CN115801897A - Dynamic message processing method for edge proxy - Google Patents
Dynamic message processing method for edge proxy Download PDFInfo
- Publication number
- CN115801897A CN115801897A CN202211643197.1A CN202211643197A CN115801897A CN 115801897 A CN115801897 A CN 115801897A CN 202211643197 A CN202211643197 A CN 202211643197A CN 115801897 A CN115801897 A CN 115801897A
- Authority
- CN
- China
- Prior art keywords
- port
- message
- priority
- messages
- type
- 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.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims abstract description 71
- 238000012545 processing Methods 0.000 claims abstract description 32
- 238000000034 method Methods 0.000 claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 14
- 210000002569 neuron Anatomy 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 238000003062 neural network model Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 5
- 230000014509 gene expression Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 210000002364 input neuron Anatomy 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A dynamic message processing method for an edge proxy comprises the following steps of S1: establishing an initial weight matrix to express initial weight values when each port in the edge proxy equipment processes different types of messages; s2: establishing an influence factor matrix to express influence ratios of different influence factors on normal receiving of the ports when the ports process different types of messages; s3: calculating a priority weight matrix based on the initial weight matrix established in the step S1 and the influence factor matrix established in the step S2 so as to re-express the priority weight value when each port processes different types of messages; s4: and (4) adjusting the priority of each port when processing the message through the priority weight matrix established in the step (S3). By the scheme, the mode of adjusting the priority weights of different types of message processing according to the change of the condition factors better supports the service application to carry out service function processing according to the priority order.
Description
Technical Field
The invention relates to the technical field of digital informatization, in particular to a dynamic message processing method for an edge proxy.
Background
For edge proxy equipment (equipment providing an entry point for an enterprise core network), an objective evaluation method needs to be established for the sequence of processing different types of messages by different receiving ports, and the method has strong background significance if the processing priority is judged according to the importance degree of the different types of messages. However, in the actual operation of the power distribution scene, whether the port receives all types of messages, whether the communication flow is stable, and whether the message arrival timestamp is correct or not all make the judgment of the processing priority level incomplete and inaccurate only according to the important programs of different types of messages. Therefore, a way of dynamically adjusting the priority weights of different types of message processing according to the change of the condition factors needs to be established.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a dynamic message processing method for an edge proxy; the method realizes the mode of adjusting the priority weights of different types of message processing according to the change of the condition factors, and better supports the service application to carry out service function processing according to the priority order.
In order to achieve the purpose, the invention adopts the following technical scheme:
a dynamic message processing method for an edge proxy comprises the following steps:
s1: establishing an initial weight matrix to express initial weight values when each port in the edge proxy equipment processes different types of messages;
s2: establishing an influence factor matrix to express influence ratios of different influence factors on normal receiving of the ports when the ports process different types of messages;
s3: calculating a priority weight matrix based on the initial weight matrix established in the step S1 and the influence factor matrix established in the step S2 so as to re-express the priority weight value when each port processes different types of messages;
s4: and (4) adjusting the priority of each port when processing the message through the priority weight matrix established in the step (S3).
In order to optimize the technical scheme, the specific measures adopted further comprise:
further, in step S1, the expression of the initial weight matrix is:
in the formula (d) 1 ～d p Representing each port number in the edge proxy equipment; a is 1 ～a n Indicating the type of the received message; beta is a beta 1 ～β n The initial weight values corresponding to different types of messages processed by different port numbers are shown, and the initial weight values of the ports aiming at the same type of messages are consistent under the initial condition.
Further, in step S2, the expression of the impact factor matrix is:
in the formula, c 1 ～c m The factors are factors which can influence the normal message receiving of the port, namely influence factors; a is 1 ～a n Indicating the type of the received message; omega 1k1 ～ω mkn Representing the proportion of the impact of different impact factors on different types of message reception under each port, where ω is 1k1 The first parameter 1 in the subscript indicates the impact shadow c 1 The second parameter k represents the port number, which is d 1 ～d p The third parameter 1 represents the message type a 1 Thus ω 1k1 What is meant is that at port k the factor c is affected 1 For message type a 1 Proportion of influence of reception, likewise omega mkn What is meant is that factor c is affected at port k m For message type a n The received impact ratio.
