CN112672368A - Method and system for dynamically deploying CU under CU Pool - Google Patents
Method and system for dynamically deploying CU under CU Pool Download PDFInfo
- Publication number
- CN112672368A CN112672368A CN202011495168.6A CN202011495168A CN112672368A CN 112672368 A CN112672368 A CN 112672368A CN 202011495168 A CN202011495168 A CN 202011495168A CN 112672368 A CN112672368 A CN 112672368A
- Authority
- CN
- China
- Prior art keywords
- data
- centralized unit
- platform
- machine learning
- pool
- 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
Images
Abstract
The invention discloses a method and a system for dynamically deploying CUs under CU Pool, wherein the method comprises the following steps: s1, constructing a machine learning model by adopting a preset method; s2, collecting sample data of the centralized unit platform extracted by the machine learning model and reporting information; s3, processing the reported sample data by adopting a control center to realize machine learning, prediction and decision; s4, realizing the data synchronization of the main and standby centralization units and the platform switching of the main and standby centralization units. Has the advantages that: by extracting key data of the system, a reasonable machine learning model can be constructed; through continuous machine training, the influence of user services on the CU service platform can be accurately predicted and evaluated, so that the optimal CU service platform in the CU POOL is selected to carry out dynamic CU module deployment; the performance and reliability of the system are effectively improved, the abnormal risk of the CUs is reduced, and reasonable distribution of the CU resources is realized.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to a method and a system for dynamically deploying CUs under CU Pool.
Background
With the development of wireless communication technology, 3GPP divides the 5G baseband unit into two logical nodes, CU and DU. Its CU is a centralized unit node responsible for handling higher layer protocol functions and managing multiple DUs. The DU is a distributed unit node and is responsible for completing part of the underlying baseband protocol processing functions. The CU is connected to the 5GC via the NG interface and to the DU via the F1 interface. The CU and DU together perform the entire NR baseband processing function. 3GPP discusses 8 candidate schemes for CU and DU split options, and the standardization work is mainly focused on Option2, namely CU mainly completes the upper layer processing function of RRC/PDCP baseband, and DU completes the functions of RLC, MAC layer and physical layer protocol.
From the base station deployment scene, the base station can be divided into a split architecture and an integrated architecture. For the base station separation architecture, CU and DU can be deployed in the same device (internal interface communication) or separately in different devices (F1 interface communication), and DU is connected to RU through frontaul interface. For the base station integration architecture, the DU and the RU are deployed in the same hardware entity, and communicate with the CU through an F1 interface.
Therefore, the logic splitting architecture of the 5G RAN brings great convenience to the rapid deployment of the 5G network system, and simultaneously puts higher requirements on the reliability of CUs. In one aspect, CUs can be flexibly deployed at HUs (host units) or cloud servers based on user needs. On the other hand, a CU is a centralized node that manages a plurality of DU units, and as the number of users and the traffic volume of a plurality of cells managed by the CU increase, the CU is limited by the CU platform resource capability, which may cause a rapid decrease in system performance (e.g., an increase in transmission delay, network congestion, a user drop, etc.), and may have a large impact on user experience.
In order to solve the above problems, in the prior art, load information of a plurality of CU platforms is detected, and when a CU platform with a high load is higher than a high threshold and a CU platform with a low load is lower than a low threshold, a CU function carried by the CU platform with the high load is transferred to the CU platform with the low load, so that multi-CU load balancing is achieved. The prior art has the following defects:
1) the CU load statistics is calculated by weighting a plurality of load influencing factors, and weights are configured by artificial parameters, so that inaccurate weights can cause inaccurate CU load estimation.
2) The difference of the CU unit operation platform resources (CPU, memory, bandwidth, etc.) is not considered, and the CU load high-low threshold decision may cause the low-performance CU platform to be crushed when the CU unit load of the high-performance CU platform is migrated to the low-performance CU platform.
An effective solution to the problems in the related art has not been proposed yet.
