Label-based multilayer container arrangement method and system

Abstract

The invention provides a label-based multilayer container arrangement method and system, and the method comprises the steps: hierarchical label definition: defining multilayer labels for containers or functional modules, embedding layer information into each layer of label to generate a label database containing the layer information, and the layer information is a binary numerical value or a character string and is used for identifying a hierarchical relationship; performing multi-layer arrangement execution to generate an executable cross-layer label arrangement strategy; performing cross-cloud protocol adaptation according to whether the target cloud supports a label protocol to generate a data transmission instruction compatible with the heterogeneous cloud environment; mapping the cross-layer label sequence into a thinking chain, a thinking tree or a thinking network structure; based on a data transmission instruction, an autonomously programmed neural network model is generated in combination with a routing weight strategy, and computing power distribution and target function selection and use are realized through containerized deployment. The method can be expanded to cover cross-cloud arrangement, multi-layer arrangement and other containers and function arrangement, and realizes conversion from single-layer arrangement to multi-layer arrangement.

Description

Technical Field

[0001] The present invention relates to the field of cloud platforms, and in particular, to a multi-layer container orchestration method and system based on tags. Background Art

[0002] Currently, the way to implement container orchestration is mainly to execute and deploy through Kubernetes. Through the orchestration of containers, not only the elastic deployment, load balancing, and self-healing of containers are realized, but also the hardware resources are maximally utilized, and further agile development or DevOps methods are realized.

[0003] However, there are still some problems in the existing container orchestration technology. This technology can only be used for single-layer container orchestration, and the orchestration results cannot be applied deeper, multi-layer orchestration cannot be achieved, and container orchestration cannot be combined with artificial intelligence. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a multi-layer container orchestration method and system based on tags, which can achieve multi-layer orchestration of containers.

[0005] A multi-layer container orchestration method based on tags includes:

[0006] Step 1: Hierarchical label definition: Define multi-layer tags for containers or functional modules, and embed layer information in each layer of tags to generate a tag database containing layer information. The layer information is a binary value or a string, which is used to identify the hierarchical relationship;

[0007] Step 2: Multi-layer orchestration execution:

[0008] Manual orchestration: Based on the tag database in Step 1, label the tag sequence layer by layer from bottom to top, and establish an associated mapping of upper and lower layer tags through layer information;

[0009] Automatic orchestration: Based on the tag database in Step 1 and the upper and lower layer tag mapping relationship, generate a cross-layer tag orchestration sequence from top to bottom or from bottom to top;

[0010] Intelligent orchestration: Based on the orchestration sequence generated by automatic orchestration, compare the deviation between the actual execution result and the expected result through a verification module, and dynamically adjust the tag sequence of each layer;

[0011] Generate an executable cross-layer tag orchestration strategy based on the above multi-layer orchestration execution steps;

[0012] Step 3: Cross-cloud protocol adaptation:

[0013] If the target cloud supports the tag protocol, encapsulate the orchestration strategy generated in Step 2 into a data frame with a binary tag extension header;

[0014] If the target cloud does not support the label protocol, reconstruct the orchestration strategy in Step 2 into a standardized data payload according to the target cloud API specification;

[0015] Generate a data transmission instruction that is compatible with heterogeneous cloud environments through the above cross-cloud protocol adaptation;

[0016] Step 4. Artificial Intelligence Integration:

[0017] Map the cross-layer label sequence generated in Step 2 to a thought chain, thought tree, or thought network structure;

[0018] Based on the data transmission instruction in Step 3, generate a self-programming neural network model in combination with the routing weight strategy, and realize computing power allocation, destination function selection and use through containerized deployment.

[0019] Furthermore, the introduction method of the layer information in Step 1 is as follows: the same-level labels are uniformly marked with the same layer identifier; single labels are marked layer by layer in a stacked manner, and the number of layers can be dynamically expanded; the upper-layer labels are an aggregated and simplified representation of the lower-layer associated labels, and the label sequence length of the upper-layer labels is reduced to 30%-50% of the lower-layer associated labels.

[0020] Furthermore, the automatic orchestration in Step 2 includes:

[0021] Top-down generation: After manually defining the label sequence at the top layer, parse layer by layer to generate the lower-layer label sequence, and the parsing is implemented through a label mapping table;

[0022] Bottom-up generation: After completing container orchestration at the bottom layer, generate the upper-layer simplified label sequence layer by layer through the correlation algorithm.