Further, the specific content of step S3 is:
s31: based on the initial weight matrix established in the step S1 and the influence factor matrix established in the step S2, calculating the priority weight values of different ports for processing different types of messages:
for any port k, the processing message type is a n The priority weight calculation formula is:
in the formula, beta nk Indicates that the type of the processed message under the port k is a n A priority weight value of; beta is a n Indicating message type a n The initial weight value of (2) is obtained by the initial weight matrix in the step (S1); omega mkn Denotes the impact factor c at port k m For message type a n The influence proportion of (2) is obtained by the influence factor matrix in the step (S2); m below the sum character represents the impact factor c m When m =1, the influence factor c is expressed 1 When m =2, the influence factor c is expressed 2 And so on; n represents the total number and the type of factors influencing the normal message receiving of the port k in a period of time, namely the total number and the type of the influencing factors under the port k;
similarly, the priority weight of any port k when processing other different types of messages is obtained;
s32: according to the content of step S31, a priority weight matrix is established to re-express the priority weight value when each port processes different types of messages, and the priority weight matrix is as follows:
in the formula, beta 11 ～β np Represents priority weighted values corresponding to different types of messages processed by different port numbers, wherein beta 11 The first parameter 1 in the subscript indicates the message type a 1 The second parameter 1 represents the port number d 1 Thus beta 11 Meaning indicated as port d 1 Processing type message a 1 Corresponding priority weight value, likewise beta np Meaning indicated as port d p Handling message type a n The corresponding priority weight value.
Further, the specific content of step S4 is:
initializing a buffer pool of an FPGA module corresponding to each port according to the message type, wherein the initial weight values corresponding to the initialized buffer pool are the same;
carrying out matching updating on the initialization weighted value corresponding to each FPGA module according to the priority weight matrix;
and the FPGA module of each port extracts the messages in various message buffer pools according to the priority weight matrix, and transmits the messages to upper-layer application to finish the priority adjustment when each port processes the messages.
Further, the method also comprises the step S5: and the chain neural network is adopted to dynamically adjust the priority weight matrix at proper time so as to further optimize each parameter in the priority weight matrix and better support each service in the upper application.
Further, the specific content of step S5 is:
establishing a stack type self-coding neural network model structure, which comprises 1 input layer, 3 hidden layers and 1 output layer, wherein the number of neurons in the input layer is 30, the number of neurons in the 3 hidden layers is respectively 25, 20 and 15, and the number of neurons in the output layer is 18;
after the self-coding neural network model is successfully established, taking a flow sample and a protocol sample set (x) of each port in a period of time 1 ,x 2 ,x 3 ...x n ) T Training the model as an input signal, and carrying out quantitative perception on the input signal by each neuron in the model and carrying out gradient learning layer by layer; in the formula x 1 ～x n Representing a flow sample or protocol sample;
returning a mark of whether each type of message is successfully received in the upper application as information to a self-coding neural network model for feedback training, and updating the model by taking covariance as a loss function, so that the model realizes the characteristic learning of normal operation data, and an output vector is obtained, wherein the output vector contains effective protocol type messages obtained by various influence factors under different port conditions; therefore, according to the feedback requirement of the upper layer in a period of time, the parameters in the priority weight matrix are further dynamically adjusted, and for the message type with obvious feedback requirement in a period of time, the priority weight parameter of the corresponding port for processing the message type is increased, and otherwise, the priority weight parameter is reduced.
A computer-readable storage medium storing a computer program for causing a computer to execute the dynamic message processing method according to any one of the above.
An electronic device, comprising: the message dynamic processing method comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein when the processor executes the computer program, the message dynamic processing method is realized according to any one of the above items.
