High-efficiency dual-protocol intelligent agent data interaction system

Abstract

The application relates to the technical field of data interaction, and discloses a high-efficiency dual-protocol intelligent agent data interaction system, which comprises a dual-stack deployment and channel mapping module, a cross-protocol weight distribution module, a collaborative perception logical channel construction module, a cross-protocol flow control coupling and transparent migration module and a packaging and delivery module. Dual-protocol communication components containing MCP protocol stacks and A2A protocol stacks are deployed on each intelligent agent node, and a virtual channel mapping table is established. The channel distribution weight of data packets between MCP channels and A2A channels is determined. Dual-protocol logical channels are constructed, and data packets with cross-protocol channels and high business collaboration degrees are mapped to the same logical topology path. When the dual-protocol logical channels are congested, the cross-protocol flow control coupling channel distribution weight is distributed, and protocol transparent switching is performed on the data flow on the corresponding logical topology path. The migrated data packets after adaptive packaging are transmitted to target intelligent agent nodes. The application can improve the efficiency of intelligent agent data interaction.

Description

Technical Field

[0001] This invention relates to the field of data interaction technology, and in particular to a high-efficiency dual-protocol intelligent agent data interaction system. Background Technology

[0002] In multi-agent network data interaction scenarios, existing communication architectures generally adopt a single protocol transmission mode, with fixed protocol channels. This makes it impossible to coordinate and allocate multi-protocol resources based on business characteristics and real-time network status. Network bandwidth resources are not used in a balanced manner, and it is difficult to meet the differentiated transmission requirements of different services for latency and packet loss tolerance at the same time. The adaptability and flexibility of data transmission between agent nodes have obvious shortcomings.

[0003] Existing intelligent agent data interaction systems have significant shortcomings in cross-protocol logical channel construction and business collaborative perception. They also lack cross-protocol flow control coupling mechanisms and the ability to adapt to transparent data flow migration. When faced with dynamic changes such as network link congestion and latency jitter, they are unable to autonomously complete dynamic switching and weight adjustment of transmission channels. Highly collaborative business data flows are prone to path fragmentation and transmission blockage, resulting in high data interaction latency, insufficient transmission stability, and difficulty in improving overall data interaction throughput performance. Therefore, how to improve the efficiency of intelligent agent data interaction has become an urgent problem to be solved. Summary of the Invention

[0004] In view of the above problems, the purpose of this invention is to provide an efficient dual-protocol intelligent agent data interaction system to solve the problems of current technologies that mostly adopt single-protocol transmission mode, lack dual-protocol collaborative scheduling, business flow collaborative path planning and data flow adaptive transparent migration capability under congestion, low network resource utilization, high data interaction latency and poor stability.

[0005] This invention provides a high-efficiency dual-protocol intelligent agent data interaction system, characterized in that the system includes a dual-stack deployment and channel mapping module, a cross-protocol weight allocation module, a collaborative perception logic channel construction module, a cross-protocol flow control coupling and transparent migration module, and an encapsulation and delivery module, wherein: The dual-stack deployment and channel mapping module is used to deploy a dual-protocol communication component containing an MCP protocol stack and an A2A protocol stack on each agent node, and to establish a virtual channel mapping table. The cross-protocol weight allocation module is used to determine the channel allocation weight of the data packet between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators. The collaborative sensing logical channel construction module is used to construct a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, and jointly map cross-protocol channel data packets with high service coordination to the same logical topology path; The cross-protocol flow control coupling and transparent migration module is used to assign weights to the channel through cross-protocol flow control when the dual-protocol logical channel becomes congested, and to perform transparent protocol switching on the data streams on the corresponding logical topology path, so as to realize the active migration of data streams between the MCP channel and the A2A channel. The encapsulation and delivery module is used to adaptively encapsulate the migrated data packets and transmit the encapsulated frames to the target intelligent agent node through the dual-protocol logic channel.

[0006] In a preferred embodiment, when the dual-stack deployment and channel mapping module deploys a dual-protocol communication component including an MCP protocol stack and an A2A protocol stack on each agent node and establishes a virtual channel mapping table, it is specifically used for: In the operating system kernel of the intelligent agent node, the MCP protocol stack kernel component and the A2A protocol stack kernel component are loaded to obtain a dual protocol stack kernel instance; Based on the dual-protocol stack kernel instance, register the MCP virtual network interface and the A2A virtual network interface; Each of the MCP virtual network interface and the A2A virtual network interface is assigned an independent intra-node communication port number, and a mapping relationship is established between the intra-node communication port number and the remote intelligent agent node identifier to obtain a virtual channel mapping table.

[0007] In a preferred embodiment, before the cross-protocol weight allocation module determines the channel allocation weight of the data packet between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators, it is specifically used for: Extract the service coordination degree identifier, end-to-end latency requirement level, and packet loss tolerance identifier of the data stream to be transmitted, and generate a normalized service feature vector; The available bandwidth and round-trip time jitter of the MCP channel are obtained through the MCP virtual network interface, and the available bandwidth and round-trip time jitter of the A2A channel are obtained through the A2A virtual network interface.

[0008] In a preferred embodiment, the cross-protocol weight allocation module, when determining the channel allocation weight of data packets between the MCP channel and the A2A channel based on the characteristic parameter set of the data stream to be transmitted and the current network status indicators, specifically performs the following: Based on the normalized service feature vector, the available bandwidth of the MCP channel, and the available bandwidth of the A2A channel, a basic allocation bias factor is determined, wherein the basic allocation bias factor is used to characterize the degree to which the data flow is initially biased towards the MCP channel allocation without considering the real-time congestion of the channel. The availability index of the MCP channel is determined based on the available bandwidth and round-trip time jitter of the MCP channel. The availability index of the MCP channel increases with the increase of the available bandwidth of the MCP channel and decreases with the increase of the round-trip time jitter of the MCP channel. The availability index of the A2A channel is determined based on the available bandwidth and round-trip time jitter of the A2A channel. The MCP channel allocation weight is calculated using the basic allocation tendency factor, the MCP channel availability index, and the A2A channel availability index. The value obtained by subtracting the MCP channel allocation weight is then used as the A2A channel allocation weight.