[0023] In intelligent orchestration, the condition for the verification module to trigger policy adjustment is that the deviation between the actual result and the expected result exceeds the set threshold.

[0024] Furthermore, the label extension header in Step 3 includes the following fields: layer identifier, label priority, target cloud protocol type, source cloud identifier, call target label, call target layer identifier, and other control contents; the API specification reconstruction includes: converting the label sequence into a JSON format or YAML format supported by the target cloud.

[0025] Furthermore, the thought network structure in Step 4 is a three-dimensional network, which aggregates and classifies two-dimensional neural network nodes through layer information to reduce the complexity of the neural network; the routing weight strategy dynamically adjusts the node connection path based on the label priority, target node resource occupancy ratio, and target node function.

[0026] A multi-layer container orchestration system based on labels, including:

[0027] Hierarchical label definition module, used to define multiple layers of labels for containers or functional modules, embedding layer information in each layer of labels to generate a label database containing layer information, where the layer information is a binary value or a string, used to identify the hierarchical relationship;

[0028] Multi-layer orchestration execution module, used to execute multi-layer orchestration to generate an executable cross-layer label orchestration strategy. The multi-layer orchestration includes manual orchestration, automatic orchestration, and intelligent orchestration. Among them, manual orchestration is based on the label database, layer-by-layer marking the label sequence from bottom to top, and establishing an associated mapping of upper and lower layer labels through layer information; automatic orchestration is based on the label database and the upper and lower layer label mapping relationship, generating a cross-layer label orchestration sequence from top to bottom or from bottom to top; intelligent orchestration is based on the orchestration sequence generated by automatic orchestration, comparing the deviation between the actual execution result and the expected result through the verification module, and dynamically adjusting the label sequence of each layer;

[0029] Cross-cloud protocol adaptation module, used to encapsulate the orchestration strategy generated by the multi-layer orchestration execution module into a data frame with a binary label extension header if the target cloud supports the label protocol; if the target cloud does not support the label protocol, reconstruct the orchestration strategy generated by the multi-layer orchestration execution module according to the target cloud API specification into a standardized data payload; generate a data transmission instruction compatible with heterogeneous cloud environments based on the cross-cloud protocol adaptation result;

[0030] Artificial intelligence integration module, used to map the cross-layer label sequence generated by the multi-layer orchestration execution module into a thinking chain, thinking tree, or thinking network structure; based on the data transmission instruction generated by the cross-cloud protocol adaptation module, combine the routing weight strategy to generate a self-programming neural network model, and realize computing power allocation, destination function selection and use through containerized deployment.

[0031] Further, the introduction method of the layer information in the hierarchical label definition module is as follows: the same layer of labels are uniformly marked with the same layer identifier; single labels are marked layer by layer and stacked, and the number of layers can be dynamically expanded; the upper layer label is an aggregated and simplified representation of the lower layer associated label, and the length of the label sequence of the upper layer label is reduced to 30%-50% of the lower layer associated label.

[0032] Further, when the multi-layer orchestration execution module executes automatic orchestration, it includes:

[0033] Generating from top to bottom: After manually defining the label sequence at the top layer, parsing layer by layer to generate the lower layer label sequence, and the parsing is realized through the label mapping table;

[0034] Generating from bottom to top: After completing container orchestration at the bottom layer, generating the upper layer simplified label sequence layer by layer through the correlation algorithm;

[0035] When executing intelligent orchestration, the condition for the verification module to trigger strategy adjustment is that the deviation between the actual result and the expected result exceeds the set threshold.

[0036] Further, the label extension header includes the following fields: layer identifier, label priority, target cloud protocol type, source cloud identifier, call target label, call target layer identifier, and other control contents; the API specification reconstruction includes: converting the label sequence into a JSON format or a YAML format supported by the target cloud.

[0037] Further, the label extension header includes the following fields: layer identifier, label priority, target cloud protocol type, source cloud identifier, call target label, call target layer identifier, and other control contents; the API specification reconstruction includes: converting the label sequence into a JSON format or a YAML format supported by the target cloud.

[0038] The present invention has the following beneficial effects:

[0039] First, it realizes the localization of cloud orchestration technology and provides another option for domestic cloud operators.

