a block flow diagram of an embodiment of the present invention.
Detailed Description
The invention is further illustrated below with reference to fig. 1:
the invention discloses a digital twin technology-based NFV network element full-enclosure testing method, which comprises the following specific processes:
s1, constructing an NFV network element surrounding test digital twin model, and establishing a virtual-real mapping relation between the service system module and the digital twin model. The service system architecture hardware, the cloud operating system and the service software based on the NFV realize complete decoupling and can be independently provided by different providers. Therefore, entities of all layers in the NFV architecture system need to be digitally mapped, and are divided into COTS commercial hardware layers (mainly computer nodes, storage and network switch objects), CloudOS cloud operation layers (mainly virtual machine objects), service layers (mainly different service network elements and system objects), and finally, a virtual twin of the NFV service system is completed.
S2, in the surrounding test module of the digital twin model, network element simulation, flow customization, message simulation, dynamic traffic model construction, index statistical analysis are carried out, and the counted indexes are transmitted to the algorithm intelligent scoring module.
S3, the network element is comprehensively evaluated through the intelligent algorithm grading module, the network element evaluation result is obtained, and the prediction result is fed back to the surrounding test module.
S4, the network element scoring data and the network element testing effect obtained from the digital twin model are applied to the operation and maintenance of the current network element, and the operation and maintenance work of the actual service is completed.
Further, step S2 implementation is as follows:
s21 capability simulation of peripheral network elements, in order to construct a network enclosing test, the tested network elements need to be in butt joint with the simulation network elements according to the difference of tested environments, so that the simulation network elements need to have partial capability of real network elements, and the aim of butt joint of the network elements is fulfilled. For the purpose of surrounding test, in an EPC core network, eNodeB, HSS, SAEGW and MME network elements need to be simulated, link capacity of establishing S1-MME, S1-U, S6a, S11-C, S11-U, S5/S8-C, S5/S8-U is provided, and single-node networking architecture the same as that of the existing network is realized. The simulation butt joint of the peripheral network elements is the basis for realizing the service test. In the VoLTE domain, the network elements of P-CSCF/SBC, I/S-CSCF, VoLTE AS, SCC AS and HSS need to be simulated, and the network elements have the communication capacity of Mw, ISC, Cx and Sh links.
S22, customizing the service flow, analyzing the current network service scene and service flow, determining and customizing the service message flow, wherein the service message flow should cover all service interfaces of the service network. For example, in an EPC core network, the attachment scene of a 4G user is simulated, and the service flow shall cover interfaces of S1, S6a, S11/S11-U, S5/S8, S5/S8-U and the like. Meanwhile, according to the difference between the service function and the test environment, service parameter information, such as TAC, APN, etc., needs to be flexibly configured in the message flow. The purpose of the customized flow is to realize comprehensive coverage test networking and ensure the test accuracy.
And S23 message simulation, after determining the test networking and service flow, the simulation network element should have the ability to initiate simulation service messages, the simulation messages should conform to the service protocol specification, and the tested network element actively identifies and triggers the service function. In an EPC core network, a service simulation message should follow S1AP/NAS, GTP and Diameter protocol specifications, SIP and Diameter protocol specifications are required to be provided in an IMS domain, a service flow and message parameters are converted into network data flow to carry out service communication between network elements, and a service function is triggered.
S24 dynamic traffic model, when really performing service test, it needs to determine traffic model of service according to different service scenes, dynamically adjust model parameters, realize dynamic injection service flow in the surrounding test process, and achieve the test purpose that the final traffic is the same as the current network service. The traffic model calculates the system design capacity (and the number of users) and the busy hour traffic, and finally obtains the average traffic initiated per second and the concurrent number per second. The calculation method comprises the following steps: busy hour traffic is equal to the number of concurrent services × 3600/system design capacity is equal to the average traffic per user × 3600/average occupied duration.
And S25 index analysis, wherein the index analysis comprises a service index and the network element overall performance capital. The service index is different according to different service scenes and service flows, so that the detection points of the service index in a specific simulation message flow are different, different service monitoring indexes exist, and the service index analysis is statistical analysis based on the index detection of the service flow, is the final output result of the test and is used as a data basis for evaluating the capability of the tested network element. For example: in VoLTE service, flow nodes of SIP 180RING are recorded as call completion number index, flow nodes of SIP INVITE 200 OK are recorded as call response number index, and parameter variation in service flow is recorded. In the EPC, the following indexes are found according to different statistical angles according to different message nodes in the attach flow: for example, the MME network element has statistical indicators of the number of S1 mode attach requests, the number of S1 mode IMEI check requests, the number of S1 mode IMEI check successes, the number of S6a interface location update requests, the number of S6a interface location update successes, the number of packets for sending mobility management type messages, the number of packets for receiving mobility management type messages, the number of SAE bearer establishment requests, the number of S1 mode attach successes, and the number of SAE bearer establishment successes. In addition, network element system indexes based on the NFV architecture are all-around, and all-around index monitoring of a physical layer, a virtual layer, a network and storage is included, for example, the physical layer and the virtual layer mainly monitor indexes such as computer nodes and virtual machine system performance loads, for example, memory, CPU, system IO and network card data packets, the network monitors indexes such as switch equipment loads and data packets, and the storage mainly monitors indexes such as transmission rate, active time, read-write and response.
TABLE 1 Business index System
TABLE 2 System index System
Further, step S3 implementation is as follows:
s31, since the number of the service indexes and the system indexes affecting the performance of the network element is 53, data is slightly redundant, and the feature set is firstly reduced through Principal Component Analysis (PCA) to obtain a feature set with 4 dimensions.
S32, for sample set D ═ x
1,x
2,x
3…x
mConsists of m samples, each x consisting of 53 features, x ∈ R
53×1. All samples are centered, i.e., each element is subtracted by the mean of all samples of its same feature.
S33: from the samples, a covariance matrix a is calculated, which is 53 × 53 in size.
S34: the characteristic value lambda of the method can be obtained by calculation through A
1,λ
2...λ
53And taking the maximum four eigenvalues and obtaining four corresponding eigenvectors to form a projection matrix W, wherein the size of the matrix W is 4 multiplied by 53.
S35: by performing x on each sample
*An operation of W.x associates x with R
53×1Is mapped to x
*∈R
4×1On the feature vector space.
S36: and (4) solving 10 clustering points of each characteristic dimension in the converted characteristic set by utilizing K-means clustering, respectively corresponding to 10 values, and sequencing the 10 clustering points.
S37: in actual calculation, firstly, the nearest clustering point is found in the converted characteristic index dimension, and the numerical value of the point is subtracted to obtain a value which is the distance T from the actual characteristic value to the clustering point.
S38: converting distance into score S in certain range by using sigmoid function
Wherein S
0And (3) the score corresponding to the clustering point, T is a distance value difference, R is a value range, and R is a constant constraint parameter. Assume that the clustering point value of a certain feature dimension is [0.02.. 0.6, 0.7, 0.75, 0.9 ]]Corresponding to a score of [ 5.. 65, 75, 85, 95 respectively]R is 0.01, R is 5, and when the actual value of the converted feature is 1, the clustering point closest to 1 is 0.9. Specific scores can be found to be:
s39: by the method, the corresponding score can be calculated in each dimension through the converted characteristics, 4 scores are obtained in 4 dimensions, the average value of the 4 scores is taken to obtain the final corresponding network element performance comprehensive score, and the score becomes the final network element evaluation result.
Claims (7)
1. A NFV network element full-enclosure testing method based on a digital twin technology is characterized by comprising the following steps:
s1, constructing an NFV network element surrounding test module, and establishing a virtual-real mapping relation between an NFV service system module and the NFV network element surrounding test module;
s2, in the NFV network element surrounding test module, performing network element simulation, service flow customization, message simulation, dynamic traffic model construction, index statistical analysis, and transmitting the statistical index to the intelligent scoring module;
s3, comprehensively evaluating the network element through the intelligent evaluation module to obtain a network element evaluation result, and feeding the result back to the NFV network element surrounding test module;
s4, applying the network element scoring data and the network element testing effect to the operation and maintenance of the current network element to help complete the operation and maintenance work of the actual service;
the index statistical analysis comprises a service index and a network element integral performance index;
the service index makes the detection points of the service index in the specific simulation message flow different according to different service scenes and service flows, and the service index analysis is statistical analysis based on index detection of the service flow, is the final output result of the test and is used as a data basis for evaluating the capability of the tested network element;
the network element overall performance index comprises a physical layer, a virtual layer, a network and a stored all-round monitoring index;
if the total number of the service indexes and the overall performance indexes of the network element is 53, the intelligent scoring module executes the following procedures:
s31: reducing the dimension of the service index and the overall performance index of the network element to obtain a feature set with 4 dimensions, wherein the dimension reduction adopts a principal component analysis method;
s32: for sample set D ═ x
1,x
2,x
3...x
mConsists of m samples, each x consisting of 53 features, x ∈ R
53×1(ii) a Centralizing all samples, i.e. subtracting the mean of all samples from each element;
s33: calculating a covariance matrix A according to the samples, wherein the size of the covariance matrix A is 53 multiplied by 53;
s34: the eigenvalue lambda of the covariance matrix A is obtained through calculation
1,λ
2...λ
53Taking the maximum four eigenvalues and obtaining four corresponding eigenvectors to form a projection matrix W, wherein the size of the matrix W is 4 multiplied by 53;
s35: by performing x on each sample
*An operation of W.x associates x with R
53×1Is mapped to x
*∈R
4×1On the feature vector space of (2);
s36: calculating 10 clustering points of each characteristic dimension in the converted characteristic set by using K-means clustering, respectively corresponding to 10 values, and sequencing the 10 clustering points;
s37: finding the nearest clustering point from the transformed feature dimension, and subtracting the numerical value of the point to obtain a value which is the distance T from the actual feature value to the clustering point;
s38: converting the distance into a score S in a certain range by using a sigmoid function;
s39: and calculating to obtain corresponding scores in each dimension through the converted characteristics, obtaining 4 scores in 4 dimensions, and obtaining the final corresponding network element performance comprehensive score by taking the average value of the 4 scores, wherein the score becomes the final network element evaluation result.
2. The NFV network element full enclosure testing method based on the digital twin technology as claimed in claim 1, wherein: in the network element simulation, a tested network element is in butt joint with a simulation network element, and the simulation network element has partial capacity of a real network element so as to achieve the aim of butt joint of the real network element; the EPC core network simulates eNodeB, HSS, SAEGW and MME network elements, has the link capacity of establishing S1-MME, S1-U, S6a, S11-C, S11-U, S5/S8-C, S5/S8-U, and realizes the single-node networking architecture which is the same as that of the existing network; in the VoLTE domain, simulating P-CSCF/SBC, I/S-CSCF, VoLTE AS, SCC AS and HSS network elements, and having the communication capability of establishing Mw, ISC, Cx and Sh links.
3. The NFV network element full enclosure testing method based on the digital twin technology as claimed in claim 1, wherein: the service flow customization is to analyze the current network service scene and service flow, determine and customize the service message flow, wherein the service message flow is to cover all service interfaces of the service networking; in the EPC core network, the attachment scene of 4G users is simulated, and the service flow covers S1, S6a, S11/S11-U, S5/S8 and S5/S8-U interfaces.
4. The NFV network element full enclosure testing method based on the digital twin technology as claimed in claim 1, wherein: in message simulation, a simulation network element initiates a simulation service message, the simulation message conforms to a service protocol specification, and a tested network element actively identifies and triggers a service function; in EPC core network, the service simulation message follows S1AP/NAS, GTP and Diameter protocol specification, and the IMS domain has SIP and Diameter protocol specification.
5. The NFV network element full enclosure testing method based on the digital twin technology as claimed in claim 1, wherein: and determining a traffic model of the service according to different service scenes in the construction of the dynamic traffic model, and dynamically adjusting model parameters to realize dynamic injection of a service flow in the surrounding test process so as to achieve the test purpose of the same operation flow as the current network service.
6. The method for testing the full-enclosure of the NFV network element based on the digital twin technology as claimed in claim 5, wherein: the traffic model calculates according to the system design capacity and the busy hour traffic, and finally obtains the average traffic initiated per second and the concurrency per second.
7. A NFV network element fully-surrounding testing device based on a digital twin technology comprises an NFV service system module, an NFV network element surrounding testing module and an intelligent scoring module, and is characterized in that:
the NFV business system module and the NFV network element surrounding test module form a virtual-real mapping relation;
the NFV network element surrounding test module obtains network element test data and then sends the network element test data to the intelligent scoring module;
the intelligent scoring module comprehensively evaluates the network element, obtains a network element evaluation result and feeds the result back to the NFV network element surrounding test module;
the NFV network element surrounding test module comprises a network element simulation unit, a business process customizing unit, a message simulation unit, a dynamic telephone traffic model building unit and an index statistical analysis unit;
the network element simulation unit is used for being in butt joint with a network element to be tested; the business process customizing unit is used for analyzing the current network business scene and business process, and determining and customizing the business message process; the message simulation unit is used for simulating a simulation network element in the unit to initiate a simulation service message; the dynamic traffic model building unit determines a traffic model of a service according to different service scenes, dynamically adjusts model parameters, and dynamically injects a service flow into the surrounding test process to achieve the test purpose of the same operation flow as the current network service; the index statistical analysis unit is used for acquiring a service index and a network element overall performance index as a data basis for evaluating the capability of the tested network element;
the index statistical analysis unit comprises a service index and a network element overall performance index;
the service index makes the detection points of the service index in the specific simulation message flow different according to different service scenes and service flows, and the service index analysis is statistical analysis based on index detection of the service flow, is the final output result of the test and is used as a data basis for evaluating the capability of the tested network element;
the network element overall performance index comprises a physical layer, a virtual layer, a network and a stored all-round monitoring index;
if the total number of the service indexes and the overall performance indexes of the network element is 53, the intelligent scoring module executes the following procedures:
s31: reducing the dimension of the service index and the overall performance index of the network element to obtain a feature set with 4 dimensions, wherein the dimension reduction adopts a principal component analysis method;
s32: for sample set D ═ x
1,x
2,x
3...x
mConsists of m samples, each x consisting of 53 features, x ∈ R
53×1(ii) a Centralizing all samples, i.e. subtracting the mean of all samples from each element;
s33: calculating a covariance matrix A according to the samples, wherein the size of the covariance matrix A is 53 multiplied by 53;
s34: the eigenvalue lambda of the covariance matrix A is obtained through calculation
1,λ
2...λ
53Taking the maximum four eigenvalues and obtaining four corresponding eigenvectors to form a projection matrix W, wherein the size of the matrix W is 4 multiplied by 53;
s35: by performing x on each sample
*An operation of W.x associates x with R
53×1Is mapped to x
*∈R
4×1On the feature vector space of (2);
s36: calculating 10 clustering points of each characteristic dimension in the converted characteristic set by using K-means clustering, respectively corresponding to 10 values, and sequencing the 10 clustering points;
s37: finding the nearest clustering point from the transformed feature dimension, and subtracting the numerical value of the point to obtain a value which is the distance T from the actual feature value to the clustering point;
s38: converting the distance into a score S in a certain range by using a sigmoid function;
s39: and calculating to obtain corresponding scores in each dimension through the converted characteristics, obtaining 4 scores in 4 dimensions, and obtaining the final corresponding network element performance comprehensive score by taking the average value of the 4 scores, wherein the score becomes the final network element evaluation result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910674326.5A CN110247827B (en) | 2019-07-25 | 2019-07-25 | NFV network element full-surrounding test method and device based on digital twin technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910674326.5A CN110247827B (en) | 2019-07-25 | 2019-07-25 | NFV network element full-surrounding test method and device based on digital twin technology |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110247827A CN110247827A (en) | 2019-09-17 |
CN110247827B true CN110247827B (en) | 2020-02-11 |
Family
ID=67893404
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910674326.5A Active CN110247827B (en) | 2019-07-25 | 2019-07-25 | NFV network element full-surrounding test method and device based on digital twin technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110247827B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021155579A1 (en) * | 2020-02-07 | 2021-08-12 | 华为技术有限公司 | Data analysis method, apparatus and system |
CN112084646B (en) * | 2020-09-03 | 2022-08-19 | 江西农业大学 | Modular product customization method based on digital twinning |
CN112367109B (en) * | 2020-09-28 | 2022-02-01 | 西北工业大学 | Incentive method for digital twin-driven federal learning in air-ground network |
CN115484173A (en) * | 2021-05-28 | 2022-12-16 | 中国移动通信有限公司研究院 | Flow simulation method of digital twin network and digital twin network |
CN115426280B (en) * | 2022-07-18 | 2024-04-09 | 华为技术有限公司 | Network analysis method and related device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107623596A (en) * | 2017-09-15 | 2018-01-23 | 郑州云海信息技术有限公司 | Start the method for testing network element positioning investigation failure in a kind of NFV platforms |
CN107832497A (en) * | 2017-10-17 | 2018-03-23 | 广东工业大学 | A kind of intelligent workshop fast custom design method and system |
US10027569B1 (en) * | 2014-08-07 | 2018-07-17 | Amdocs Development Limited | System, method, and computer program for testing virtual services |
CN109150678A (en) * | 2018-08-07 | 2019-01-04 | 中国航空无线电电子研究所 | Distributed information physical system intelligence assembly shop topological model |
CN109791516A (en) * | 2016-08-02 | 2019-05-21 | 西门子公司 | For unit to be monitored and controlled used in autonomous system from X characteristic having |
CN110048904A (en) * | 2019-03-25 | 2019-07-23 | 北京天地互连信息技术有限公司 | A kind of test macro and method for user-plane function network element in 5G core net |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10572650B2 (en) * | 2016-02-29 | 2020-02-25 | Intel Corporation | Technologies for independent service level agreement monitoring |
CN106254178B (en) * | 2016-08-03 | 2019-12-17 | 陈鸣 | network test platform NFVNTP based on NFV and test method thereof |
KR102105683B1 (en) * | 2017-04-28 | 2020-05-29 | 한국전자통신연구원 | Integrated Platform Management Device And Method For Wire and Mobile communication Service |
-
2019
- 2019-07-25 CN CN201910674326.5A patent/CN110247827B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10027569B1 (en) * | 2014-08-07 | 2018-07-17 | Amdocs Development Limited | System, method, and computer program for testing virtual services |
CN109791516A (en) * | 2016-08-02 | 2019-05-21 | 西门子公司 | For unit to be monitored and controlled used in autonomous system from X characteristic having |
CN107623596A (en) * | 2017-09-15 | 2018-01-23 | 郑州云海信息技术有限公司 | Start the method for testing network element positioning investigation failure in a kind of NFV platforms |
CN107832497A (en) * | 2017-10-17 | 2018-03-23 | 广东工业大学 | A kind of intelligent workshop fast custom design method and system |
CN109150678A (en) * | 2018-08-07 | 2019-01-04 | 中国航空无线电电子研究所 | Distributed information physical system intelligence assembly shop topological model |
CN110048904A (en) * | 2019-03-25 | 2019-07-23 | 北京天地互连信息技术有限公司 | A kind of test macro and method for user-plane function network element in 5G core net |
Non-Patent Citations (1)
Title |
---|
"网络功能虚拟化系统测试技术研究";杨健,;《中国优秀硕士学位论文全文数据库-信息科技辑》;20150815;全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN110247827A (en) | 2019-09-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110247827B (en) | NFV network element full-surrounding test method and device based on digital twin technology | |
Baştuğ et al. | Big data meets telcos: A proactive caching perspective | |
CN106845820B (en) | NFV system reliability evaluation method based on performance margin | |
CN103051500A (en) | Test method and system of service network system | |
US11811844B2 (en) | Product validation based on simulated enhanced calling or messaging communications services in telecommunications network | |
US11849492B2 (en) | Unified query tool for network function virtualization architecture | |
US11770323B2 (en) | Unified interface and tracing tool for network function virtualization architecture | |
CN108390775A (en) | A kind of user experience quality evaluation method and system based on SPICE | |
WO2014040646A1 (en) | Determining the function relating user-centric quality of experience and network performance based quality of service | |
CN107113356A (en) | Improve audio call performance test | |
CN115802398A (en) | Interference optimization method and device, storage medium and electronic equipment | |
CN115705190A (en) | Method and device for determining dependence degree | |
US20230153685A1 (en) | Methods, apparatus and machine-readable media relating to data analytics in a communications network | |
CN111935769B (en) | Poor cell identification method, device and equipment | |
CN104284353B (en) | A kind of wireless local area network service performance test methods and system | |
WO2023186090A1 (en) | Verification method, apparatus and device | |
Ahmed et al. | Monitoring quality-of-experience for operational cellular networks using machine-to-machine traffic | |
Volvach et al. | Method of generalized quality index calculation in mobile networks | |
Wehbi et al. | Network monitoring challenges in the evolved packet core | |
CN117176623B (en) | Pressure testing method and system based on flow playback | |
Wang et al. | A general QoE assessment framework for applications and services | |
CN117370127A (en) | Scheduling test method and device for content distribution network and computer equipment | |
CN114697219B (en) | Network control method, equipment and system for live network | |
WO2024012235A1 (en) | Method, apparatus and device for analyzing artificial intelligence request | |
US20230045754A1 (en) | Trust related management of artificial intelligence or machine learning pipelines in relation to the trustworthiness factor explainability |
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 |