[0009] In a preferred embodiment, the formula for calculating the MCP channel allocation weight is as follows: ; In the formula, Assign weights to the MCP channels. The basic allocation tendency factor, This is the preset amplification factor for channel state differences. This refers to the availability index of the MCP channel. This is the availability index of the A2A channel.

[0010] In a preferred embodiment, the collaborative sensing logical channel construction module, when performing the construction of a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, is specifically used for: Obtain the link connectivity and link costs between each intelligent agent node to obtain real-time network topology information; Based on the virtual channel mapping table, from the real-time network topology information, node pairs that have both MCP and A2A reachable paths between the source agent node and the target agent node are selected to obtain a set of dual-protocol reachable node pairs. On the topology subgraph formed by the set of dual-protocol reachable node pairs, search for logical paths that satisfy the dual-protocol bandwidth constraints to construct a dual-protocol logical channel.

[0011] In a preferred embodiment, when the collaborative sensing logical channel construction module performs the joint mapping of cross-protocol channel data packets with high service coordination to the same logical topology path, it is specifically used for: Data packets belonging to the same business collaboration group in the data stream to be transmitted are regarded as a collaboration data packet cluster; The service coordination degree is obtained by adding the entropy of the destination address sequence of the MCP data sub-stream and the entropy of the destination address sequence of the A2A data sub-stream in the coordinated data packet cluster, and then subtracting the entropy of the joint address sequence of the two. When the business synergy exceeds a preset synergy threshold, the synergy data packet cluster is determined to have high business synergy, and the MCP data sub-stream and A2A data sub-stream in the synergy data packet cluster with high business synergy are jointly mapped to the same logical topology path.

[0012] In a preferred embodiment, when the cross-protocol flow control coupling and transparent migration module performs the cross-protocol flow control coupling of the channel weight allocation when congestion occurs in the dual-protocol logical channel, it is specifically used for: In the dual-protocol logical channel, the sub-channel whose queue depth exceeds the corresponding preset congestion threshold is designated as the congested sub-channel. When a congested sub-channel occurs, freeze the update of the current channel allocation weight pair and take the most recently determined channel allocation weight pair before the congestion as the snapshot of the optimal allocation weight. In the optimal allocation weight snapshot, the channel allocation weights of the congested sub-channels are shifted and adjusted towards the non-congested sub-channels to obtain the coupled adjusted channel allocation weight pairs.

[0013] In a preferred embodiment, when the cross-protocol flow control coupling and transparent migration module performs protocol transparent switching on the data streams on the corresponding logical topology path to achieve active migration of the data streams between the MCP channel and the A2A channel, it is specifically used for: In the protocol stack send buffer of the congested sub-channel, data packets that have completed flow control confirmation but have not yet been sent are treated as data packets to be migrated; The current protocol header of the data packet to be migrated is replaced with the protocol header of the non-congested sub-channel, while the application layer payload and cooperative flow label are retained to obtain a transparent migration data packet; The transparent migration data packet is inserted into the transmission queue of the non-congested sub-channel, and the channel mapping entry of the transparent migration data packet in the virtual channel mapping table is updated.

[0014] In a preferred embodiment, when the encapsulation and delivery module performs adaptive frame encapsulation on the migrated data packets and transmits the encapsulated frames to the target agent node through the dual-protocol logical channel, it is specifically used for: The transparent migration data packet is processed using a unified encapsulation format to obtain an encapsulated data packet; Select the corresponding verification mechanism based on the protocol type of the non-congested sub-channel to which the transparent migration data packet currently belongs; When the non-congested sub-channel is an MCP channel, a lightweight check mechanism is used to generate the first check code. When the non-congested sub-channel is an A2A channel, a high-strength check mechanism is used to generate a second check code. The encapsulated data packet and the corresponding checksum are encapsulated into an adaptive frame, and the adaptive frame is sent to the target agent node through the non-congested sub-channel.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention deploys a dual protocol stack of MCP and A2A on the agent node and constructs a virtual channel mapping table. It dynamically allocates the weights of the dual protocol channels based on service characteristic parameters and network status indicators. At the same time, it builds a dual protocol logical channel based on the real-time network topology, and aggregates data packets with high service coordination to the same logical topology path. This enables the coordinated scheduling of dual protocol communication resources, adapts to differentiated service transmission requirements, and effectively improves network bandwidth utilization and multi-agent data transmission adaptation capabilities.

[0016] 2. This invention can perceive the congestion status of logical channels in real time, optimize channel allocation weights through cross-protocol flow control coupling, and complete transparent switching of data flow protocols and active migration across channels. At the same time, it adopts a matching verification mechanism for different protocols to perform adaptive frame encapsulation and delivery, which can adapt to dynamic changes such as network latency jitter and link congestion, avoid the problems of blocking and path fragmentation in the transmission of collaborative business data flow, reduce data interaction latency, improve transmission stability and reliability, and significantly optimize the overall efficiency of multi-agent data interaction. Attached Figure Description

[0017] Figure 1 This is a system architecture diagram of an efficient dual-protocol intelligent agent data interaction system provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments belong to some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “said” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0020] Depending on the context, the word "if" or "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0021] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0022] In practice, the server-side equipment deployed in an efficient dual-protocol intelligent agent data interaction system may consist of one or more devices. This efficient dual-protocol intelligent agent data interaction system can be implemented as: a business instance, a virtual machine, or hardware devices. For example, this efficient dual-protocol intelligent agent data interaction system can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, this efficient dual-protocol intelligent agent data interaction system can be understood as software deployed on a cloud node, used to provide an efficient dual-protocol intelligent agent data interaction system for each user terminal. Alternatively, this efficient dual-protocol intelligent agent data interaction system can also be implemented as a virtual machine deployed on one or more devices in a cloud node. This virtual machine contains application software for managing each user terminal. Alternatively, this efficient dual-protocol intelligent agent data interaction system can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide an efficient dual-protocol intelligent agent data interaction system for each user terminal.