The beneficial effects of the invention are:
1. the method includes the steps that an initial weight matrix and an influence factor matrix are established, a priority weight matrix is designed based on the initial weight matrix, various influence factors in actual operation of a power distribution scene are fully considered in the matrix, and the weight value of a message processed by each port is dynamically adjusted. The application is different from the traditional technology, and only depends on the difference of message types, and different priority concepts are set.
2. The application further provides the capability of identifying and judging the background requirement by using the neural network, for example, in a period of time, the requirement of the upper layer application on a certain message is increased, and then the corresponding parameters in the priority weight matrix are increased, so that the effect of dynamic adjustment is further realized.
3. The edge proxy can configure the accurate arrival of the required message entering the upper application internal priority according to the service scene.
4. The edge proxy can be set according to factors influencing the stability of the message on the actual site, and the factors influencing the message receiving can be set according to the strategy of the actual site.
5. The method and the device can be used for training through the chain neural network, so that the optimal response weight factors can be formed by different ports under different influence factor conditions according to application requirements.
6. According to the method and the device, the service application can acquire the required message more preferentially according to the requirement, and the service application processing can be performed more rapidly.
Drawings
Fig. 1 is a schematic diagram of a neuron model in an existing neural network proposed in the present application.
Detailed Description
The present invention will now be described in further detail with reference to the accompanying drawings.
The whole technical scheme of the application is as follows:
a dynamic message processing method for an edge proxy comprises the following steps:
s1, designing a weighted graph to establish a weight matrix to express the initial weight value of each port for processing different types of messages;
s2, establishing an influence factor matrix according to influence factors which may influence the priority of message processing of the equipment port. The dynamic real-time update of the impact factor matrix, wherein the impact factors include but are not limited to communication flow, message arrival time stamp, time window and the like, and the item 4 in the beneficial effects can be known and set and considered according to the factors influencing the message stability in the actual field;
s3, calculating a final priority weighting matrix based on the initial weighting matrix and the influence factor matrix;
and S4, dynamically adjusting the priority weight matrix parameters by adopting a chain neural network, so that the priority weight matrix parameters reach an optimal distribution state under various interference conditions to support various service application functions.
Further explanation is as follows:
1) Firstly, a weighting matrix is established by utilizing a weighting graphTo express the initial weight value of each port for processing different types of messages. The initial weight value is designed according to the importance degree of the message, wherein d 1 ～d p Indicating the device port number, a 1 ～a n Indicating the type of message received, beta 1 ～β n Indicating the initial different port d p Handling different types of messages a n The weight value of (3). As follows:
description of the invention: under the initial condition, because the information of any message type received by the port is unknown, each port is set to the same initial weight value for each type of message. Wherein beta is n N of (2) represents a message type.
2) And establishing an influence factor matrix according to influence factors which may influence the priority of message processing of the equipment port, and dynamically updating the influence factor matrix in real time. Wherein a is 1 ～a n Indicating the type of message received, c 1 ～c m Representing influencing factors (e.g. traffic flow, message arrival timestamp, time window, etc.), ω mkn And (3) calculating the scale factor of the message type (subscript n represents the message type) by representing different influence factors (subscript m represents the influence factor type) under the k ports. As follows:
description of the invention: in the matrix of influence factors, ω 1k1 The first subscript 1 denotes the influence factor c 1 The second index k indicates the port number, and the third index 1 indicates the type a of the packet 1 。
3) And calculating a final priority weighting matrix according to the initial weighting matrix and the influence factor matrix. As follows:
for any port k, the type of the processed message is a n The priority weight calculation formula of (2) is:
similarly, the priority weight of any port k when processing other different types of messages is obtained;
thus, the priority weight matrix is as follows:
description of the drawings: in the final calculated priority matrix, in beta 11 For example: if the number of the influencing factors is 4 in a period of time, and the influencing factors are respectively c 1 、c 2 、c 3 、c 4 In this case, β 11 Has a value of (beta) 1 *ω 111 +β 1 *ω 211 +β 1 *ω 311 +β 1 *ω 411 ) And (3) is (a). Wherein beta is 11 The first 1 in (1) indicates the packet type and the second 1 indicates the port number. Finally, the weight value of each port is a value obtained after dynamic adjustment. If the number of the influence factors is 3 in another period of time, the corresponding weight value is calculated in the same way, so that the dynamic implementation adjustment is realized.
