CN114764632A

CN114764632A - Joint learning client side controller based on Internet of things

Info

Publication number: CN114764632A
Application number: CN202110049110.7A
Authority: CN
Inventors: 张敏; 高庆
Original assignee: Xinzhi Cloud Data Service Co ltd
Current assignee: Xinzhi Cloud Data Service Co ltd
Priority date: 2021-01-14
Filing date: 2021-01-14
Publication date: 2022-07-19

Abstract

The invention discloses a joint learning client side controller based on the Internet of things, which comprises a client side, a cloud management side and a joint learning side, wherein the client side comprises a client management control console, a client side controller, an Internet of things layer, a cloud mixing layer and a cloud mixing controller, the client side controller in the client side is used as a management access node, the client side is connected with the cloud management side and the joint learning side by the client side controller control node, the client side controller and the Internet of things layer execute tasks mutually, the client side controller provides a scheduling execution inlet to butt joint a cloud mixing nano tube and an equipment layer on the cloud management side to execute corresponding operations, the client side controller and a resource node are communicated through a mutual communication authentication mode to issue a script, and the script execution and file issue channel is communicated. According to the invention, the client side controller is used as a management access node, the control node is connected with the cloud management side and the client side and is accessed into an environment isolated from the new intelligent cloud, so that data butt joint of the cloud management side and the client side is ensured, and the progress of joint learning is ensured.

Description

Joint learning client side controller based on Internet of things

Technical Field

The invention relates to the technical field of joint learning, in particular to a joint learning client side controller based on the Internet of things.

Background

Joint Learning (Federated Learning) is a new artificial intelligence basic technology, was proposed by Google in 2016, and is originally used for solving the problem of local model updating of android mobile phone terminal users, and the design goal is to develop efficient machine Learning among multiple parties or multiple computing nodes on the premise of guaranteeing information security during big data exchange, protecting terminal data and personal data privacy and guaranteeing legal compliance. The machine learning algorithm which can be used for federal learning is not limited to a neural network, and also comprises important algorithms such as a random forest. The joint learning is expected to become the basis of the next generation artificial intelligence cooperative algorithm and cooperative network.

Federal learning is a machine learning setting in which many clients co-train models under the coordination of a central server while maintaining decentralization and decentralization of training data.

Long-term goals for federal learning: data from multiple data owners is analyzed and learned without exposing the data.

The system framework of the joint learning is as follows: the system architecture of joint learning is described by taking a scenario involving two data owners (i.e., enterprises A and B) as an example. The framework is extensible to scenarios involving multiple data owners. Suppose enterprises A and B want to jointly train a machine learning model, and their business systems respectively have relevant data of their respective users. In addition, enterprise B also owns tag data that the model needs to predict. In consideration of data privacy protection and safety, A and B cannot directly exchange data, and a joint learning system can be used for establishing a model. The joint learning system framework is composed of three parts.

The federal Learning classification is divided into horizontal federal Learning (horizontal federal fed Learning), vertical federal Learning (vertical federal Learning), and federal Transfer Learning (FmL).

In the case of transverse federated learning, when the user features of two data sets overlap more and the user overlap less, the data sets are divided according to the transverse direction (namely the user dimension), and the data with the same user features but not identical users is taken out for training. This method is called horizontal federal learning. Compared with banks in two different regions, the user groups of the banks come from the regions where the banks are respectively located, and the intersection of the user groups is very small. However, their services are very similar, and therefore the recorded user characteristics are the same. At this point, we can use horizontal federal learning to build the federated model. Google in 2016 proposed a data federation modeling scheme for android phone model updates: when a single user uses the android mobile phone, model parameters are continuously updated locally and uploaded to the android cloud, and therefore all data owners with the same characteristic dimension can establish a combined model.