[0023] In terms of implementation, the efficient dual-protocol intelligent agent data interaction system and the user terminal are mutually compatible. That is, if the efficient dual-protocol intelligent agent data interaction system is implemented as an application installed on a cloud service platform, then the user terminal is implemented as a client that establishes a communication connection with the application; or if the efficient dual-protocol intelligent agent data interaction system is implemented as a website, then the user terminal is implemented as a webpage; or if the efficient dual-protocol intelligent agent data interaction system is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.

[0024] like Figure 1 The figure shown is a system architecture diagram of a high-efficiency dual-protocol intelligent agent data interaction system provided in an embodiment of the present invention.

[0025] The efficient dual-protocol intelligent agent data interaction system 100 described in this invention can be set up in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed as a website. Depending on the implemented functions, the efficient dual-protocol intelligent agent data interaction system 100 may include a dual-stack deployment and channel mapping module 101, a cross-protocol weight allocation module 102, a collaborative perception logic channel construction module 103, a cross-protocol flow control coupling and transparent migration module 104, and an encapsulation and delivery module 105. The modules described in this invention can also be called units, referring to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0026] In this embodiment of the invention, in a high-efficiency dual-protocol intelligent agent data interaction system, each of the above-mentioned modules can be implemented independently and can call other modules. Here, "calling" can be understood as a module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the high-efficiency dual-protocol intelligent agent data interaction system provided by this embodiment of the invention, the applicable scope of the system architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the high-efficiency dual-protocol intelligent agent data interaction system. In practical applications, the above-mentioned modules can be set in the same device or different devices, or they can be set in virtual devices, such as service instances in a cloud server.

[0027] The following describes, with reference to specific embodiments, each component and its specific workflow of a high-efficiency dual-protocol intelligent agent data interaction system: The dual-stack deployment and channel mapping module 101 is used to deploy a dual-protocol communication component containing an MCP protocol stack and an A2A protocol stack on each agent node, and to establish a virtual channel mapping table. In this embodiment of the invention, when the dual-stack deployment and channel mapping module deploys a dual-protocol communication component containing an MCP protocol stack and an A2A protocol stack on each agent node and establishes a virtual channel mapping table, it is specifically used for: In the operating system kernel of the intelligent agent node, the MCP protocol stack kernel component and the A2A protocol stack kernel component are loaded to obtain a dual protocol stack kernel instance; Based on the dual-protocol stack kernel instance, register the MCP virtual network interface and the A2A virtual network interface; Each of the MCP virtual network interface and the A2A virtual network interface is assigned an independent intra-node communication port number, and a mapping relationship is established between the intra-node communication port number and the remote intelligent agent node identifier to obtain a virtual channel mapping table.

[0028] The operating system kernel allocates a dedicated kernel component loading storage space. This storage space is set with a fixed memory addressing range, which is only suitable for communication protocol stack kernel components. The complete packaged MCP protocol stack kernel component program file and A2A protocol stack kernel component program file are written into this memory addressing range to complete the mounting of the component's underlying driver and the registration of the kernel communication interface. After the mounting and registration are completed, the scheduling path and data transmission and reception path of the two types of protocol stack kernel components are automatically merged. After merging, an independent and complete running entity is generated. This running entity is the dual protocol stack kernel instance.

[0029] The dual-protocol stack kernel instance calls the virtual interface registration service built into the operating system kernel. This service sets a unified virtual interface identity coding standard, assigns a dedicated identity code to the MCP service and an independent identity code to the A2A service according to the coding standard, and initiates a virtual interface creation request to the operating system kernel with their respective dedicated identity codes. After receiving the request, the kernel allocates dedicated virtual link resources and hardware interrupt response resources to the two types of interfaces respectively. After the resource configuration is completed, MCP virtual network interfaces and A2A virtual network interfaces with independent data reception, forwarding and sending capabilities are generated respectively.

[0030] According to the port number segmentation standard preset by the operating system, the available communication port numbers are divided into two independent port segments that do not overlap. A fixed number is selected from one port segment as the exclusive intra-node communication port number for the MCP virtual network interface, and a fixed number is selected from the other port segment as the exclusive intra-node communication port number for the A2A virtual network interface. The unique identification code corresponding to all remote intelligent agent nodes is collected, and the intra-node communication port number matched by each virtual network interface is bound one by one with the corresponding remote intelligent agent node identifier. All binding and association information is recorded one by one in a fixed row list format. All association records are integrated to form a structured data table, which is the virtual channel mapping table.

[0031] The beneficial effects are as follows: by deploying MCP and A2A dual-protocol stack kernel instances in the kernel, the two types of protocols can run in parallel and independently at the underlying level without competing for kernel scheduling and transmission resources; registering dedicated virtual network interfaces can realize link isolation and independent sending and receiving of data streams of different protocols, avoiding data crosstalk; the constructed virtual channel mapping table can accurately establish the binding relationship between local ports and remote intelligent agent identifiers, providing a regular underlying mapping basis for subsequent cross-protocol channel allocation and logical path mapping, ensuring the orderly conduct of dual-protocol data interaction between intelligent agents.

[0032] The cross-protocol weight allocation module 102 is used to determine the channel allocation weight of the data packet between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators. In this embodiment of the invention, before the cross-protocol weight allocation module determines the channel allocation weight of the data packet between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators, it is specifically used for: Extract the service coordination degree identifier, end-to-end latency requirement level, and packet loss tolerance identifier of the data stream to be transmitted, and generate a normalized service feature vector; The available bandwidth and round-trip time jitter of the MCP channel are obtained through the MCP virtual network interface, and the available bandwidth and round-trip time jitter of the A2A channel are obtained through the A2A virtual network interface.