4) And initializing a buffer pool by the FPGA module corresponding to each port according to the message type, wherein the initial weights corresponding to the initialized buffer pool are the same.
5) And carrying out matching updating on the initialization weight corresponding to each FPGA according to the priority weight matrix.
6) And the FPGA module extracts various message buffer pool messages according to the priority weight matrix and uploads the messages to the service application.
The scheme can realize the problem proposed in the background technology, namely a mode of adjusting the priority weights of different types of message processing according to the change of condition factors; however, based on the above steps 1) to 6), the problem of the requirement of the upper layer application is not considered, that is, in step 6), each time the packet is extracted according to a certain proportion and uploaded to the upper layer application to wait for the upper layer application to process, but in practical situations, the upper layer application may need more packets in a certain time period, and if the packet is only uploaded according to a certain proportion, the requirement of the upper layer application is not considered, so the following sub-steps are continuously proposed:
7) The invention provides a stacked self-coding neural network structure comprising 3 hidden layers, 1 input layer and 1 output layer (5 layers in total), wherein the number of input neurons n =30, the number of hidden layers is S1=25, S2=20, S3=15 and the number of output layers is k = 18.
8) Each neuron in the network model inputs a signal x i (i =1 to n) is subjected to quantitative perception (the existing neuron model refers to fig. 1).
9) Network parameter P of each layer i Is denoted as P i ＝(w i ,b i ) (i = 1-3), where w, b are mathematical expressions of the weight matrix and intercept vector, respectively, in the network model.
10 After the network model is successfully established, taking each port flow sample and protocol sample set (x) 1 ,x 2 ,x 3 ...x n ) T And training the model, and learning the gradient layer by layer.
11 The application layer feedback signal 'this type of message is successfully received' is adopted as a mark to be given to feedback type training, and therefore normal operation data characteristic learning is achieved.
12 The loss function updates the network parameters by taking the covariance to finally obtain the output vector. The output vector comprises effective protocol type message matrixes obtained by various influence factors under different port conditions, so that the corresponding priority weight matrixes are matched, dynamic updating is carried out, and the optimal distribution state is achieved under various interference conditions. The concrete explanation is as follows: the message transmitted by the port of the edge proxy equipment is successfully received and processed within a period of time by the upper layer application, the message requirement, namely the service requirement, within the period of time of the upper layer application is judged, and then the parameters in the priority weight matrix are dynamically adjusted.
In conclusion, the overall technical scheme of the application is completed.
It should be noted that the terms "upper", "lower", "left", "right", "front", "back", etc. used in the present invention are for clarity of description only, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not limited by the technical contents of the essential changes.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.
Claims (9)
1. A dynamic message processing method for an edge proxy is characterized by comprising the following steps:
s1: establishing an initial weight matrix to express initial weight values when each port in the edge proxy equipment processes different types of messages;
s2: establishing an influence factor matrix to express the influence proportion of different influence factors on the normal receiving of the ports when the ports process different types of messages;
s3: calculating a priority weight matrix based on the initial weight matrix established in the step S1 and the influence factor matrix established in the step S2 so as to re-express the priority weight value when each port processes different types of messages;
s4: and (4) adjusting the priority of each port when processing the message through the priority weight matrix established in the step (S3).
2. The method according to claim 1, wherein in step S1, the expression of the initial weight matrix is:
in the formula, d 1 ～d p Representing each port number in the edge proxy equipment; a is 1 ～a n Indicating the type of the received message; beta is a 1 ～β n The initial weight values corresponding to different types of messages processed by different port numbers are shown, and the initial weight values of the ports for the same type of messages are consistent under the initial condition.