[0040] Second, it expands the existing orchestration technology content, expands the original orchestration technology that only realizes basic functions such as container arrangement and service discovery into container and function orchestration covering cross-cloud orchestration, multi-layer orchestration, etc., and realizes the transformation from single-level orchestration to multi-level orchestration.

[0041] Third, it combines with artificial intelligence, and through artificial intelligence technology, realizes the self-orchestration of cloud containers and functions, gets rid of the traditional manual orchestration method, and feeds back to artificial intelligence technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic diagram of the information annotation method for the multi-layer orchestration layer in the embodiment of the present invention.

[0043] Figure 2 It is a schematic diagram of multi-layer orchestration in the embodiment of the present invention.

[0044] Figure 3 It is a schematic diagram of manual multi-layer orchestration in multi-layer orchestration in the embodiment of the present invention.

[0045] Figure 4 It is a schematic diagram of automatic multi-layer orchestration in layer orchestration in the embodiment of the present invention.

[0046] Figure 5 It is a schematic diagram of intelligent multi-layer orchestration in multi-layer orchestration in the embodiment of the present invention.

[0047] Figure 6 It is a schematic diagram of multi-layer orchestration of a thinking tree and a thinking network in multi-layer orchestration in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0049] Please refer to Figure 1-6 , the embodiments of the present invention provide a multi-layer container orchestration method based on tags, including the following steps:

[0050] Step 1. Hierarchical tag definition: Define multi-layer tags for containers or functional modules, and embed layer information in each layer of tags to generate a tag database containing layer information. The layer information is a binary value or a string, and is used to identify the hierarchical relationship;

[0051] Multi-layer orchestration is based on single-layer orchestration. By introducing layer information represented by numerical values or strings in tags, the multi-layer orchestration of functions or containers is realized. The upper-layer tags are re-tagged for the tag sequences with high correlation in the lower layer, shortening the length of the orchestration sequence. The layer information introduction method can be unified annotation for the same-level tags, or individual annotation for each tag. The number of layers can be continuously stacked as needed, and the maximum number of layers is not limited. As Figure 1 shown, Figure 1 and the subsequent schematic diagrams only use three layers to simplify and illustrate the content. Figure 1 In which L1, L2, and L3 are the layer information of the corresponding layers.

[0052] In the tag forwarding communication protocol system, the layer information is a binary value, which is introduced during manual or automatic orchestration. By using the binary tag and real-time forwarding switching characteristics of the communication protocol, dynamic orchestration and multi-layer automatic orchestration are realized. In a cloud environment without a tag forwarding communication protocol system, if tags are added to containers or functional modules, multi-layer orchestration can also be introduced. The layer information is a string or a binary value, which is introduced during manual or automatic orchestration.

[0053] Multi-layer orchestration can also realize the combined orchestration of different hierarchical systems and different tag systems. The upper layer and the lower layer can be composed of different software and hardware systems, using different orchestration methods and different tag systems for orchestration. The orchestration result sequence of the upper-layer system can be directly used or translated to become the orchestration content that can be used by the lower-layer system, guiding the lower-layer system for orchestration and execution. There is no dependency relationship between the upper-layer system and the lower-layer system, and they can belong to different parts of the cloud system or different cloud systems. As Figure 2 shown, Figure 2The container layer, function layer, and behavior layer in it are only logical names, respectively indicating different hierarchical systems, and there can be several sub-layers within each system. The top-level orchestration result describes the real business process with a tag sequence; the middle-level orchestration result is the correspondence between the current layer's tags and the lower-layer tag sequence, used for the system to parse the lower-layer tag sequence; the bottom layer is the actual execution process, and the system uses it to call containers or functions. The top-level orchestration result sequence can be made public as the already orchestrated result for other automated programs or cloud management servers to call.

[0054] Step 2: Multi-layer orchestration execution: The orchestration methods in the embodiments of the present invention are divided into manual orchestration, automatic orchestration, and intelligent orchestration. Among them:

[0055] Manual orchestration: Based on the tag database in Step 1, label the tag sequence layer by layer from bottom to top, and establish the associated mapping of upper and lower layer tags through layer information;

[0056] Manual orchestration is that manually, according to the actual business requirements, using the orchestration management tool, relying on manual work to perform orchestration layer by layer from bottom to top. First, label the tag sequence with a high degree of association at the bottom layer, arrange the current layer's tag sequence according to functional requirements and add layer information, and then complete it in turn according to the same steps and methods, and finally form the layer-by-layer orchestration result. As Figure 3 shown, in the figure, it is indicated by unified information annotation, and the annotation order starts from layer 0 to layer 2.