[0033] The cross-protocol weight allocation module, when determining the channel allocation weight of data packets between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators, is specifically used for: Based on the normalized service feature vector, the available bandwidth of the MCP channel, and the available bandwidth of the A2A channel, a basic allocation bias factor is determined, wherein the basic allocation bias factor is used to characterize the degree to which the data flow is initially biased towards the MCP channel allocation without considering the real-time congestion of the channel. The availability index of the MCP channel is determined based on the available bandwidth and round-trip time jitter of the MCP channel. The availability index of the MCP channel increases with the increase of the available bandwidth of the MCP channel and decreases with the increase of the round-trip time jitter of the MCP channel. The availability index of the A2A channel is determined based on the available bandwidth and round-trip time jitter of the A2A channel. The MCP channel allocation weight is calculated using the basic allocation tendency factor, the MCP channel availability index, and the A2A channel availability index. The value obtained by subtracting the MCP channel allocation weight is then used as the A2A channel allocation weight.

[0034] The formula for calculating the weight allocation of the MCP channel is as follows: ; In the formula, Assign weights to the MCP channels. The basic allocation tendency factor, This is the preset amplification factor for channel state differences. This refers to the availability index of the MCP channel. This is the availability index of the A2A channel.

[0035] The three types of identifier codes recorded in the fixed storage area at the header of the data stream to be transmitted are read. The business coordination degree identifier is predefined into one hundred continuous level intervals from zero to one hundred. The end-to-end latency requirement level is predefined into five fixed latency interval exclusive codes. The packet loss tolerance identifier is predefined into six packet loss threshold exclusive codes. The level values ​​corresponding to the three types of identifiers are uniformly converted into standard values ​​in the range of zero to one. The three standard values ​​after conversion are integrated into an ordered set of values ​​according to a fixed order. This ordered set of values ​​is the normalized business feature vector.

[0036] The function of reading the channel status built into the MCP virtual network interface is invoked. This function continuously collects the idle transmission capacity value of the physical link connected to the MCP virtual network interface as the available bandwidth of the MCP channel. It also synchronously collects the time difference value generated by ten consecutive round-trip transmissions of data packets and calculates the fluctuation difference between all time differences as the round-trip delay jitter of the MCP channel. The function of reading the channel status built into the A2A virtual network interface is invoked. This function continuously collects the idle transmission capacity value of the physical link connected to the A2A virtual network interface as the available bandwidth of the A2A channel. It also synchronously collects the time difference value generated by ten consecutive round-trip transmissions of data packets and calculates the fluctuation difference between all time differences as the round-trip delay jitter of the A2A channel.

[0037] The normalized service feature vector is decomposed into three sets of standard values ​​with fixed order. The service channel preference scores corresponding to the three sets of values ​​are retrieved by referring to a pre-compiled service matching table. The retrieved service channel preference scores are matched and converted with the available bandwidth of the MCP channel and the available bandwidth of the A2A channel, respectively. The conversion process strictly follows the preset bandwidth preference conversion rules to complete the numerical conversion. The two sets of converted values ​​are substituted into the pre-defined interval mapping rules for processing. Finally, a fixed value in the range of zero to one is output. This value is the basic allocation tendency factor. The basic allocation tendency factor is equipped with a fixed bias judgment standard. A value of 0.5 means that there is no initial allocation bias for the two types of channels after the real-time congestion state of the data flow stripping channel. A value higher than 0.5 means that the initial allocation bias is towards the MCP channel after the real-time congestion state of the data flow stripping channel. A value lower than 0.5 means that the initial allocation bias is towards the A2A channel after the real-time congestion state of the data flow stripping channel.

[0038] The system retrieves a pre-stored channel status baseline data table. This table contains tiered available bandwidth standard ranges and tiered round-trip delay jitter standard ranges. The system matches the currently acquired available bandwidth of the MCP channel to the corresponding bandwidth standard range and retrieves the corresponding positive baseline score. Similarly, it matches the currently acquired round-trip delay jitter of the MCP channel to the corresponding delay jitter standard range and retrieves the corresponding negative correction score. The positive baseline score is subtracted from the negative correction score to obtain an intermediate value. This intermediate value is then converted to a value between zero and one using a preset value compression rule. The converted value is the MCP channel availability index. Each upward shift of the bandwidth standard range increases the corresponding positive baseline score by a fixed value, and each upward shift of the delay jitter standard range increases the corresponding negative correction score by a fixed value. This score variation rule ensures that the MCP channel availability index increases synchronously with the increase in the available bandwidth value of the MCP channel and decreases synchronously with the increase in the round-trip delay jitter value.

[0039] Retrieve the pre-stored channel status baseline data table shared with the MCP channel. The data table contains the available bandwidth standard range and the round-trip delay jitter standard range divided into different tiers. Match the available bandwidth of the currently acquired A2A channel to the corresponding bandwidth standard range and retrieve the positive baseline score for that range. Match the round-trip delay jitter of the currently acquired A2A channel to the corresponding delay jitter standard range and retrieve the negative correction score for that range. Subtract the negative correction score from the positive baseline score to obtain the intermediate value. Convert the intermediate value to the zero to one range using a preset value compression rule. The converted range value is the A2A channel availability index.

[0040] The basic allocation preference factor is generated by combining the normalized service feature vector with the available bandwidth of the two types of channels, without introducing real-time channel latency jitter data throughout the process; the MCP channel availability index is obtained by converting the available bandwidth and round-trip latency jitter of the channel read from the MCP virtual network interface with a benchmark score and interval compression; the A2A channel availability index is obtained by converting the available bandwidth and round-trip latency jitter of the channel read from the A2A virtual network interface with a benchmark score and interval compression.

[0041] The channel status difference amplification coefficient is determined by filtering and calibrating based on the massive historical data of synchronous operation of dual-protocol channels accumulated by the intelligent agent nodes. The complete filtering process involves batch retrieving tens of thousands of synchronously collected MCP channel availability indices and A2A channel availability indices, substituting different constant values ​​one by one to calculate the output difference amplitude of the two indices, and comparing it with the fixed threshold of channel switching response sensitivity preset in the business data stream. Only the unique fixed constant that can make the index difference amplitude match the sensitivity threshold is retained. This constant is pre-stored in the read-only storage area of ​​the operating system kernel, and the value of this constant will not be modified in a single data stream weight calculation.