Disclosure of Invention
The invention provides a method and a system for dynamically deploying a CU under a CU Pool, aiming at the problems in the related art and solving the technical problems in the prior art.
Therefore, the invention adopts the following specific technical scheme:
according to an aspect of the present invention, there is provided a method for dynamically deploying a CU under a CU Pool, the method including the steps of:
s1, constructing a machine learning model by adopting a preset method;
s2, collecting sample data of the centralized unit platform extracted by the machine learning model and reporting information;
s3, processing the reported sample data by adopting a control center to realize machine learning, prediction and decision;
s4, realizing the data synchronization of the main and standby centralization units and the platform switching of the main and standby centralization units.
Further, the construction of the machine learning model by using the preset method comprises the following steps:
s11, simplifying a data model through system analysis, and extracting key data when the centralized unit operates;
and S12, completing the construction of the machine learning model by using a preset rule.
Further, the key data comprises service characteristic data and centralized unit platform target data;
the service characteristic data includes, but is not limited to, the number of service users, the service type and the throughput of a packet data convergence protocol, and the centralized unit platform target data includes, but is not limited to, the occupancy rate of a central processing unit, the occupancy rate of a memory, the occupancy rate of a link, and the average delay of the link.
Further, the construction of the machine learning model by using the preset rule comprises the following steps:
s121, describing a random linear relation between the centralized unit platform target data and the service characteristic data by the system through a multiple linear regression model, wherein the relation is as follows:
yi＝β0+β1xi1+β2xi2+β3xi3+εi，i＝1，2，…，n；
wherein y represents the performance evaluation target data index of the centralized unit platform, x1 is the number of service users, x2 is the service type, x3 is the service packet data convergence protocol, the throughput of the protocol data unit, and beta0，β1，β2，β3Denotes the regression coefficient,. epsiloniRepresenting random error terms, each epsiloniIndependent of each other and obey N (0, sigma)2) Distribution, n represents sample volume, and n sample observations are as follows:
(yi，xi1，xi2，xi3)i＝1，2，…，n；
s122, solving the multiple linear regression model by using a least square method or a gradient descent method, and defining a loss function, wherein the formula of the loss function is as follows:
wherein the content of the first and second substances,
to predict value, yiIs an observed value;
s123, performing partial derivation on the regression parameters to obtain a normal equation set, wherein the normal equation set has the following formula:
s124, converting the normal equation set into a matrix form to obtain a partial derivative parameter solution of the regression parameters, wherein the matrix form is as follows: x 'X β ═ X' Y, the partial derivative parameter of the regression parameters is solved as:
further, the collecting and information reporting of the sample data of the centralized unit platform extracted by the machine learning model comprises the following steps:
s21, 5G base station system opens the trial run stage, according to the priority strategy, the centralized unit deploys and runs in the centralized unit pool all the centralized unit platforms in turn, and collects the sample data of the current network on each centralized unit platform through the preset period;
and S22, averaging the sampled characteristic quantity data according to a preset period by using the running central unit in the central unit platform, and sending the sample characteristic data to the control center point according to a reporting period.
Further, the step of performing data processing on the reported sample data by using the control center to realize machine learning, prediction and decision includes the following steps:
s31, the control center receives the feature data reported by all the working concentration units in the concentration unit pool, and carries out the numeralization and normalization processing on the feature data;
s32, enabling all the centralized unit platforms in the centralized unit pool to become working centralized unit platforms in turn for sample data collection, and training the corresponding evaluation models;
s33, reporting sample data to the centralized unit model, and training and updating by the control center;
and S34, predicting all the centralized unit platforms in the centralized unit pool and making an optimal selection.