Longitudinal federated learning divides the data sets according to the longitudinal direction (namely feature dimension) under the condition that the users of the two data sets overlap more and the user features overlap less, and takes out the part of data which is the same for both users and the user features are not completely the same for training. This method is called longitudinal federal learning. For example, if there are two different establishments, the home is a bank in one location and the other home is an e-commerce in the same location. Their user population is likely to contain a large proportion of the inhabitants of the site and therefore the intersection of users is large. However, the bank records the user's balance and credit rating, and the e-commerce maintains the user's browsing and purchasing history, so the intersection of their user features is small. Longitudinal federal learning is to aggregate these different features in an encrypted state to enhance model capabilities. At present, a plurality of machine learning models such as a logistic regression model, a tree structure model and a neural network model are gradually proved to be capable of being established on the federal system.

In the case of federate migration learning, under the condition that the overlap of users and user features of two data sets is less, the users do not segment data, and the migration learning country is utilized to overcome the condition of insufficient data or labels. This method is called federal migration learning. For example, there are two different institutions, one being a bank located in china and the other being an e-commerce located in the united states. Due to regional limitation, the user population intersection of the two organizations is small. Meanwhile, due to the difference of mechanism types, the data characteristics of the two are only partially overlapped. Under the condition, migration learning must be introduced to solve the problems of small scale of unilateral data and few label samples so as to improve the effect of the model for effective federal learning.

The joint learning has the following advantages:

(1) data isolation is realized, so that data cannot be leaked to the outside, and the requirements of user privacy protection and data security are met;

(2) the quality of the model can be ensured to be lossless, negative migration cannot occur, and the effect of the federal model is better than that of a split independent model;

(3) the participants have equal positions, so that fair cooperation can be realized;

(4) the method can ensure that the participating parties carry out encryption exchange of information and model parameters under the condition of keeping independence, and can grow simultaneously.

However, in the joint learning, there is no controller for one docking schedule among the client side 100, the cloud management side 200, and the joint learning side 300, and it is difficult to maintain the progress of the joint learning.

Disclosure of Invention

The invention aims to provide a joint learning client side controller based on the Internet of things, wherein the client side controller is connected with a cloud management side and a client side in an abutting mode, data are fully scheduled so that the joint learning can be maintained to be in progress, and the problems in the background technology can be solved.

In order to achieve the purpose, the invention provides the following technical scheme:

the client side controller is used as a management access node, the client side is connected with the client side and connected with the cloud management side and the combined learning side through the client side controller control node, the client side controller and the internet of things layer execute tasks mutually, the client side controller provides a scheduling execution inlet to be connected with a cloud storage pipe and a device layer of the cloud management side to execute corresponding operations, the client side controller and the resource node are communicated through a mutual communication authentication mode to be issued with a script, and the script execution and file issuing channel is formed.

Furthermore, the client management console comprises a resource management module, a learning management module and an operation and maintenance management module, wherein the learning management module is accessed to an input interface of the client controller, and the learning management module inputs data to be managed to the client controller.

Further, the client side controller comprises an arranging and executing module, a configuration checking module, a configuration collecting module, an information forwarding module, a performance collecting module and an equipment receiving and managing module, wherein the arranging and executing module is controlled by the cloud management side and receives an instruction of the cloud management side to complete an arranging and executing task.

Furthermore, the internet of things layer comprises an edge module, the edge module uploads results to the information forwarding module, the information forwarding module sends instructions to the edge module to execute tasks, and mutual trust authentication is performed between the edge module and the equipment management module.

Further, the cloud layer comprises a mixed cloud non-management module and a mixed cloud management module, the mixed cloud non-management module uploads results to the cloud controller, and the cloud controller sends instructions to the mixed cloud non-management module to execute tasks.

Further, the cloud mixing controller comprises an intelligent management module, a safety protection module, a service management module and a basic function module, wherein:

the intelligent management module is used for intelligent management of the cloud mixing controller, and ensures smooth progress of joint learning;

the safety protection module is used for safety protection of a client side, a cloud management side and a joint learning side in the joint learning;

the service management module is used for performing service experience in the joint learning;

the basic function module is used for having multiple basic functions in the joint learning.