[0042] This set of numerical conversion logic synchronously incorporates two types of information: the inherent service allocation preferences of the data stream to be transmitted and the real-time link operation quality of the two types of channels. It uses a channel status difference amplification coefficient to increase the numerical differentiation of the link quality of the two channels. Through a three-layer fixed numerical conversion process of exponential mapping, summation, and division, it outputs the MCP channel allocation weight. This weight value directly determines the proportion of data stream capacity diverted to the MCP virtual network interface. The difference between the numerical constant and this weight determines the proportion of data stream capacity of the A2A virtual network interface, realizing traffic distribution control that coordinates the matching of service requirements and real-time channel conditions.

[0043] The beneficial effects include: synchronously collecting the business attribute information inherent in the data stream and the real-time channel operation data output from the two types of virtual network interfaces; calculating a basic allocation tendency factor that fits the inherent transmission requirements of the data stream based on the business information; generating a dual-channel availability index that objectively reflects link quality based on channel bandwidth and latency jitter; and obtaining precise dual-channel allocation weights through a unified conversion logic. This can simultaneously take into account the inherent transmission preferences of the business and the real-time load conditions of the channels, dynamically adjust the data splitting ratio of the two types of protocol channels, alleviate the problem of excessive load on single-channel data transmission, reduce data stream transmission stuttering and packet loss, and improve the overall smoothness of data interaction under dual-protocol parallel transmission.

[0044] The collaborative sensing logical channel construction module 103 is used to construct a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, and jointly map data packets with high service coordination across protocol channels to the same logical topology path; In this embodiment of the invention, the collaborative sensing logical channel construction module, when performing the construction of a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, is specifically used for: Obtain the link connectivity and link costs between each intelligent agent node to obtain real-time network topology information; Based on the virtual channel mapping table, from the real-time network topology information, node pairs that have both MCP and A2A reachable paths between the source agent node and the target agent node are selected to obtain a set of dual-protocol reachable node pairs. On the topology subgraph formed by the set of dual-protocol reachable node pairs, search for logical paths that satisfy the dual-protocol bandwidth constraints to construct a dual-protocol logical channel.

[0045] The collaborative sensing logical channel construction module, when executing the joint mapping of cross-protocol channel data packets with high service coordination to the same logical topology path, is specifically used for: Data packets belonging to the same business collaboration group in the data stream to be transmitted are regarded as a collaboration data packet cluster; The service coordination degree is obtained by adding the entropy of the destination address sequence of the MCP data sub-stream and the entropy of the destination address sequence of the A2A data sub-stream in the coordinated data packet cluster, and then subtracting the entropy of the joint address sequence of the two. When the business synergy exceeds a preset synergy threshold, the synergy data packet cluster is determined to have high business synergy, and the MCP data sub-stream and A2A data sub-stream in the synergy data packet cluster with high business synergy are jointly mapped to the same logical topology path.

[0046] Each agent node continuously pushes local link connectivity status messages to neighboring nodes to summarize the link connectivity relationships between all nodes. It synchronously collects the real-time round-trip delay and idle bandwidth values ​​of each link and integrates them into a unified evaluation value as the link cost. It integrates the connectivity status records and link evaluation values ​​of all links in the entire network into a complete structured information set, which is the real-time network topology information.

[0047] The previously generated and stored virtual channel mapping table is retrieved. This table fully records the binding relationship between the communication port number within the node and the identifier of the remote intelligent agent node. All remote intelligent agent node identifiers in the table are extracted one by one. The entire network path record stored in the real-time network topology information is compared with the records. The transmission path from any source intelligent agent node to the corresponding target intelligent agent node is checked one by one to see if there are independent transmission paths that are adapted to both the MCP protocol and the A2A protocol. All the verified source node and target node pairing records are collected and organized into an ordered record set. This ordered record set is the set of dual-protocol reachable node pairs.

[0048] Extract all source nodes, target nodes, and corresponding link records between nodes from the set of reachable nodes for both protocols. Based on the extracted node and link data, a local topology graph independent of the complete real-time network topology is delineated. This local topology graph is called a topology subgraph. Pre-set fixed values ​​for the minimum idle bandwidth of MCP links and A2A links as dual-protocol bandwidth constraints. Completely traverse all node connection paths within the topology subgraph, verifying whether the available bandwidth values ​​for both protocols on each link in the path all reach the minimum fixed bandwidth value of the corresponding protocol. Completely retain the connection paths where both types of bandwidth values ​​meet the constraint standards. Encapsulate the retained paths into a dedicated transmission path structure adapted for parallel transmission of dual-protocol data streams. This dedicated transmission path structure is the dual-protocol logical channel.

[0049] The dedicated code for the business collaboration group is read from the fixed storage area at the header of a single data stream to be transmitted. The dedicated code for the business collaboration group is predefined as a continuous coding range from zero to two hundred. All data packets with completely identical header coding numbers are uniformly collected and stored. The collected data packets form a complete data packet combination, which is the collaborative data packet cluster.

[0050] The process involves splitting the MCP data substream from the collaborative data packet cluster, counting the frequency of each address code, and calculating the disorder of the address sequence based on the frequency distribution. The output disorder result is the entropy of the destination address sequence of the MCP data substream. Similarly, it involves splitting the A2A data substream from the collaborative data packet cluster, counting the frequency of each address code, and calculating the disorder of the address sequence based on the frequency distribution. The output disorder result is the entropy of the destination address sequence of the A2A data substream. Simultaneously, it involves counting the joint frequency of all paired destination address codes after mixing the two types of data substreams, and calculating the disorder of the mixed address sequence based on the joint frequency distribution. The output disorder result is the entropy of the joint address sequence. Finally, the two sets of disorder statistics for the MCP and A2A substreams are added together. The sum is then subtracted from the disorder statistics for the joint address sequence. The final numerical output is the business collaboration degree.