Further, the implementation of data synchronization and platform switching between the main and standby concentration units includes the following steps:
s341, the control center respectively performs filtering processing on various types of target data reported by the centralized unit platform;
s342, when the filtering value of any one target data item continuously exceeds a preset high threshold for a preset time, starting the prediction evaluation of the backup centralized unit platform in the centralized unit pool;
s343, when the running characteristic data exists in the backup centralized unit, merging the characteristic data of the working centralized unit and the characteristic data of the backup centralized unit;
s344, when the running characteristic data does not exist in the backup centralized unit, taking the characteristic data of the working centralized unit as the combined characteristic data;
s345, substituting the combined feature data into a regression model of each backup centralized unit platform, and predicting each item of target data after the centralized unit module is switched to the backup centralized unit platform;
s346, carrying out average weighting processing on all predicted target data to obtain a comprehensive evaluation result;
s347, finding out a backup centralized unit platform with the minimum predicted value from the predicted comprehensive evaluation results of all the backup centralized units;
and S348, when the comprehensive evaluation result is lower than the set lower threshold, selecting the backup centralized unit platform as the optimal switching platform, otherwise, all the backup centralized unit platforms do not support switching.
Further, the data synchronization comprises cell-level data and user-level data;
the cell level data includes, but is not limited to, configuration information and state information of the local cell and neighboring cells, and the user level data includes, but is not limited to, radio resource control state information and terminal context information.
According to another aspect of the present invention, there is also provided a system for dynamically deploying CUs under a CU Pool, the system including a centralized unit Pool and a control center;
and the control center is used for realizing prediction evaluation by machine learning, extracting key business data and centralized unit service platform data, and performing machine learning and training by combining a regression model.
Further, the hub pool includes a plurality of hub service platforms, and the hub service platforms include a host unit and a server.
The invention has the beneficial effects that: the reliability of the system can be improved by adopting the CU Pool technology under the 5G RAN CU-DU segmentation architecture. According to the invention, a reasonable machine learning model is constructed by extracting key data of the system. Through continuous machine training, the influence of user services on the CU service platforms can be accurately predicted and evaluated, and therefore the optimal CU service platform is selected from the CU POOL to carry out dynamic CU module deployment. The performance and reliability of the system are effectively improved, the abnormal risk of the CUs is reduced, and reasonable distribution of the CU resources is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a block flow diagram of a method for dynamically deploying CUs under a CU Pool according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a system for dynamically deploying CUs under CU Pool according to an embodiment of the present invention.
Detailed Description
For further explanation of the various embodiments, the drawings which form a part of the disclosure and which are incorporated in and constitute a part of this specification, illustrate embodiments and, together with the description, serve to explain the principles of operation of the embodiments, and to enable others of ordinary skill in the art to understand the various embodiments and advantages of the invention, and, by reference to these figures, reference is made to the accompanying drawings, which are not to scale and wherein like reference numerals generally refer to like elements.
According to the embodiment of the invention, a method and a system for dynamically deploying CUs under CU Pool are provided.
Referring to the drawings and the detailed description, the invention will be further described, as shown in fig. 1, a method for dynamically deploying CUs under a CU Pool according to an embodiment of the invention includes the following steps:
s1, constructing a machine learning model by adopting a preset method;
s2, collecting sample data of the centralized unit platform extracted by the machine learning model and reporting information;
s3, processing the reported sample data by adopting a control center to realize machine learning, prediction and decision;
and S4, realizing data synchronization of the main and standby CU units and platform switching of the main and standby CU units.
In one embodiment, the constructing of the machine learning model by using the preset method includes the following steps:
s11, simplifying a data model through system analysis, and extracting key data when the centralized unit operates;
the CU (central unit) unit is a centralized node that manages a plurality of DU (distribution unit) units, and as the number of users and the traffic volume of a plurality of cells under CU management increase, the CU load level will continuously increase, and is limited by the CU platform hardware capability, which will cause the system performance to rapidly decrease.
And S12, completing the construction of the machine learning model by using a preset rule.
In one embodiment, the key data includes business feature data and centralized unit platform target data;
the service characteristic data includes, but is not limited to, the number of service users, the service type and the throughput of a packet data convergence protocol, and the centralized unit platform target data includes, but is not limited to, the occupancy rate of a central processing unit, the occupancy rate of a memory, the occupancy rate of a link, and the average delay of the link.
