Multi-modal network architecture, multi-modal data transmission method, storage medium and electronic device

Abstract

The invention discloses a multi-modal network architecture, a multi-modal data transmission method, a storage medium and an electronic device, and relates to the technical field of network communication, the architecture comprises a modal deployment management and control system, a multi-modal controller, a multi-modal network element and a cloud platform, the modal deployment management and control system is connected with the multi-modal controller and the multi-modal network element, and the cloud platform is connected with the multi-modal controller and the multi-modal network element. The multi-modal controller and the multi-modal network element are deployed on the cloud platform; the modal deployment management and control system is used for performing centralized control and modal issuing on the cloud platform which is in network intercommunication with the modal deployment management and control system; the multi-modal controller is used for managing a local multi-modal network element in the cloud platform; the multi-modal network element is used for forwarding the multi-modal message on the data level in the multi-modal network architecture, and by adopting the technical scheme, the technical problem that a traditional cloud service network protocol is single and cannot flexibly adapt to diversified network service requirements is solved.

Description

Technical Field

[0001] This application relates to the field of network communication technology, and more specifically, to a multimodal network architecture, a multimodal data transmission method, a storage medium, and an electronic device. Background Technology

[0002] With the deep integration of information networks and the production sector, personalized and specialized new network services are emerging in an endless stream. Multimodal network environment theory, as a new paradigm for network development, has opened up a new way to resolve the contradiction between diverse network applications and services and the construction / operation of unified infrastructure. It allows users to dynamically switch between different communication protocols, routing mechanisms, or security policies according to their needs, fundamentally changing the "services adapt to the network" to "the network adapts to services on demand." However, current cloud service architectures and cloud intranets are still generally based on IP network architectures, often only able to conduct services oriented towards IP protocols. When application needs arise based on multimodal (non-IP protocol suites), the cloud intranet cannot provide the appropriate network protocols, resulting in a disconnect between cloud and network service carrying capacity, limiting the deployment and use of new network services in the cloud. Furthermore, high-speed networks in data centers only solve the problems of high bandwidth and low latency under specific conditions, failing to achieve diversified network modes and service carrying capacity.

[0003] There is still no effective solution to the technical problem that traditional cloud service network protocols are singular and cannot flexibly adapt to the diverse network service needs. Summary of the Invention

[0004] This application provides a multimodal network architecture, a multimodal data transmission method, a storage medium, and an electronic device to at least solve the technical problem in the related art that traditional cloud service network protocols are singular and cannot flexibly adapt to diverse network service needs.

[0005] According to one embodiment of this application, a multimodal network architecture is provided, including: a modal deployment and control system, a multimodal controller, multimodal network elements, and a cloud platform. The modal deployment and control system is connected to the multimodal controller and the multimodal network elements, respectively. The multimodal controller and the multimodal network elements are deployed on the cloud platform. The modal deployment and control system is used for centralized control and modal distribution of the cloud platform, which has network interconnection with the modal deployment and control system. The multimodal controller is used to manage the local multimodal network elements in the cloud platform. The multimodal network elements are used to forward multimodal packets at the data layer in the multimodal network architecture.

[0006] In an exemplary embodiment, the modal deployment control system includes: a modal deployment subsystem, a deployment control subsystem, and a data management subsystem; wherein, the modal deployment subsystem is used to convert the operational intent of the target object received by the modal deployment control system into a target program, and send the target program to the data management subsystem; the deployment control subsystem is used to collect full-domain information from the cloud platform connected to the modal deployment control system, and generate a configuration policy corresponding to the full-domain information, wherein the configuration policy is used to adjust the data transmission parameters on different cloud platforms, and the data transmission parameters include at least one of the following: network parameters, status parameters, and traffic parameters; the data management subsystem is used to store the target program generated by the modal deployment subsystem and adjust the timing of the target program's distribution.

[0007] In an exemplary embodiment, the modal deployment subsystem comprises a modal environment module and an intent translation module; wherein the modal environment module is used to build a development environment for receiving operational intents and to generate code text; the intent translation module is connected to the modal environment module and is used to receive the code text output by the modal environment module and compile the code text into a target program that allows multimodal network elements to run.