[0051] The preset collaboration threshold is determined by setting a fixed value based on tens of thousands of historical business data packet transmission records. This fixed value serves as the benchmark for determining the strength of business interaction dependencies among data packet clusters. A value higher than this fixed value indicates that the data interaction correlation strength of data packets within the data packet cluster has reached the predetermined transmission standard. The calculated business collaboration value is compared with the fixed value of the preset collaboration threshold. When the business collaboration value is greater than this fixed value, the corresponding collaborative data packet cluster is determined to have high business collaboration. The complete record of the logical topology path corresponding to the dual-protocol logical channel built in the early stage is retrieved. The MCP data sub-stream forwarding rules and A2A data sub-stream forwarding rules obtained from the split collaborative data packet cluster are synchronously bound to the same logical topology path to complete the joint mapping operation of the two types of data sub-streams to the same topology path.

[0052] The beneficial effects include the ability to accurately collect the link connectivity status and link costs of all intelligent agent nodes in the network, forming complete real-time network topology information, accurately selecting node pairs with dual-protocol transmission conditions, and constructing compliant and usable dual-protocol logical channels based on fixed bandwidth constraints, ensuring the resource adaptability of the transmission path. It accurately calculates the degree of service coordination by quantifying the degree of address sequence disorder, and standardizes the determination of high-service-coordination data packet clusters based on fixed preset coordination thresholds. It uniformly maps the dual-protocol data sub-streams of related services to the same logical topology path, avoiding timing deviations and resource waste caused by the dispersed transmission of collaborative service data streams, strengthening the synchronization and integrity of related data transmission, and maximizing the adaptation of network topology link resources, thereby improving the stability and transmission efficiency of multi-agent collaborative service data interaction.

[0053] The cross-protocol flow control coupling and transparent migration module 104 is used to perform cross-protocol flow control coupling of the channel allocation weights and to perform protocol transparent switching of the data stream on the corresponding logical topology path when the dual-protocol logical channel is congested, so as to realize the active migration of the data stream between the MCP channel and the A2A channel. In this embodiment of the invention, when the cross-protocol flow control coupling and transparent migration module performs the cross-protocol flow control coupling of the channel weight allocation when congestion occurs in the dual-protocol logical channel, it is specifically used for: In the dual-protocol logical channel, the sub-channel whose queue depth exceeds the corresponding preset congestion threshold is designated as the congested sub-channel. When a congested sub-channel occurs, freeze the update of the current channel allocation weight pair and take the most recently determined channel allocation weight pair before the congestion as the snapshot of the optimal allocation weight. In the optimal allocation weight snapshot, the channel allocation weights of the congested sub-channels are shifted and adjusted towards the non-congested sub-channels to obtain the coupled adjusted channel allocation weight pairs.

[0054] The cross-protocol flow control coupling and transparent migration module, when performing protocol-transparent switching on data streams along the corresponding logical topology path to achieve active migration of data streams between the MCP channel and the A2A channel, is specifically used for: In the protocol stack send buffer of the congested sub-channel, data packets that have completed flow control confirmation but have not yet been sent are treated as data packets to be migrated; The current protocol header of the data packet to be migrated is replaced with the protocol header of the non-congested sub-channel, while the application layer payload and cooperative flow label are retained to obtain a transparent migration data packet; The transparent migration data packet is inserted into the transmission queue of the non-congested sub-channel, and the channel mapping entry of the transparent migration data packet in the virtual channel mapping table is updated.

[0055] Based on the historical data carried by the sub-channel queues collected by the intelligent agent nodes over a long period of time, independent fixed data packet count values ​​are set for the MCP sub-channel and A2A sub-channel as their respective preset congestion thresholds. The cumulative number of data packets temporarily stored and waiting to be forwarded in the buffer queue of each sub-channel under the dual protocol logic channel is counted in real time to obtain the queue depth. The queue depth value counted in real time for each sub-channel is compared with its own exclusive preset congestion threshold data packet count value. When the number of data packets temporarily stored in the sub-channel queue exceeds the corresponding fixed threshold value, the sub-channel is determined to be a congested sub-channel.

[0056] The channel allocation weight values ​​corresponding to the dual channels are paired to form channel allocation weight pairs. These weight pairs are automatically updated according to a fixed time period. Once a congested sub-channel is detected, the weight value iteration process for all subsequent periods is terminated immediately, and no new allocation values ​​are generated. The most recent set of channel allocation weight pairs that has been verified and taken effect before the congestion determination time is accurately captured and archived at a fixed point. The archived weight allocation data is the optimal allocation weight snapshot.

[0057] Based on historical experience in handling massive network congestion traffic, a fixed weight share transfer standard is established. The weight share of congested sub-channels is reduced from the ratio of the optimal weight allocation snapshot. All the reduced weight shares are then fully allocated to non-congested sub-channels that are in normal operating condition. Throughout the process, the total weight of the two channels is kept at a fixed ratio. After the share transfer is completed, a brand-new pairing data of dual-channel weights is formed. This brand-new pairing data is the channel allocation weight pair after coupling adjustment.

[0058] The protocol stack send buffer corresponding to the congested sub-channel is divided into independent state storage partitions. The flow control acknowledgment message returned by the receiver during the communication interaction is used as the sole criterion for judgment. All data packets in the buffer that have received the flow control acknowledgment message but have not yet started the link sending process are filtered out. All the filtered and collected data packets are uniformly classified and collected to form a fixed data packet set, which is the data packet to be migrated.

[0059] The original protocol header structure adapted to the congested sub-channel transmission format is stripped from the outer layer of the data packet to be migrated. The application layer payload content carrying the original service data and the cooperative flow label used to mark the service home link are not changed throughout the process. A new protocol header structure adapted to the non-congested sub-channel communication specification is re-encapsulated on the outer layer of the data packet. The complete data packet with the protocol header replaced and the core data and identifiers inside are completely preserved is the transparent migration data packet.

[0060] Following the established first-come-first-served queuing rules of the non-congested sub-channel sending queue, transparent migration data packets are sequentially and orderly placed into the sending queue's waiting sequence position. The original record rows in the virtual channel mapping table that match the original channel identifier of this type of data packet are retrieved one by one. The corresponding protocol channel number and node port mapping association information in the record row are modified, and the mapping parameters of the currently assigned non-congested sub-channel are synchronously written to complete the full update of the corresponding channel mapping entry in the data table.