In one embodiment, the constructing of the machine learning model by using the preset rule includes the following steps:
s121, describing a random linear relation between the centralized unit platform target data and the service characteristic data by the system through a multiple linear regression model, wherein the relation is as follows:
yi＝β0+β1xi1+β2xi2+β3xi3+εi，i＝1，2，…，n；
wherein y represents the performance evaluation target data index of the centralized unit platform, x1 is the number of service users, x2 is the service type, x3 is the service packet data convergence protocol, the throughput of the protocol data unit, and beta0，β1，β2，β3Denotes the regression coefficient,. epsiloniRepresenting random error terms, each epsiloniIndependent of each other and obey N (0, sigma)2) Distribution, n represents sample volume, and n sample observations are as follows:
(yi，xi1，xi2，xi3) i＝1，2，…，n；
s122, solving the multiple linear regression model by using a least square method or a gradient descent method, and defining a loss function, wherein the formula of the loss function is as follows:
wherein the content of the first and second substances,
to predict value, yiIs an observed value;
s123, performing partial derivation on the regression parameters to obtain a normal equation set, wherein the normal equation set has the following formula:
s124, converting the normal equation set into a matrix form to obtain a partial derivative parameter solution of the regression parameters, wherein the matrix form is as follows: x 'X β ═ X' Y,the partial derivative parameter solution of the regression parameters is:
in one embodiment, the collecting and information reporting of the sample data of the centralized unit platform extracted by the machine learning model includes the following steps:
s21, 5G base station system opens the trial run stage, according to the priority strategy, the centralized unit deploys and runs in the centralized unit pool all the centralized unit platforms in turn, and collects the sample data of the current network on each centralized unit platform through the preset period;
and S22, averaging the sampled characteristic quantity data according to a preset period (one week or one month) by using the running central unit in the central unit platform, and sending the sample characteristic data to the control center point according to a reporting period.
In one embodiment, the performing data processing on the reported sample data by using the control center to realize machine learning, prediction and decision includes the following steps:
s31, the control center receives the feature data reported by all the working concentration units in the concentration unit pool, and carries out the numeralization and normalization processing on the feature data;
after data processing, the dimensional expression is changed into a dimensionless expression, and the model precision is effectively improved.
S32, enabling all the centralized unit platforms in the centralized unit pool to become working centralized unit platforms in turn for sample data collection, and training the corresponding evaluation models;
after formal operation, the working CU continuously reports sample data, the control center regularly trains and updates the sample data, and the most suitable sample data is solved
Further improving the effect of the model.
S33, reporting sample data to the centralized unit model, and training and updating by the control center;
and S34, predicting the centralized unit platform and making an optimal selection.
In one embodiment, the implementing of data synchronization and platform switching between the main CU unit and the standby CU unit includes the following steps:
s341, the control center respectively performs filtering processing on various types of target data reported by the centralized unit platform;
and each target data item of the CU platform is configured with a fixed high threshold for triggering the prediction evaluation of the CU platform. And the control center respectively filters the various types of target data reported by the CU platform. And if the filtered value of any target data item (such as the CPU occupancy rate) continuously exceeds a preset high threshold for a certain time, starting the prediction evaluation of the backup CU platform in the Pool.
S342, when the filtering value of any one target data item continuously exceeds a preset high threshold for a preset time, starting the prediction evaluation of the backup centralized unit platform in the centralized unit pool;
s343, when the running characteristic data exists in the backup centralized unit, merging the characteristic data of the working centralized unit and the characteristic data of the backup centralized unit;
s344, when the running characteristic data does not exist in the backup centralized unit, taking the characteristic data of the working centralized unit as the combined characteristic data;
s345, substituting the combined feature data into a regression model of each backup centralized unit platform, and predicting each item of target data after the centralized unit module is switched to the backup centralized unit platform;
s346, carrying out average weighting processing on all predicted target data to obtain a comprehensive evaluation result;
s347, finding out a backup centralized unit platform with the minimum predicted value from the prediction comprehensive evaluation results of all the backup centralized units (the lower the predicted value is, the stronger the residual capacity of the platform is);
and S348, when the comprehensive evaluation result is lower than the set lower threshold, selecting the backup centralized unit platform as the optimal switching platform, otherwise, all the backup centralized unit platforms do not support switching.
