CN111884854B - Virtual network traffic migration method based on multi-mode hybrid prediction - Google Patents
Virtual network traffic migration method based on multi-mode hybrid prediction Download PDFInfo
- Publication number
- CN111884854B CN111884854B CN202010741775.XA CN202010741775A CN111884854B CN 111884854 B CN111884854 B CN 111884854B CN 202010741775 A CN202010741775 A CN 202010741775A CN 111884854 B CN111884854 B CN 111884854B
- Authority
- CN
- China
- Prior art keywords
- link
- prediction
- flow
- virtual
- time sequence
- 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
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0896—Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2111—Selection of the most significant subset of features by using evolutionary computational techniques, e.g. genetic algorithms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/08—Computing arrangements based on specific mathematical models using chaos models or non-linear system models
-
- 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
- H04L41/12—Discovery or management of network topologies
-
- 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
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
- H04L43/0882—Utilisation of link capacity
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/125—Shortest path evaluation based on throughput or bandwidth
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/38—Flow based routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/70—Routing based on monitoring results
-
- 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
- H04L47/41—Flow control; Congestion control by acting on aggregated flows or links
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Abstract
The invention relates to a multi-mode hybrid prediction-based virtual network traffic migration method, wherein a virtual network control platform senses the current traffic of a virtual link at any moment, predicts the traffic value of the link in the next period by using historical data, and calculates the bandwidth utilization rate u of each link ij . Then the link utilization u ij Respectively with the upper threshold value W H And a lower threshold value W L A comparison is made. If u is present ij ≥W H Calculating the overload flow of the link and increasing the bandwidth value B of the link l . If the bandwidth utilization rate of the link in one prediction period satisfies u ij ≤W L And if other constraint conditions are met, deleting the virtual link and ensuring the link utilization rate u ij ≤W H Will be present on the basis ofTraffic on the link is migrated to the target virtual link. The method ensures the prediction accuracy, has shorter operation time and higher prediction efficiency, improves the virtual network flow migration efficiency and can save more bandwidth resources.
Description
Technical Field
The invention relates to a virtual network traffic migration method, in particular to a virtual network traffic migration method based on multi-mode hybrid prediction.
Background
The document "Din D, Chou C. virtual topology reconfiguration for mixed-line-rate optical WDM Network under dynamic traffic, 2015,30(2): 1-19" discloses a virtual Network topology reconfiguration method. Aiming at the problem of virtual network topology reconstruction under the dynamic flow demand, the method provides a topology reconstruction method to track the change of flow by monitoring the communication traffic of links, and optimizes the resource utilization rate and the network flow performance by adding or deleting one or more links. However, this method has the following problems:
(1) the method disclosed by the literature has the phenomenon of 'topology reconstruction lag', and because the method passively performs topology reconstruction by detecting network traffic and does not predict future network traffic information, the problem of virtual network topology reconstruction lag can occur.
(2) The method provides that the newly added virtual link cannot be deleted in the next period of time, although frequent jitter of the link is avoided to a certain extent, the newly added virtual link may occupy a large amount of bandwidth resources for a long time, resulting in resource waste.
Disclosure of Invention
Technical problem to be solved
In order to solve the problems of reconstruction lag and resource waste existing in the conventional Virtual Network topology reconstruction method, the invention provides a Virtual Network Traffic Migration method (VNTM-MHP) Based on Multi-mode Hybrid Prediction.
