CN118200257B - Signal high-speed transmission method and system based on high-speed connector - Google Patents
Signal high-speed transmission method and system based on high-speed connector Download PDFInfo
- Publication number
- CN118200257B CN118200257B CN202410603298.9A CN202410603298A CN118200257B CN 118200257 B CN118200257 B CN 118200257B CN 202410603298 A CN202410603298 A CN 202410603298A CN 118200257 B CN118200257 B CN 118200257B
- Authority
- CN
- China
- Prior art keywords
- index
- fuzzy
- application demand
- monitoring
- application
- 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
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
- H04L47/803—Application aware
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
-
- 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
-
- 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
- H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/83—Admission control; Resource allocation based on usage prediction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Genetics & Genomics (AREA)
- Algebra (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Fuzzy Systems (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Automation & Control Theory (AREA)
- Physiology (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to the field of data transmission, in particular to a signal high-speed transmission method and system based on a high-speed connector. A high-speed connector-based signal high-speed transmission system, comprising: the system comprises a monitoring index data set acquisition module, an application demand index time sequence prediction set output module, a bandwidth allocation strategy analysis module and a bandwidth allocation module. According to the invention, the bandwidth allocation strategy corresponding to the next preset period can be output by carrying out fuzzy decision analysis based on the application demand prediction index set corresponding to the current monitoring time point and the monitoring index data set acquired by the current monitoring time point, and the bandwidth allocation strategy can improve the overall throughput of the high-speed connector under different bandwidth demands and different network states, so that the high-speed transmission demand based on the high-speed connector is realized.
Description
Technical Field
The invention relates to the field of data transmission, in particular to a signal high-speed transmission method and system based on a high-speed connector.
Background
In the information age today, the rate requirements for data communication are increasing. In the fields of personal computers, data center servers, high-performance storage systems, and the like, there is a strong demand for realizing high-speed and reliable data transmission. The high-speed transmission can not only improve the system performance, but also meet the requirements of users on real-time response and high-capacity content.
In application scenarios such as the internet, the internet of things and the like, rapid transmission of mass data is more important. Taking cloud computing as an example, large-scale data resources need to be efficiently exchanged inside and outside a data center, so that real-time and smooth service experience is provided for users. Meanwhile, with the rising of emerging technologies such as artificial intelligence, big data analysis and the like, high-speed transmission becomes a key for realizing rapid calculation and calculation separation.
In addition, in the multimedia fields of live video broadcast, 4K/8K ultra-high definition video transmission and the like, the high-speed data transmission of Gbps level directly influences the image quality and the smoothness. Under the demands of industrial control, real-time monitoring and the like, the system also requires millisecond-level low-delay high-speed transmission.
It can be seen that the high-speed data transmission method and technology are crucial to the development of the information society, and are fundamental for realizing high-bandwidth and low-delay data communication.
Disclosure of Invention
The method comprises the steps of obtaining historical application demand index sets of all application data streams, and carrying out time sequence analysis on the historical application demand index sets of all application data streams to obtain application demand prediction index sets corresponding to current monitoring time points; and carrying out fuzzy decision analysis based on the application demand prediction index set corresponding to the current monitoring time point and the monitoring index data set acquired by the current monitoring time point, and outputting a bandwidth allocation strategy corresponding to the next preset period, wherein the bandwidth allocation strategy can improve the overall throughput of the high-speed connector under different bandwidth demands and different network states, so that the high-speed transmission demand based on the high-speed connector is realized.
A high-speed connector-based signal high-speed transmission method, comprising:
acquiring a monitoring index data set of a physical link at a current monitoring time point, and recording a time period between adjacent monitoring time points as a preset detection period;
Simultaneously acquiring an application demand index set F n corresponding to all current application data flows at a monitoring time point, wherein n=1, 2,3, & gtN, wherein N is the total number of the application data flows; for each application data stream, performing the following operation, forming an application demand index time sequence prediction set S n by the currently acquired application demand index set F n and the application demand index set F n acquired M-1 times before, then sending the demand index time sequence prediction set S n into an application demand index prediction model delta n for processing, and outputting an application demand prediction index set FR n corresponding to the next monitoring time point;
The method comprises the steps that a fuzzy decision set is formed by a monitoring index data set obtained at a current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, the fuzzy decision set is sent to a bandwidth allocation strategy analysis model to be processed, a bandwidth allocation strategy in the next preset period is output, the bandwidth allocation strategy analysis model is built based on a fuzzy decision tree, the bandwidth allocation strategy analysis model is trained through monitoring analysis training samples, the monitoring analysis training samples are marked based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and the target allocation duty fuzzy language value and the target transmission priority fuzzy language value are obtained through a group optimization algorithm;
And controlling bandwidth allocation in the next preset period based on the bandwidth allocation strategy.
As a preferred aspect, the application demand index prediction model δ n is built based on a time sequence prediction model, and training for any application demand index prediction model δ n includes the following steps:
Acquiring historical application demand index training samples of application data streams corresponding to a plurality of application demand index prediction models delta n, wherein M+1 historical application demand index sets arranged according to time sequence are stored in the historical application demand index training samples, the first M historical application demand index sets in the historical application demand index training samples are used as input data of an application demand index prediction model delta n, and the last 1 historical application demand index sets in the historical application demand index training samples are used as target data of an application demand index prediction model delta n; all the historical application demand index training samples form a historical application demand index training set, the historical application demand index training set is sent into an application demand index prediction model delta n with initialized parameters for training, an application demand index loss value is calculated, whether the application demand index loss value is in a first confidence range is judged, and if the application demand index loss value is in the first confidence range, a trained application demand index prediction model delta n is output; otherwise, continuing to train the application demand index prediction model delta n through the historical application demand index training set.
As a preferred aspect, the bandwidth allocation policy analysis model includes an input data blurring layer, a blurring decision analysis layer and a bandwidth allocation policy output layer, where the input data blurring layer is configured to blur the monitoring index data set and the application demand prediction index set FR n to construct the monitoring index blurring data set and the application demand prediction blurring index set FR n; the fuzzy decision analysis layer is used for outputting a corresponding allocation duty ratio fuzzy language value mu n and a transmission priority fuzzy language value epsilon n to each application data stream based on a fuzzy data set and an application demand prediction fuzzy index set FR n, and is established based on a fuzzy decision tree and comprises two leaf nodes which are respectively marked as allocation duty ratio leaf nodes and transmission priority leaf nodes; the bandwidth allocation policy output layer is used for outputting the bandwidth allocation policy based on the corresponding fuzzy language values output by all the application data streams.
As a preferred aspect, a fuzzy decision set is formed by a monitoring index data set obtained at a current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, and the fuzzy decision set is sent to a bandwidth allocation policy analysis model for processing, and a bandwidth allocation policy in a next preset period is output, which specifically includes the following steps:
Carrying out fuzzification treatment on the monitoring index data set by a Gaussian fuzzification method to obtain a monitoring index fuzzification data set;
Aiming at each application demand prediction index set FR n, carrying out fuzzification processing on the application demand prediction index set FR n by a Gaussian fuzzification method to obtain an application demand prediction fuzzification index set FR n, splicing the application demand prediction fuzzification index set FR n and the monitoring index fuzzification data set, then sending the spliced application demand prediction fuzzification index set FR n and the monitoring index fuzzification data set into a fuzzification decision analysis layer for analysis, and marking the output leaf node as a minimum comprehensive fuzzification value in all paths of the distribution proportion leaf node as a fuzzification language value mu n; meanwhile, the output leaf node is the minimum comprehensive fuzzy value in all paths of the transmission priority leaf node and is marked as a transmission priority fuzzy language value epsilon n;
in a bandwidth allocation strategy output layer, obtaining allocation duty ratio fuzzy language values mu n and transmission priority fuzzy language values epsilon n corresponding to all application data streams; performing normalization on the allocation duty ratio fuzzy language values mu n corresponding to all the application data flows, and outputting the bandwidth duty ratio X n corresponding to all the application data flows; normalizing the transmission priority fuzzy language values epsilon n corresponding to all the application data streams, and outputting the transmission priorities Y n corresponding to all the application data streams; and arranging the bandwidth occupation ratios X n corresponding to all the application data flows according to the sequence from the large to the small of the corresponding transmission priority Y n, and constructing a bandwidth allocation strategy.
As a preferred aspect, training for the fuzzy decision analysis layer inside the bandwidth allocation policy analysis model comprises the steps of:
obtaining a plurality of monitoring analysis training samples, wherein the monitoring analysis training samples comprise a historical monitoring index data set and a historical application demand index set which are obtained at any monitoring time point, the historical application demand prediction index set is a historical application demand index set corresponding to any application data stream, the monitoring analysis training samples are subjected to fuzzification processing to obtain fuzzy analysis training samples, the fuzzy analysis training samples comprise a historical monitoring index fuzzy data set and a historical application demand fuzzy index set, the fuzzy analysis training samples are marked through a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value, and the target allocation proportion fuzzy language value and the target transmission priority fuzzy language value are obtained through a group optimization algorithm; all fuzzy analysis training samples marked by the target allocation duty fuzzy language value and the target transmission priority fuzzy language value form a fuzzy analysis training set, the fuzzy analysis training set is sent to a fuzzy decision analysis layer in a bandwidth allocation strategy analysis model initialized by parameters for training, a comprehensive loss value is calculated, whether the comprehensive loss value is in a second confidence range is judged, and if the comprehensive loss value is in the second confidence range, the fuzzy decision analysis layer in the trained bandwidth allocation strategy analysis model is output; otherwise, continuing to train the fuzzy decision analysis layer in the bandwidth allocation strategy analysis model through the fuzzy analysis training set.
As a preferred aspect, the method specifically includes, for any monitored environmental condition data, the monitored environmental condition data includes a historical monitoring index data set and a historical application demand index set corresponding to all application data flows, and executing the following operations:
Setting the maximum iteration times, generating a plurality of simulation individuals, and forming a population set by all the simulation individuals; generating a simulation individual, wherein the simulation allocation duty ratio fuzzy language value and the simulation transmission priority fuzzy language value corresponding to each application data stream are randomly generated based on the allocation duty ratio fuzzy language value interval and the simulation transmission priority interval for all application data streams, and the simulation allocation duty ratio fuzzy language value and the simulation transmission priority fuzzy language value corresponding to all application data streams are formed into the simulation individual;
Calculating fitness according to any simulation individual, outputting a corresponding simulation distribution strategy through the operation of a bandwidth distribution strategy output layer based on all simulation distribution proportion fuzzy language values and all simulation transmission priority fuzzy language values corresponding to the simulation individual, and performing simulation on the transmission of all application data streams through the simulation distribution strategy under the constraint of monitoring environmental condition data, wherein the transmission rate in the transmission process is used as the fitness of the simulation individual;
simulating and updating a population set through a population optimization algorithm;
and outputting a simulation individual with highest fitness degree as an optimal simulation individual until the iteration number reaches the maximum iteration number, selecting a historical monitoring index data set and a historical application demand index set corresponding to any application data stream from the monitoring environment condition data to form a monitoring analysis training sample, and selecting a simulation allocation proportion fuzzy language value and a simulation transmission priority fuzzy language value corresponding to the application data stream in the monitoring analysis training sample from the optimal simulation individual as a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value.
As a preferred aspect, the population optimization algorithm is a genetic algorithm, including selection, recombination and mutation operations.
A high-speed connector-based signal high-speed transmission system, comprising:
The monitoring index data set acquisition module is used for acquiring a monitoring index data set of the physical link at the current monitoring time point, and the time period between adjacent monitoring time points is recorded as a preset detection period;
The application demand index set acquisition module is used for simultaneously acquiring application demand index sets corresponding to all current application data streams at the monitoring time point;
The application demand index time sequence prediction set output module is used for executing the following operations aiming at each application data stream, forming an application demand index time sequence prediction set S n by the currently acquired application demand index set F n and the application demand index set F n acquired M-1 times before, sending the demand index time sequence prediction set S n into an application demand index prediction model delta n for processing, and outputting an application demand prediction index set FR n corresponding to the next monitoring time point;
The bandwidth allocation strategy analysis module is used for forming a fuzzy decision set from a monitoring index data set acquired at the current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, sending the fuzzy decision set into the bandwidth allocation strategy analysis model for processing, outputting a bandwidth allocation strategy in the next preset period, establishing the bandwidth allocation strategy analysis model based on the fuzzy decision tree, training the bandwidth allocation strategy analysis model through a monitoring analysis training sample, marking the monitoring analysis training sample based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and acquiring the target allocation duty fuzzy language value and the target transmission priority fuzzy language value through a group optimization algorithm;
and the bandwidth allocation module is used for controlling bandwidth allocation in the next preset period based on the bandwidth allocation strategy.
The invention has the following advantages:
1. The method comprises the steps of obtaining historical application demand index sets of all application data streams, and carrying out time sequence analysis on the historical application demand index sets of all application data streams to obtain application demand prediction index sets corresponding to current monitoring time points; and carrying out fuzzy decision analysis based on the application demand prediction index set corresponding to the current monitoring time point and the monitoring index data set acquired by the current monitoring time point, and outputting a bandwidth allocation strategy corresponding to the next preset period, wherein the bandwidth allocation strategy can improve the overall throughput of the high-speed connector under different bandwidth demands and different network states, so that the high-speed transmission demand based on the high-speed connector is realized.
2. The invention carries out simulation training on the labeling conditions of the training set of the fuzzy decision analysis layer through a genetic algorithm, so that the fuzzy decision analysis layer obtained by training can be more attached to different bandwidth requirements and different network states, and the bandwidth allocation strategy can reach higher transmission rate.
Drawings
Fig. 1 is a schematic structural diagram of a signal high-speed transmission system based on a high-speed connector according to an embodiment of the present invention.
Detailed Description
In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Examples
A high-speed connector-based signal high-speed transmission method, comprising:
The method comprises the steps that a monitoring index data set of a physical link is obtained at a current monitoring time point, wherein the physical link refers to a copper cable or an optical fiber cable commonly used in a high-speed connector, each cable line forms an independent physical transmission channel, namely a physical link, and the monitoring index data set obtained at each physical link comprises a transmission rate, a time delay, a packet loss rate, an error rate and the like, and the obtaining means can be a hardware counter, an embedded probe and the like; the monitoring time points are determined by the configuration file, and the time period between the adjacent monitoring time points is recorded as a preset detection period;
Simultaneously acquiring an application demand index set F n corresponding to all current application data flows at a monitoring time point, wherein n=1, 2,3, N is the total number of the application data flows, the application data flows refer to data flows transmitted by different application programs through a high-speed connector, such as video data flows from video-on-demand application and HTTP request response data flows from web browsing; the application demand index set comprises bandwidth demand, time delay demand, loss tolerance and the like of the application data stream during transmission, the demand of the current application program on the network link performance can be reflected, the acquisition of the application demand index set is acquired at an application layer, namely, a logic link between the end and the end, for example, a probe is arranged at a client, and the real-time code rate corresponding to the application data stream can be acquired (the bandwidth demand of the application data stream can be reflected); for each application data stream, the following operation is executed, the application demand index set F n obtained currently and the application demand index set F n obtained M-1 times before are formed into an application demand index time sequence prediction set S n, the description and the depiction of the application demand change condition in the past period are reflected, then the demand index time sequence prediction set S n is sent into an application demand index prediction model delta n for processing, the application demand index set FR n corresponding to the next monitoring time point is output, and the application demand index generally shows a certain time correlation and periodicity rule, for example, the demand of certain applications in a specific time period is obviously increased (such as daily working time peak of a video conference). Capturing and exploiting this intrinsic law helps to predict future demands more accurately;
The method comprises the steps that a fuzzy decision set is formed by a monitoring index data set obtained at a current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, the fuzzy decision set is sent to a bandwidth allocation strategy analysis model to be processed, a bandwidth allocation strategy in the next preset period is output, the bandwidth allocation strategy analysis model is built based on a fuzzy decision tree, the bandwidth allocation strategy analysis model is trained through monitoring analysis training samples, the monitoring analysis training samples are marked based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and the target allocation duty fuzzy language value and the target transmission priority fuzzy language value are obtained through a group optimization algorithm; the bandwidth allocation strategy analysis model established through the fuzzy decision tree can well process the uncertainty of the bandwidth requirement and the network state; the bandwidth allocation policy includes bandwidth duty ratio and transmission priority corresponding to each application data stream, for example, the bandwidth allocation policy is { a:40%, B:30%, C:30% }, indicating that the bandwidth corresponding to the application data stream a is 40%, the bandwidth corresponding to the application data stream B is 30%, the bandwidth corresponding to the application data stream C is 30%, and the transmission priority is { a > B > C };
the bandwidth allocation strategy is controlled in the next preset period based on the bandwidth allocation strategy, specifically, the bandwidth allocation strategy is issued to the high-speed connector, the resource access port is called, and the bandwidth allocation strategy is executed, for example, the bandwidth allocation strategy is { A:40%, B:30%, C:30% }, which means that in all time segments corresponding to each clock cycle, the application data stream a is transmitted at the rate of full-load physical link and occupies 40% of the time segments, and similarly, the application data stream B is transmitted at the rate of full-load physical link and occupies 30% of the time segments, the application data stream C is transmitted at the rate of full-load physical link and occupies 30% of the time segments, and in one time period, the application data stream a is transmitted first, the application data stream B is transmitted, and finally, the application data stream C is transmitted, so as to realize the link multiplexing of the physical link in the high-speed connector;
The method comprises the steps of obtaining historical application demand index sets of all application data streams, and carrying out time sequence analysis on the historical application demand index sets of all application data streams to obtain application demand prediction index sets corresponding to current monitoring time points; and carrying out fuzzy decision analysis based on the application demand prediction index set corresponding to the current monitoring time point and the monitoring index data set acquired by the current monitoring time point, and outputting a bandwidth allocation strategy corresponding to the next preset period, wherein the bandwidth allocation strategy can improve the overall throughput of the high-speed connector under different bandwidth demands and different network states, so that the high-speed transmission demand based on the high-speed connector is realized.