Furthermore, the intelligent management module comprises a channel intelligent switching module, a power adjusting module and a load balancing module;

the safety protection module comprises an equipment connection safety protection module, a node access protection module and an illegal equipment detection module;

the service management module comprises a business service module, a position information service module, an authority management module and a time application control module;

the basic function module comprises a data forwarding module, an alarm module, an automatic operation and maintenance module and a cloud account management module.

Further, the cloud controller further comprises a probe module and a management authentication module.

Further, the cloud management side comprises a cloud platform, the cloud platform comprises a usage statistic module, an arranging and scheduling center, an entity data module and a monitoring and managing platform, the usage statistic module is connected with the monitoring and managing platform, the monitoring and managing platform is connected with the cloud mixing controller, the entity data module transmits data to the arranging and scheduling center, and the arranging and scheduling center sends an execution instruction to the arranging and executing module.

Furthermore, the joint learning side comprises a joint learning server and a joint learning operation time module, the joint learning server receives working instructions from the learning management module and the arrangement execution module, the joint learning server and the joint learning operation time module transmit data to each other, and the joint learning operation time module feeds back the data to the arrangement scheduling center.

Compared with the prior art, the invention has the beneficial effects that:

1. the client side controller is used as a management access node, the control node is connected with the cloud management side and the client side and accesses an environment isolated from the new intelligent cloud, and the client side controller provides a piece of software and hardware equipment which is provided for the new intelligent cloud or provided for the client side and can be connected with the new intelligent cloud, so that data butt joint of the cloud management side and the client side is guaranteed, and the progress of joint learning is guaranteed.

2. The client side controller provides a scheduling execution inlet to butt the cloud nano-tubes and the equipment layer on the cloud to execute corresponding operations. And providing performance acquisition and configuration acquisition functions, acquiring performance data in the joint learning process, providing a data forwarding channel to forward a joint learning model result and an intermediate result, providing configuration acquisition and inspection, and verifying resource configuration participating in joint learning.

3. The cloud mixing controller executes the data matching task of the cloud mixing layer, optimizes data resources, and executes the task with the monitoring management platform of the cloud management side, so that the client side and the cloud management side are more friendly.

Drawings

Fig. 1 is an architecture diagram of a joint learning client-side controller based on the internet of things according to a first embodiment of the present invention;

FIG. 2 is an architecture diagram of the intelligent management module of the present invention;

FIG. 3 is a schematic diagram of the load balancing algorithm of the present invention;

FIG. 4 is an architecture diagram of the safety module of the present invention;

FIG. 5 is an architecture diagram of a service management module of the present invention;

FIG. 6 is an architecture diagram of the basic functional module of the present invention;

fig. 7 is an architecture diagram of a client-side controller for joint learning based on the internet of things according to a second embodiment of the present invention;

in the figure: 100. a client side; 101. a client management console; 1011. a resource management module; 1012. A learning management module; 1013. an operation and maintenance management module; 102. a client-side controller; 1021. a layout execution module; 1022. configuring a checking module; 1023. configuring an acquisition module; 1024. an information forwarding module; 1025. a performance acquisition module; 1026. an equipment nanotube module; 103. an Internet of things layer; 1031. an edge module; 104. a cloud layer; 1041. a mixed cloud non-management module; 1042. a hybrid cloud management module; 105. A cloud mixing controller; 1051. an intelligent management module; 1052. a safety protection module; 1053. a service management module; 1054. a basic function module; 1055. a probe module; 1056. a management authentication module; 200. A cloud management side; 201. a cloud platform; 2011. a usage statistics module; 2012. arranging a scheduling center; 2013. An entity data module; 2014. monitoring and managing the platform; 300. a joint learning side; 301. a joint learning server; 302. and a joint learning runtime module.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Example 1