In one embodiment, the data synchronization includes cell-level data and user-level data;
the cell level data includes, but is not limited to, configuration information and state information of the local cell and neighboring cells, and the user level data includes, but is not limited to, radio resource control state information and terminal context information.
According to another embodiment of the present invention, as shown in fig. 2, there is further provided a system for dynamically deploying CUs under a CU Pool, the system including a Pool formed by a plurality of CU service platforms and a control center, wherein the CU service platforms include a dedicated HU host unit, a high performance server or a cloud server. The main CU unit (working CU) periodically reports the statistical data to the control center, the control center extracts key business data (such as the number of users) and CU service platform data (such as CPU occupancy rate), and machine learning and training are carried out by combining a regression model.
And when the target data reported by the working CU is higher than the high threshold, the control center respectively merges all backup CU characteristic data of the CU Pool with the working CU characteristic data, substitutes the merged data into a regression model of the backup CU, and predicts the target data condition of the working CU module migrating to the backup CU platform, so that the optimal backup CU platform meeting the conditions is selected as a new CU deployment platform.
The control center is a logic function unit and is mainly used for machine learning to realize prediction evaluation. Can be independently deployed or can be co-deployed with the CU platform.
In summary, by means of the above technical solution of the present invention, the system reliability can be improved by adopting the CU Pool technology under the 5G RAN CU-DU segmentation architecture. According to the invention, a reasonable machine learning model is constructed by extracting key data of the system. Through continuous machine training, the influence of user services on the CU service platforms can be accurately predicted and evaluated, and therefore the optimal CU service platform is selected from the CU POOL to carry out dynamic CU module deployment. Not only effectively improving the performance and reliability of the system and reducing the abnormal risk of the CU, but also realizing reasonable distribution of CU resources
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A method for dynamically deploying CUs under a CU Pool is characterized by comprising the following steps:
s1, constructing a machine learning model by adopting a preset method;
s2, collecting sample data of the centralized unit platform extracted by the machine learning model and reporting information;
s3, processing the reported sample data by adopting a control center to realize machine learning, prediction and decision;
s4, realizing the data synchronization of the main and standby centralization units and the platform switching of the main and standby centralization units.
2. The method of claim 1, wherein the constructing the machine learning model by using the predetermined method comprises the following steps:
s11, simplifying a data model through system analysis, and extracting key data when the centralized unit operates;
and S12, completing the construction of the machine learning model by using a preset rule.
3. The method of claim 2 wherein the key data includes service feature data and centralized unit platform object data;
the service characteristic data includes, but is not limited to, the number of service users, the service type and the throughput of a packet data convergence protocol, and the centralized unit platform target data includes, but is not limited to, the occupancy rate of a central processing unit, the occupancy rate of a memory, the occupancy rate of a link, and the average delay of the link.
4. The method of claim 1, wherein the step of constructing the machine learning model using the predetermined rule comprises the steps of:
s121, describing a random linear relation between the centralized unit platform target data and the service characteristic data by the system through a multiple linear regression model, wherein the relation is as follows:
yi＝β0+β1xi1+β2xi2+β3xi3+εi，i＝1，2，…，n；
wherein y represents the performance evaluation target data index of the centralized unit platform, x1 is the number of service users, x2 is the service type, x3 is the service packet data convergence protocol, the throughput of the protocol data unit, and beta0，β1，β2，β3Denotes the regression coefficient,. epsiloniRepresenting random error terms, each epsiloniIndependent of each other and obey N (0, sigma)2) Distribution, n represents sample volume, and n sample observations are as follows:
(yi，xi1，xi2，xi3)i＝1，2，…，n；
s122, solving the multiple linear regression model by using a least square method or a gradient descent method, and defining a loss function, wherein the formula of the loss function is as follows:
wherein the content of the first and second substances,
to predict value, yiIs an observed value;
s123, performing partial derivation on the regression parameters to obtain a normal equation set, wherein the normal equation set has the following formula:
s124, converting the normal equation set into a matrix form to obtain a partial derivative parameter solution of the regression parameters, wherein the matrix form is as follows: x 'X β ═ X' Y,
the partial derivative parameter solution of the regression parameters is:
5. the method of claim 1, wherein the collecting and reporting the sample data of the centralized unit platform extracted by the machine learning model comprises:
s21, 5G base station system opens the trial run stage, according to the priority strategy, the centralized unit deploys and runs in the centralized unit pool all the centralized unit platforms in turn, and collects the sample data of the current network on each centralized unit platform through the preset period;
and S22, averaging the sampled characteristic quantity data according to a preset period by using the running central unit in the central unit platform, and sending the sample characteristic data to the control center point according to a reporting period.