3. The method according to claim 2, wherein in step S2, the expression of the impact factor matrix is:
in the formula, c 1 ～c m The factors are factors which can influence the normal message receiving of the port, namely influence factors; a is a 1 ～a n Indicating the type of the received message; omega 1k1 ～ω mkn Representing the proportion of the impact of different impact factors on different types of message reception under each port, where ω is 1k1 The first parameter 1 in the subscript indicates the impact shadow c 1 The second parameter k represents the port number, which is d 1 ～d p The third parameter 1 represents the message type a 1 Thus ω 1k1 What is meant is that at port k the factor c is affected 1 For message type a 1 Proportion of influence received, for the same reason omega mkn What is meant is that factor c is affected at port k m For message type a n The received impact ratio.
4. The method for dynamically processing the packet of the edge proxy according to claim 3, wherein the specific content of the step S3 is:
s31: based on the initial weight matrix established in the step S1 and the influence factor matrix established in the step S2, calculating the priority weight values of different ports for processing different types of messages:
for any port k, the type of the processed message is a n The priority weight calculation formula is:
in the formula, beta nk Indicates that the type of the processed message under the port k is a n A priority weight value of; beta is a n Indicating message type a n The initial weight value of (2) is obtained by the initial weight matrix in the step (S1); omega mkn Denotes the impact factor c at port k m For message type a n The influence proportion of (2) is obtained by the influence factor matrix in the step (S2); m below the sum character represents the impact factor c m When m =1, the influence factor c is expressed 1 When m =2, the influence factor c is expressed 2 And so on; n represents the total number and the type of factors influencing the normal message receiving of the port k in a period of time, namely the total number and the type of the influencing factors under the port k;
similarly, the priority weight of any port k when processing other different types of messages is obtained;
s32: according to the content of step S31, a priority weight matrix is established to re-express the priority weight value when each port processes different types of messages, and the priority weight matrix is as follows:
in the formula, beta 11 ～β np Represents priority weighted values corresponding to different types of messages processed by different port numbers, wherein beta 11 The first parameter 1 in the subscript indicates the message type a 1 The second parameter 1 represents the port number d 1 Thus beta 11 Meaning indicated as port d 1 Processing type message a 1 Corresponding priority weight values, likewise beta np Meaning indicated as port d p Processing message type a n The corresponding priority weight value.
5. The method for dynamically processing the packet of the edge proxy according to claim 1, wherein the specific content of the step S4 is:
initializing a buffer pool of an FPGA module corresponding to each port according to the message type, wherein the initial weight values corresponding to the initialized buffer pool are the same;
carrying out matching updating on the initialization weighted value corresponding to each FPGA module according to the priority weight matrix;
and the FPGA module of each port extracts the messages in various message buffer pools according to the priority weight matrix, and transmits the messages to upper-layer application to finish the priority adjustment when each port processes the messages.
6. The method for dynamically processing packets of an edge proxy according to claim 1, further comprising the step S5: and the chain neural network is adopted to dynamically adjust the priority weight matrix at proper time so as to further optimize each parameter in the priority weight matrix and better support each service in the upper application.
7. The method for dynamically processing the packet of the edge proxy according to claim 6, wherein the specific content of step S5 is:
establishing a stack type self-coding neural network model structure, which comprises 1 input layer, 3 hidden layers and 1 output layer, wherein the number of neurons in the input layer is 30, the number of neurons in the 3 hidden layers is respectively 25, 20 and 15, and the number of neurons in the output layer is 18;
after the self-coding neural network model is successfully established, taking a flow sample and a protocol sample set (x) of each port in a period of time 1 ,x 2 ,x 3 ...x n ) T Training the model as an input signal, and carrying out quantitative perception on the input signal by each neuron in the model and carrying out gradient learning layer by layer; in the formula x 1 ～x n Representing a flow sample or protocol sample;
returning a mark of whether each type of message is successfully received in the upper application as information to a self-coding neural network model for feedback training, and updating the model by taking covariance as a loss function, so that the model realizes the characteristic learning of normal operation data, and an output vector is obtained, wherein the output vector contains effective protocol type messages obtained by various influence factors under different port conditions; therefore, according to the feedback requirement of the upper layer in a period of time, the parameters in the priority weight matrix are further dynamically adjusted, and for the message type with obvious feedback requirement in a period of time, the priority weight parameter of the corresponding port for processing the message type is increased, and otherwise, the priority weight parameter is reduced.
8. A computer-readable storage medium storing a computer program, the computer program causing a computer to execute the dynamic message processing method according to any one of claims 1 to 7.
9. An electronic device, comprising: memory, processor and computer program stored on the memory and executable on the processor, the processor implementing the message dynamic processing method according to any one of claims 1 to 7 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211643197.1A CN115801897B (en) | 2022-12-20 | Message dynamic processing method of edge proxy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211643197.1A CN115801897B (en) | 2022-12-20 | Message dynamic processing method of edge proxy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115801897A true CN115801897A (en) | 2023-03-14 |
| CN115801897B CN115801897B (en) | 2024-05-24 |
Family
ID=
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116684483A (en) * | 2023-08-02 | 2023-09-01 | 北京中电普华信息技术有限公司 | Method for distributing communication resources of edge internet of things proxy and related products |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007002443A2 (en) * | 2005-06-23 | 2007-01-04 | Network Appliance, Inc. | Control of service workload management |
| US20110158248A1 (en) * | 2009-12-24 | 2011-06-30 | Juniper Networks, Inc. | Dynamic prioritized fair share scheduling scheme in over-subscribed port scenario |
| CN104092622A (en) * | 2014-07-04 | 2014-10-08 | 杭州华三通信技术有限公司 | Message processing method and device |
| CN105490961A (en) * | 2014-09-19 | 2016-04-13 | 杭州迪普科技有限公司 | Message processing method, and device and network device |
| WO2016082603A1 (en) * | 2014-11-25 | 2016-06-02 | 深圳市中兴微电子技术有限公司 | Scheduler and dynamic multiplexing method for scheduler |
| CN107786469A (en) * | 2017-11-06 | 2018-03-09 | 烟台慧彦网络科技有限公司 | It is a kind of based on it is preceding to task estimate Internet of Things response type bandwidth adjustment method |
| CN107896192A (en) * | 2017-11-20 | 2018-04-10 | 电子科技大学 | The QoS control method of differentiated service priority in a kind of SDN |
| CN108600118A (en) * | 2018-03-30 | 2018-09-28 | 新华三技术有限公司 | Message processing method, device and electronic equipment |
| CN110417683A (en) * | 2019-07-24 | 2019-11-05 | 新华三大数据技术有限公司 | Message processing method, device and server |
| CN111985845A (en) * | 2020-09-02 | 2020-11-24 | 浙江工业大学 | Node priority tuning method for heterogeneous Spark cluster |
| CN115277888A (en) * | 2022-09-26 | 2022-11-01 | 中国电子科技集团公司第三十研究所 | Method and system for analyzing message type of mobile application encryption protocol |