Fig. 1 shows an architecture diagram of a joint learning client-side controller based on the internet of things according to the present embodiment, the joint learning client-side controller based on the internet of things includes a client-side 100, a cloud management side 200, and a joint learning side 300, the client-side 100 includes a client management console 101, a client-side controller 102, an association layer 103, a cloud layer 104, and a cloud controller 105, the client-side controller 102 in the client-side 100 serves as a management access node, the client-side controller 102 controls the node connection to connect the client-side 100 to the cloud management side 200 and the joint learning side 300, so that the client-side 100, the cloud management side 200, and the joint learning side 300 can be linked with each other, thereby ensuring normal operation of the joint learning, the client-side controller 102 and the association layer 103 perform tasks with each other, the client-side controller 102 provides a scheduling execution entrance to perform corresponding operations with a cloud nanotube and an equipment layer on the cloud management side, the client side controller 102 and the resource node are communicated through a mutual communication authentication mode to issue the script, and the script execution and file issuing channels are formed.

The client management console 101 includes a resource management module 1011, a learning management module 1012, and an operation and maintenance management module 1013, where the learning management module 1012 accesses an input interface of the client controller 102, and the learning management module 1012 inputs data to be managed to the client controller 102.

The resource management module 1011 manages the client resources of the client side 100, maintains the clients, and improves the satisfaction of the clients.

The learning management module 1012 is used for managing the learning status of the client 100, such as learning status, learning progress, learning effect, etc. of the client for unified management, so as to arrange the subsequent learning plan.

The operation and maintenance management module 1013 performs operation and maintenance scheduling on the whole client learning process of the client side 100, and constantly grasps the learning problem of the client.

The client-side controller 102 includes an arrangement executing module 1021, a configuration checking module 1022, a configuration collecting module 1023, an information forwarding module 1024, a performance collecting module 1025, and an equipment receiving module 1026, wherein the arrangement executing module 1021 is controlled by the cloud management side 200, and receives an instruction of the cloud management side 200 to complete an arrangement executing task.

The orchestration execution module 1021 may communicate with the orchestration scheduling center 2012 of the cloud management side 200, receive an orchestration execution instruction of the orchestration scheduling center 2012, and complete an orchestration operation.

The configuration checking module 1022 provides configuration collection and checking for the result of forwarding the joint learning model and the intermediate result through the data forwarding channel, and verifies the resource configuration participating in the joint learning.

The configuration acquisition module 1023 provides performance acquisition and configuration acquisition functions to acquire performance data in the combined learning process.

The information forwarding module 1024 performs data transmission with the edge module 1031 of the internet of things layer 103 to ensure task execution.

The performance acquisition module 1025 acquires the performance of the equipment, so that the operation parameters and functions of the equipment can be further mastered.

The internet of things layer 103 includes an edge module 1031, the edge module 1031 uploads the result to the information forwarding module 1024, the information forwarding module 1024 sends an instruction to the edge module 1031 to perform task execution, and mutual trust authentication is performed between the edge module 1031 and the equipment housing and management module 1026.

The cloud layer 104 includes a cloud-to-cloud non-management module 1041 and a cloud-to-cloud management module 1042, the cloud-to-cloud non-management module 1041 uploads the result to the cloud controller 105, and the cloud controller 105 sends an instruction to the cloud-to-cloud non-management module 1041 to perform task execution.

The cloud controller 105 includes an intelligent management module 1051, a security protection module 1052, a service management module 1053, and a base function module 1054, wherein:

the intelligent management module 1051 is used for intelligent management of the cloud controller 105, and ensures smooth proceeding of joint learning.

The security module 1052 is used for security protection of the client side 100, the cloud management side 200, and the joint learning side 300 in the joint learning.

The service management module 1053 is used for performing service experience in joint learning.

The basic function module 1054 is used to provide multiple basic functions in the joint learning.

Fig. 2 shows an architecture diagram of the intelligent management module of this embodiment, where the intelligent management module 1051 includes a channel intelligent switching module, a power adjustment module, and a load balancing module.