[0057] Automatic orchestration: Based on the tag database in Step 1 and the upper and lower layer tag mapping relationship, generate a cross-layer tag orchestration sequence from top to bottom or from bottom to top;

[0058] Automatic orchestration is mainly a supplement to manual orchestration. The automated program can obtain the correspondence of tags between each level, and generate a tag orchestration sequence layer by layer from top to bottom or from bottom to top according to needs. In the top-down mode, at the top layer ( Figure 4 layer 2 in it), the tag orchestration sequence is formulated manually; in the middle layer ( Figure 4 layer 1 in it), the automated program completes the tag orchestration sequence of each level from top to bottom according to the tag correspondence; at the bottom layer ( Figure 4 layer 0 in it), according to the middle orchestration sequence, the corresponding tag sequence is automatically generated to complete the orchestration of containers or functions. In the bottom-up mode, at the bottom layer ( Figure 4 layer 0 in it), the container or function orchestration is completed manually; in the middle layer ( Figure 4 layer 1 in it), the automated program completes the tag orchestration sequence of each level from bottom to top according to the tag correspondence, and finally generates the tag sequence of the top layer ( Figure 4 layer 2 in it). As Figure 4 shown, in the figure, it is indicated by unified information annotation.

[0059] Intelligent Orchestration: Based on the orchestration sequence generated by automatic orchestration, the deviation between the actual execution result and the expected result is compared through a verification module, and the label sequences of each layer are dynamically adjusted;

[0060] After container orchestration is completed, a verification module is added to the orchestration system. On the one hand, it verifies the rationality of the label sequence of each layer layer by layer, and on the other hand, it verifies the data result processed by the container processing sequence at the bottom layer. Through a mode similar to artificial neural network training, the label sequence of each layer is checked and adjusted layer by layer from top to bottom or from bottom to top. When verifying in the top-down direction, the top layer generates a label orchestration sequence manually or by an automated program, and uses the automatic orchestration method to generate the label sequences of each layer layer by layer. The verification module verifies the deviation between the simulated or actual data processing result at the bottom layer and the expected result, continuously optimizes and adjusts the orchestration sequences of all levels, and obtains the optimal sequence combination. When verifying in the bottom-up direction, at the bottom layer, the verification module selects the optimal automatically generated orchestration sequence according to the target result, and then uses the automatic orchestration method to complete the streamlining of the orchestration sequence layer by layer upward, or automatically calculates the correlation degree between each level and each label, generates the label orchestration sequence level by level, and the verification module verifies and adjusts the label sequences of each level, and finally generates the streamlined orchestration result that meets the requirements. The verification module can be manual verification or machine automatic verification. In the label forwarding protocol, using the real-time forwarding and switching characteristics of the protocol, the orchestration result can be quickly applied to the production environment. For example Figure 5 , which is indicated by unified information annotation in the figure.

[0061] Utilize the function of intelligent orchestration to automatically generate label sequences of each layer, generate label sequences layer by layer from top to bottom or from bottom to top, cooperate with the routing weight activation strategy, construct the function of automatically generating and adjusting the processing path based on the target result, and then combine with the existing artificial intelligence code generation ability to build the function of generating and adjusting the programs of each container node, and finally realize a new type of artificial neural network that can program independently and route functions independently.

[0062] Generate an executable cross-layer label orchestration strategy based on the above multi-layer orchestration execution steps.

[0063] Step 3: Cross-cloud protocol adaptation:

[0064] If the target cloud supports the label protocol, encapsulate the orchestration strategy generated in Step 2 into a data frame with a binary label extension header; the label extension header contains the following fields: layer identifier (4bit), label priority (2bit), target cloud protocol type (2bit), source cloud identifier (4Byte), call target label (8Byte), call target layer identifier (4Byte), other control content (8Byte).