[0061] The beneficial effects include: accurately identifying congested sub-channels under dual protocols based on preset congestion thresholds; promptly freezing iterative updates of channel allocation weights and retaining snapshots of optimal allocation weights; achieving load balancing of dual channels through reasonable weight offset adjustments; accurately screening compliant data packets to be migrated in the buffer; completing lossless replacement of protocol headers while fully preserving application layer payloads and collaborative flow tags; achieving transparent migration of data streams across sub-channels; synchronously updating the virtual channel mapping table to ensure accurate and effective link mapping relationships; quickly resolving sub-channel congestion; avoiding data queuing and transmission interruptions; ensuring the continuity of business data stream transmission; and improving the overall operational stability and data transmission throughput of dual-protocol logical channels.

[0062] The encapsulation and delivery module 105 is used to perform adaptive frame encapsulation on the migrated data packets and transmit the encapsulated frames to the target intelligent agent node through the dual-protocol logic channel.

[0063] In this embodiment of the invention, when the encapsulation and delivery module performs adaptive frame encapsulation on the migrated data packets and transmits the encapsulated frames to the target intelligent agent node through the dual-protocol logic channel, it is specifically used for: The transparent migration data packet is processed using a unified encapsulation format to obtain an encapsulated data packet; Select the corresponding verification mechanism based on the protocol type of the non-congested sub-channel to which the transparent migration data packet currently belongs; When the non-congested sub-channel is an MCP channel, a lightweight check mechanism is used to generate the first check code. When the non-congested sub-channel is an A2A channel, a high-strength check mechanism is used to generate a second check code. The encapsulated data packet and the corresponding checksum are encapsulated into an adaptive frame, and the adaptive frame is sent to the target agent node through the non-congested sub-channel.

[0064] Transparent migration data packets are processed according to a predefined fixed encapsulation and layout specification. This specification uniformly defines a fixed frame header placeholder area, a data carrying area, and a reserved padding area. The transparent migration data packets are placed completely within the data carrying area, and fixed placeholder characters are filled in to fill in the empty fields according to the specification requirements, so that all processed data packets have a completely consistent external encapsulation structure. After being standardized and formatted, they form a regular and independent data unit, which is the encapsulated data packet.

[0065] A fixed binding relationship between protocol type and verification mechanism is established in advance in the kernel configuration table of the agent node. A unique verification execution rule is bound to the MCP protocol type and A2A protocol type respectively. The protocol category corresponding to the non-congested sub-channel to which the transparent migration data packet currently belongs is identified. The kernel configuration table is retrieved to find the verification execution rule that matches the protocol category, and the accurate matching and selection of the verification mechanism is completed.

[0066] The lightweight verification mechanism pre-defines the content traversal and feature verification only for the fixed preceding byte segment of the data packet, without traversing the entire payload byte content. After determining that the non-congested sub-channel belongs to the MCP channel, it strictly reads the encapsulated data packet content bit by bit according to the limited byte segment range and generates a unique verification identifier code. The verification identifier code generated according to the lightweight rules is the first check code.

[0067] The high-strength verification mechanism is pre-set to perform a complete traversal and verification of all bytes of the encapsulated data packet from beginning to end, covering all fields of the protocol header, application layer payload, and cooperative flow label. After determining that the non-congested sub-channel belongs to the A2A channel, the content is verified bit by bit according to the full-domain traversal rule and a unique verification identifier code is generated. The verification identifier code generated according to the high-strength rule is the second check code.

[0068] The adaptive frame is pre-defined with a fixed partitioning structure, dividing the data storage partition and the checksum storage partition. The encapsulated data packet is completely placed into the data storage partition, and the first or second checksum generated by matching is placed separately into the checksum storage partition. The two partitions are combined and encapsulated according to a fixed splicing order. The complete data frame structure formed by splicing is the adaptive frame. According to the predetermined message forwarding sequence and link transmission rules of the non-congested sub-channel, the formed adaptive frame is directly delivered to the corresponding target intelligent agent node along the current channel link.

[0069] The beneficial effects include: standardizing transparent migration data packets using a unified encapsulation format, which can standardize the data transmission structure of dual-protocol channels, avoid parsing anomalies caused by messy formats, match differentiated verification mechanisms according to different protocol types, reduce data verification processing time and hardware resource consumption by relying on a lightweight verification mechanism for the MCP channel, and realize full data integrity verification by relying on a high-strength verification mechanism for the A2A channel. The adaptive frame integrates business data and verification codes, which can adapt to the different latency and reliability requirements of the two types of channels. While taking into account the transmission processing efficiency, it effectively avoids packet loss and error problems in the data transmission process, and improves the regularity, real-time and reliability of cross-channel data delivery of dual-protocol intelligent agents.

[0070] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0071] The embodiments of this application can acquire and process relevant data based on an artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A high-efficiency dual-protocol intelligent agent data interaction system, characterized in that, The system includes a dual-stack deployment and channel mapping module, a cross-protocol weight allocation module, a collaborative awareness logic channel construction module, a cross-protocol flow control coupling and transparent migration module, and an encapsulation and delivery module, wherein: The dual-stack deployment and channel mapping module is used to deploy a dual-protocol communication component containing an MCP protocol stack and an A2A protocol stack on each agent node, and to establish a virtual channel mapping table. The cross-protocol weight allocation module is used to determine the channel allocation weight of the data packet between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators. The collaborative sensing logical channel construction module is used to construct a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, and jointly map cross-protocol channel data packets with high service coordination to the same logical topology path; The cross-protocol flow control coupling and transparent migration module is used to assign weights to the channel through cross-protocol flow control when the dual-protocol logical channel becomes congested, and to perform transparent protocol switching on the data streams on the corresponding logical topology path, so as to realize the active migration of data streams between the MCP channel and the A2A channel. The encapsulation and delivery module is used to adaptively encapsulate the migrated data packets and transmit the encapsulated frames to the target intelligent agent node through the dual-protocol logic channel.

2. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, The dual-stack deployment and channel mapping module, when deploying a dual-protocol communication component including an MCP protocol stack and an A2A protocol stack on each agent node and establishing a virtual channel mapping table, is specifically used for: In the operating system kernel of the intelligent agent node, the MCP protocol stack kernel component and the A2A protocol stack kernel component are loaded to obtain a dual protocol stack kernel instance; Based on the dual-protocol stack kernel instance, register the MCP virtual network interface and the A2A virtual network interface; Each of the MCP virtual network interface and the A2A virtual network interface is assigned an independent intra-node communication port number, and a mapping relationship is established between the intra-node communication port number and the remote intelligent agent node identifier to obtain a virtual channel mapping table.

3. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, Before determining the channel allocation weight between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators, the cross-protocol weight allocation module is specifically used for: Extract the service coordination degree identifier, end-to-end latency requirement level, and packet loss tolerance identifier of the data stream to be transmitted, and generate a normalized service feature vector; The available bandwidth and round-trip time jitter of the MCP channel are obtained through the MCP virtual network interface, and the available bandwidth and round-trip time jitter of the A2A channel are obtained through the A2A virtual network interface.

4. The efficient dual-protocol intelligent agent data interaction system as described in claim 3, characterized in that, The cross-protocol weight allocation module, when determining the channel allocation weight of data packets between the MCP channel and the A2A channel based on the feature parameter set of the data stream to be transmitted and the current network status indicators, is specifically used for: Based on the normalized service feature vector, the available bandwidth of the MCP channel, and the available bandwidth of the A2A channel, a basic allocation bias factor is determined, wherein the basic allocation bias factor is used to characterize the degree to which the data flow is initially biased towards the MCP channel allocation without considering the real-time congestion of the channel. The availability index of the MCP channel is determined based on the available bandwidth and round-trip time jitter of the MCP channel. The availability index of the MCP channel increases with the increase of the available bandwidth of the MCP channel and decreases with the increase of the round-trip time jitter of the MCP channel. The availability index of the A2A channel is determined based on the available bandwidth and round-trip time jitter of the A2A channel. The MCP channel allocation weight is calculated using the basic allocation tendency factor, the MCP channel availability index, and the A2A channel availability index. The value obtained by subtracting the MCP channel allocation weight is then used as the A2A channel allocation weight.

5. The efficient dual-protocol intelligent agent data interaction system as described in claim 4, characterized in that, The formula for calculating the weight allocation of the MCP channel is as follows: ； In the formula, Assign weights to the MCP channels. For the aforementioned basic allocation tendency factor, This is the preset amplification factor for channel state differences. This refers to the availability index of the MCP channel. This is the availability index of the A2A channel.

6. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, The collaborative sensing logical channel construction module, when executing the construction of a dual-protocol logical channel based on the virtual channel mapping table and real-time network topology information, is specifically used for: Obtain the link connectivity and link costs between each intelligent agent node to obtain real-time network topology information; Based on the virtual channel mapping table, from the real-time network topology information, node pairs that have both MCP and A2A reachable paths between the source agent node and the target agent node are selected to obtain a set of dual-protocol reachable node pairs. On the topology subgraph formed by the set of dual-protocol reachable node pairs, search for logical paths that satisfy the dual-protocol bandwidth constraints to construct a dual-protocol logical channel.

7. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, The collaborative sensing logical channel construction module, when executing the joint mapping of cross-protocol channel data packets with high service coordination to the same logical topology path, is specifically used for: Data packets belonging to the same business collaboration group in the data stream to be transmitted are regarded as a collaboration data packet cluster; The service coordination degree is obtained by adding the entropy of the destination address sequence of the MCP data sub-stream and the entropy of the destination address sequence of the A2A data sub-stream in the coordinated data packet cluster, and then subtracting the entropy of the joint address sequence of the two. When the business synergy exceeds a preset synergy threshold, the synergy data packet cluster is determined to have a high business synergy, and the MCP data sub-stream and A2A data sub-stream in the synergy data packet cluster with a high business synergy are jointly mapped to the same logical topology path.

8. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, When the cross-protocol flow control coupling and transparent migration module performs the channel weight allocation in the event of congestion in the dual-protocol logical channel, it is specifically used for: In the dual-protocol logical channel, the sub-channel whose queue depth exceeds the corresponding preset congestion threshold is designated as the congested sub-channel. When a congested sub-channel occurs, freeze the update of the current channel allocation weight pair and take the most recently determined channel allocation weight pair before the congestion as the snapshot of the optimal allocation weight. In the optimal allocation weight snapshot, the channel allocation weights of the congested sub-channels are shifted and adjusted towards the non-congested sub-channels to obtain the coupled adjusted channel allocation weight pairs.

9. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, The cross-protocol flow control coupling and transparent migration module, when performing protocol-transparent switching on data streams along the corresponding logical topology path to achieve active migration of data streams between the MCP channel and the A2A channel, is specifically used for: In the protocol stack send buffer of the congested sub-channel, data packets that have completed flow control confirmation but have not yet been sent are treated as data packets to be migrated; The current protocol header of the data packet to be migrated is replaced with the protocol header of the non-congested sub-channel, while the application layer payload and cooperative flow label are retained to obtain a transparent migration data packet; The transparent migration data packet is inserted into the transmission queue of the non-congested sub-channel, and the channel mapping entry of the transparent migration data packet in the virtual channel mapping table is updated.

10. The efficient dual-protocol intelligent agent data interaction system as described in claim 1, characterized in that, When the encapsulation and delivery module performs adaptive frame encapsulation on the migrated data packets and transmits the encapsulated frames to the target agent node through the dual-protocol logical channel, it is specifically used for: The transparent migration data packet is processed using a unified encapsulation format to obtain an encapsulated data packet; Select the corresponding verification mechanism based on the protocol type of the non-congested sub-channel to which the transparent migration data packet currently belongs; When the non-congested sub-channel is an MCP channel, a lightweight check mechanism is used to generate the first check code. When the non-congested sub-channel is an A2A channel, a high-strength check mechanism is used to generate a second check code. The encapsulated data packet and the corresponding checksum are encapsulated into an adaptive frame, and the adaptive frame is sent to the target agent node through the non-congested sub-channel.

Images