The channel intelligent switching module ensures that the channel allocated by each wireless access point is optimized through self-adaptive channel selection, and reduces and avoids adjacent/co-channel interference as much as possible.

The power adjusting module can automatically adjust the power based on the radio frequency environment of the equipment, automatically enlarge or reduce the radio frequency power of the equipment, realize the automatic detection and compensation of the coverage area and achieve the optimal coverage effect.

Fig. 3 shows an algorithm diagram of load balancing according to the embodiment, where the load balancing module provides balancing based on the number of access users (when the number of users in the wireless access points exceeds a set threshold, the uniform distribution of users in different wireless access points can be dynamically adjusted), provides load balancing based on traffic (when the traffic value in the wireless access points exceeds a set threshold, the uniform distribution of users in different wireless access points can be dynamically adjusted), and provides load balancing based on frequency bands.

Fig. 4 shows an architecture diagram of the security module of this embodiment, and the security module 1052 includes a device connection security module, a node access security module, and an illegal device detection module.

The equipment connection safety protection module is used for deploying wireless access point equipment in an open environment of the Internet, the access of a fake wireless access point and the theft of the belonged wireless access point by a binding frame are two main equipment safety problems, and the equipment identity authentication problem on the Internet of things is reasonably solved by utilizing a TLS/DTLS network safety protocol.

The node access protection module supports the access control of a black and white list of a user, supports real-name system online authentication based on short messages, and intelligently identifies the identity characteristics of the user according to the mode of accessing the user to a network, so that the user can share customized strategies and services.

The illegal device detection module provides a wireless intrusion detection WIPS function, searches illegal access points and temporary networks by continuously monitoring a wireless space, and provides an alarm and an attack of illegal devices/users.

Fig. 5 shows an architecture diagram of the service management module of this embodiment, and the service management module 1053 includes a business service module, a location information service module, a rights management module, and a time application control module.

And the business service module can provide remote office service.

The position information service module is internally provided with a wireless access point position information identifier, and different network services can be set based on physical position information, such as the wireless access point at the position of the foreground provides a visitor Internet function, and simultaneously limits to access Internet and specified internal resources.

The authority management module can define the admission/discharge authority of the roles, define one or more roles for the user, and each role can have different admission authorities, control the uplink and downlink bandwidths of the user and provide differentiated management service for the user.

The time application control module may set time-based service control.

Fig. 6 shows an architecture diagram of the service management module of this embodiment, where the basic function module 1054 includes a data forwarding module, an alarm module, an automation operation and maintenance module, and a cloud account management module.

The data forwarding module is used for determining that the user traffic is forwarded locally by the wireless access point based on the policy, and under the same wireless access point, one data traffic is forwarded locally, and one or more data traffics are forwarded centrally.

And the alarm module is used for providing interface state alarm, equipment disconnection alarm and illegal wireless access point/client end alarm, and can self-define the range and service of triggering an early warning line.

The automatic operation and maintenance module guarantees the reliability of service; when the network is idle, network upgrading or fault detection is realized, all the operations can be automatically completed without human real-time participation, and the network operation and maintenance difficulty is simplified.

The cloud account management module is used for managing and operating in a cloud service mode, the management right and the operation right do not depend on ownership of equipment, are shared by operation entities at all levels, and are remotely realized in a cloud account mode.

Cloud management side 200 includes cloud platform 201, and cloud platform 201 includes quantity statistics module 2011, arrangement scheduling center 2012, entity data module 2013, control management platform 2014, quantity statistics module 2011 is connected with control management platform 2014, and to the statistics of control data through quantity statistics module 2011 of control management platform 2014, control management platform 2014 is connected with mixed cloud controller 105, and entity data module 2013 transmits data to arrangement scheduling center 2012, and arrangement scheduling center 2012 sends executive instruction to arrangement execution module 1021.