6. The method according to claim 1, wherein the step of performing data processing on the reported sample data by using the control center to realize machine learning, prediction and decision includes the following steps:
s31, the control center receives the feature data reported by all the working concentration units in the concentration unit pool, and carries out the numeralization and normalization processing on the feature data;
s32, enabling all the centralized unit platforms in the centralized unit pool to become working centralized unit platforms in turn for sample data collection, and training the corresponding evaluation models;
s33, reporting sample data to the centralized unit model, and training and updating by the control center;
and S34, predicting the centralized unit platform and making an optimal selection.
7. The method according to claim 6, wherein the step of implementing data synchronization of the active/standby central units and platform switching of the active/standby central units comprises the steps of:
s341, the control center respectively performs filtering processing on various types of target data reported by the centralized unit platform;
s342, when the filtering value of any one target data item continuously exceeds a preset high threshold for a preset time, starting the prediction evaluation of the backup centralized unit platform in the centralized unit pool;
s343, when the running characteristic data exists in the backup centralized unit, merging the characteristic data of the working centralized unit and the characteristic data of the backup centralized unit;
s344, when the running characteristic data does not exist in the backup centralized unit, taking the characteristic data of the working centralized unit as the combined characteristic data;
s345, substituting the combined feature data into a regression model of each backup centralized unit platform, and predicting each item of target data after the centralized unit module is switched to the backup centralized unit platform;
s346, carrying out average weighting processing on all predicted target data to obtain a comprehensive evaluation result;
s347, finding out a backup centralized unit platform with the minimum predicted value from the predicted comprehensive evaluation results of all the backup centralized units;
and S348, when the comprehensive evaluation result is lower than the set lower threshold, selecting the backup centralized unit platform as the optimal switching platform, otherwise, all the backup centralized unit platforms do not support switching.
8. The method of claim 1, wherein the data synchronization comprises cell-level data and user-level data;
the cell level data includes, but is not limited to, configuration information and state information of the local cell and neighboring cells, and the user level data includes, but is not limited to, radio resource control state information and terminal context information.
9. A system for dynamically deploying CUs under CU Pool, for implementing the steps of the method for dynamically deploying CUs under CU Pool as claimed in any one of claims 1 to 8, wherein the system comprises a centralized unit Pool and a control center;
and the control center is used for realizing prediction evaluation by machine learning, extracting key business data and centralized unit service platform data, and performing machine learning and training by combining a regression model.
