Multi-mode message processing method and device, equipment and storage medium

Abstract

The invention discloses a multi-mode message processing method and device, equipment and a storage medium. The method comprises the following steps: analyzing a multi-modal mixed message based on a preset entry static pipeline, generating a descriptor according to a message analysis result, and judging whether dynamic enhancement processing needs to be performed or not based on the descriptor; if yes, screening a target dynamic assembly line from the preset dynamic assembly lines, and performing dynamic enhancement processing on the multi-modal mixed message according to the target dynamic assembly line to obtain a processed message; and according to the descriptor, screening a target exit assembly line from preset exit static assembly lines to forward the processed message, wherein the preset exit static assembly line comprises an exit assembly line corresponding to at least one mode of the multi-mode network message to be forwarded. According to the method and the device, the mode identification is performed on the mixed message which does not pass through the network mode, and the static and dynamic dual-pipeline cooperative processing and forwarding are combined, so that the differential processing of the multi-mode message is realized.

Description

Technical Field

[0001] This application relates to the field of computing network technology, and in particular to a multimodal message processing method, apparatus, device and storage medium. Background Technology

[0002] With the deep integration of computing networks and multimodal networks, new computing network services (such as latency-sensitive edge computing, computing-intensive intelligent computing tasks, and general cloud computing services) place stringent demands on the network's high-throughput transmission, highly deterministic scheduling, and agile service adaptation capabilities. Traditional network data plane processing technologies rely on fixed IP routing protocols for packet forwarding, partially employing software-defined static pipeline processing mechanisms. They only support stateless forwarding of single-modality packets, depend on independent computing nodes for computing power, and are not deeply integrated with the network data plane. Different network modalities require different physical devices or completely isolated logical planes for support, resulting in complex infrastructure, high costs, and a lack of inter-modality collaboration. Consequently, they cannot identify and forward packets of multiple modalities, making it impossible to meet the differentiated needs of diverse services. Summary of the Invention

[0003] The main objective of this application is to provide a multimodal message processing method, apparatus, device, and storage medium, which aims to solve the technical problem that traditional message forwarding only supports stateless forwarding of single-modal messages, and cannot realize the identification, forwarding, and processing of multiple modal messages, thus failing to meet the differentiated needs of diverse services.

[0004] To achieve the above objectives, this application provides a multimodal message processing method, which includes the following steps: The multimodal mixed message is parsed based on a preset inlet static pipeline, and a descriptor is generated based on the message parsing result. The preset inlet static pipeline includes an inlet pipeline corresponding to at least one mode of the multimodal mixed message. Determine whether there is an enhancement requirement based on the descriptor; If so, then select the target dynamic pipeline from the preset dynamic pipeline, and perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message; The target exit pipeline is selected from the preset static exit pipelines according to the descriptor to forward the processed message. The preset static exit pipelines include the exit pipelines corresponding to at least one mode of the multimodal network message to be forwarded.

[0005] Optionally, the step of parsing multimodal mixed messages based on a preset inlet static pipeline and generating descriptors based on the message parsing results includes: The modal characteristics of multimodal mixed messages are analyzed based on the preset inlet static pipeline to determine the modal type; The service characteristics of the multimodal mixed message are analyzed based on the preset entry static pipeline to obtain the enhanced requirement type; A descriptor is generated based on the request priority, modality type, and enhanced requirement type corresponding to the multimodal mixed message.

[0006] Optionally, after the step of determining whether dynamic enhancement processing is needed based on the descriptor, the method further includes: If the enhancement requirement type in the descriptor is no enhancement requirement, then the target static pipeline is determined from the preset static pipeline according to the modality type; The multimodal mixed message is scheduled to the target static pipeline to execute the forwarding logic corresponding to the modality type.

[0007] Optionally, if so, the step of selecting a target dynamic pipeline from a preset dynamic pipeline and dynamically enhancing the multimodal mixed message according to the target dynamic pipeline to obtain the processed message includes: If the enhancement requirement type in the descriptor is "enhancement requirement", then the target dynamic pipeline is selected from the preset dynamic pipelines based on the modality type and the enhancement requirement type. The load status corresponding to the target dynamic pipeline is monitored to obtain load status information; Based on the load status information and the target dynamic pipeline, the multimodal mixed message is dynamically enhanced to obtain the processed message.

[0008] Optionally, the step of dynamically enhancing the multimodal mixed message based on the load status information and the target dynamic pipeline to obtain the processed message includes: If the load status information is in an uncongested state, then the multimodal mixed message is dynamically enhanced through the target dynamic pipeline to obtain the processed message. If the load status information indicates a congestion state, the multimodal mixed message is buffered according to the input queue depth corresponding to the target dynamic pipeline until the congestion state is resolved. If the load status is congestion-relieved, the cached multimodal mixed packets are sent to the target dynamic pipeline for dynamic enhancement processing to obtain the processed packets.

[0009] Optionally, the step of forwarding the processed message to a target exit pipeline from a preset static exit pipeline based on the descriptor includes: Target exit pipelines are selected from preset static exit pipelines based on the modality type in the descriptor; The port priority is determined based on the request priority corresponding to the multimodal mixed message; The processed packets are forwarded according to the preset hybrid forwarding table, the port priority, and the target egress pipeline.

[0010] Optionally, the step of forwarding the processed packets according to the preset hybrid forwarding table, the port priority, and the target egress pipeline includes: Obtain the queue load information corresponding to the target output pipeline; The processed packets are forwarded based on the queue load information, the preset hybrid forwarding table, the port priority, and the target egress pipeline.