The joint learning side 300 comprises a joint learning server 301 and a joint learning runtime module 302, wherein the joint learning server 301 receives work instructions from a learning management module 1012 and an orchestration execution module 1021, the joint learning server 301 and the joint learning runtime module 302 transmit data to each other, and the joint learning runtime module 302 feeds back the data to an orchestration scheduling center 2012.

Example 2

Fig. 7 shows an architecture diagram of a joint learning client-side controller based on the internet of things according to this embodiment, the joint learning client-side controller based on the internet of things includes a client side 100, a cloud management side 200, and a joint learning side 300, the client side 100 includes a client management console 101, a client side controller 102, an physical layer 103, a cloud layer 104, and a cloud controller 105, the client side controller 102 in the client side 100 serves as a management access node, the client side controller 102 controls the node connection to connect the client side 100 to the cloud management side 200 and the joint learning side 300, so that the client side 100, the cloud management side 200, and the joint learning side 300 can be linked with each other to ensure normal operation of the joint learning, the client side controller 102 and the physical layer 103 execute tasks mutually, the client side controller 102 provides a scheduling execution entrance to interface with the cloud nanotube and the equipment layer on the cloud management side to execute corresponding operations, the client side controller 102 and the resource node are communicated through a mutual communication authentication mode to issue the script, and the script execution and file issuing channels are formed.

The client management console 101 includes a resource management module 1011, a learning management module 1012, and an operation and maintenance management module 1013, wherein the learning management module 1012 accesses an input interface of the client controller 102, and the learning management module 1012 inputs data to be managed to the client controller 102.

The learning management module 1012 is used for managing the learning conditions of the client side 100, such as the learning state, learning progress, learning effect, etc. of the client side for uniform management, so as to arrange the subsequent learning plan.

The configuration checking module 1022 provides configuration collection and checking for the data forwarding channel to forward the results of the joint learning model and the intermediate results, and verifies the resource configuration participating in the joint learning.

The information forwarding module 1024 ensures the task execution by performing data transmission with the edge module 1031 of the internet of things layer 103.

The cloud layer 104 includes a mixed cloud non-management module 1041 and a mixed cloud management module 1042, the mixed cloud non-management module 1041 uploads a result to the cloud controller 105, and the cloud controller 105 sends an instruction to the mixed cloud non-management module 1041 to perform task execution.

The clouding controller 105 includes a probe module 1055 and a management authentication module 1056.

The probe module 1055 supports the probe function, can set the acquisition frequency of the probe according to the application scenario, and simultaneously supports the wireless access point to switch to a pure probe mode. As long as the WiFi client side starts WiFi connection, the wireless access point can acquire information such as MAC of the mobile terminal, the wireless access point of the associated third party, signal strength and the like, and meanwhile, the acquired information can be submitted to a data acquisition server (VDS) under an ABLOOMY line, and services such as user positioning, historical track and the like can be realized by combining the VDS.

The management authentication module can realize integrated management and authentication of a wired network and a wireless network. And for the user, whether the user logs in a wired network or a wireless network, managing and authenticating the account of the user.

According to the invention, the client side controller is used for connecting the cloud management side and the client side, and the client side exchanges data with the joint learning side through the cloud management side, so that optimization service is provided for the joint learning, the joint learning efficiency is improved, and the coordination of the joint learning is increased.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the equivalent replacement or change according to the technical solution and the inventive concept of the present invention should be covered by the scope of the present invention.

Claims

1. A joint learning client side controller based on the Internet of things is characterized by comprising a client side (100), a cloud management side (200) and a joint learning side (300), wherein the client side (100) comprises a client side management console (101), a client side controller (102), an Internet of things layer (103), a cloud mixing layer (104) and a cloud mixing controller (105), the client side controller (102) in the client side (100) serves as a management access node, the client side controller (102) controls the node connection to connect the client side (100) to the cloud management side (200) and the joint learning side (300), the client side controller (102) and the Internet of things layer (103) execute tasks mutually, the client side controller (102) provides a scheduling execution entrance to be in butt joint with a cloud nano-tube and an equipment layer on the cloud management side to execute corresponding operations, and the client side controller (102) and the resource node communicate and issue scripts in a mutual communication authentication mode, script execution and file issuing channels.