[0065] If the target cloud does not support the label protocol, reconstruct the orchestration strategy in step two into a standardized data payload according to the target cloud API specification; the API specification reconstruction includes: converting the label sequence into a JSON format or YAML format supported by the target cloud.

[0066] Generate a data transmission instruction compatible with heterogeneous cloud environments through the above cross-cloud protocol adaptation.

[0067] In the label forwarding communication protocol system, it is also possible to utilize features such as zoning, real-time forwarding switching, and cross-cloud forwarding, combined with technologies such as miniaturization and containerization of neural network models, to generate a large-scale general-purpose artificial neural network with functional partitioning, real-time adjustment, multi-cloud collaboration, and less computing power requirements.

[0068] Step Four: Artificial Intelligence Integration:

[0069] Map the cross-layer label sequence generated in step two into a thought chain, thought tree, or thought network structure; the thought network structure is a three-dimensional network that aggregates and classifies two-dimensional neural network nodes through layer information, reducing the complexity of the neural network; the routing weight strategy dynamically adjusts the node connection path based on label priority, target node resource occupancy ratio, and target node function.

[0070] The label sequence conforms to the thought chain structure, and a large number of label sequences can form a thought tree and a thought network. Through intelligent orchestration or an artificial intelligence algorithm with intelligent orchestration capabilities, by streamlining, thinning, and adding layer information, the artificial neural network is upgraded from a two-dimensional network to a three-dimensional network. On the one hand, the scale of the artificial neural network is expanded, and on the other hand, the complexity of the artificial neural network is reduced, reducing the computing power requirements, as Figure 6 shown.

[0071] Based on the data transmission instruction in step three, combined with the routing weight strategy, generate an autonomously programmed neural network model, and achieve computing power allocation, destination function selection and use through containerized deployment. The routing weight strategy dynamically adjusts the node connection path based on label priority, target node resource occupancy ratio, and target node function.

[0072] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A label-based multi-layer container orchestration method, characterized in that It includes the following steps: Step 1, Hierarchical label definition: Define multiple layers of labels for containers or functional modules, and embed layer information in each layer of labels to generate a label database containing layer information. The layer information is a binary value or a string, which is used to identify the hierarchical relationship; Step 2, Multi-layer orchestration execution: Manual orchestration: Based on the label database in Step 1, label the label sequence layer by layer from bottom to top, and establish an associated mapping between the upper and lower layer labels through layer information; Automatic orchestration: Based on the label database in Step 1 and the mapping relationship between the upper and lower layer labels, generate a cross-layer label orchestration sequence from top to bottom or from bottom to top; Intelligent orchestration: Based on the orchestration sequence generated by automatic orchestration, compare the deviation between the actual execution result and the expected result through a verification module, and dynamically adjust the label sequence of each layer; Generate an executable cross-layer label orchestration strategy based on the above multi-layer orchestration execution steps; Step 3, Cross-cloud protocol adaptation: If the target cloud supports the label protocol, encapsulate the orchestration strategy generated in Step 2 into a data frame with a binary label extension header; If the target cloud does not support the label protocol, reconstruct the orchestration strategy in Step 2 into a standardized data payload according to the target cloud API specification; Generate a data transmission instruction compatible with heterogeneous cloud environments through the above cross-cloud protocol adaptation; Step 4, Artificial intelligence integration: Map the cross-layer label sequence generated in Step 2 to a thinking chain, thinking tree or thinking network structure; Based on the data transmission instruction in Step 3, combine the routing weight strategy to generate a self-programming neural network model, and realize computing power allocation, destination function selection and use through containerized deployment.

2. The method according to claim 1, characterized in that, The introduction method of the layer information in Step 1 is: The same-level labels are uniformly marked with the same layer identifier; Single labels are marked layer by layer, and the number of layers can be dynamically expanded; The upper layer label is an aggregated and simplified representation of the lower layer associated labels, and the label sequence length of the upper layer label is reduced to 30%-50% of the lower layer associated labels.

3. The method according to claim 1, wherein The automatic orchestration in Step 2 includes: Generate from top to bottom: After manually defining the label sequence at the top layer, parse and generate the lower layer label sequence layer by layer, and the parsing is realized through a label mapping table; Generate from bottom to top: After completing the container orchestration at the bottom layer, generate the upper layer simplified label sequence layer by layer through a correlation algorithm; In intelligent orchestration, the condition for the verification module to trigger strategy adjustment is that the deviation between the actual result and the expected result exceeds the set threshold.

4. The method according to claim 1, characterized in that, The label extension header in Step 3 includes the following fields: layer identifier, label priority, target cloud protocol type, source cloud identifier, call target label, call target layer identifier, and other control contents; The API specification reconstruction includes: converting the label sequence into a JSON format or YAML format supported by the target cloud.

5. The method according to claim 1, characterized in that, The thinking network structure in Step 4 is a three-dimensional network, which aggregates and classifies two-dimensional neural network nodes through layer information to reduce the complexity of the neural network; The routing weight strategy dynamically adjusts the node connection path based on the label priority, the resource occupancy ratio of the target node, and the function of the target node.

6. A label-based multi-layer container orchestration system, characterized in that, It includes: A hierarchical label definition module, which is used to define multiple layers of labels for containers or functional modules, and embed layer information in each layer of labels to generate a label database containing layer information. The layer information is a binary value or a string, which is used to identify the hierarchical relationship; The multi - layer orchestration execution module is used to execute multi - layer orchestration to generate an executable cross - layer label orchestration strategy. The multi - layer orchestration includes manual orchestration, automatic orchestration, and intelligent orchestration. Among them, manual orchestration is based on a label database, layer - by - layer annotates the label sequence from bottom to top, and establishes an associated mapping of upper and lower layer labels through layer information; automatic orchestration is based on the label database and the upper and lower layer label mapping relationship, and generates a cross - layer label orchestration sequence from top to bottom or from bottom to top; intelligent orchestration is based on the orchestration sequence generated by automatic orchestration, and dynamically adjusts the label sequences of each layer by comparing the deviation between the actual execution result and the expected result through a verification module. The cross - cloud protocol adaptation module is used to encapsulate the orchestration strategy generated by the multi - layer orchestration execution module into a data frame with a binary label extension header if the target cloud supports the label protocol. If the target cloud does not support the label protocol, it reconstructs the orchestration strategy generated by the multi - layer orchestration execution module into a standardized data payload according to the target cloud API specification; generates a data transmission instruction compatible with heterogeneous cloud environments based on the cross - cloud protocol adaptation result. The artificial intelligence integration module is used to map the cross - layer label sequence generated by the multi - layer orchestration execution module into a thinking chain, thinking tree, or thinking network structure; based on the data transmission instruction generated by the cross - cloud protocol adaptation module, combines the routing weight strategy to generate a self - programming neural network model, and realizes computing power allocation, destination function selection and use through containerized deployment.

7. The tag-based multi-layer container orchestration system according to claim 6, wherein The way of introducing the layer information in the hierarchical label definition module is as follows: the same - level labels are uniformly marked with the same layer identifier; single labels are marked layer - by - layer and the number of layers can be dynamically expanded; the upper - layer label is an aggregated and simplified representation of the lower - layer associated label, and the length of the label sequence of the upper - layer label is reduced to 30% - 50% of the lower - layer associated label.

8. The tag-based multi-layer container orchestration system according to claim 6, characterized in that When the multi - layer orchestration execution module executes automatic orchestration, it includes: Generating from top to bottom: After defining the label sequence manually at the top layer, it parses layer by layer to generate the lower - layer label sequence, and the parsing is realized through a label mapping table. Generating from bottom to top: After completing container orchestration at the bottom layer, it generates the upper - layer simplified label sequence layer by layer through a correlation algorithm. When executing intelligent orchestration, the condition for the verification module to trigger strategy adjustment is that the deviation between the actual result and the expected result exceeds the set threshold.

9. The tag-based multi-layer container orchestration system according to claim 6, characterized in that The label extension header includes the following fields: layer identifier, label priority, target cloud protocol type, source cloud identifier, call target label, call target layer identifier, and other control contents. The API specification reconstruction includes: converting the label sequence into a JSON format or YAML format supported by the target cloud.

10. The tag-based multi-layer container orchestration system according to claim 6, characterized in that The thinking network structure in the artificial intelligence integration module is a three - dimensional network, which aggregates and classifies two - dimensional neural network nodes through layer information to reduce the complexity of the neural network; the routing weight strategy dynamically adjusts the node connection path based on label priority, target node resource occupancy ratio, and target node function.

Images