[0011] Furthermore, to achieve the above objectives, this application also provides a multimodal message processing apparatus, which includes: The message parsing module is used to parse multimodal mixed messages based on a preset inlet static pipeline and generate descriptors based on the message parsing results. The preset inlet static pipeline includes an inlet pipeline corresponding to at least one mode of the multimodal mixed message. The message pipeline processing module is used to determine whether there is an enhancement requirement based on the descriptor; The dynamic enhancement processing module is used to select a target dynamic pipeline from the preset dynamic pipeline if the condition is met, and to perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message. The message forwarding module is used to select a target exit pipeline from the preset exit static pipeline according to the descriptor to forward the processed message. The preset exit static pipeline is a sequence of network mode messages to be forwarded.

[0012] In addition, to achieve the above objectives, this application also proposes a multimodal message processing apparatus, the apparatus comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the multimodal message processing method described above.

[0013] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the multimodal message processing method described above.

[0014] One or more technical solutions proposed in this application have at least the following technical effects: This application parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it determines whether there is an enhancement requirement. If so, it selects a target dynamic pipeline from a preset dynamic pipeline and performs dynamic enhancement processing on the multimodal mixed packets according to the target dynamic pipeline to obtain the processed packets. Based on the descriptors, it selects a target egress pipeline from a preset egress static pipeline to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded. Compared to traditional packet forwarding, which only supports stateless forwarding of single-mode packets and cannot identify and forward multiple-mode packets, thus failing to meet the differentiated needs of diverse services, this application achieves differentiated processing of multimodal packets by performing modal identification on mixed packets that do not pass through network modes and combining static and dynamic dual-pipeline collaborative processing and forwarding. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the first embodiment of the multimodal message processing method of this application. Figure 2 A schematic diagram of the multimodal network data plane processing logic provided in the first embodiment of the multimodal message processing method of this application; Figure 3 This is a flowchart illustrating the second embodiment of the multimodal message processing method of this application. Figure 4 This is a schematic diagram of multimodal network data plane processing provided in the second embodiment of the multimodal message processing method of this application; Figure 5 This is a schematic diagram of the module structure of the multimodal message processing device according to an embodiment of this application; Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the multimodal message processing method in the embodiments of this application.

[0018] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0020] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0021] The main solution of this application embodiment is as follows: This application parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the packet parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it is determined whether there is an enhancement requirement. If so, a target dynamic pipeline is selected from the preset dynamic pipelines, and the multimodal mixed packets are dynamically enhanced based on the target dynamic pipelines to obtain the processed packets. Based on the descriptors, a target egress pipeline is selected from the preset egress static pipelines to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded.

[0022] In this embodiment, for ease of description, the following description uses a computing service device as the execution subject.

[0023] Because traditional message forwarding only supports stateless forwarding of single-modal messages, it cannot identify and forward messages of multiple modalities, thus failing to meet the differentiated needs of diverse services.

[0024] This application provides a solution that achieves differentiated processing of multimodal packets by performing modality recognition on mixed packets that do not pass through network modalities, and combining static and dynamic dual pipeline collaborative processing and forwarding.

[0025] As can be seen from the above embodiments, this application parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the packet parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it determines whether there is an enhancement requirement. If so, it selects a target dynamic pipeline from the preset dynamic pipelines and performs dynamic enhancement processing on the multimodal mixed packets according to the target dynamic pipelines to obtain the processed packets. Based on the descriptors, it selects a target egress pipeline from the preset egress static pipelines to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded. Compared to traditional packet forwarding, which only supports stateless forwarding of single-mode packets and cannot achieve multimodal packet identification and forwarding processing, resulting in the inability to meet the differentiated needs of diverse services, this application achieves differentiated processing of multimodal packets by performing modal identification on mixed packets that do not pass through network modes and combining static and dynamic dual pipelines for collaborative processing and forwarding.

[0026] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, including a multimodal message processing system. The following description uses a computer as an example to illustrate this embodiment and the subsequent embodiments.

[0027] Based on this, embodiments of this application provide a multimodal message processing method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the multimodal message processing method of this application.

[0028] In this embodiment, the multimodal message processing method includes steps S10 to S30: Step S10: Parse the multimodal mixed message based on the preset entry static pipeline, and generate a descriptor based on the message parsing result.

[0029] It should be noted that the proposed solution involves static pipeline programming before implementation: defining the parsing rules and forwarding logic for multimodal packets in the static pipeline. Dynamic pipeline construction involves logically combining the acceleration units provided by various heterogeneous computing platforms (CPU, DPU, FPGA) in the enhanced capability resource pool according to business requirements, forming a customized processing pipeline for specific businesses. For the current embodiment, the system pre-constructs the following dynamic pipelines: IP deep packet inspection pipeline (built on DPU resources); NDN enhanced processing pipeline (built on CPU resources); and computing power service distributed scheduling decision pipeline (built on CPU+FPGA resource collaboration). Simultaneously, the construction of the dynamic pipeline supports runtime reprogramming, meaning that when business requirements change, the combination of heterogeneous resources can be dynamically adjusted online to rebuild a new processing pipeline. After the above configuration, the input processing, entry preprocessing and descriptor generation, pipeline allocation and differentiated processing, and post-processing and forwarding output in this application are executed. This application proposes a multimodal network data plane processing architecture and method that supports network-computer convergence enhanced computing. To further illustrate the architecture of this solution, please refer to... Figure 2 The diagram illustrates the multimodal network data plane processing logic of compute-network converged enhanced computing. Through input recognition, data pipelines corresponding to different modes, and dynamic pipeline execution of service enhancement computing, the processed packets are forwarded and scheduled via compute-network modality hybrid forwarding. This architecture follows the paradigm of "separation of technical system and physical platform" and consists of three main parts: modality recognition, dynamic packet pipeline processing, and modality hybrid forwarding and scheduling. Through software-defined modality recognition, "static + dynamic" dual pipeline processing, compute-network converged hybrid forwarding and scheduling, and a heterogeneous compatible enhanced computing architecture, differentiated processing of multimodal packets, joint scheduling of compute-network resources, and intra-network enhanced computing are achieved. This application endows the multimodal network data plane with compute-network converged processing capabilities, supporting high-throughput, low-latency, and highly reliable compute-network service delivery. Multimodal mixed packets can include mixed packets from different network modes, such as IP mode, NDN mode, and computing-network converged mode. The mixed packets from different network modes are input from the input port to the data plane and parsed. A software-defined parser performs deep parsing and identification of the packets to distinguish the specific mode to which the packets belong and their service characteristics.

[0030] Understandably, the preset ingress static pipeline must contain at least one ingress pipeline corresponding to at least one modality of multimodal mixed messages. The ingress pipeline refers to the sequence of ingress pipelines, which can be constructed based on multiple match-action stages of a programmable switching chip (such as a chip using the RMT architecture). It forms the hardware foundation for "input recognition" and "static pipeline" in the logical functions. Its main functions include: 1) message parsing, extracting key fields, and generating message descriptors upon which subsequent scheduling depends. The content of these descriptors can be defined through software programming and typically includes data packet length, modality type, service characteristics, and the key "enhanced processing type" field, thereby tagging each message with processing requirements at the hardware level; 2) completing standard, high-speed header processing for each modality.

[0031] It should be understood that this application parses multimodal mixed messages based on a preset entry static pipeline to obtain message parsing results. The message parsing results may include data packet length, request priority, modality type, business characteristics, and the key "enhanced processing type" field, and generate descriptors based on the message parsing results.

[0032] Furthermore, step S10 further includes: parsing the modal features of the multimodal mixed message based on a preset inlet static pipeline to determine the modal type; parsing the service features of the multimodal mixed message based on the preset inlet static pipeline to obtain the enhanced requirement type; and generating a descriptor according to the request priority corresponding to the multimodal mixed message, the modal type, and the enhanced requirement type.

[0033] It should be noted that by inputting mixed packets of multiple network modes into the data plane for modality recognition, the modality type is determined based on the extracted modality features, and the mixed modality packets are parsed according to a preset packet parser. Based on the extracted service features, it is determined whether enhancement processing is required, and the enhancement processing type is determined based on the service features. Finally, the enhancement requirement type is determined based on the enhancement processing type and the enhancement processing requirements.

[0034] Understandably, descriptors are generated based on the request priority, modality type, and enhanced requirement type corresponding to the multimodal mixed message.

[0035] In specific implementation, the following example illustrates the use of mixed-modal packets A (IP mode ordinary data packet), B (NDN data packet), C (IP mode security detection data packet), and D (computing-network converged mode task request data packet). These packets are input into the data plane and modality is identified to determine the modality types as IP mode, NDN mode, and computing-network converged mode, respectively. The software-defined packet parser then parses the mixed-modal packets to determine the corresponding modality type and enhancement requirement type (including enhancement processing requirements and enhancement processing types) for each modality. For example: A: marked as "IP mode" + "no enhancement requirement"; B: marked as "NDN mode" + "enhancement requirement" + "internal caching enhancement"; C: marked as "IP mode" + "enhancement requirement" + "deep detection enhancement"; D: marked as "computing-network converged mode" + "enhancement requirement" + "scheduling decision". The descriptor generation process for four types of packets (A, B, C, and D) is performed through a pre-defined static pipeline. Packet A: enters the IP modal pipeline for processing; Packet B: enters the NDN modal pipeline for preprocessing and generates the descriptor "NDN + Intra-network Cache Enhancement + Medium Priority"; Packet C: enters the IP modal pipeline for preprocessing and generates the descriptor "IP + Deep Detection Enhancement + Medium-High Priority"; Packet D: enters the computing-network convergence modal pipeline for preprocessing, parses the computing power request parameters, and generates the descriptor "Computing-Network Convergence + Scheduling Decision + High Priority".

[0036] Step S20: Determine whether there is an enhancement requirement based on the descriptor.

[0037] It should be noted that the presence of an enhancement requirement is determined based on the enhancement requirement type in the descriptor. If the enhancement requirement type indicates an enhancement requirement, then enhancement processing is required, and the packet is sent to the dynamic pipeline. If the enhancement requirement type indicates no enhancement requirement, then enhancement processing is not required, and packets that do not require enhancement processing are directly transmitted to the static egress pipeline for forwarding. Finally, both enhanced and unenhanced packets undergo mixed forwarding at the egress point.

[0038] Step S30: If yes, then select the target dynamic pipeline from the preset dynamic pipelines, and perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message.

[0039] It should be noted that the preset dynamic pipeline can be a pre-set sequence of enhanced capability pipelines, which is the core hardware carrier for realizing the logical function of "dynamically enhanced computing paths". It is a dynamically buildable and reconfigurable collection of pipelines, independent of fixed hardware logic. Through a programmable interface, it can chain together acceleration units provided by various heterogeneous computing platforms (CPU, DPU, FPGA) in the enhanced capability resource pool as needed to form customized processing pipelines for specific services. For example, for the computing-network converged mode, this pipeline sequence can instantiate a dedicated "distributed scheduling decision for computing power services" enhanced pipeline, and execute real-time scheduling decisions using the computing-network status information parsed and reported by the static pipeline.

[0040] Understandably, the dynamic enhanced computing path guides service-enhanced packets (such as service flows requiring complex calculations like intra-network aggregation, intelligent caching, and scheduling decisions) to the software-defined dynamic pipeline. This path doesn't correspond to a fixed network mode but is dynamically constructed based on the service-enhanced computing needs. It can utilize heterogeneous computing resources to perform enhanced processing on packets beyond traditional forwarding. For computing task scheduling requirements in the computing-network convergence mode, this dynamic pipeline can perform "control plane function decentralization" based on real-time computing power and network status perceived by the static pipeline. This means that rapid decision-making for computing-network convergence scheduling is completed in a distributed manner on the data plane, thereby achieving low-latency, highly self-consistent computing task scheduling.

[0041] It should be understood that the determination of whether additional calculation is needed is based on the enhancement requirement type contained in the descriptor. If so, the target dynamic pipeline is selected from the preset dynamic pipeline, and the multimodal mixed message is dynamically enhanced according to the target dynamic pipeline to obtain the processed message.

[0042] In its implementation, this application can read the descriptors generated by the ingress pipeline through a scheduler, and is responsible for efficient scheduling and load balancing between the static forwarding platform and the enhancement processing platform. The dynamic pipeline is determined through standard descriptors, and the packet orientation is determined based on the corresponding fields in the standard descriptors. Simultaneously, the load balancing mechanism of the co-caching mechanism completes the dynamic enhancement processing flow of the dynamic pipeline.

[0043] Step S40: Select the target exit pipeline from the preset exit static pipeline according to the descriptor and forward the processed message.

[0044] It should be noted that the forwarding of packets is accomplished by reading the descriptors generated by the ingress pipeline and using a preset static egress pipeline. The preset static egress pipeline must contain at least one egress pipeline corresponding to at least one mode of each network mode packet to be forwarded. The egress pipeline can be a sequence of pipelines responsible for post-processing and forwarding the packets of each mode to be sent out, specifically executing the "post-forwarding processing" step in the logical function, such as performing traffic statistics, encapsulation / decapsulation, etc.

[0045] Understandably, the target exit pipeline is selected from the preset static exit pipeline by descriptor to forward the processed packets. Both the enhanced packets and the packets that do not require enhancement are mixed and forwarded at the static exit.

[0046] This embodiment parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it determines whether there is an enhancement requirement. If so, a target dynamic pipeline is selected from the preset dynamic pipelines, and the multimodal mixed packets are dynamically enhanced according to the target dynamic pipeline to obtain the processed packets. Based on the descriptors, a target egress pipeline is selected from the preset egress static pipelines to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded. Compared to traditional packet forwarding, which only supports stateless forwarding of single-mode packets and cannot identify and forward multiple-mode packets, thus failing to meet the differentiated needs of diverse services, this embodiment achieves differentiated processing of multimodal packets by performing modal identification on mixed packets that do not pass through network modes and combining static and dynamic dual-pipeline collaborative processing and forwarding.

[0047] Based on the above Figure 1 The first embodiment shown illustrates a second embodiment of the multimodal message processing method of this application; refer to... Figure 3 , Figure 3 This is a flowchart illustrating a second embodiment of the multimodal message processing method of this application. Based on the first embodiment of this application, the same or similar content as the first embodiment described above can be referred to the above description and will not be repeated hereafter.

[0048] In this embodiment, after step S20, the method further includes: if the enhancement requirement type in the descriptor is no enhancement requirement, then determine the target egress static pipeline from the preset egress static pipeline according to the modality type; schedule the multimodal mixed message to the target egress static pipeline to execute the forwarding logic corresponding to the modality type.

[0049] It should be noted that the static pipeline is divided into an inlet static pipeline and an outlet static pipeline, which work together to complete the static pipeline's tasks. The inlet static pipeline performs modality recognition, and the recognized modal packets undergo standard, high-speed header processing for each modality in subsequent stages of the inlet static pipeline; the outlet static pipeline completes the aforementioned "packet forwarding logic".

[0050] Understandably, the preset static pipeline includes preset inlet static pipelines and preset outlet static pipelines, referring to the static data processing paths corresponding to the inlet and outlet. Based on the identified specific modal packets, the corresponding predefined modal inlet and outlet data pipelines can be determined. These pipelines can be implemented based on programmable switching chips. The inlet static pipeline is responsible for completing the standard, high-speed packet header processing for each modality, while the outlet static pipeline is responsible for packet forwarding logic. If the enhancement requirement type is "no enhancement requirement," the target outlet static pipeline is determined from the preset outlet static pipelines based on the modal type in the standard descriptor. Multimodal mixed packets are then scheduled to the target outlet static pipeline to execute the forwarding logic corresponding to the modal type.

[0051] Understandably, for general IP mode, a series of header processing steps are performed in the ingress static pipeline. Since there is no need for enhanced processing, it is directly scheduled to the egress static pipeline to perform the forwarding actions defined in the rule table. For computing-network converged mode, its ingress static pipeline supports key capabilities of computing-network converged mode, such as computing power awareness, through software-defined protocol parsing capabilities. It can parse the computing power demand identifier embedded in the computing power request message and the status information (such as CPU / GPU load) embedded in the computing power announcement message sent by the local or near-end computing power node in real time and send it to the dynamic pipeline, providing a decision basis for subsequent dynamic scheduling. After being processed by the ingress static pipeline, this type of mode message, due to its enhanced processing requirements, is sent to the dynamic pipeline. After completing the dynamic pipeline processing, it enters the egress static pipeline to perform forwarding actions.

[0052] In the specific implementation, the pipeline for subsequent processing is determined based on the incremental type field. If there is no need for enhanced processing, a static pipeline is selected to execute the forwarding logic, and then the process jumps to the exit point. If there is a need for enhanced processing, the corresponding dynamic pipeline is selected based on the enhancement type, and the congestion status of the dynamic pipeline is checked.

[0053] In this embodiment, step S30 further includes: Step S301: If the enhancement requirement type in the descriptor is "there is enhancement requirement", then select a target dynamic pipeline from the preset dynamic pipelines based on the modal type and the enhancement requirement type.

[0054] It should be noted that if the enhancement requirement type in the descriptor is "there is enhancement requirement", then the target dynamic pipeline is selected from the preset dynamic pipelines based on the modality type and the enhancement requirement type. The preset dynamic pipelines include enhancement computing pipelines corresponding to at least one enhancement requirement type. For example, the dynamic pipelines corresponding to messages B, C, and D are respectively the NDN enhancement processing pipeline, the NDN enhancement processing pipeline, and the computing power service distributed scheduling decision pipeline.

[0055] In the specific implementation, the pipeline determination based on descriptors is as follows: The scheduler parses the corresponding fields in the descriptor to determine the destination of the packet: A: Static pipeline, executes the actions defined in the rule table; B: Dynamic pipeline, executes the "NDN Enhanced Processing Pipeline"; C: Dynamic pipeline, executes the "NDN Enhanced Processing Pipeline"; D: Dynamic pipeline, executes the "Distributed Scheduling Decision Pipeline for Computing Power Services", including: reading the computing network status view and request information, running the scheduling algorithm, generating scheduling results and flow table entries, and feeding them back to the static pipeline.

[0056] Step S302: Monitor the load status corresponding to the target dynamic pipeline to obtain load status information.

[0057] It should be noted that, since the processing throughput of dynamic pipelines (based on heterogeneous computing resources) is generally lower than that of static pipelines (based on ASICs), and the processing time of different services varies significantly, a load balancing mechanism using a collaborative cache is required to execute congestion control logic to achieve load balancing. Load status information is determined by real-time monitoring of the input queue depth and congestion status of each dynamic pipeline. Congestion status includes non-congested state, congested state, and congestion-relieved state.

[0058] Step S303: Based on the load status information and the target dynamic pipeline, perform dynamic enhancement processing on the multimodal mixed message to obtain the processed message.

[0059] It should be noted that by monitoring the load status information, it is determined whether the packets on the target dynamic pipeline need to be cached, thereby avoiding congestion and packet loss due to excessive throughput, which would prevent efficient data processing.

[0060] Understandably, by monitoring load status information and coordinating with a pre-set cache load balancing mechanism, a target dynamic pipeline is scheduled to dynamically enhance multimodal mixed packets to obtain the processed packets.

[0061] Furthermore, step S303 further includes: if the load status information is in an uncongested state, then the multimodal mixed message is dynamically enhanced through the target dynamic pipeline to obtain the processed message; if the load status information is in a congested state, then the multimodal mixed message is cached according to the input queue depth corresponding to the target dynamic pipeline until the congestion is resolved; if the load status is in a congestion-resolved state, then the cached multimodal mixed message is sent to the target dynamic pipeline for dynamic enhancement processing to obtain the processed message.

[0062] It should be noted that packets are allocated to either a static pipeline or a dynamic pipeline (enhanced processing platform) based on standard descriptors. Dynamic pipelines may experience congestion. In such cases, load balancing is achieved through a collaborative caching mechanism. If congestion occurs, packets are temporarily stored in a cache unit, awaiting rescheduling once congestion is resolved. If there is no congestion, packets directly enter the corresponding dynamic pipeline (such as NDN enhanced processing, IP deep packet inspection, and distributed scheduling decisions for computing power services). For further illustration of the pipeline processing diagram in this application, please refer to... Figure 4 The diagram shown illustrates the multimodal network data plane processing of the computing-network fusion augmented computing, which includes an inlet pipeline sequence, an outlet pipeline sequence, and an augmented capability pipeline sequence (dynamic pipeline sequence).

[0063] Understandably, modal packets requiring no enhancement are sent directly to the egress pipeline; modal packets requiring enhancement are scheduled to the corresponding enhancement pipeline for processing via configurable switching logic; and load balancing is performed across heterogeneous platforms to avoid congestion. The configurable switching logic organizes and controls the flow of data packets among various core components based on the scheduler's scheduling decisions, achieving line-rate data exchange. Configurable data exchange logic includes loopback, bypass, cross-connect, and buffering. If the load status is uncongested, the multimodal mixed packets are dynamically enhanced through the target dynamic pipeline to obtain the processed packets. If the load status is congested, the multimodal mixed packets are buffered according to the input queue depth of the target dynamic pipeline until the congestion is resolved. If the load status is decongested, the buffered multimodal mixed packets are sent to the target dynamic pipeline for dynamic enhancement processing to obtain the processed packets. The system monitors the input queue depth of each dynamic processing pipeline in real time. When a target dynamic processing pipeline is found to be congested, the system stores the messages to be scheduled to that pipeline and their descriptors in the cache unit and records the queue position. The system continuously monitors the status of the target dynamic processing pipeline. Once the congestion is relieved, the system retrieves the messages and their descriptors from the cache unit and submits them for processing.

[0064] In practice, the input queue depth and congestion status of each dynamic pipeline are monitored in real time. Congestion handling (taking packet C as an example): When the scheduler prepares to send packet C to the "IP Depth Packet Inspection Pipeline," if it detects that the pipeline is currently congested (e.g., the queue is full or the load is too high), the scheduler does not drop the packet directly. Instead, it instructs the buffer unit to temporarily store packet C and record its position in the queue. After detecting that the congestion in the dynamic pipeline has been relieved (resources are released), the scheduler retrieves packet C from the buffer unit and sends it to the pipeline for processing. If the target pipeline is not congested, it is directly sent via configurable switching logic.

[0065] This embodiment parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the packet parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it determines whether there is an enhancement requirement. If the enhancement requirement type in the descriptor is "enhancement requirement exists," a target dynamic pipeline is selected from the preset dynamic pipelines based on the mode type and the enhancement requirement type. The load status corresponding to the target dynamic pipeline is monitored to obtain load status information. Based on the load status information and the target dynamic pipeline, the multimodal mixed packets are dynamically enhanced to obtain processed packets. A target egress pipeline is selected from the preset egress static pipelines based on the descriptors to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded. Compared to traditional packet forwarding, which only supports stateless forwarding of single-modality packets and cannot identify and forward multiple modal packets, thus failing to meet the differentiated needs of diverse services, this embodiment achieves differentiated processing of multimodal packets by performing modal identification on mixed packets that do not pass through network modalities and combining static and dynamic dual pipeline collaborative processing and forwarding.

[0066] Based on the above Figure 1 The first embodiment shown presents a third embodiment of the multimodal message processing method of this application. Based on the first embodiment of this application, the same or similar content as the first embodiment described above can be referred to the above description and will not be repeated hereafter.

[0067] In this embodiment, step S40 further includes: selecting a target egress pipeline from a preset static egress pipeline based on the modality type in the descriptor; determining a port priority based on the request priority corresponding to the multimodal mixed message; and forwarding the processed message according to the preset mixed forwarding table, the port priority, and the target egress pipeline.

[0068] It should be noted that after all packets are processed, they enter the egress processing stage: each packet enters the corresponding egress pipeline according to its modality, queries the hybrid forwarding table (for packets in the computing-network converged mode, the forwarding table has been updated according to computing power routing), and forwards them according to port priority. This application, through the above-mentioned static and dynamic pipelines, converges all packets into a unified post-processing stage. First, the packets enter the computing-network hybrid forwarding module. This module queries the matching conditions of the packets according to the unified multimodal hybrid forwarding table (generated by the control plane combining computing power routing and multimodal network routing) and executes corresponding actions (such as specifying port forwarding, adding / removing routing headers, etc.). Subsequently, the packets enter the computing-network hybrid scheduling module. The scheduler performs queue management and scheduling decisions based on the packet's service priority, latency requirements, and other quality of service parameters to ensure the performance of critical service flows.

[0069] Understandably, a pre-defined hybrid forwarding table can be a unified set of forwarding entries generated by jointly calculating multimodal network routing information and computing power routing information. It integrates routing and scheduling information from different sources and dimensions, enabling the data plane to simultaneously consider network reachability, computing power resource status, and business policies when forwarding a packet, thereby making a globally optimal forwarding decision. The forwarding table can contain matching items from multiple modalities, such as IP address prefixes, NDN name prefixes, computing power task identifiers, and geographical coordinates.

[0070] Furthermore, the step of forwarding the processed packets according to the preset hybrid forwarding table, the port priority, and the target egress pipeline includes: obtaining the queue load information corresponding to the target egress pipeline; and forwarding the processed packets based on the queue load information, the preset hybrid forwarding table, the port priority, and the target egress pipeline.

[0071] In specific implementation, different forwarding ports are allocated by combining queue load balancing, such as: A: Entering the IP mode egress pipeline to perform post-processing, querying the hybrid forwarding table and forwarding according to port priority; B: Entering the NDN mode egress pipeline to perform post-processing, querying the hybrid forwarding table and forwarding according to port priority; C: Entering the IP mode egress pipeline to perform post-processing, querying the hybrid forwarding table and forwarding according to port priority; D: Entering the computing-network converged mode egress pipeline to perform post-processing, querying the hybrid forwarding table (updated according to the latest computing power routing calculation results) and forwarding according to port priority.

[0072] To further illustrate, a hybrid pipeline processing example of static / dynamic collaboration (taking the computing-network convergence mode in the above working example as an example) is described in the following scenario: For message D (computing-network convergence mode - computing power request message), after parsing, identification, and scheduling by the static pipeline and scheduler, it is sent to the dynamic pipeline of "distributed scheduling decision for computing power services" for in-depth processing. 1. Information Acquisition (Static / Dynamic Pipeline Collaboration): Request Feature Extraction: The computing-network convergence mode processing unit in the static pipeline has completed the parsing of message D, extracting the computing power demand vector (including but not limited to the required computing power resource specifications, QoS latency / bandwidth requirements, etc.), and encapsulating the above information in a packet descriptor and transmitting it with the message. Computing Network View Maintenance: Prior to this, the static pipeline has continuously parsed periodic "computing power announcement messages" and "network measurement messages" to report and update the latest network and computing power status to the local computing network status view of the dynamic pipeline of "distributed scheduling decision for computing power services". 2. Decision Execution (Dynamic Pipeline): After message D and its descriptor are scheduled into the "Distributed Scheduling Decision for Computing Power Services" dynamic pipeline, the following steps are executed: Information Reading: The pipeline first reads the real-time topology and resource information in the local computing network status view, and reads the computing power request information from the message descriptor; Algorithm Execution: Based on the above information, a preset scheduling algorithm (e.g., LeastLoadFirst, LLF) is executed; Result Generation and Feedback: A response message is generated based on the scheduling decision result, the corresponding forwarding logic and flow table entries are constructed, and they are loaded back into the static pipeline to guide the forwarding behavior of subsequent messages.

[0073] This embodiment parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the packet parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. It then determines whether there is an enhancement requirement based on the descriptors. If so, it selects a target dynamic pipeline from the preset dynamic pipelines and performs dynamic enhancement processing on the multimodal mixed packets according to the target dynamic pipeline to obtain the processed packets. It then selects a target egress pipeline from the preset egress static pipelines based on the mode type in the descriptors. It determines the port priority based on the request priority corresponding to the multimodal mixed packets. Finally, it forwards the processed packets according to the preset mixed forwarding table, port priority, and target egress pipeline. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded. Compared to traditional packet forwarding, which only supports stateless forwarding of single-modality packets and cannot identify and forward multiple modal packets, thus failing to meet the differentiated needs of diverse services, this embodiment achieves differentiated processing of multimodal packets by performing modal identification on mixed packets that do not pass through network modalities and combining static and dynamic dual pipeline collaborative processing and forwarding.

[0074] This application also provides a multimodal message processing apparatus; please refer to... Figure 5 The multimodal message processing device includes: The message parsing and preprocessing module 10 is used to parse multimodal mixed messages based on a preset inlet static pipeline and generate descriptors according to the message parsing results. The preset inlet static pipeline includes an inlet pipeline corresponding to at least one mode of the multimodal mixed message. Message pipeline processing module 20 is used to determine whether there is an enhancement requirement based on the descriptor; The dynamic enhancement processing module 30 is used to, if so, select a target dynamic pipeline from the preset dynamic pipelines, and perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message. The message forwarding module 40 is used to select a target exit pipeline from the preset exit static pipeline according to the descriptor to forward the processed message. The preset exit static pipeline is a sequence of network mode messages to be forwarded.

[0075] Furthermore, the message parsing module 10 is also used to parse the modal features of the multimodal mixed message based on the preset inlet static pipeline to determine the modal type; parse the service features of the multimodal mixed message based on the preset inlet static pipeline to obtain the enhanced requirement type; and generate a descriptor according to the request priority, the modal type, and the enhanced requirement type corresponding to the multimodal mixed message.

[0076] Furthermore, the dynamic enhancement processing module 30 is also used to determine the target egress static pipeline from the preset egress static pipeline according to the modality type if the enhancement requirement type in the descriptor is no enhancement requirement; and to schedule the multimodal mixed message to the target egress static pipeline to execute the forwarding logic corresponding to the modality type.

[0077] Furthermore, the dynamic enhancement processing module 30 is also configured to, if the enhancement requirement type in the descriptor is enhancement requirement, select a target dynamic pipeline from the preset dynamic pipelines based on the modality type and the enhancement requirement type; monitor the load status corresponding to the target dynamic pipeline to obtain load status information; and perform dynamic enhancement processing on the multimodal mixed message based on the load status information and the target dynamic pipeline to obtain the processed message.

[0078] Furthermore, the dynamic enhancement processing module 30 is also configured to: if the load status information is in an uncongested state, perform dynamic enhancement processing on the multimodal mixed message through the target dynamic pipeline to obtain the processed message; if the load status information is in a congested state, cache the multimodal mixed message according to the input queue depth corresponding to the target dynamic pipeline until the congestion is resolved; if the load status is in a congestion-resolved state, send the cached multimodal mixed message to the target dynamic pipeline for dynamic enhancement processing to obtain the processed message.

[0079] Furthermore, the message forwarding module 40 is also used to filter target exit pipelines from preset static exit pipelines according to the modality type in the descriptor; determine port priority according to the request priority corresponding to the multimodal mixed message; and forward the processed message according to the preset mixed forwarding table, the port priority, and the target exit pipeline.

[0080] Furthermore, the message forwarding module 40 is also used to obtain the queue load information corresponding to the target egress pipeline; and to forward the processed message based on the queue load information, the preset hybrid forwarding table, the port priority, and the target egress pipeline.

[0081] The multimodal packet processing apparatus provided in this application, employing the multimodal packet processing method described in the above embodiments, can solve the technical problem that traditional packet forwarding only supports stateless forwarding of single-modal packets, failing to identify and forward multiple modal packets, thus hindering the fulfillment of differentiated requirements for diverse services. Compared with the prior art, the beneficial effects of the multimodal packet processing apparatus provided in this application are the same as those of the multimodal packet processing method provided in the above embodiments, and other technical features in the multimodal packet processing apparatus are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0082] This application provides a multimodal message processing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the multimodal message processing method described in the first embodiment above.

[0083] The following is for reference. Figure 6This document illustrates a structural schematic diagram of a multimodal message processing device suitable for implementing embodiments of this application. The multimodal message processing device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The multimodal message processing device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0084] like Figure 6 As shown, the multimodal message processing device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the multimodal message processing device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the multimodal message processing device to communicate wirelessly or wiredly with other devices to exchange data. Although a multimodal message processing device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented alternatively.

[0085] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0086] The multimodal packet processing device provided in this application, employing the multimodal packet processing method described in the above embodiments, solves the technical problem that traditional packet forwarding only supports stateless forwarding of single-modal packets, failing to identify and process multiple modal packets, thus hindering the fulfillment of differentiated requirements for diverse services. Compared with the prior art, the beneficial effects of the multimodal packet processing device provided in this application are the same as those of the multimodal packet processing method described in the above embodiments, and other technical features of this multimodal packet processing device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0087] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0088] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0089] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the multimodal message processing method described in the above embodiments.

[0090] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0091] The aforementioned computer-readable storage medium may be included in a multimodal message processing device; or it may exist independently and not assembled into a multimodal message processing device.

[0092] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the multimodal packet processing device, the multimodal packet processing device performs the following actions: It parses multimodal mixed packets based on a preset ingress static pipeline and generates descriptors based on the parsing results. The preset ingress static pipeline includes ingress pipelines corresponding to at least one mode of the multimodal mixed packets. Based on the descriptors, it determines whether there is an enhancement requirement. If so, it selects a target dynamic pipeline from a preset dynamic pipeline and performs dynamic enhancement processing on the multimodal mixed packets based on the target dynamic pipeline to obtain processed packets. Based on the descriptors, it selects a target egress pipeline from a preset egress static pipeline to forward the processed packets. The preset egress static pipeline includes egress pipelines corresponding to at least one mode of the multimodal network packets to be forwarded.

[0093] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0094] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0095] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0096] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described multimodal message processing method. This solves the technical problem that traditional message forwarding only supports stateless forwarding of single-modal messages, failing to recognize and forward multiple modal messages, thus hindering the fulfillment of diverse service requirements. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the multimodal message processing method provided in the above embodiments, and will not be repeated here.

[0097] The above are only some embodiments of this application and do not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A multimodal message processing method, characterized in that, The method includes: The multimodal mixed message is parsed based on a preset inlet static pipeline, and a descriptor is generated based on the message parsing result. The preset inlet static pipeline includes an inlet pipeline corresponding to at least one mode of the multimodal mixed message. Determine whether there is an enhancement requirement based on the descriptor; If so, then select the target dynamic pipeline from the preset dynamic pipeline, and perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message; The target exit pipeline is selected from the preset static exit pipelines according to the descriptor to forward the processed message. The preset static exit pipelines include the exit pipelines corresponding to at least one mode of the multimodal network message to be forwarded.

2. The multimodal message processing method as described in claim 1, characterized in that, The step of parsing multimodal mixed messages based on a preset inlet static pipeline and generating descriptors based on the message parsing results includes: The modal characteristics of multimodal mixed messages are analyzed based on the preset inlet static pipeline to determine the modal type; The service characteristics of the multimodal mixed message are analyzed based on the preset entry static pipeline to obtain the enhanced requirement type; A descriptor is generated based on the request priority, modality type, and enhanced requirement type corresponding to the multimodal mixed message.

3. The multimodal message processing method as described in claim 2, characterized in that, After the step of determining whether dynamic enhancement processing is needed based on the descriptor, the method further includes: If the enhancement requirement type in the descriptor is no enhancement requirement, then the target exit static pipeline is determined from the preset exit static pipeline according to the modal type. The multimodal mixed message is scheduled to the target egress static pipeline to execute the forwarding logic corresponding to the modality type.

4. The multimodal message processing method according to any one of claims 1 to 3, characterized in that, If so, the step of selecting a target dynamic pipeline from the preset dynamic pipelines and performing dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message includes: If the enhancement requirement type in the descriptor is "enhancement requirement", then the target dynamic pipeline is selected from the preset dynamic pipelines based on the modality type and the enhancement requirement type. The load status corresponding to the target dynamic pipeline is monitored to obtain load status information; Based on the load status information and the target dynamic pipeline, the multimodal mixed message is dynamically enhanced to obtain the processed message.

5. The multimodal message processing method as described in claim 4, characterized in that, The step of dynamically enhancing the multimodal mixed message based on the load status information and the target dynamic pipeline to obtain the processed message includes: If the load status information is in an uncongested state, then the multimodal mixed message is dynamically enhanced through the target dynamic pipeline to obtain the processed message. If the load status information indicates a congestion state, the multimodal mixed message is buffered according to the input queue depth corresponding to the target dynamic pipeline until the congestion state is resolved. If the load status is congestion-relieved, the cached multimodal mixed packets are sent to the target dynamic pipeline for dynamic enhancement processing to obtain the processed packets.

6. The multimodal message processing method as described in claim 5, characterized in that, The step of forwarding the processed message by selecting a target exit pipeline from a preset static exit pipeline based on the descriptor includes: Target exit pipelines are selected from preset static exit pipelines based on the modality type in the descriptor; The port priority is determined based on the request priority corresponding to the multimodal mixed message; The processed packets are forwarded according to the preset hybrid forwarding table, the port priority, and the target egress pipeline.

7. The multimodal message processing method as described in claim 6, characterized in that, The step of forwarding the processed packets according to the preset hybrid forwarding table, the port priority, and the target egress pipeline includes: Obtain the queue load information corresponding to the target output pipeline; The processed packets are forwarded based on the queue load information, the preset hybrid forwarding table, the port priority, and the target egress pipeline.

8. A multimodal message processing apparatus, characterized in that, The device includes: The message parsing and preprocessing module is used to parse multimodal mixed messages based on a preset inlet static pipeline and generate descriptors based on the message parsing results. The preset inlet static pipeline includes an inlet pipeline corresponding to at least one mode of the multimodal mixed message. The message pipeline processing module is used to determine whether there is an enhancement requirement based on the descriptor; The dynamic enhancement processing module is used to select a target dynamic pipeline from the preset dynamic pipeline if the condition is met, and to perform dynamic enhancement processing on the multimodal mixed message according to the target dynamic pipeline to obtain the processed message. The message forwarding module is used to select a target exit pipeline from the preset exit static pipeline according to the descriptor to forward the processed message. The preset exit static pipeline is a sequence of network mode messages to be forwarded.

9. A multimodal message processing device, characterized in that, The multimodal message processing device includes: a memory, a processor, and a multimodal message processing program stored in the memory and executable on the processor, wherein the multimodal message processing program is configured to implement the multimodal message processing method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a multimodal message processing program, which, when executed by a processor, implements the steps of the multimodal message processing method as described in any one of claims 1 to 7.

Images