Technical scheme
A virtual network traffic migration method based on multi-mode hybrid prediction is characterized by comprising the following steps:
step 1: initializing network parameters, and respectively setting an upper limit threshold and a lower limit threshold of link utilization rate as W H And W L ；
Step 2: sensing the current flow of the virtual link, predicting the flow value of the link in the next period by using a multi-mode mixed flow prediction method based on parameter optimization selection, and calculating the bandwidth utilization rate u of each link ij If u is ij ≥W H Then store the link into set L h (ii) a If u ij ≤W L Then the link is stored in the set L v ；
And step 3: for set L h Each virtual link l in ij Calculating the link overload traffic T l ＝u ij ·B ij -W H ·B ij And increasing the link bandwidth value B l ＝T l /W H (ii) a Wherein, B ij For a virtual link l ij The bandwidth value of (a);
and 4, step 4: for set L v Each virtual link l in (b) ij Selecting other shortest paths between nodes i and j and storing them in set L r ；
And 5: sequentially computing the set L r The residual available bandwidth resources of each link in the system, if the residual bandwidth resources of the kth link exist, the condition B is satisfied k >u ij ·B ij ′，B i ′ j For the current link bandwidth to be migrated, flow will flow through the virtual link l ij Traffic of (2) to set L r And deleting the virtual link l ij Recovering bandwidth resources;
and 6: and updating the virtual network topology and executing the step 2.
The multi-mode mixed flow prediction method based on parameter optimization selection comprises the following steps: firstly, decomposing flow data into a high-frequency detail time sequence and a low-frequency approximate time sequence by adopting a wavelet decomposition method; then, performing feature extraction on the decomposed time sequence by using a phase space reconstruction and optimal sample selection method based on particle swarm optimization to construct a training sample; then, training and predicting the detailed time sequence and the approximate time sequence by adopting a chaotic model and an extreme learning machine neural network respectively; and finally, judging the prediction error by using a threshold value, and adaptively triggering a combination parameter selection algorithm based on particle swarm optimization.
Advantageous effects
The invention provides a Multi-mode Hybrid Prediction-Based virtual network Traffic migration method, which predicts the network Traffic of the next period by using a Multi-mode Hybrid Prediction Approach on Parameter Optimization Selection (MHTP-POS) Based on Parameter Optimization Selection, and migrates the Traffic in real time according to the Traffic Prediction result, thereby saving more bandwidth resources while avoiding ping-pong effect. The method ensures the prediction accuracy, has shorter operation time and higher prediction efficiency, improves the virtual network flow migration efficiency and can save more bandwidth resources.
Drawings
FIG. 1 is a flow chart of a VNTM-MHP method proposed by the present invention.
FIG. 2 is an exemplary diagram of a VNTM-MHP method proposed by the present invention.
FIG. 3 is a flow chart of the MHTP-POS method proposed by the present invention.
FIG. 4 is a comparison of the flow prediction sequence of the present invention with the original sequence.
FIG. 5 shows the prediction error of the MHTP-POS method in the invention for the network traffic sequence.
Figure 6 is a comparison graph of bandwidth resources saved by the VNTM-MHP method and the IW method in the present invention.
FIG. 7 is a diagram of the VNTM-MHP method of the present invention at different upper threshold W of link utilization H A saved bandwidth resource scenario.
FIG. 8 is a diagram of the VNTM-MHP method of the present invention at different lower thresholds W for link utilization L A saved bandwidth resource scenario.
Detailed Description
The invention will now be further described with reference to the following examples and drawings:
1. method flow for constructing VNTM-MHP
The invention provides a VNTM-MHP method for solving the problems of reconstruction lag and resource waste existing in the conventional virtual network topology reconstruction method, and the specific flow is shown in figure 1.
Firstly, sensing the current flow of a virtual link at any moment by a virtual network control platform, predicting the flow value of the link in the next period by using historical data, and calculating the bandwidth utilization rate u of each link ij . Then the link utilization u ij Respectively with the upper threshold value W H And a lower threshold value W L A comparison is made. If u is present ij ≥W H Calculating the overload traffic of the link and increasing the bandwidth value B of the link l . If the bandwidth utilization rate of the link in one prediction period satisfies u ij ≤W L And if other constraint conditions are met, deleting the virtual link and ensuring the link utilization rate u ij ≤W H And migrating the traffic on the current link to the target virtual link on the basis.
As shown in FIG. 2, which is an example of virtual network traffic migration, when the next periodic link l is predicted ad Bandwidth utilization of u ad ≥W H Calculating the overload flow of the link and increasing the bandwidth value B of the link ad . When the next periodic link l is predicted ab Bandwidth utilization of u ab ≤W L And another link l from node a to node b acb Has enough bandwidth resources and can ensure the link l ab The bandwidth utilization rate of the traffic on the network still satisfies u after the traffic on the network is migrated acb ≤W H . At this time, it will flow through the link l ab Traffic migration to link l acb And deletes link l ab And recovering bandwidth resources.
2. Method for predicting network flow at next moment by using MHTP-POS (Mobile high-definition multimedia Point-of-sale)
The accuracy of flow prediction and the time required by the prediction have important influence on the result and the efficiency of virtual network flow migration, and in order to improve the accuracy and the efficiency of flow prediction, the invention introduces a multi-mode mixed flow prediction method based on parameter optimization selection. The method comprises the steps of firstly, decomposing flow data into a high-frequency detail time sequence and a low-frequency approximate time sequence by adopting a wavelet decomposition method; then, performing feature extraction on the decomposed time sequence by using a phase space reconstruction and optimal sample selection method based on particle swarm optimization to construct a training sample; then, a chaotic model and an extreme learning machine neural network are adopted to respectively train and predict the detailed time sequence and the approximate time sequence so as to improve the prediction precision and the prediction efficiency; and finally, judging the prediction error by using a threshold value, and adaptively triggering a combination parameter selection algorithm based on particle swarm optimization.
The flow of the multi-mode hybrid flow prediction method based on parameter optimization selection is shown in fig. 3, and the method mainly comprises five steps of wavelet decomposition, phase space reconstruction, optimal sample selection, combined parameter optimization selection and flow prediction.
(1) Wavelet decomposition
Because the small-scale network flow sequence has chaotic characteristics, and larger prediction errors are easy to generate when the flow prediction is directly carried out, the invention adopts a wavelet decomposition method to decompose the network flow into an approximate time sequence and a detail time sequence and then respectively adopts different models for processing, thereby improving the prediction precision. When wavelet decomposition is performed, the larger the number of decomposition layers is, the more detailed characteristics of the observed network traffic become, but when the number of decomposition layers is too large, the calculation amount is also increased rapidly, and the prediction efficiency is reduced. According to the research, when the number of decomposition layers is 3, the prediction error can basically achieve the expected effect, and meanwhile, the calculation complexity is low. The invention therefore breaks down the network traffic sequence into one approximate time series AT3 and three detailed time series DT1, DT2 and DT 3.
(2) Phase space reconstruction
The network flow chaotic time sequence has complex dynamic characteristics, and the traditional low-dimensional coordinates cannot accurately depict the network flow chaotic time sequence. The phase space reconstruction theory is used as an important theory for predicting the chaotic time sequence, and the conventional data is brought into a describable framework by accurately describing the evolution rule of the hidden chaotic attractor. Based on the characteristics of chaos and mutability of small-scale network traffic, the invention selects the phase space reconstruction of a delay coordinate state based on an embedding theorem, converts the small-scale network traffic prediction into a nonlinear time sequence prediction problem, namely selects appropriate parameters m and tau, wherein m is an embedding dimension, tau is delay time and is a positive integer, and a certain mapping f exists, so that Y is f (X). For the flow prediction problem:
x n+1 ＝f(x n-(m-1)τ ,x n-(m-2)τ ,…,x n-τ ,x n ) (1)
wherein x is n+1 Is the predicted future time network flow value, { x n-(m-1)τ ,x n-(m-2)τ ,...,x n-τ ,x n Is the historical network traffic sample. Assuming that the length of the flow data is n, through phase space reconstruction, a learning sample can be obtained:
(3) optimal sample selection
The nonlinear time series prediction method is divided into a global prediction method and a local prediction method. The global prediction method utilizes all historical data to predict future values, and has large calculation amount and complex realization. The local prediction method only selects part of proper historical data to predict future values, the complexity is low, and the performance of the local prediction method is superior to that of the global prediction method under the same embedded dimension. Therefore, the method and the device predict the future network flow by using a local prediction method.
By performing phase space reconstruction on historical flow data, in a chaotic time sequence, the most relevant information to a prediction sample is in a sample point closest to the Euclidean distance of the most relevant information to the prediction sample, and the two have great distance correlation. Therefore, k training samples with the Euclidean distance to the prediction sample are selected to predict the flow sequence.
x t Is a stand-by in phase spacePredicted point, x n Is x in phase space t Of the neighboring points. X is then t And x n The euclidean distance s between is expressed as:
s＝||x t -x n || 2 (3)
the number k of the training samples has important significance in the local area prediction method, and not only influences the precision of a prediction result, but also influences the complexity of the prediction method. When the number k of training samples is too small, the prediction accuracy is reduced, but if the number k of training samples is too large, an "overfitting" phenomenon may occur, which may not only result in reduction of the prediction accuracy, but also increase the prediction complexity.
(4) Combined parameter selection based on particle swarm optimization
When partial historical information is used for predicting network flow, parameters m, tau and the number k of training samples in phase space reconstruction all have important influence on flow prediction results. In order to better select combination parameters, the invention provides an optimal combination parameter selection algorithm based on particle swarm optimization. First the particle-related parameters and operations are defined.
1) Particle position: position vector D of particle i ＝[d i1 ,d i2 ,d i3 ]Is defined as a parameter selection scheme, d i1 ，d i2 And d i3 Respectively representing the values of the parameters m, τ and k, d i1 ∈M，d i2 ∈N，d i3 ∈K。
2) Particle velocity: particle velocity vector V i ＝[v i1 ,v i2 ,v i3 ]Adjustment decisions defined as parameters, v i1 ,v i2 ,v i3 Is a binary variable, if v i1 ,v i2 ,v i3 A value of 0 indicates that the parameter needs to be reselected.
3) Fitness f (D) i ) Representing the flow prediction error of the particle selection scheme.
4) Subtraction Θ: when the two position vectors are subtracted, if the values in the corresponding dimensions are the same, the difference is 1, otherwise, the difference is 0. For example, (1,2,3) Θ (1,2,4) ═ 1,1, 0.
5) Addition ≦: p i V i ⊕P j V j For obtaining an adjustment decision for the parameter selection scheme. Wherein, P i V i And P j V j Respectively represent by P i Probability maintenance of V i The sum of the values of the dimensions is P j Probability of maintaining V j Value of each dimension, and P i +P j ＝1(0≤P i ≤1，0≤P j Less than or equal to 1). For example, 0.1(1,0,0) · 0.9(1,1,0) · (1, · 0), where · indicates that the dimension is not certain 0 or 1. In this example, this dimension is represented by a probability of 0.1 taking 0 and a probability of 0.9 taking 1.
6) Multiplication
For obtaining a new parameter selection scheme. Parameter selection scheme D i According to adjustment decision V i And (6) adjusting. For example,
indicating that the second parameter in the protocol needs to be adjusted.
Defining a position and speed updating basic formula of a particle swarm optimization algorithm as follows:
wherein, P 1 ，P 2 And P 3 Is a constant number, and P 1 +P 2 +P 3 ＝1。D pb And D gb Respectively the self historical optimum position and the neighborhood historical optimum position of the particle.
Therefore, the optimal parameter selection algorithm based on particle swarm optimization comprises the following specific steps:
step 1: initialChanging network flow information, setting value ranges of parameters M, tau and K as M, N and K respectively, calculating maximum iteration number MG, and randomly generating initial position parameter D by particles i And a speed parameter V i 。
Step 2: calculating the fitness f (D) of all the particles i ) To obtain D pb And D gb 。
And 3, step 3: the particle position and velocity parameters are updated according to equations (4) and (5).
And 4, step 4: for each particle, if f (D) i )<f(D i ) Then D is pb ＝D i (ii) a If f (D) pb )<f(D gb ) Then D is gb ＝D pb 。
And 5: if the iteration number is smaller than that of the MG, executing the step 3; otherwise, step 6 is executed.
Step 6: and outputting an optimization parameter selection scheme.
(5) Flow prediction
For the approximate time sequence and the detail time sequence after the phase space reconstruction is carried out, different training prediction models are respectively adopted to carry out analysis processing on the approximate time sequence and the detail time sequence. And aiming at the detail time sequence, a chaos model is adopted to train and predict the detail time sequence to obtain a prediction sequence. And for the approximate time sequence, training and predicting by adopting an extreme learning machine neural network. The extreme learning machine neural network is a single hidden layer feedforward neural network, based on a function approximation theory, an input weight is randomly selected during training, an output weight is determined by an analytic method, the convergence rate of the artificial neural network can be greatly improved, and the extreme learning machine neural network has the advantages of simplicity in training, simple structure, high learning convergence rate and the like.
And linearly superposing the detail time sequence processed by the chaotic model and the approximate time sequence processed by the neural network of the extreme learning machine to obtain a prediction sequence.
3. Performance evaluation and analysis
The invention takes Matlab as a simulation platform and designs two groups of simulation experiments. The first set of simulation experiments verified the performance of the MHTP-POS. A second set of simulation experiments verified the performance of VNTM-MHP.
(1) Experimental Environment settings
The method adopts the actual flow LBL-tcp-3.tcp as simulation data, the original sampling time is 2 hours, and the data is 1789995 in total. Resampling the original flow at intervals of 1 second to obtain a flow sequence with the length of 7199, normalizing to obtain an actual network flow time sequence for simulation, setting the value range of the embedding dimension m to be [5,25], setting the value range of the delay time tau to be [1,10], and setting the value range of the sample number k to be [10,400 ].
(2) MHTP-POS method performance verification
Firstly, the MHTP-POS method provided by the invention is used for predicting the flow within 200-800 s, and the performance of the MHTP-POS method is analyzed, and the results are shown in fig. 4 and fig. 5.
As can be seen from fig. 4, the flow prediction sequence can be accurately fitted to the original sequence. The MHTP-POS method decomposes the chaotic sequence into an approximate sequence and a detail sequence through wavelet decomposition, and respectively processes the two sequences by using different training models, so that accurate prediction of network flow is realized, and a good prediction result is obtained.
As can be seen from fig. 5, although the training samples are subjected to phase space reconstruction and parameter optimization, there still exists a certain prediction error when affected by noise or a sudden traffic sequence occurs. As can be seen from fig. 5, the MHTP-POS method predicts a maximum network traffic error value of about 0.035, which is within an acceptable range.
(3) Virtual network traffic migration method performance simulation
1) Performance comparison of different virtual network traffic migration methods
As shown in fig. 6, which is a comparison of bandwidth resources saved by different virtual network traffic migration methods, it can be seen from fig. 6 that, in order to avoid a ping-pong effect when the IW method performs traffic migration, newly added virtual links are not deleted within a period of time, and some virtual links with lower bandwidth utilization rate occupy the bandwidth resources for a long time, which results in resource waste and less saved bandwidth resources. The VNTM-MHP method introduces a flow prediction function, predicts the flow of the next period by using a mixed flow prediction model, and performs virtual network flow migration according to the prediction result, so that more bandwidth resources are saved.
2) Impact of link utilization threshold setting on performance of VNTM-MHP method
Fig. 7 and fig. 8 show bandwidth resource changes saved by the VNTM-MHP method when different upper and lower thresholds of link utilization are set.
FIG. 7 shows an upper link utilization threshold limit W H The VNTM-MHP method saves bandwidth resource changes at 0.7, 0.8, and 0.9, respectively. As can be seen from FIG. 7, with W H The saved bandwidth resources are gradually increased. When W is H When the bandwidth utilization rate is increased, the number of the congested virtual links is reduced, and each virtual link has more resources which can be used for bearing the traffic migrated from the virtual link with the low bandwidth utilization rate, so that more bandwidth resources are saved.
FIG. 8 shows a lower threshold W for link utilization L The VNTM-MHP method saves bandwidth resource changes at 0.1, 0.2, and 0.3, respectively. As can be seen from FIG. 8, with W L The saved bandwidth resources are increased along with the increase of the bandwidth. With W L And gradually increasing, wherein more virtual links meet the link deletion condition, the number of the deleted virtual links is increased, and the released bandwidth resources are increased.
Claims (1)
1. A virtual network traffic migration method based on multi-mode hybrid prediction is characterized by comprising the following steps:
step 1: initializing network parameters, and respectively setting an upper limit threshold and a lower limit threshold of link utilization rate as W H And W L ；
Step 2: sensing the current flow of the virtual link, predicting the flow value of the link in the next period by using a multi-mode mixed flow prediction method based on parameter optimization selection, and calculating the bandwidth utilization rate u of each link ij If u is ij ≥W H Then the link is stored in the set L h (ii) a If u ij ≤W L Then store the link into set L v ；
And step 3: for setsHetero L h Each virtual link l in ij Calculating the link overload traffic T l ＝u ij ·B ij -W H ·B ij And increasing the link bandwidth value B l ＝T l /W H (ii) a Wherein, B ij For a virtual link l ij The bandwidth value of (a);
and 4, step 4: for the set L v Each virtual link l in ij Selecting other shortest paths between nodes i and j and storing them in set L r ；
And 5: sequentially computing the set L r If the residual bandwidth resource of the kth link exists, the residual bandwidth resource of each link in the system meets B k ＞u ij ·B′ ij ，B′ ij For the current link bandwidth to be migrated, flow will flow through the virtual link l ij Traffic of (2) to set L r And deleting the virtual link l ij Recovering bandwidth resources;
step 6: updating the virtual network topology and executing the step 2;
the multi-mode mixed flow prediction method based on parameter optimization selection comprises the following steps:
firstly, decomposing flow data into a high-frequency detail time sequence and a low-frequency approximate time sequence by adopting a wavelet decomposition method comprises decomposing a network flow sequence into an approximate time sequence AT3 and three detail time sequences DT1, DT2 and DT 3; the phase space reconstruction and optimal sample selection method based on particle swarm optimization comprises the steps of selecting a delay coordinate state phase space reconstruction based on an embedding theorem, converting small-scale network flow prediction into a nonlinear time sequence prediction problem, namely selecting appropriate parameters m and tau, wherein m is an embedding parameter, tau is delay time and is a positive integer, and a certain mapping f exists so that Y is f (X), and for the flow prediction problem:
x n+1 ＝f(x n-(m-1)τ ,x n-(m-2)τ ,L,x n-τ ,x n )
wherein x is n+1 Is the predicted future time network flow value, { x n-(m-1)τ ,x n-(m-2)τ ,...,x n-τ ,x n The } is a historical network traffic sample; assuming that the length of the flow data is n, through phase space reconstruction, a learning sample can be obtained:
predicting future network flow by using a local prediction method: selecting k training samples closest to the Euclidean distance of the prediction samples to predict the flow sequence; the optimal combination parameter selection algorithm based on particle swarm optimization comprises the following specific steps:
step S1: initializing network flow information, setting the value ranges of parameters M, tau and K as M, N and K, calculating maximum iteration number MG, and randomly generating initial position parameter D by particles i And a speed parameter V i ；
Step S2: calculating the fitness f (D) of all the particles i ) To obtain D pb And D gb ；
Step S3: according to the formula
And
updating the particle position and the speed parameter; p 1 ，P 2 And P 3 Is a constant number, and P 1 +P 2 +P 3 ＝1；
Step S4: for each particle, if f (D) pb )<f(D i ) Then D is pb ＝D i (ii) a If f (D) pb )<f(D gb ) Then D is gb ＝D pb ；D pb And D gb Respectively being the self historical optimal position and the neighborhood historical optimal position of the particle;
step S5: if the iteration number is less than the MG, executing the step 3; otherwise, executing step 6;
step 6: outputting an optimized parameter selection scheme;
respectively adopting different training prediction models to analyze and process the approximate time sequence and the detail time sequence subjected to phase space reconstruction, and adopting a chaotic model to train and predict the detail time sequence to obtain a prediction sequence; for the approximate time sequence, training and predicting by adopting an extreme learning machine neural network; and linearly superposing the detail time sequence processed by the chaotic model and the approximate time sequence processed by the neural network of the extreme learning machine to obtain a prediction sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010741775.XA CN111884854B (en) | 2020-07-29 | 2020-07-29 | Virtual network traffic migration method based on multi-mode hybrid prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010741775.XA CN111884854B (en) | 2020-07-29 | 2020-07-29 | Virtual network traffic migration method based on multi-mode hybrid prediction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111884854A CN111884854A (en) | 2020-11-03 |
| CN111884854B true CN111884854B (en) | 2022-09-02 |
Family
ID=73200982
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010741775.XA Active CN111884854B (en) | 2020-07-29 | 2020-07-29 | Virtual network traffic migration method based on multi-mode hybrid prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111884854B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113411819B (en) * | 2021-05-30 | 2022-07-12 | 西安电子科技大学 | 5G NFV slice reconfiguration method, system and wireless communication system |
| CN115037642B (en) * | 2022-03-30 | 2023-11-21 | 武汉烽火技术服务有限公司 | Method and device for identifying flow bottleneck |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101729315A (en) * | 2009-12-24 | 2010-06-09 | 北京邮电大学 | Network flow-predicting method and device based on wavelet package decomposition and fuzzy neural network |
| WO2012148038A1 (en) * | 2011-04-26 | 2012-11-01 | 서울대학교산학협력단 | Method and device for allocating a resource in a virtual network environment |
| CN103095496A (en) * | 2013-01-10 | 2013-05-08 | 周亚建 | Prediction method and device for network flow |
| CN106357456A (en) * | 2016-10-11 | 2017-01-25 | 广东工业大学 | Prediction method of network traffic and device thereof |
| CN110210658A (en) * | 2019-05-22 | 2019-09-06 | 东南大学 | Prophet and Gaussian process user network method for predicting based on wavelet transformation |
Family Cites Families (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011093396A1 (en) * | 2010-01-27 | 2011-08-04 | 日本電信電話株式会社 | Virtual network control method and system based on fluctuations |
| EP2378722B1 (en) * | 2010-02-16 | 2012-11-28 | Siemens Aktiengesellschaft | A method for data transmission in a communication network |
| CN102130824B (en) * | 2010-10-30 | 2014-09-17 | 华为技术有限公司 | Joint optimization method and device |
| CN102082732A (en) * | 2011-02-23 | 2011-06-01 | 中国人民解放军信息工程大学 | Virtual network energy saving method based on virtual router on the move (VROOM) |
| US8694644B2 (en) * | 2011-09-29 | 2014-04-08 | Nec Laboratories America, Inc. | Network-aware coordination of virtual machine migrations in enterprise data centers and clouds |
| US20180006893A1 (en) * | 2015-01-21 | 2018-01-04 | Telefonaktiebolaget Lm Ericsson (Publ) | Elasticity in a Virtualised Network |
| CN107995027B (en) * | 2017-11-23 | 2021-06-25 | 东北大学 | Improved quantum particle swarm optimization algorithm and method applied to predicting network flow |
| CN108399744A (en) * | 2018-02-24 | 2018-08-14 | 上海理工大学 | Short-time Traffic Flow Forecasting Methods based on grey wavelet neural network |
| CN108900358B (en) * | 2018-08-01 | 2021-05-04 | 重庆邮电大学 | Virtual network function dynamic migration method based on deep belief network resource demand prediction |
| CN110858973B (en) * | 2018-08-23 | 2023-04-28 | 中国移动通信集团山东有限公司 | Cell network flow prediction method and device |
| CN110708318A (en) * | 2019-10-10 | 2020-01-17 | 国网湖北省电力有限公司电力科学研究院 | Network abnormal flow prediction method based on improved radial basis function neural network algorithm |
| CN111245701B (en) * | 2020-01-20 | 2021-08-31 | 中国电子科技集团公司第五十四研究所 | Link priority virtual network mapping method based on maximum weighted matching |
-
2020
- 2020-07-29 CN CN202010741775.XA patent/CN111884854B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101729315A (en) * | 2009-12-24 | 2010-06-09 | 北京邮电大学 | Network flow-predicting method and device based on wavelet package decomposition and fuzzy neural network |
| WO2012148038A1 (en) * | 2011-04-26 | 2012-11-01 | 서울대학교산학협력단 | Method and device for allocating a resource in a virtual network environment |
| CN103095496A (en) * | 2013-01-10 | 2013-05-08 | 周亚建 | Prediction method and device for network flow |
| CN106357456A (en) * | 2016-10-11 | 2017-01-25 | 广东工业大学 | Prediction method of network traffic and device thereof |
| CN110210658A (en) * | 2019-05-22 | 2019-09-06 | 东南大学 | Prophet and Gaussian process user network method for predicting based on wavelet transformation |
Non-Patent Citations (3)
| Title |
|---|
| Forecasting energy fluctuation model by wavelet decomposition and stochastic recurrent wavelet neural network;Lili Huang，Jun Wang;《Neurocomputing》;20180510;全文 * |
| 基于提升小波变换的网络流量混合预测模型;邹威等;《计算机工程》;20150115(第01期);全文 * |
| 基于混合算法优化小波神经网络的短时交通流量预测;唐瑞;《中国优秀硕士学位论文全文数据库》;20200315;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111884854A (en) | 2020-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113313947B (en) | Road condition evaluation method of short-term traffic prediction graph convolution network | |
| CN106897254B (en) | Network representation learning method | |
| CN111884854B (en) | Virtual network traffic migration method based on multi-mode hybrid prediction | |
| CN108287808A (en) | A kind of general dynamic tracing sequential sampling method of structure-oriented fail-safe analysis | |
| CN113852432B (en) | Spectrum Prediction Sensing Method Based on RCS-GRU Model | |
| CN113784410B (en) | Heterogeneous wireless network vertical switching method based on reinforcement learning TD3 algorithm | |
| JP2001236337A (en) | Predicting device using neural network | |
| CN111260124A (en) | Chaos time sequence prediction method based on attention mechanism deep learning | |
| JP2010134863A (en) | Control input determination means of control object | |
| CN110213784B (en) | Flow prediction method and device | |
| CN113537580B (en) | Public transportation passenger flow prediction method and system based on self-adaptive graph learning | |
| CN116366453A (en) | Self-adaptive dynamic deployment method for heterogeneous network element service demand characterization and virtual network element | |
| CN115862319A (en) | Traffic flow prediction method for space-time diagram self-encoder | |
| CN114419884B (en) | Self-adaptive signal control method and system based on reinforcement learning and phase competition | |
| CN116935649A (en) | Urban traffic flow prediction method for multi-view fusion space-time dynamic graph convolution network | |
| CN103209005B (en) | The pre-examining system of frequency hop sequences of a kind of graphic based model | |
| CN114708479A (en) | Self-adaptive defense method based on graph structure and characteristics | |
| Nie et al. | Digital twin for transportation Big data: A reinforcement learning-based network traffic prediction approach | |
| CN114461931A (en) | User trajectory prediction method and system based on multi-relation fusion analysis | |
| CN116976405A (en) | Variable component shadow quantum neural network based on immune optimization algorithm | |
| CN115438588A (en) | Temperature prediction method, system, equipment and storage medium of lithium battery | |
| CN114444922A (en) | Hybrid traffic efficiency evaluation method under group intelligent control | |
| CN113641496A (en) | DIDS task scheduling optimization method based on deep reinforcement learning | |
| Lahiany et al. | PTEENet: Post-Trained Early-Exit Neural Networks Augmentation for Inference Cost Optimization | |
| CN112508220A (en) | Traffic flow prediction method and device |
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 |