The application demand index prediction model delta n is built based on a time sequence prediction model, and an LSTM model or a transducer model can be selected, the application demand index prediction model delta n is trained according to any application demand index prediction model delta n, and the training method comprises the following steps:
acquiring historical application demand index training samples of application data streams corresponding to a plurality of application demand index prediction models delta n, wherein M+1 historical application demand index sets arranged according to time sequence are stored in the historical application demand index training samples, the first M historical application demand index sets in the historical application demand index training samples are used as input data of an application demand index prediction model delta n, and the last 1 historical application demand index sets in the historical application demand index training samples are used as target data of an application demand index prediction model delta n; all the historical application demand index training samples form a historical application demand index training set, the historical application demand index training set is sent into an application demand index prediction model delta n with initialized parameters for training, an application demand index loss value is calculated, whether the application demand index loss value is in a first confidence range or not is judged, the first confidence range is set by professionals according to experience, and if the application demand index loss value is in the first confidence range, a trained application demand index prediction model delta n is output; otherwise, continuing to train the application demand index prediction model delta n through the historical application demand index training set.
The bandwidth allocation strategy analysis model comprises an input data fuzzy layer, a fuzzy decision analysis layer and a bandwidth allocation strategy output layer, wherein the input data fuzzy layer is used for fuzzifying a monitoring index data set and an application demand prediction index set FR n so as to construct the monitoring index fuzzy data set and the application demand prediction fuzzy index set FR n; the fuzzy decision analysis layer is used for outputting a corresponding allocation duty ratio fuzzy language value mu n and a transmission priority fuzzy language value epsilon n for each application data stream based on a fuzzy data set and an application demand prediction fuzzy index set FR n, is established based on a fuzzy decision tree and comprises two leaf nodes which are respectively marked as allocation duty ratio leaf nodes and transmission priority leaf nodes, the allocation duty ratio fuzzy language value mu n is used for representing bandwidth allocation suggestions of different degrees, and the transmission priority fuzzy language value epsilon n is used for representing different transmission priority suggestions; the bandwidth allocation policy output layer is used for outputting the bandwidth allocation policy based on the corresponding fuzzy language values output by all the application data streams.
The method comprises the steps of forming a fuzzy decision set by a monitoring index data set acquired at a current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, sending the fuzzy decision set into a bandwidth allocation strategy analysis model for processing, and outputting a bandwidth allocation strategy in a next preset period, wherein the method specifically comprises the following steps:
Carrying out fuzzification treatment on the monitoring index data set by a Gaussian fuzzification method to obtain a monitoring index fuzzification data set, wherein the specific operation is to calculate each monitoring index in the monitoring index data set according to a corresponding Gaussian function to obtain membership degrees, and combining the membership degrees corresponding to all the monitoring indexes to construct the monitoring index fuzzification data set;
Aiming at each application demand prediction index set FR n, carrying out fuzzification processing on the application demand prediction index set FR n by a Gaussian fuzzification method to obtain an application demand prediction fuzzification index set FR n, splicing the application demand prediction fuzzification index set FR n and the monitoring index fuzzification data set, then sending the obtained result to a fuzzy decision analysis layer for analysis, and recording an output leaf node as a minimum comprehensive fuzzification value in all paths of the distribution ratio leaf node and as a fuzzification language value mu n, wherein the comprehensive fuzzification value is the sum of all node fuzzification values in any path, and the node fuzzification value is the membership degree of an input fuzzification value at a node; meanwhile, the output leaf node is the minimum comprehensive fuzzy value in all paths of the transmission priority leaf node and is marked as a transmission priority fuzzy language value epsilon n;
in a bandwidth allocation strategy output layer, obtaining allocation duty ratio fuzzy language values mu n and transmission priority fuzzy language values epsilon n corresponding to all application data streams; performing normalization on the allocation duty ratio fuzzy language values mu n corresponding to all the application data flows, and outputting the bandwidth duty ratio X n corresponding to all the application data flows; normalizing the transmission priority fuzzy language values epsilon n corresponding to all the application data streams, and outputting the transmission priorities Y n corresponding to all the application data streams; and arranging the bandwidth occupation ratios X n corresponding to all the application data flows according to the sequence from the large to the small of the corresponding transmission priority Y n, and constructing a bandwidth allocation strategy.
Training of the fuzzy decision analysis layer inside the analysis model of the bandwidth allocation strategy comprises the following steps:
acquiring a plurality of monitoring analysis training samples, wherein the monitoring analysis training samples comprise a historical monitoring index data set and a historical application demand index set which are acquired at any monitoring time point, the historical application demand prediction index set is a historical application demand index set corresponding to any application data stream, the historical application demand index set can be set by a professional according to the requirement of the application data stream, the monitoring analysis training samples are subjected to fuzzification processing to obtain a fuzzy analysis training sample, the fuzzy analysis training sample comprises a historical monitoring index fuzzy data set and a historical application demand fuzzy index set, the fuzzy analysis training sample is marked by a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value, and the target allocation proportion fuzzy language value and the target transmission priority fuzzy language value are acquired by a group optimization algorithm; all fuzzy analysis training samples marked by the target allocation duty fuzzy language value and the target transmission priority fuzzy language value form a fuzzy analysis training set, the fuzzy analysis training set is sent to a fuzzy decision analysis layer in a bandwidth allocation strategy analysis model initialized by parameters for training, a comprehensive loss value is calculated, the comprehensive loss value comprises a loss value corresponding to the target allocation duty fuzzy language value and the target transmission priority fuzzy language value, whether the comprehensive loss value is located in a second confidence range is judged, a second confidence range leaf is set by a professional according to experience, and if the comprehensive loss value is located in the second confidence range, the fuzzy decision analysis layer in the trained bandwidth allocation strategy analysis model is output; otherwise, continuing to train the fuzzy decision analysis layer in the bandwidth allocation strategy analysis model through the fuzzy analysis training set.
The fuzzy language value of the target allocation duty ratio and the fuzzy language value of the target transmission priority are obtained through a group optimization algorithm, specifically comprising the following steps of, aiming at any monitoring environmental condition data, the monitoring environmental condition data comprises a historical monitoring index data set and a historical application demand index set corresponding to all application data streams, executing the following operations:
setting the maximum iteration times, generating a plurality of simulation individuals, and forming a population set by all the simulation individuals; generating a simulation individual, wherein the simulation allocation duty ratio fuzzy language value and the simulation transmission priority fuzzy language value corresponding to each application data stream are randomly generated based on the allocation duty ratio fuzzy language value interval and the simulation transmission priority interval for all application data streams, the allocation duty ratio fuzzy language value interval and the simulation transmission priority interval are set by a professional user, and the simulation allocation duty ratio fuzzy language value corresponding to all application data streams and the simulation transmission priority fuzzy language value corresponding to all application data streams form the simulation individual;
Calculating fitness according to any simulation individual, outputting a corresponding simulation distribution strategy through the operation of a bandwidth distribution strategy output layer based on all simulation distribution proportion fuzzy language values and all simulation transmission priority fuzzy language values corresponding to the simulation individual, and performing simulation on the transmission of all application data streams through the simulation distribution strategy under the constraint of monitoring environmental condition data, wherein the transmission rate in the transmission process is used as the fitness of the simulation individual;
Simulation updating is carried out on the population set through a genetic algorithm, selection, recombination and mutation operations are executed in the process, the selection operations are based on the fitness of all simulation individuals, the selection of the simulation individuals is carried out through a roulette algorithm, and all the selected simulation individuals form a father set; the recombination operation is that any two simulation individuals in the father set are selected, a part of the simulation distribution proportion fuzzy language value in the simulation individuals is selected at random, the simulation distribution proportion fuzzy language value of the two simulation individuals at the position is replaced, a part of the simulation transmission priority fuzzy language value is also selected at random, and the simulation transmission priority fuzzy language value of the two simulation individuals at the position is replaced; the mutation operation is that a mutation value is randomly generated for each simulated individual after the recombination operation, and is judged with a mutation threshold value which is set in advance, if the mutation value is higher than the mutation threshold value, a position is randomly selected for the simulated allocation proportion fuzzy language value part in the simulated individual, the simulated allocation proportion fuzzy language value of the simulated individual in the position is randomly set, a position is randomly selected for the simulated transmission priority fuzzy language value part in the simulated individual, and the simulated transmission priority fuzzy language value of the simulated individual in the position is randomly set;
and outputting a simulation individual with highest fitness degree as an optimal simulation individual until the iteration number reaches the maximum iteration number, selecting a historical monitoring index data set and a historical application demand index set corresponding to any application data stream from the monitoring environment condition data to form a monitoring analysis training sample, and selecting a simulation allocation proportion fuzzy language value and a simulation transmission priority fuzzy language value corresponding to the application data stream in the monitoring analysis training sample from the optimal simulation individual as a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value.
The application carries out simulation training on the labeling conditions of the training set of the fuzzy decision analysis layer through a genetic algorithm, so that the fuzzy decision analysis layer obtained by training can be more attached to different bandwidth requirements and different network states, and the bandwidth allocation strategy can reach higher transmission rate.
Examples
A high-speed connector-based signal high-speed transmission system, as shown in fig. 1, comprising:
The monitoring index data set acquisition module is used for acquiring a monitoring index data set of a physical link at a current monitoring time point, and it is to be noted that the physical link refers to a copper cable or an optical fiber cable commonly used in a high-speed connector, each cable line forms an independent physical transmission channel, namely a physical link, and the monitoring index data set acquired in each physical link comprises a transmission rate, a time delay, a packet loss rate, an error rate and the like, and the acquisition means can be a hardware counter, an embedded probe and the like; the monitoring time points are determined by the configuration file, and the time period between the adjacent monitoring time points is recorded as a preset detection period;
The application demand index set acquisition module is used for simultaneously acquiring application demand index sets corresponding to all current application data flows at a monitoring time point, wherein the application data flows refer to data flows transmitted by different application programs through a high-speed connector, such as video data flows from video-on-demand application and HTTP request response data flows from web browsing; the application demand index set comprises bandwidth demand, time delay demand, loss tolerance and the like of the application data stream during transmission, the demand of the current application program on the network link performance can be reflected, the acquisition of the application demand index set is acquired at an application layer, namely, a logic link between the end and the end, for example, a probe is arranged at a client, and the real-time code rate corresponding to the application data stream can be acquired (the bandwidth demand of the application data stream can be reflected);
The application demand index time sequence prediction set output module is used for executing the following operations aiming at each application data stream, forming an application demand index time sequence prediction set S n by the currently acquired application demand index set F n and the previously acquired application demand index set F n for M-1 times, representing the description and the depiction of the application demand change condition in the past period, sending the demand index time sequence prediction set S n into an application demand index prediction model delta n for processing, outputting an application demand index prediction set FR n corresponding to the next monitoring time point, wherein the application demand index generally shows a certain time correlation and periodicity rule, such as that the demand of certain applications is obviously increased in a specific time period (such as daily working time peak of a video conference). Capturing and exploiting this intrinsic law helps to predict future demands more accurately;
The bandwidth allocation strategy analysis module is used for forming a fuzzy decision set from a monitoring index data set acquired at the current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, sending the fuzzy decision set into the bandwidth allocation strategy analysis model for processing, outputting a bandwidth allocation strategy in the next preset period, establishing the bandwidth allocation strategy analysis model based on the fuzzy decision tree, training the bandwidth allocation strategy analysis model through a monitoring analysis training sample, marking the monitoring analysis training sample based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and acquiring the target allocation duty fuzzy language value and the target transmission priority fuzzy language value through a group optimization algorithm; the bandwidth allocation strategy analysis model established through the fuzzy decision tree can well process the uncertainty of the bandwidth requirement and the network state;
The bandwidth allocation module is configured to control bandwidth allocation in a next preset period based on a bandwidth allocation policy, specifically, issue the bandwidth allocation policy to the high-speed connector, and invoke the resource access port to execute the bandwidth allocation policy, for example, the bandwidth allocation policy is { a:40%, B:30%, C:30% }, it is stated that in all time segments corresponding to each clock cycle, the application data stream a is transmitted at the rate of fully loading the physical link, and occupies 40% of the time segments, and similarly, the application data stream B is transmitted at the rate of fully loading the physical link, and occupies 30% of the time segments, the application data stream C is transmitted at the rate of fully loading the physical link, and occupies 30% of the time segments, and in one time period, the application data stream a is transmitted first, then the application data stream B is transmitted, and finally the application data stream C is transmitted, so as to realize link multiplexing of the physical link in the high-speed connector.
It will be understood that modifications and variations will be apparent to those skilled in the art from the foregoing description, and it is intended that all such modifications and variations be included within the scope of the following claims. Parts of the specification not described in detail belong to the prior art known to those skilled in the art.
Claims (8)
1. A high-speed signal transmission method based on a high-speed connector, comprising:
acquiring a monitoring index data set of a physical link at a current monitoring time point, and recording a time period between adjacent monitoring time points as a preset detection period;
Simultaneously acquiring an application demand index set F n corresponding to all current application data flows at a monitoring time point, wherein n=1, 2,3, & gtN, wherein N is the total number of the application data flows; for each application data stream, performing the following operation, forming an application demand index time sequence prediction set S n by the currently acquired application demand index set F n and the application demand index set F n acquired M-1 times before, then sending the demand index time sequence prediction set S n into an application demand index prediction model delta n for processing, and outputting an application demand prediction index set FR n corresponding to the next monitoring time point;
The method comprises the steps that a fuzzy decision set is formed by a monitoring index data set obtained at a current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, the fuzzy decision set is sent to a bandwidth allocation strategy analysis model to be processed, a bandwidth allocation strategy in the next preset period is output, the bandwidth allocation strategy analysis model is built based on a fuzzy decision tree, the bandwidth allocation strategy analysis model is trained through monitoring analysis training samples, the monitoring analysis training samples are marked based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and the target allocation duty fuzzy language value and the target transmission priority fuzzy language value are obtained through a group optimization algorithm;
And controlling bandwidth allocation in the next preset period based on the bandwidth allocation strategy.
2. The high-speed signal transmission method based on a high-speed connector according to claim 1, wherein the application demand index prediction model δ n is established based on a time sequence prediction model, and training for any application demand index prediction model δ n comprises the following steps:
Acquiring historical application demand index training samples of application data streams corresponding to a plurality of application demand index prediction models delta n, wherein M+1 historical application demand index sets arranged according to time sequence are stored in the historical application demand index training samples, the first M historical application demand index sets in the historical application demand index training samples are used as input data of an application demand index prediction model delta n, and the last 1 historical application demand index sets in the historical application demand index training samples are used as target data of an application demand index prediction model delta n; all the historical application demand index training samples form a historical application demand index training set, the historical application demand index training set is sent into an application demand index prediction model delta n with initialized parameters for training, an application demand index loss value is calculated, whether the application demand index loss value is in a first confidence range is judged, and if the application demand index loss value is in the first confidence range, a trained application demand index prediction model delta n is output; otherwise, continuing to train the application demand index prediction model delta n through the historical application demand index training set.
3. The high-speed signal transmission method based on the high-speed connector according to claim 2, wherein the bandwidth allocation policy analysis model comprises an input data blurring layer, a blurring decision analysis layer and a bandwidth allocation policy output layer, wherein the input data blurring layer is used for blurring the monitoring index dataset and the application demand prediction index dataset FR n to construct the monitoring index blurring dataset and the application demand prediction blurring index dataset FR n; the fuzzy decision analysis layer is used for outputting a corresponding allocation duty ratio fuzzy language value mu n and a transmission priority fuzzy language value epsilon n to each application data stream based on a fuzzy data set and an application demand prediction fuzzy index set FR n, and is established based on a fuzzy decision tree and comprises two leaf nodes which are respectively marked as allocation duty ratio leaf nodes and transmission priority leaf nodes; the bandwidth allocation policy output layer is used for outputting the bandwidth allocation policy based on the corresponding fuzzy language values output by all the application data streams.
4. The high-speed signal transmission method based on high-speed connector as claimed in claim 3, wherein the fuzzy decision set is formed by a monitoring index data set obtained at the current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, and the fuzzy decision set is sent into a bandwidth allocation strategy analysis model for processing, and a bandwidth allocation strategy in the next preset period is output, and the method specifically comprises the following steps:
Carrying out fuzzification treatment on the monitoring index data set by a Gaussian fuzzification method to obtain a monitoring index fuzzification data set;
Aiming at each application demand prediction index set FR n, carrying out fuzzification processing on the application demand prediction index set FR n by a Gaussian fuzzification method to obtain an application demand prediction fuzzification index set FR n, splicing the application demand prediction fuzzification index set FR n and the monitoring index fuzzification data set, then sending the spliced application demand prediction fuzzification index set FR n and the monitoring index fuzzification data set into a fuzzification decision analysis layer for analysis, and marking the output leaf node as a minimum comprehensive fuzzification value in all paths of the distribution proportion leaf node as a fuzzification language value mu n; meanwhile, the output leaf node is the minimum comprehensive fuzzy value in all paths of the transmission priority leaf node and is marked as a transmission priority fuzzy language value epsilon n;
in a bandwidth allocation strategy output layer, obtaining allocation duty ratio fuzzy language values mu n and transmission priority fuzzy language values epsilon n corresponding to all application data streams; performing normalization on the allocation duty ratio fuzzy language values mu n corresponding to all the application data flows, and outputting the bandwidth duty ratio X n corresponding to all the application data flows; normalizing the transmission priority fuzzy language values epsilon n corresponding to all the application data streams, and outputting the transmission priorities Y n corresponding to all the application data streams; and arranging the bandwidth occupation ratios X n corresponding to all the application data flows according to the sequence from the large to the small of the corresponding transmission priority Y n, and constructing a bandwidth allocation strategy.
5. The high-speed connector-based signal high-speed transmission method according to claim 4, wherein training of the fuzzy decision analysis layer inside the bandwidth allocation policy analysis model comprises the steps of:
obtaining a plurality of monitoring analysis training samples, wherein the monitoring analysis training samples comprise a historical monitoring index data set and a historical application demand index set which are obtained at any monitoring time point, the historical application demand prediction index set is a historical application demand index set corresponding to any application data stream, the monitoring analysis training samples are subjected to fuzzification processing to obtain fuzzy analysis training samples, the fuzzy analysis training samples comprise a historical monitoring index fuzzy data set and a historical application demand fuzzy index set, the fuzzy analysis training samples are marked through a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value, and the target allocation proportion fuzzy language value and the target transmission priority fuzzy language value are obtained through a group optimization algorithm; all fuzzy analysis training samples marked by the target allocation duty fuzzy language value and the target transmission priority fuzzy language value form a fuzzy analysis training set, the fuzzy analysis training set is sent to a fuzzy decision analysis layer in a bandwidth allocation strategy analysis model initialized by parameters for training, a comprehensive loss value is calculated, whether the comprehensive loss value is in a second confidence range is judged, and if the comprehensive loss value is in the second confidence range, the fuzzy decision analysis layer in the trained bandwidth allocation strategy analysis model is output; otherwise, continuing to train the fuzzy decision analysis layer in the bandwidth allocation strategy analysis model through the fuzzy analysis training set.
6. The method for high-speed signal transmission based on high-speed connector according to claim 5, wherein the method for obtaining the target allocation duty ratio fuzzy language value and the target transmission priority fuzzy language value by a group optimization algorithm specifically comprises the following steps, for any monitored environmental condition data, the monitored environmental condition data comprises a historical monitoring index data set and a historical application demand index set corresponding to all application data streams, and the following operations are performed:
Setting the maximum iteration times, generating a plurality of simulation individuals, and forming a population set by all the simulation individuals; generating a simulation individual, wherein the simulation allocation duty ratio fuzzy language value and the simulation transmission priority fuzzy language value corresponding to each application data stream are randomly generated based on the allocation duty ratio fuzzy language value interval and the simulation transmission priority interval for all application data streams, and the simulation allocation duty ratio fuzzy language value and the simulation transmission priority fuzzy language value corresponding to all application data streams are formed into the simulation individual;
Calculating fitness according to any simulation individual, outputting a corresponding simulation distribution strategy through the operation of a bandwidth distribution strategy output layer based on all simulation distribution proportion fuzzy language values and all simulation transmission priority fuzzy language values corresponding to the simulation individual, and performing simulation on the transmission of all application data streams through the simulation distribution strategy under the constraint of monitoring environmental condition data, wherein the transmission rate in the transmission process is used as the fitness of the simulation individual;
simulating and updating a population set through a population optimization algorithm;
and outputting a simulation individual with highest fitness degree as an optimal simulation individual until the iteration number reaches the maximum iteration number, selecting a historical monitoring index data set and a historical application demand index set corresponding to any application data stream from the monitoring environment condition data to form a monitoring analysis training sample, and selecting a simulation allocation proportion fuzzy language value and a simulation transmission priority fuzzy language value corresponding to the application data stream in the monitoring analysis training sample from the optimal simulation individual as a target allocation proportion fuzzy language value and a target transmission priority fuzzy language value.
7. The high-speed connector-based signal transmission method according to claim 6, wherein the population optimization algorithm is a genetic algorithm including selection, recombination and mutation operations.
8. A high-speed connector-based signal high-speed transmission system, wherein the system applies the high-speed connector-based signal high-speed transmission method according to any one of the preceding claims 1 to 7, and comprises:
The monitoring index data set acquisition module is used for acquiring a monitoring index data set of the physical link at the current monitoring time point, and the time period between adjacent monitoring time points is recorded as a preset detection period;
The application demand index set acquisition module is used for simultaneously acquiring application demand index sets corresponding to all current application data streams at the monitoring time point;
The application demand index time sequence prediction set output module is used for executing the following operations aiming at each application data stream, forming an application demand index time sequence prediction set S n by the currently acquired application demand index set F n and the application demand index set F n acquired M-1 times before, sending the demand index time sequence prediction set S n into an application demand index prediction model delta n for processing, and outputting an application demand prediction index set FR n corresponding to the next monitoring time point;
The bandwidth allocation strategy analysis module is used for forming a fuzzy decision set from a monitoring index data set acquired at the current monitoring time point and all application demand prediction index sets FR n corresponding to the current monitoring time point, sending the fuzzy decision set into the bandwidth allocation strategy analysis model for processing, outputting a bandwidth allocation strategy in the next preset period, establishing the bandwidth allocation strategy analysis model based on the fuzzy decision tree, training the bandwidth allocation strategy analysis model through a monitoring analysis training sample, marking the monitoring analysis training sample based on a target allocation duty fuzzy language value and a target transmission priority fuzzy language value, and acquiring the target allocation duty fuzzy language value and the target transmission priority fuzzy language value through a group optimization algorithm;
and the bandwidth allocation module is used for controlling bandwidth allocation in the next preset period based on the bandwidth allocation strategy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410603298.9A CN118200257B (en) | 2024-05-15 | 2024-05-15 | Signal high-speed transmission method and system based on high-speed connector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410603298.9A CN118200257B (en) | 2024-05-15 | 2024-05-15 | Signal high-speed transmission method and system based on high-speed connector |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118200257A CN118200257A (en) | 2024-06-14 |
| CN118200257B true CN118200257B (en) | 2024-07-16 |
Family
ID=91399139
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410603298.9A Active CN118200257B (en) | 2024-05-15 | 2024-05-15 | Signal high-speed transmission method and system based on high-speed connector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118200257B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118885817B (en) * | 2024-09-29 | 2025-01-21 | 南昌交通学院 | A sports data intelligent analysis method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101248619A (en) * | 2005-12-23 | 2008-08-20 | 香港应用科技研究院有限公司 | A Distributed Wireless Network with Dynamic Bandwidth Allocation |
| CN105897612A (en) * | 2016-06-06 | 2016-08-24 | 中国电子科技集团公司第三十研究所 | Multi-service dynamic bandwidth allocation method and system based on SDN (Software Defined Network) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7499453B2 (en) * | 2000-05-19 | 2009-03-03 | Cisco Technology, Inc. | Apparatus and methods for incorporating bandwidth forecasting and dynamic bandwidth allocation into a broadband communication system |
| US9398346B2 (en) * | 2007-05-04 | 2016-07-19 | Time Warner Cable Enterprises Llc | Methods and apparatus for predictive capacity allocation |
| EP3132640B1 (en) * | 2014-04-14 | 2018-07-04 | Hughes Network Systems, LLC | Apparatus and method for a bandwidth allocation approach in a shared bandwidth communications system |
| WO2020013884A1 (en) * | 2018-07-13 | 2020-01-16 | Google Llc | Machine-learned prediction of network resources and margins |
| CN115086244B (en) * | 2022-06-16 | 2024-06-18 | 平安付科技服务有限公司 | Bandwidth allocation method, device, computer equipment and storage medium |
| US12170608B2 (en) * | 2022-06-21 | 2024-12-17 | Juniper Networks, Inc. | Link behavior prediction for use in path selection |
| CN117376989A (en) * | 2023-11-07 | 2024-01-09 | 北京红山信息科技研究院有限公司 | Wireless network resource management system capable of accessing network on line |
-
2024
- 2024-05-15 CN CN202410603298.9A patent/CN118200257B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101248619A (en) * | 2005-12-23 | 2008-08-20 | 香港应用科技研究院有限公司 | A Distributed Wireless Network with Dynamic Bandwidth Allocation |
| CN105897612A (en) * | 2016-06-06 | 2016-08-24 | 中国电子科技集团公司第三十研究所 | Multi-service dynamic bandwidth allocation method and system based on SDN (Software Defined Network) |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118200257A (en) | 2024-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113726545B (en) | Network traffic generation method and device based on knowledge-enhanced generative confrontation network | |
| WO2022028304A1 (en) | Multimedia data processing method and apparatus, device and readable storage medium | |
| Alomar et al. | {CausalSim}: A causal framework for unbiased {Trace-Driven} simulation | |
| Liu et al. | When wireless video streaming meets AI: A deep learning approach | |
| Mei et al. | Realtime mobile bandwidth prediction using lstm neural network | |
| Cui et al. | TCLiVi: Transmission control in live video streaming based on deep reinforcement learning | |
| Wang et al. | A multitask learning-based network traffic prediction approach for SDN-enabled industrial Internet of Things | |
| CN118200257B (en) | Signal high-speed transmission method and system based on high-speed connector | |
| CN112733043B (en) | Comment recommendation method and device | |
| CN113011886B (en) | Method and device for determining account type and electronic equipment | |
| Meng et al. | Practically deploying heavyweight adaptive bitrate algorithms with teacher-student learning | |
| CN115348551A (en) | Lightweight service identification method and device, electronic equipment and storage medium | |
| CN114297338A (en) | Text matching method, apparatus, storage medium and program product | |
| Li et al. | High-precision cluster federated learning for smart home: An edge-cloud collaboration approach | |
| Lisas et al. | Sequential learning for modeling video quality of delivery metrics | |
| Yin et al. | Ant: Learning accurate network throughput for better adaptive video streaming | |
| US12380359B2 (en) | Constraint sampling reinforcement learning for recommendation systems | |
| Chen et al. | ASTPPO: A proximal policy optimization algorithm based on the attention mechanism and spatio–temporal correlation for routing optimization in software-defined networking | |
| CN118802602A (en) | Computing network service execution method, device, readable storage medium and program product | |
| Wu et al. | A user behavior-aware multi-task learning model for enhanced short video recommendation | |
| CN113971454A (en) | Deep learning model quantification method and related device | |
| Elwerghemmi et al. | An efficient stacked deep incremental model for online streaming video QoE prediction | |
| CN116743605A (en) | Network service quality prediction method and device | |
| CN116644180A (en) | Training method and training system for text matching model and text label determining method | |
| Wang et al. | Imitation Learning for Adaptive Video Streaming with Future Adversarial Information Bottleneck Principle |
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 |