10. The system of claim 9 wherein the Pool of hub units comprises a plurality of hub unit services platforms, the hub unit services platforms comprising a host unit and a server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011495168.6A CN112672368B (en) | 2020-12-17 | 2020-12-17 | Method and system for dynamically deploying CU under CU Pool |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011495168.6A CN112672368B (en) | 2020-12-17 | 2020-12-17 | Method and system for dynamically deploying CU under CU Pool |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112672368A true CN112672368A (en) | 2021-04-16 |
| CN112672368B CN112672368B (en) | 2021-12-21 |
Family
ID=75404694
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011495168.6A Active CN112672368B (en) | 2020-12-17 | 2020-12-17 | Method and system for dynamically deploying CU under CU Pool |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112672368B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106792992A (en) * | 2016-12-12 | 2017-05-31 | 上海掌门科技有限公司 | A kind of method and apparatus for providing WAP information |
| CN109548037A (en) * | 2017-08-01 | 2019-03-29 | 中国移动通信有限公司研究院 | A kind of control method, distributed unit, centralized unit and storage medium |
| CN110366134A (en) * | 2019-07-12 | 2019-10-22 | 北京邮电大学 | A kind of vehicle connection network-building method and device |
| US20200195495A1 (en) * | 2019-02-25 | 2020-06-18 | Intel Corporation | 5g network slicing with distributed ledger traceability and resource utilization inferencing |
| CN111582494A (en) * | 2020-04-17 | 2020-08-25 | 浙江大学 | Hybrid distributed machine learning updating method based on delay processing |
-
2020
- 2020-12-17 CN CN202011495168.6A patent/CN112672368B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106792992A (en) * | 2016-12-12 | 2017-05-31 | 上海掌门科技有限公司 | A kind of method and apparatus for providing WAP information |
| CN109548037A (en) * | 2017-08-01 | 2019-03-29 | 中国移动通信有限公司研究院 | A kind of control method, distributed unit, centralized unit and storage medium |
| US20200195495A1 (en) * | 2019-02-25 | 2020-06-18 | Intel Corporation | 5g network slicing with distributed ledger traceability and resource utilization inferencing |
| CN110366134A (en) * | 2019-07-12 | 2019-10-22 | 北京邮电大学 | A kind of vehicle connection network-building method and device |
| CN111582494A (en) * | 2020-04-17 | 2020-08-25 | 浙江大学 | Hybrid distributed machine learning updating method based on delay processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112672368B (en) | 2021-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20170163540A1 (en) | Systems and methods for generating a virtual network topology for m2m communications | |
| CN101686497B (en) | Cell load equalization method, and cell load evaluation method and device | |
| US9538401B1 (en) | Cellular network cell clustering and prediction based on network traffic patterns | |
| EP3972339A1 (en) | Handover success rate prediction and management using machine learning for 5g networks | |
| EP2871804A1 (en) | A method for access points scheduling for backhaul aggregation in a telecommunications network and a device | |
| WO2018120802A1 (en) | Collaborative content cache control system and method | |
| CN108076486A (en) | A kind of dynamic radio resource allocation algorithm based on load balancing | |
| CN109121164B (en) | Load balancing method and device | |
| CN106211206A (en) | The analysis method of the abnormal release of a kind of E-RAB, Apparatus and system | |
| CN102695218B (en) | Mode identification-based switching self-optimizing method | |
| Mirzaei Somarin et al. | Big data based self-optimization networking in next generation mobile networks | |
| CN112672368B (en) | Method and system for dynamically deploying CU under CU Pool | |
| CN106572475B (en) | Access node management method, access network management entity, equipment and access node | |
| CN113676357A (en) | Decision method for edge data processing in power internet of things and application thereof | |
| CN108271216B (en) | Ad hoc network load balancing method and device and cloud platform disaster recovery system | |
| CN105916178A (en) | Load balancing method and apparatus thereof | |
| WO2023011371A1 (en) | Method and system for configuring a threshold value for a handover parameter of a wireless communication system | |
| Liya et al. | A MADM-based handover management in software defined 5G network | |
| CN103260270B (en) | A kind of base station | |
| CN107770816A (en) | A kind of LTE load-balancing methods, call access control module and system | |
| CN104349347B (en) | A kind of compensation method of customer service experience and controller | |
| Semov et al. | Autonomous learning model for achieving multi cell load balancing capabilities in HetNet | |
| CN117596175B (en) | Hierarchical monitoring method, device, equipment, system and storage medium for switch | |
| Stoliarchuk et al. | Methods of common resources management of independent heterogeneous networks by using big data technology | |
| Glabowski et al. | Modelling load balancing mechanisms in self-optimising 4G mobile networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |