CN115496175A - Newly-built edge node access evaluation method and device, terminal equipment and product - Google Patents
Newly-built edge node access evaluation method and device, terminal equipment and product Download PDFInfo
- Publication number
- CN115496175A CN115496175A CN202110682690.3A CN202110682690A CN115496175A CN 115496175 A CN115496175 A CN 115496175A CN 202110682690 A CN202110682690 A CN 202110682690A CN 115496175 A CN115496175 A CN 115496175A
- Authority
- CN
- China
- Prior art keywords
- edge node
- newly
- built
- bandwidth
- accessed
- 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.)
- Pending
Links
Images
Abstract
The invention discloses a method, a device, terminal equipment and a product for evaluating the access of a newly-built edge node, wherein the method comprises the following steps: the edge computing management platform receives an access request sent by a newly-built edge node; acquiring time sequence data of the bandwidth used by each accessed edge node in a first preset time, inputting the time sequence data into a bandwidth use predictor of the accessed edge node after pre-training is finished, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the consumed total bandwidth obtained by prediction from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within a second preset time in the future; and inputting a performance requirement text of the newly-built edge node and a residual available total bandwidth time sequence into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
Description
Technical Field
The invention relates to the technical field of communication, in particular to a method, a device, terminal equipment and a product for evaluating access of a newly-built edge node.
Background
With the rapid development of mobile communication technology, 5G technology has come up. In the 5G era, everything is interconnected, mass internet of things equipment extends upwards, bottlenecks are generated in the aspects of cloud computing mode data processing and cost and energy consumption, and meanwhile, extreme user experience also needs cloud contents to extend to users, so that the rapid development of the MEC (edge computing) technology is inevitable in technical evolution.
The MEC technology provides flexible network access capability and edge computing service at the edge of a mobile network, reduces network transmission and service delivery time delay, improves data security, and provides new development kinetic energy for the vertical industry. The edge computing nodes can be hierarchically deployed in a core computer room, an important convergence computer room, a common convergence computer room, an access park computer room and the like of a city according to the requirements of industry customers.
At present, the evaluation of the access of the newly-built 5G edge node to the bearer network is mainly realized through expert experience. However, the edge node serves industry customers, faces thousands of industries and hundreds of industries, has the characteristics of demand diversity, demand burstiness and the like, and evaluates the feasibility of accessing the newly-built 5G edge node to the bearer network in a manual experience mode, so that the problems of low efficiency, time and labor waste, insufficient precision in evaluation and the like exist.
Disclosure of Invention
The invention mainly aims to provide a method, a device, terminal equipment and a product for access evaluation of a newly-built edge node, and aims to improve the accuracy and efficiency of access evaluation of the newly-built edge node.
In order to achieve the above object, an embodiment of the present invention provides a method for evaluating access to a newly-built edge node, where the method is applied to an edge computing management platform, and the method includes the following steps:
the edge computing management platform receives an access request which is sent by a newly-built edge node and is accessed to a bearer network, wherein the access request carries a performance requirement text of the newly-built edge node;
according to the access request, acquiring the time sequence data of the bandwidth used by each accessed edge node in a first preset time, inputting the time sequence data of the used bandwidth to a bandwidth usage predictor of the accessed edge node which is trained in advance, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future;
subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
Optionally, the step of predicting to obtain a predicted value of a total bandwidth consumed by each accessed edge node within a second preset time in the future includes:
extracting time dynamic characteristics of the used bandwidth time sequence data by using a long-short term memory layer through the accessed edge node bandwidth usage predictor;
and coding the time dynamic characteristics into a plurality of characteristic vectors with fixed lengths, merging the characteristic vectors, decoding the merged characteristic vectors, and mapping a predicted value of the total bandwidth consumed by each accessed edge node in the second preset time in the future through a long-short term memory layer.
Optionally, the step of obtaining a feasibility evaluation result of the newly-built edge node accessing the bearer network by evaluation includes:
accessing an evaluator through the newly-built edge node, extracting text characteristics of the performance requirement text by using a long-term and short-term memory layer, and extracting a time sequence dynamic rule of a time sequence of the remaining available total bandwidth of the bearer network;
and respectively coding the text characteristics and the time sequence dynamic rules to obtain two characteristic vectors with fixed lengths, merging the two characteristic vectors, and decoding the merged characteristic vectors through a fully-connected neural network to obtain a feasibility evaluation result of the newly-built edge node accessing the carrier network.
Optionally, before the step of receiving, by the edge computing management platform, an access request for accessing to the bearer network sent by a newly-built edge node, the method further includes:
training to obtain an accessed edge node bandwidth usage predictor specifically comprises:
collecting historical accessed bandwidth time sequence data sets used by each edge node within first preset time, and marking the used bandwidth time sequence data within the first preset time of each accessed edge node with real data of the used bandwidth within next second preset time of the accessed edge node to obtain a data set of an accessed edge node bandwidth usage predictor model;
preprocessing data in the data set of the accessed edge node bandwidth usage predictor model;
dividing a preprocessed data set of the accessed edge node bandwidth usage predictor model into a predictor training set and a predictor testing set;
training the accessed edge node bandwidth by using a predictor model through the predictor training set;
and verifying the trained bandwidth usage predictor model of the accessed edge node through the predictor test set, and obtaining the finally trained bandwidth usage predictor of the accessed edge node after the model is converged.
Optionally, the step of receiving, by the edge computing management platform, an access request for accessing to the bearer network sent by the newly-built edge node further includes:
training to obtain a newly-built edge node access evaluator specifically comprises:
collecting a performance requirement set of the historical newly-built edge node;
subtracting the marked real data of the bandwidth used in the next second preset time of the accessed edge node from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence set of the bearer network in the second preset time in the future;
obtaining a marked feasibility evaluation result set of the newly-built edge node accessing the bearer network;
taking the performance demand set, the residual available total bandwidth time sequence set and the feasibility evaluation result set as a data set of the newly-built edge node access evaluator model;
preprocessing data in the data set of the newly-built edge node access evaluator model;
dividing a preprocessed data set of the newly-built edge node access evaluator model into an evaluator training set and an evaluator testing set;
training the newly-built edge node access evaluator model through the evaluator training set;
and verifying the trained newly-built edge node access evaluator model through the evaluator test set, and obtaining the finally-trained newly-built edge node access evaluator after the model is converged.
Optionally, the accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator are both formed by a coding and decoding neural network.
Optionally, before the step of inputting the performance requirement text of the newly-built edge node and the remaining available total bandwidth time series of the bearer network into the pre-trained newly-built edge node access evaluator, the method further includes:
preprocessing the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network; and/or
The step of obtaining the bandwidth time sequence data used by each accessed edge node within the first preset time according to the access request further comprises:
and preprocessing the used bandwidth time sequence data.
The embodiment of the present invention further provides a newly-built edge node access evaluation apparatus, where the newly-built edge node access evaluation apparatus includes:
the request receiving module is used for receiving an access request which is sent by a newly-built edge node and is used for accessing the bearer network, wherein the access request carries a performance requirement text of the newly-built edge node;
the prediction module is used for acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time according to the access request, inputting the used bandwidth time sequence data to a pre-trained accessed edge node bandwidth usage predictor, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and the evaluation module is used for inputting the performance requirement text of the newly-built edge node and the residual available total bandwidth time sequence of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
The embodiment of the present invention further provides a terminal device, where the terminal device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and when the computer program is executed by the processor, the method for evaluating access to a newly-built edge node is implemented.
An embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program, and when executed by a processor, the computer program implements the method for evaluating access to a newly-built edge node as described above.
According to the access evaluation method, the access evaluation device, the terminal equipment and the product of the newly-built edge node provided by the embodiment of the invention, an access request for accessing a bearer network, which is sent by the newly-built edge node, is received through an edge computing management platform, wherein the access request carries a performance requirement text of the newly-built edge node; according to the access request, acquiring the time sequence data of the bandwidth used by each accessed edge node in a first preset time, inputting the time sequence data of the used bandwidth to a bandwidth usage predictor of the accessed edge node which is trained in advance, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future; and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is pre-trained, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
Drawings
FIG. 1 is a schematic diagram of the structure of an LSTM neuron in an embodiment of the present invention;
fig. 2 is a schematic diagram of functional modules of a terminal device to which a newly-built edge node access evaluation device belongs;
fig. 3 is a flowchart illustrating an exemplary embodiment of a method for evaluating access to a newly-built edge node according to the present invention;
fig. 4 is a flowchart illustrating a method for evaluating access to a newly-built edge node according to another exemplary embodiment of the present invention;
FIG. 5 is a diagram illustrating a predictor model for bandwidth usage of an accessed edge node according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a model of a newly-built edge node access evaluator in the embodiment of the present invention.
The implementation, functional features and advantages of the present invention will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The main solution of the embodiment of the invention is as follows: receiving an access request which is sent by a newly-built edge node and is accessed to a bearer network through an edge computing management platform, wherein the access request carries a performance requirement text of the newly-built edge node; according to the access request, acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data into a pre-trained bandwidth use predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future; and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is pre-trained, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
The technical terms related to the embodiment of the invention are as follows:
MEC (Mobile Edge Computing), moving Edge Computing.
The coding and decoding neural network is a mode for organizing a cyclic neural network, is mainly used for solving the problem of sequence prediction with a plurality of inputs or a plurality of outputs, and comprises an encoder and a decoder. The encoder is responsible for encoding the input sequence word by word into a vector with fixed length, namely a context vector (context vector); the decoder is responsible for reading the context vector output by the encoder and generating the output sequence. In the embodiment of the invention, the encoder and the decoder are both composed of long-term and short-term memory neurons.
Wherein, the long-short-term memory (LSTM) is a special type of recurrent neural network, and by controlling the time for storing the values in the cache, the long-term information can be memorized, and the method is suitable for predicting the time sequence. Each neuron has four inputs and one output, and each neuron stores a memorized value. The LSTM neuron is shown in fig. 1, and the calculation formula involved is as follows:
Y t ＝σ(W′h t ) (7)
wherein, each LSTM neuron contains three gates: forgetting gate, input gate, output gate. Equation (1) represents a forgetting gate, new information is added in equations (2), (3), equation (4) fuses new and old information, and equations (5), (6) output information about the next timestamp that the LSTM unit has learned so far.
Long and short term memory spiritHas good effect on learning long-time sequence via network, and each connecting line in LSTM unit has corresponding weight, wherein x t Represents the input vector, h t Representing a hidden state, C t Representing the state of the neuron at time t, Y t Represents the output of the neuron, W is a trainable weight matrix, b is a bias vector, and σ represents the sigmoid activation function, omicron t Indicating the output result of the output gate, f t Output result of forgetting gate, i t Shows the output result of the input gate,
Respectively, representing the current temporal state.
In the embodiment of the invention, the feasibility of accessing the newly-built 5G edge node to the bearer network is evaluated in a manual experience mode in the conventional related scheme, and the related scheme has the problems of low efficiency, time and labor waste, inaccurate evaluation and the like.
Therefore, the embodiment of the invention provides a solution, and the predictor for bandwidth usage of the accessed edge node and the access evaluator of the newly-built edge node are respectively formed by building two different coding and decoding neural networks, so that feasibility evaluation of the newly-built edge node, especially the feasibility evaluation of the 5G edge node access bearer network can be automatically realized, and the accuracy and efficiency of operation and maintenance personnel on the access evaluation of the newly-built edge node are improved. The embodiment of the invention relates to a communication core network and an artificial intelligence technology, in particular to deep learning and 5G edge calculation.
In the scheme of the embodiment of the invention, an accessed edge node bandwidth usage predictor and a newly-built edge node access evaluator are configured, wherein the accessed edge node bandwidth usage predictor consists of a coding and decoding neural network, the predictor utilizes a long and short term memory layer to extract the time dynamic characteristics of the time sequence data of the used bandwidth within the latest T time of each accessed edge node, the time dynamic characteristics are coded into N characteristic vectors with fixed length by a coder and are combined and then decoded, and a decoder utilizes the long and short term memory layer to finally map the predicted value of the total bandwidth consumed by each accessed edge node within the future M time.
The estimator extracts the text characteristics of the performance requirement of the newly-built edge node i by using a long-short term memory layer and extracts the dynamic law of the time sequence of the remaining available total bandwidth sequence of the bearer network within the future M time output by the last predictor, the text characteristics are coded into two characteristic vectors with fixed lengths and then combined into one characteristic vector, the characteristic vector is decoded by the full-connection neural network, and finally the feasibility evaluation result of the newly-built edge node i accessing the bearer network is output. And feeding back the result to the edge computing management platform, thereby improving the accuracy and efficiency of operation and maintenance personnel for access evaluation of the newly-built edge node.
Specifically, referring to fig. 2, fig. 2 is a schematic diagram of functional modules of a terminal device to which the newly-built edge node access evaluation apparatus belongs. The access evaluation device for the newly-built edge node may be a device which is independent of the terminal device and can implement access evaluation for the newly-built edge node, and may be carried on the terminal device in a form of hardware or software. The terminal equipment can be an intelligent mobile terminal such as a mobile phone and a tablet personal computer, and can also be network equipment such as a server.
In this embodiment, the terminal device to which the new edge node access evaluation apparatus belongs at least includes an output module 110, a processor 120, a memory 130, and a communication module 140.
The memory 130 stores an operating system and a newly-built edge node access evaluation program; the output module 110 may be a display screen, a speaker, etc. The communication module 140 may include a WIFI module, a mobile communication module, a bluetooth module, and the like, and communicates with an external device or a server through the communication module 140.
As an embodiment, the newly created edge node access evaluation program in the memory 130 implements the following steps when being executed by the processor:
receiving an access request for accessing a bearer network, which is sent by a newly-built edge node, through an edge computing management platform, wherein the access request carries a performance requirement text of the newly-built edge node;
according to the access request, acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data into a pre-trained bandwidth use predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future;
subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
Further, the new edge node access evaluation program in the memory 130 when executed by the processor further implements the following steps:
extracting time dynamic characteristics of the used bandwidth time sequence data by using a long-short term memory layer through the accessed edge node bandwidth usage predictor;
and coding the time dynamic characteristics into a plurality of characteristic vectors with fixed lengths, merging the characteristic vectors, decoding the merged characteristic vectors, and mapping a predicted value of the total bandwidth consumed by each accessed edge node in the second preset time in the future through a long-short term memory layer.
Further, the new edge node access evaluation program in the memory 130 when executed by the processor further implements the following steps:
accessing an evaluator through the newly-built edge node, extracting text characteristics of the performance requirement text by using a long-term and short-term memory layer, and extracting a time sequence dynamic rule of a remaining available total bandwidth time sequence of the bearer network;
and respectively coding the text characteristics and the time sequence dynamic rule to obtain two characteristic vectors with fixed lengths, merging the two characteristic vectors, and decoding the merged characteristic vectors through a full-connection neural network to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
Further, the new edge node access evaluation program in the memory 130 when executed by the processor further implements the following steps:
training to obtain an accessed edge node bandwidth usage predictor specifically comprises:
collecting historical accessed bandwidth time sequence data sets used by each edge node within first preset time, and marking the used bandwidth time sequence data within the first preset time of each accessed edge node with real data of the used bandwidth within next second preset time of the accessed edge node to obtain a data set of an accessed edge node bandwidth usage predictor model;
preprocessing data in the data set of the accessed edge node bandwidth usage predictor model;
dividing a preprocessed data set of the accessed edge node bandwidth usage predictor model into a predictor training set and a predictor testing set;
training the accessed edge node bandwidth by using a predictor model through the predictor training set;
and verifying the trained bandwidth usage predictor model of the accessed edge node through the predictor test set, and obtaining the finally trained bandwidth usage predictor of the accessed edge node after the model is converged.
Further, the new edge node access evaluation program in the memory 130 when executed by the processor further implements the following steps:
training to obtain a newly-built edge node access evaluator specifically comprises:
collecting a performance requirement set of the historical newly-built edge node;
subtracting the marked real data of the bandwidth used in the next second preset time of the accessed edge node from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence set of the bearer network in the second preset time in the future;
obtaining a marked feasibility evaluation result set of the newly-built edge node accessing the bearer network;
taking the performance demand set, the residual available total bandwidth time sequence set and the feasibility evaluation result set as a data set of the newly-built edge node access evaluator model;
preprocessing the data in the data set of the newly-built edge node access evaluator model;
dividing a preprocessed data set of the newly-built edge node access evaluator model into an evaluator training set and an evaluator testing set;
training the newly-built edge node access evaluator model through the evaluator training set;
and verifying the trained newly-built edge node access evaluator model through the evaluator test set, and obtaining the finally-trained newly-built edge node access evaluator after the model is converged.
Further, the new edge node access evaluation program in the memory 130 when executed by the processor further implements the following steps:
preprocessing the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network; and/or
The step of obtaining the bandwidth time sequence data used by each accessed edge node within the first preset time according to the access request further comprises:
preprocessing the used bandwidth time sequence data;
according to the scheme, an edge computing management platform receives an access request which is sent by a newly-built edge node and is accessed to a bearer network, wherein the access request carries a performance requirement text of the newly-built edge node; according to the access request, acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data into a pre-trained bandwidth use predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future; and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is pre-trained, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network, so that the feasibility of accessing the newly-built edge node into the bearer network is evaluated by an accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator, and the accuracy and the efficiency of access evaluation of the newly-built edge node by operation and maintenance personnel are improved.
Based on the above terminal device architecture, but not limited to the above architecture, the method embodiment of the present invention is proposed.
Referring to fig. 3, fig. 3 is a flowchart illustrating an exemplary embodiment of a method for evaluating access to a newly-built edge node according to the present invention. The method is applied to an edge computing management platform and comprises the following steps:
step S101, an edge computing management platform receives an access request which is sent by a newly-built edge node and is accessed to a bearer network, wherein the access request carries a performance requirement text of the newly-built edge node;
specifically, in this embodiment, the feasibility of accessing the newly-built edge node to the bearer network is evaluated through the edge computing management platform, and in other embodiments, the feasibility of accessing the newly-built edge node to the bearer network may also be evaluated by using a corresponding network device, a server, or a network platform.
In specific implementation, when a new edge node is accessed to a bearer network, an access request for accessing the bearer network is submitted to an edge computing management platform, and a performance requirement text of the new edge node is carried in the request message, where the performance requirement text includes bandwidth, time delay, application scenario, application industry, SLA (Service-Level Agreement) Level, and the like.
In this embodiment, the performance requirement text is used for being subsequently input to a newly-built edge node access evaluator that has been pre-trained, so as to evaluate and obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
Step S102, according to the access request, obtaining the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data to a pre-trained accessed edge node bandwidth use predictor, and predicting to obtain the predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future;
after receiving an access request of accessing a bearer network sent by a newly-built edge node, an edge computing management platform acquires the bandwidth time sequence data used by each accessed edge node within the first preset time recently, inputs the bandwidth time sequence data used within the first preset time recently into a pre-trained bandwidth usage predictor of the accessed edge node, and predicts and obtains a predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future through the bandwidth usage predictor of the accessed edge node.
Specifically, as an implementation manner, after receiving an access request for accessing a bearer network sent by a newly-built edge node, an edge computing management platform acquires bandwidth timing sequence data used by each accessed edge node within a first preset time in the near term, inputs the bandwidth timing sequence data used within the first preset time in the near term to a pre-trained bandwidth usage predictor of the accessed edge node, and extracts time dynamic characteristics of the used bandwidth timing sequence data by using a long-short term memory layer through the bandwidth usage predictor of the accessed edge node; and coding the time dynamic characteristics into a plurality of characteristic vectors with fixed lengths, merging the characteristic vectors, decoding the merged characteristic vectors, and mapping a predicted value of the total bandwidth consumed by each accessed edge node in the second preset time in the future through a long-short term memory layer.
More specifically, as another embodiment, after receiving an access request sent by a newly-built edge node to access a bearer network and acquiring bandwidth timing data used by each accessed edge node within a first preset time, an edge computing management platform may further perform preprocessing on the bandwidth timing data used by each accessed edge node within the first preset time.
The preprocessing may be a normalization processing, and each dimension of the data may be processed separately when processing the calculation, where the processing includes: the data was attribute-wise (column-wise) minus its mean and divided by its variance. By standardizing the used bandwidth time sequence data, the accuracy of access feasibility evaluation of the newly-built edge node can be improved.
The specific implementation process can be as follows:
after receiving an access request for accessing the bearer network, sent by a newly-built edge node, the edge computing management platform acquires bandwidth time sequence data used within the latest T time (first preset time) of each accessed edge node, and sends the bandwidth time sequence data to a pre-configured data preprocessing module for standardized processing.
Inputting the data after standardization processing to an accessed edge node bandwidth usage predictor formed by a pre-trained coding and decoding neural network;
the accessed edge node bandwidth usage predictor utilizes the long and short term memory layer to extract the time dynamic characteristics of the time series data of the accessed bandwidth used in the latest T time (first preset time) of each edge node, the time dynamic characteristics are coded into N (wherein N is a positive integer) characteristic vectors with fixed lengths, the obtained characteristic vectors are merged and decoded, and a decoder finally maps the predicted value of the total bandwidth consumed by each accessed edge node in the future M time (second preset time) through the long and short term memory layer.
Step S103, subtracting the predicted value of the total bandwidth consumed by each accessed edge node in the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network in the second preset time in the future;
and the bandwidth usage predictor of the accessed edge node subtracts the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain the time sequence of the remaining available total bandwidth of the bearer network within the second preset time in the future, and provides the time sequence to the edge calculation management platform, so that the feasibility of accessing the bearer network by the newly-built edge node is evaluated by combining the time sequence of the remaining available total bandwidth of the bearer network with the performance requirement text of the newly-built edge node to obtain a feasibility evaluation result.
As an implementation manner, the total bandwidth of the bearer network may also be provided to an edge computing management platform by an accessed edge node bandwidth usage predictor, and the edge computing management platform subtracts a predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future, which is obtained by predicting, from the total bandwidth of the bearer network, to obtain a remaining available total bandwidth time series of the bearer network within the second preset time in the future.
And step S104, inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
Specifically, as an implementation manner, the edge computing management platform extracts a performance requirement text of the newly-built edge node carried in an access request, and inputs the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator that is pre-trained.
Accessing an evaluator through the newly-built edge node, extracting text characteristics of the performance requirement text by using a long-term and short-term memory layer, and extracting a time sequence dynamic rule of a time sequence of the remaining available total bandwidth of the bearer network; and respectively coding the text characteristics and the time sequence dynamic rule to obtain two characteristic vectors with fixed lengths, merging the two characteristic vectors, and decoding the merged characteristic vectors through a full-connection neural network to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
Further, before inputting the performance requirement text of the newly-built edge node and the remaining available total bandwidth time sequence of the bearer network into the newly-built edge node access evaluator after the pre-training, the performance requirement text of the newly-built edge node and the remaining available total bandwidth time sequence of the bearer network may be preprocessed.
The preprocessing the performance requirement text of the newly-built edge node comprises the following steps:
and performing text cleaning on the performance requirement text of the newly-built edge node, and serializing the text. Removing all punctuation marks, segmenting the text into words if the text is Chinese, unifying the letters into lower case if the text is English, and indexing (tokenize) each word at the same time, so that each segment of text is converted into a segment of index number, and zero padding is performed on sequences which do not reach the maximum text length.
More specifically, as another embodiment, a performance requirement text of the newly-built edge node i is obtained from the edge computing management platform, and the text is subjected to integer serialization processing by the data preprocessing module.
And simultaneously, respectively subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the future M time (second preset time) output by the bandwidth usage predictor of the accessed edge node from the total bandwidth of the bearer network to obtain the time sequence of the remaining available total bandwidth of the bearer network within the future M time.
And preprocessing the time sequence of the remaining available total bandwidth of the carrier network within the future M time, and inputting the preprocessed performance requirement text and the time sequence of the remaining available total bandwidth of the carrier network within the future M time into a newly-built edge node access evaluator formed by a coding and decoding neural network after the pre-training is finished.
The newly-built edge node access evaluator utilizes the long-short term memory layer to extract the text characteristics of the performance requirement text of the newly-built edge node i and extracts the bandwidth utilization forecast of the accessed edge nodeThe time sequence dynamic law of the remaining available total bandwidth time sequence of the bearing network within the future M time output by the device is respectively encoded into two feature vectors with fixed lengths, then the two feature vectors are combined into one feature vector, and then the feature vector is decoded through a full-connection neural network; finally outputting a feasibility evaluation result y of the newly-built edge node i accessing the bearer network i Wherein, y i =1 for accessible, y i =0 represents no access.
And the newly-built edge node access evaluator feeds back the feasibility evaluation result to the edge computing management platform.
According to the scheme, an access request which is sent by a newly-built edge node and is accessed to a bearer network is received through an edge computing management platform, and the access request carries a performance requirement text of the newly-built edge node; according to the access request, acquiring the time sequence data of the bandwidth used by each accessed edge node in a first preset time, inputting the time sequence data of the used bandwidth to a bandwidth usage predictor of the accessed edge node which is trained in advance, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future; and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is pre-trained, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network, so that an accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator are respectively formed by two different coding and decoding neural networks, the feasibility of accessing the newly-built edge node into the bearer network is evaluated, and the accuracy and the efficiency of access evaluation of operation and maintenance personnel on the newly-built edge node are improved.
Referring to fig. 4, fig. 4 is a flowchart illustrating a method for evaluating access to a newly-built edge node according to another exemplary embodiment of the present invention. In this embodiment, in step S101, before the edge computing management platform receives an access request for accessing to a bearer network sent by a newly-built edge node, the method further includes:
step S1001, training to obtain an accessed edge node bandwidth usage predictor;
and step S1002, training to obtain a newly-built edge node access evaluator.
In this embodiment, the accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator are formed by a codec neural network.
As an implementation manner, the training of the accessed edge node bandwidth usage predictor may specifically include:
collecting historical accessed bandwidth time sequence data sets used by each edge node within first preset time, and marking the used bandwidth time sequence data within the first preset time of each accessed edge node with real data of the used bandwidth within next second preset time of the accessed edge node to obtain a data set of an accessed edge node bandwidth usage predictor model;
preprocessing data in the data set of the accessed edge node bandwidth usage predictor model;
dividing a preprocessed data set of the accessed edge node bandwidth usage predictor model into a predictor training set and a predictor testing set;
training the accessed edge node bandwidth by using a predictor model through the predictor training set;
and verifying the trained bandwidth usage predictor model of the accessed edge node through the predictor test set, and obtaining the finally trained bandwidth usage predictor of the accessed edge node after the model is converged.
Therefore, by the scheme, the accessed edge node bandwidth usage predictor is obtained through training, so that in the process of evaluating the feasibility of accessing the newly-built edge node into the bearer network, the predicted value of the total bandwidth consumed by each accessed edge node in the second preset time in the future is predicted and obtained through the accessed edge node bandwidth usage predictor.
As an implementation manner, the training to obtain the newly-built edge node access evaluator specifically may include:
collecting a performance requirement set of the historical newly-built edge node;
subtracting the marked real data of the bandwidth used in the next second preset time of the accessed edge node from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence set of the bearer network in the second preset time in the future;
obtaining a marked feasibility evaluation result set of the newly-built edge node accessing the bearer network;
taking the performance demand set, the residual available total bandwidth time sequence set and the feasibility evaluation result set as a data set of the newly-built edge node access evaluator model;
preprocessing data in the data set of the newly-built edge node access evaluator model;
dividing a preprocessed data set of the newly-built edge node access evaluator model into an evaluator training set and an evaluator testing set;
training the newly-built edge node access evaluator model through the evaluator training set;
and verifying the trained newly-built edge node access evaluator model through the evaluator test set, and obtaining the finally-trained newly-built edge node access evaluator after the model is converged.
Therefore, through the scheme, the newly-built edge node access evaluator is obtained through training, so that in the process of evaluating the feasibility of accessing the newly-built edge node into the bearer network through the newly-built edge node access evaluator, the feasibility evaluation result of accessing the newly-built edge node into the bearer network is obtained through evaluation.
More specifically, the process of training the bandwidth usage predictor of the accessed edge node and the new edge node access evaluator in this embodiment includes: data collection and preprocessing processes, and model building and training processes. Wherein:
the data collection and preprocessing process may include:
and collecting a historical accessed bandwidth time sequence data set used by each edge node in the latest T time from an edge computing management platform (MEPM), and marking the real data of the bandwidth used by each accessed edge node in the next M time to the data in the T time of each accessed edge node as a data set of a predictor model of the bandwidth use of the accessed edge node.
Meanwhile, a performance requirement text set of the historical newly-built edge node is collected from the edge computing management platform, and meanwhile, the real data of the bandwidth used by the accessed edge node within the next M time marked by the previous model (the accessed edge node bandwidth usage predictor model) is subtracted from the total bandwidth of the bearer network, so that a time sequence set of the remaining available total bandwidth of the bearer network within the future M time is formed.
And then, manually marking the feasibility evaluation result of the newly-built edge node accessing the bearer network as a data set of the newly-built edge node accessing the evaluator model.
Standardizing a bandwidth time sequence data set used by each accessed edge node within the latest T time: (X-mean)/std. The calculation is done separately for each dimension, subtracting the mean from the data by attribute (by column) and dividing by its variance. By standardizing the data, the convergence rate of the model and the accuracy of the model can be improved.
And then, performing text cleaning on the performance requirement text set of the newly-built edge node and serializing the text. Removing all punctuation, segmenting the text into words if the text is Chinese, unifying the letters into lower case if the text is English, and indexing (tokenize) each word at the same time so that each segment of text is converted into a segment of index number, and zero-filling the sequence that does not reach the maximum text length. The longest length L is taken as the index sequence length, and the dictionary size is taken as requirement _ vocab _ size.
Wherein, manually marking the feasibility evaluation result y of the newly-built edge node accessing the bearer network i ，y i =1 for accessible, y i =0 represents no access.
Finally, the obtained total data set is divided into a training set and a test set, and as an implementation mode, 90% of the total data set can be classified into the training set and 10% of the total data set can be classified into the test set. The training set is used to train the model and the test set is used to test the model.
For the model building and training process, the following scheme can be adopted:
in the scheme of the embodiment, two models need to be built and trained: the bandwidth of the accessed edge node uses a predictor and a newly-built edge node access evaluator.
The accessed edge node bandwidth usage predictor is composed of a coding and decoding neural network, extracts the time dynamic characteristics of the accessed edge node bandwidth time sequence data in the latest T time of each edge node by using a long short-term memory layer, codes the time dynamic characteristics into N characteristic vectors with fixed lengths, combines the characteristic vectors and decodes the characteristic vectors, and finally maps the predicted value of the total bandwidth consumed by each edge node in the future M time through the long short-term memory layer. The constructed accessed edge node bandwidth usage predictor model is shown in fig. 5.
Wherein, the coding part contains N branches, N is the number of edge nodes that have been accessed, and each branch contains two layers:
the first layer is an input layer: respectively inputting indexed bandwidth time sequence data used in the latest T time of the accessed edge nodes 1-N, so that the shapes of the output data of the layer are (None, T);
the second layer is the LSTM encoding layer: each layer contains 128 LSTM neurons, the activation function is set to be 'relu', the shape of output data of the layer is (None, T, 128) respectively, and the output data are coded into a context vector with fixed length;
the third layer is a combined layer (carbonate): splicing and merging the N context vectors with fixed lengths into 1 context vector h with fixed lengths according to the column dimension;
the fourth layer is an LSTM decoding layer: 128 LSTM neurons were included, and the activation function was set to "relu";
the fifth layer is a fully connected layer (sense): the number of the neurons is set to be M, namely a predicted value of the total bandwidth consumed by each edge node in the predicted future M time is output and is expressed as { z1, z2,. And zM }, and the activation function is set to be 'relu'.
The model will train 1000 rounds (epochs = 1000), set the batch size to 32 (batch _ size = 32), select the Mean Squared Error MSE (Mean Squared Error) as the loss function, i.e. the objective function (loss = 'Mean Squared Error'):
where loss represents the loss function, y i Representing the true value.
Wherein, the gradient descent optimization algorithm selects an adam optimizer for improving the learning speed of the traditional gradient descent (optimizer = 'adam'). The neural network can find the optimal weight value which enables the target function to be minimum through gradient descent, and the neural network can learn the weight value automatically through training. Training is performed with a training set so that the smaller the objective function, the better, and the validation model is evaluated with a test set after each round of training. And after the model is converged, deriving the weight of the model to finish model training.
For the newly-built edge node access evaluator, the newly-built edge node access evaluator is composed of a coding and decoding neural network, the newly-built edge node access evaluator utilizes a long-short term memory layer to extract text characteristics of a performance requirement text of a newly-built edge node i and a time sequence dynamic rule of a time sequence of remaining available total bandwidth of the carrier network within the future M time output by a last predictor, the text characteristics are respectively coded into two feature vectors with fixed lengths and then combined into one feature vector, the feature vectors are decoded by the full-connection neural network, and finally a feasibility evaluation result of the newly-built edge node i accessing the carrier network is output.
As shown in fig. 6, the newly-built edge node access evaluator model constructed in this embodiment includes:
branch 1:
the first layer isEmbedding layer (embedding): inputting a performance requirement text sequence { s) of the newly-built edge node i 1 i 、s 2 i 、...、s L i Converting each word into a vector by using word embedding (word embedding), wherein the dimension of input data is requirement _ vocab _ size, the output is set to be a space vector which needs to convert the word into 128 dimensions, the length of an input sequence is L, and therefore the shape of output data of the layer is (None, L, 128). The layer is used for carrying out vector mapping on input words and converting the index of each word into a 128-dimensional fixed shape vector;
the second layer is the LSTM encoding layer: 128 LSTM neurons are included, an activation function is set to be 'relu', the output data of the layer are respectively (None, L, 128) in shape and are coded into a context vector with fixed length;
and branch 2:
the first layer is an input layer: inputting the standardized total bandwidth sequence b-z remaining and available for the bearer network in the future M time 1 、b-z 2 、...、b-z M }；
The second layer is the LSTM encoding layer: the method comprises two parallel LSTM layers, each layer comprises 128 LSTM neurons, an activation function is set to be 'relu', the output data of the layers are (None, L, 128) and (None, n, 128) respectively, and the output data are coded into two context vectors with fixed length;
the third layer is a combined layer (carbonate): splicing and merging two context vectors with fixed length output by the two branches into 1 context vector h with fixed length according to the column dimension;
the fourth layer is an LSTM decoding layer: 128 LSTM neurons were included, and the activation function was set to "relu";
fifth full connection (Dense) layer (output layer): the number of the neurons is set to be 1, and a feasibility evaluation result y of the newly-built edge node i accessing the bearer network is output i ，y i =1 for accessible, y i =0 for inaccessible. The activation function is set to "sigmoid".
The error between the predicted and correct results is calculated for the above model, and the training objective is to minimize this error. The objective function selects a 'binary _ cross' two-class logarithmic loss function:
wherein loss represents a loss function, y i Representing the true value.
The number of training rounds is set to 1500 (epochs = 1500), and the gradient descent optimization algorithm selects an adam optimizer for improving the learning speed of the conventional gradient descent (optizer = 'adam'). The neural network can find the optimal weight value which enables the target function to be minimum through gradient descent, and the neural network can learn the weight value automatically through training. Training is performed with a training set so that the smaller the objective function, the better, and the validation model is evaluated with a test set after each round of training. And after the model is converged, deriving the weight of the model to finish model training.
Compared with the prior art, the method and the device have the advantages that the two different coding and decoding neural networks are built to respectively form the accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator, and training is carried out, wherein the accessed edge node bandwidth usage predictor utilizes the long and short term memory layer to extract the time dynamic characteristics of the accessed bandwidth time sequence data used by each edge node within the latest T time, the time dynamic characteristics are coded into N characteristic vectors with fixed lengths and are combined and then decoded, and the decoder finally maps the predicted value of the total bandwidth consumed by each edge node within the future M time through the long and short term memory layer; the newly-built edge node access evaluator extracts text characteristics of performance requirements of a newly-built edge node i by using a long-short term memory layer, extracts a time sequence dynamic rule of a time sequence of the remaining available total bandwidth of the bearer network within the future M time output by a predictor, respectively codes the two feature vectors into two feature vectors with fixed lengths, combines the two feature vectors into one feature vector, decodes the feature vector by using a fully-connected neural network, finally outputs a feasibility evaluation result of the newly-built edge node i accessing the bearer network, and feeds the result back to an edge calculation management platform, so that the accuracy and the efficiency of operation and maintenance personnel on the access evaluation of the newly-built edge node are improved.
In addition, an embodiment of the present invention further provides a newly-built edge node access evaluation apparatus, where the newly-built edge node access evaluation apparatus includes:
the request receiving module is used for receiving an access request which is sent by a newly-built edge node and is used for accessing the bearer network, wherein the access request carries a performance requirement text of the newly-built edge node;
the prediction module is used for acquiring the time sequence data of the bandwidth used by each accessed edge node in a first preset time according to the access request, inputting the time sequence data of the used bandwidth to a pre-trained bandwidth usage predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and the evaluation module is used for inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network.
In this embodiment, please refer to the above embodiments, which are not described herein again.
In addition, an embodiment of the present invention further provides a terminal device, where the terminal device includes a memory, a processor, and a computer program that is stored in the memory and is executable on the processor, and when the computer program is executed by the processor, the method for evaluating access to a newly created edge node according to the above embodiment is implemented.
Since the newly-created edge node access evaluation program is executed by the processor, all technical solutions of all the foregoing embodiments are adopted, so that at least all beneficial effects brought by all the technical solutions of all the foregoing embodiments are achieved, and details are not repeated herein.
In addition, an embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program, and when executed by a processor, the computer program implements the access evaluation method for a newly created edge node according to the above embodiment.
Since the access evaluation program of the newly-built edge node is executed by the processor, all technical solutions of all the embodiments are adopted, so that at least all the beneficial effects brought by all the technical solutions of all the embodiments are achieved, and detailed description is omitted here.
Compared with the prior art, the access evaluation method, the access evaluation device, the terminal equipment and the product of the newly-built edge node provided by the embodiment of the invention receive an access request for accessing a bearer network, which is sent by the newly-built edge node, through an edge computing management platform, wherein the access request carries a performance requirement text of the newly-built edge node; according to the access request, acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data into a pre-trained bandwidth use predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future; and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is pre-trained, and evaluating to obtain a feasibility evaluation result of the newly-built edge node accessing the bearer network, so that the feasibility of accessing the newly-built edge node into the bearer network is evaluated by an accessed edge node bandwidth usage predictor and the newly-built edge node access evaluator, and the accuracy and the efficiency of access evaluation of the newly-built edge node by operation and maintenance personnel are improved.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the description of the foregoing embodiments, it is clear to those skilled in the art that the method of the foregoing embodiments may be implemented by software plus a necessary general hardware platform, and certainly may also be implemented by hardware, but in many cases, the former is a better implementation. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, a controlled terminal, or a network device) to execute the method of each embodiment of the present invention.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes performed by the present invention or directly or indirectly applied to other related technical fields are also included in the scope of the present invention.
Claims (10)
1. A newly-built edge node access evaluation method is applied to an edge computing management platform and comprises the following steps:
the edge computing management platform receives an access request which is sent by a newly-built edge node and is accessed to a bearer network, wherein the access request carries a performance requirement text of the newly-built edge node;
according to the access request, acquiring the used bandwidth time sequence data of each accessed edge node in a first preset time, inputting the used bandwidth time sequence data into a pre-trained bandwidth use predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future;
subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and inputting the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
2. The method for evaluating access to a newly-built edge node according to claim 1, wherein the step of predicting a predicted value of a total bandwidth consumed by each accessed edge node within a second preset time in the future comprises:
extracting time dynamic characteristics of the used bandwidth time sequence data by using a long-short term memory layer through the accessed edge node bandwidth usage predictor;
and coding the time dynamic characteristics into a plurality of characteristic vectors with fixed lengths, merging the characteristic vectors, decoding the merged characteristic vectors, and mapping a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future through a long-short term memory layer.
3. The method for evaluating access of a newly-built edge node according to claim 1, wherein the step of evaluating a feasibility evaluation result of the newly-built edge node accessing the bearer network includes:
accessing an evaluator through the newly-built edge node, extracting text characteristics of the performance requirement text by using a long-term and short-term memory layer, and extracting a time sequence dynamic rule of a time sequence of the remaining available total bandwidth of the bearer network;
and respectively coding the text characteristics and the time sequence dynamic rules to obtain two characteristic vectors with fixed lengths, merging the two characteristic vectors, and decoding the merged characteristic vectors through a fully-connected neural network to obtain a feasibility evaluation result of the newly-built edge node accessing the carrier network.
4. The method for evaluating access to a newly-built edge node according to claim 1, wherein the step of receiving, by the edge computing management platform, an access request for accessing to a bearer network sent by the newly-built edge node further includes:
training to obtain an accessed edge node bandwidth usage predictor, which specifically comprises:
collecting historical accessed bandwidth time sequence data sets used by each edge node in first preset time, and labeling the used bandwidth time sequence data of each accessed edge node in the first preset time with real data of the used bandwidth of the accessed edge node in next second preset time to obtain a data set of an accessed edge node bandwidth usage predictor model;
preprocessing data in the data set of the accessed edge node bandwidth usage predictor model;
dividing a preprocessed data set of the usage predictor model of the accessed edge node bandwidth into a predictor training set and a predictor testing set;
training the accessed edge node bandwidth by using a predictor model through the predictor training set;
and verifying the trained bandwidth usage predictor model of the accessed edge node through the predictor test set, and obtaining the finally trained bandwidth usage predictor of the accessed edge node after the model is converged.
5. The method for evaluating access to a newly-built edge node according to claim 4, wherein the step of receiving, by the edge computing management platform, an access request for accessing to a bearer network sent by the newly-built edge node further comprises:
training to obtain a newly-built edge node access evaluator specifically comprises:
collecting a performance requirement set of a historical newly-built edge node;
subtracting the marked real data of the bandwidth used in the next second preset time of the accessed edge node from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence set of the bearer network in the second preset time in the future;
obtaining a marked feasibility evaluation result set of the newly-built edge node accessing the bearer network;
taking the performance demand set, the residual available total bandwidth time sequence set and the feasibility evaluation result set as a data set of the newly-built edge node access evaluator model;
preprocessing data in the data set of the newly-built edge node access evaluator model;
dividing a preprocessed data set of the newly-built edge node access evaluator model into an evaluator training set and an evaluator testing set;
training the newly-built edge node access evaluator model through the evaluator training set;
and verifying the trained newly-built edge node access evaluator model through the evaluator test set, and obtaining the finally-trained newly-built edge node access evaluator after the model is converged.
6. The newly-built edge node access assessment method according to any one of claims 1-5, wherein the bandwidth usage predictor of the accessed edge node and the access evaluator of the newly-built edge node are both composed of codec neural networks.
7. The method for evaluating the access of a newly-built edge node according to any of claims 1 to 5, wherein the step of inputting the performance requirement text of the newly-built edge node and the time series of the remaining available total bandwidth of the bearer network into the pre-trained newly-built edge node access evaluator is preceded by the steps of:
preprocessing the performance requirement text of the newly-built edge node and the time sequence of the remaining available total bandwidth of the bearer network; and/or
The step of obtaining the bandwidth time sequence data used by each accessed edge node within the first preset time according to the access request further comprises:
and preprocessing the used bandwidth time sequence data.
8. An access evaluation device for a newly-built edge node, the access evaluation device for the newly-built edge node comprising:
a request receiving module, configured to receive an access request for accessing a bearer network, where the access request is sent by a newly-created edge node and carries a performance requirement text of the newly-created edge node;
the prediction module is used for acquiring the time sequence data of the bandwidth used by each accessed edge node in a first preset time according to the access request, inputting the time sequence data of the used bandwidth to a pre-trained bandwidth usage predictor of the accessed edge node, and predicting to obtain a predicted value of the total bandwidth consumed by each accessed edge node in a second preset time in the future; subtracting the predicted value of the total bandwidth consumed by each accessed edge node within the second preset time in the future from the total bandwidth of the bearer network to obtain a remaining available total bandwidth time sequence of the bearer network within the second preset time in the future;
and the evaluation module is used for inputting the performance requirement text of the newly-built edge node and the residual available total bandwidth time sequence of the bearer network into a newly-built edge node access evaluator which is trained in advance, and evaluating to obtain a feasibility evaluation result of accessing the newly-built edge node into the bearer network.
9. A terminal device, characterized in that the terminal device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program, when executed by the processor, implementing the newly created edge node access evaluation method according to any one of claims 1 to 7.
10. A computer program product, characterized in that the computer program product comprises a computer program which, when being executed by a processor, implements the newly created edge node access evaluation method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110682690.3A CN115496175A (en) | 2021-06-18 | 2021-06-18 | Newly-built edge node access evaluation method and device, terminal equipment and product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110682690.3A CN115496175A (en) | 2021-06-18 | 2021-06-18 | Newly-built edge node access evaluation method and device, terminal equipment and product |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115496175A true CN115496175A (en) | 2022-12-20 |
Family
ID=84464113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110682690.3A Pending CN115496175A (en) | 2021-06-18 | 2021-06-18 | Newly-built edge node access evaluation method and device, terminal equipment and product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115496175A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116321244A (en) * | 2023-02-01 | 2023-06-23 | 广州爱浦路网络技术有限公司 | Method for setting timeliness of detailed information of N3IWFs/TNGFs, computer apparatus and storage medium |
-
2021
- 2021-06-18 CN CN202110682690.3A patent/CN115496175A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116321244A (en) * | 2023-02-01 | 2023-06-23 | 广州爱浦路网络技术有限公司 | Method for setting timeliness of detailed information of N3IWFs/TNGFs, computer apparatus and storage medium |
| CN116321244B (en) * | 2023-02-01 | 2023-12-15 | 广州爱浦路网络技术有限公司 | Method for setting timeliness of detailed information of N3IWFs/TNGFs, computer apparatus and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108536679B (en) | Named entity recognition method, device, equipment and computer readable storage medium | |
| CN111382555B (en) | Data processing method, medium, device and computing equipment | |
| CN109961041B (en) | Video identification method and device and storage medium | |
| CN110069612B (en) | Reply generation method and device | |
| EP3885966B1 (en) | Method and device for generating natural language description information | |
| CN111680147A (en) | Data processing method, device, equipment and readable storage medium | |
| CN113723166A (en) | Content identification method and device, computer equipment and storage medium | |
| CN111583911B (en) | Speech recognition method, device, terminal and medium based on label smoothing | |
| CN115795038B (en) | Intent recognition method and device based on localization deep learning framework | |
| CN111767697B (en) | Text processing method and device, computer equipment and storage medium | |
| CN110955765A (en) | Corpus construction method and apparatus of intelligent assistant, computer device and storage medium | |
| CN115496175A (en) | Newly-built edge node access evaluation method and device, terminal equipment and product | |
| CN111783688B (en) | Remote sensing image scene classification method based on convolutional neural network | |
| CN115604131B (en) | Link flow prediction method, system, electronic device and medium | |
| CN111382232A (en) | Question and answer information processing method and device and computer equipment | |
| CN111161238A (en) | Image quality evaluation method and device, electronic device, and storage medium | |
| CN114638308A (en) | Method and device for acquiring object relationship, electronic equipment and storage medium | |
| CN113282821A (en) | Intelligent application prediction method, device and system based on high-dimensional session data fusion | |
| CN115705464A (en) | Information processing method, device and equipment | |
| CN113822291A (en) | Image processing method, device, equipment and storage medium | |
| CN115512693A (en) | Audio recognition method, acoustic model training method, device and storage medium | |
| CN112463964A (en) | Text classification and model training method, device, equipment and storage medium | |
| CN115102852B (en) | Internet of things service opening method and device, electronic equipment and computer medium | |
| CN117713238B (en) | Random optimization operation strategy combining photovoltaic power generation and energy storage micro-grid | |
| CN114970955B (en) | Short video heat prediction method and device based on multi-mode pre-training model |
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 |