2. The internet-of-things-based joint learning client-side controller of claim 1, wherein the client management console (101) comprises a resource management module (1011), a learning management module (1012) and an operation and maintenance management module (1013), wherein the learning management module (1012) has access to an input interface of the client-side controller (102), and the learning management module (1012) inputs data to be managed to the client-side controller (102).

3. The internet-of-things-based joint learning client-side controller of claim 2, wherein the client-side controller (102) comprises an arrangement execution module (1021), a configuration check module (1022), a configuration acquisition module (1023), an information forwarding module (1024), a performance acquisition module (1025) and a device housing module (1026), wherein the arrangement execution module (1021) is controlled by the cloud management side (200) and receives an instruction of the cloud management side (200) to complete an arrangement execution task.

4. The internet-of-things-based joint learning client-side controller according to claim 3, wherein the internet-of-things layer (103) comprises an edge module (1031), the edge module (1031) uploads the result to the information forwarding module (1024), the information forwarding module (1024) sends an instruction to the edge module (1031) for task execution, and mutual trust authentication is performed between the edge module (1031) and the equipment hosting module (1026).

5. The internet-of-things-based joint learning client-side controller of claim 4, wherein the hybrid cloud layer (104) comprises a hybrid cloud non-management module (1041) and a hybrid cloud management module (1042), the hybrid cloud non-management module (1041) uploads results to the hybrid cloud controller (105), and the hybrid cloud controller (105) sends an instruction to the hybrid cloud non-management module (1041) for task execution.

6. The internet of things-based joint learning client-side controller of claim 5, wherein the cloud controller (105) comprises an intelligent management module (1051), a security protection module (1052), a service management module (1053), and a base function module (1054), wherein:

the intelligent management module (1051) is used for intelligent management of the cloud mixing controller (105) to ensure smooth proceeding of joint learning;

the safety protection module (1052) is used for the safety protection of the client side (100), the cloud management side (200) and the joint learning side (300) in the joint learning;

the service management module (1053) is used for performing service experience in the joint learning;

the basic function module (1054) is used for possessing a plurality of basic functions in the joint learning.

7. The internet-of-things-based joint learning client-side controller of claim 6, wherein the intelligent management module (1051) comprises a channel intelligent switching module, a power adjustment module, a load balancing module;

the safety protection module (1052) comprises a device connection safety protection module, a node access protection module and an illegal device detection module;

the service management module (1053) comprises a business service module, a position information service module, a right management module and a time application control module;

the basic function module (1054) comprises a data forwarding module, an alarm module, an automatic operation and maintenance module and a cloud account management module.

8. The internet of things based joint learning client-side controller of claim 7, wherein the clouded controller (105) further comprises a probe module (1055) and a management authentication module (1056).

9. The internet-of-things-based joint learning client-side controller of claim 8, wherein the cloud management side (200) comprises a cloud platform (201), the cloud platform (201) comprises a usage statistics module (2011), an orchestration scheduling center (2012), an entity data module (2013) and a monitoring management platform (2014), the usage statistics module (2011) is connected with the monitoring management platform (2014), the monitoring management platform (2014) is connected with the hybrid cloud controller (105), the entity data module (2013) transmits data to the orchestration scheduling center (2012), and the orchestration scheduling center (2012) sends an execution instruction to the orchestration execution module (1021).

10. The internet-of-things-based joint learning client-side controller of claim 9, wherein the joint learning side (300) comprises a joint learning server (301) and a joint learning runtime module (302), the joint learning server (301) receives work instructions from the learning management module (1012) and the orchestration execution module (1021), the joint learning server (301) and the joint learning runtime module (302) mutually transmit data, and the joint learning runtime module (302) feeds back the data to the orchestration scheduling center (2012).