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007002443A2 (en) * | 2005-06-23 | 2007-01-04 | Network Appliance, Inc. | Control of service workload management |
| US20110158248A1 (en) * | 2009-12-24 | 2011-06-30 | Juniper Networks, Inc. | Dynamic prioritized fair share scheduling scheme in over-subscribed port scenario |
| CN104092622A (en) * | 2014-07-04 | 2014-10-08 | 杭州华三通信技术有限公司 | Message processing method and device |
| CN105490961A (en) * | 2014-09-19 | 2016-04-13 | 杭州迪普科技有限公司 | Message processing method, and device and network device |
| WO2016082603A1 (en) * | 2014-11-25 | 2016-06-02 | 深圳市中兴微电子技术有限公司 | Scheduler and dynamic multiplexing method for scheduler |
| CN107786469A (en) * | 2017-11-06 | 2018-03-09 | 烟台慧彦网络科技有限公司 | It is a kind of based on it is preceding to task estimate Internet of Things response type bandwidth adjustment method |
| CN107896192A (en) * | 2017-11-20 | 2018-04-10 | 电子科技大学 | The QoS control method of differentiated service priority in a kind of SDN |
| CN108600118A (en) * | 2018-03-30 | 2018-09-28 | 新华三技术有限公司 | Message processing method, device and electronic equipment |
| CN110417683A (en) * | 2019-07-24 | 2019-11-05 | 新华三大数据技术有限公司 | Message processing method, device and server |
| CN111985845A (en) * | 2020-09-02 | 2020-11-24 | 浙江工业大学 | Node priority tuning method for heterogeneous Spark cluster |
| CN115277888A (en) * | 2022-09-26 | 2022-11-01 | 中国电子科技集团公司第三十研究所 | Method and system for analyzing message type of mobile application encryption protocol |
Non-Patent Citations (1)
| Title |
|---|
| 王慧;孙志刚;戴斌;何俊峰;龚正虎;: "大规模流媒体多播系统中动态流量控制机制", 通信学报, no. 10, 25 October 2010 (2010-10-25) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116684483A (en) * | 2023-08-02 | 2023-09-01 | 北京中电普华信息技术有限公司 | Method for distributing communication resources of edge internet of things proxy and related products |
| CN116684483B (en) * | 2023-08-02 | 2023-09-29 | 北京中电普华信息技术有限公司 | Method for distributing communication resources of edge internet of things proxy and related products |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111242282B (en) | Deep learning model training acceleration method based on end edge cloud cooperation | |
| CN111130839A (en) | Flow demand matrix prediction method and system | |
| CN111399933B (en) | DNN task unloading method and terminal in edge-cloud hybrid computing environment | |
| CN114912705A (en) | Optimization method for heterogeneous model fusion in federated learning | |
| US20210056416A1 (en) | Distributed Deep Learning System | |
| CN113642653A (en) | Complex value neural network signal modulation identification method based on structure optimization algorithm | |
| WO2023124947A1 (en) | Task processing method and apparatus, and related device | |
| CN111988225A (en) | Multi-path routing method based on reinforcement learning and transfer learning | |
| WO2019159783A1 (en) | Distributed deep learning system | |
| CN107483473A (en) | A kind of low speed Denial of Service attack data-flow detection method of cloud environment | |
| US20240037408A1 (en) | Method and apparatus for model training and data enhancement, electronic device and storage medium | |
| CN111917642B (en) | SDN intelligent routing data transmission method for distributed deep reinforcement learning | |
| CN112348655A (en) | Credit evaluation method based on AFSA-ELM | |
| US11468330B2 (en) | Artificial neural network growth | |
| Poveda et al. | A class of distributed adaptive pricing mechanisms for societal systems with limited information | |
| CN108280207A (en) | A method of the perfect Hash of construction | |
| CN110929041A (en) | Entity alignment method and system based on layered attention mechanism | |
| US10853689B2 (en) | Methods for more effectively moderating one or more images and devices thereof | |
| CN102331987A (en) | Patent data mining system and method | |
| CN114048838A (en) | Knowledge migration-based hybrid federal learning method | |
| CN117151208B (en) | Asynchronous federal learning parameter updating method based on self-adaptive learning rate, electronic equipment and storage medium | |
| CN110138681A (en) | A kind of network flow identification method and device based on TCP message feature | |
| CN115801897A (en) | Dynamic message processing method for edge proxy | |
| CN115801897B (en) | Message dynamic processing method of edge proxy | |
| CN117236421A (en) | Large model training method based on federal knowledge distillation |
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 |