[0008] In an exemplary embodiment, the multimodal controller includes a modal deployment control system interface, a flow table management subsystem, a packet decision subsystem, a modal switching control subsystem, a topology management subsystem, a traffic scheduling subsystem, a load balancing subsystem, and a local resource management subsystem. The modal deployment control system interface receives management commands and configuration policies issued by the modal deployment control system, wherein the management commands are determined by a target program. The flow table management subsystem analyzes and extracts global flow table information corresponding to the cloud platform to obtain a target flow table. The traffic scheduling subsystem, connected to the flow table management subsystem, generates first configuration information for adjusting the traffic of multimodal network element ports or specific modes on the cloud platform based on the target flow table, provided the target flow table is obtained. The packet decision subsystem determines the type of special packets obtained by the multimodal network element ports, wherein the characteristic packets include at least one of the following: unknown packets, abnormal packets, and special packets are part or all of the multimodal packets to be transmitted. The system comprises three subsystems: a message dispatch control subsystem (connected to the message decision subsystem), a topology management subsystem (used to mark the target mode corresponding to a characteristic message as a mode that is not allowed to forward data when the message type is discarded, and a message type that is allowed to forward data when the message type is added), and a topology management subsystem (used to obtain traffic carrying information of other cloud platforms connected to the topology management subsystem, and adjust the number of links used by the cloud platform for data transmission based on the traffic carrying information), a load balancing subsystem (connected to the topology management subsystem, used to generate third configuration information to adjust the data transmission load on adjacent links when the number of links is fixed), and a local resource management subsystem (connected to the traffic scheduling subsystem, the topology dispatch control subsystem, and the load balancing subsystem, used to store the first configuration information, the second configuration information, and the third configuration information, and to transmit the configuration information to the multi-modal network element port for distribution to the multi-modal network element).

[0009] In an exemplary embodiment, the multimodal network element includes: a multimodal controller interaction entry, a line-speed forwarding unit, and a modal network element application; the multimodal controller interaction entry is used to receive configuration information issued by the multimodal controller and synchronize the configuration information to the line-speed forwarding unit and the modal network element application for processing; the line-speed forwarding unit is used to forward the received multimodal packets to be transmitted according to the configuration information; the modal network element application, connected to the line-speed forwarding unit, is used to determine the processing method and forwarding port of the target multimodal packets whose modality is unknown in the multimodal packets to be transmitted.

[0010] According to another embodiment of this application, a multimodal data transmission method is provided, applied to any of the above-mentioned multimodal network architectures, comprising: when it is determined that a first terminal needs to perform multimodal data transmission with a second terminal, determining a first cloud platform connected to the first terminal and a second cloud platform connected to the second terminal, and establishing a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform; when using the communication tunnels for multimodal data transmission, converting the multimodal data to be transmitted by the first terminal into multimodal packets; uploading the multimodal packets to the first cloud platform and forwarding them to the second cloud platform through the first cloud platform; and the second cloud platform sending the received forwarded packets to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

[0011] In an exemplary embodiment, converting multimodal data to be transmitted by a first terminal into a multimodal message includes: determining a message template supported by the communication tunnel, wherein the message template includes at least one of the following: an outer Ethernet header byte, an outer IP header byte, a tunnel header byte, and an inner Ethernet frame byte; filling the inner Ethernet frame byte in the message template according to the first MAC address corresponding to the service server in the first cloud platform, the second MAC address corresponding to the first terminal, and the multimodal data; filling the outer IP header byte in the message template according to the first IP address corresponding to the multimodal network element in the first cloud platform and the second IP address corresponding to the first terminal; filling the outer Ethernet header byte in the message template according to the second MAC address corresponding to the first terminal and the third MAC address corresponding to the second terminal; filling the tunnel header byte in the message template according to the mode type corresponding to the multimodal data; and obtaining a multimodal message corresponding to the multimodal data after filling is completed.

[0012] In an exemplary embodiment, after uploading the multimodal message to the first cloud platform, the method further includes: instructing the first cloud platform to parse the data packet corresponding to the multimodal message, so as to determine the modality type and destination address field when the multimodal message is forwarded from the first cloud platform to the second cloud platform based on the parsing result; and, if the parsing is completed, requesting the first cloud platform to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and destination address field.

[0013] In an exemplary embodiment, after requesting the first cloud platform to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and destination address fields, the method further includes: determining that the modality deployment control system sends configuration information required for communication with the second cloud platform to the first multimodal controller within the first cloud platform; determining a tunnel address matching table based on the configuration information, wherein each entry in the tunnel address matching table includes a matching field, an outer IP address and tunnel number field, and a priority field, and the matching field consists of a modality type byte and a destination identifier byte; and controlling the multimodal data transmission between the first cloud platform and the second cloud platform based on the tunnel address matching table.

[0014] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described multimodal data transmission method at runtime.

[0015] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described multimodal data transmission method through the computer program.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the above-described multimodal data transmission method.

[0017] In this embodiment, the provided multimodal network architecture includes: a modal deployment and control system, a multimodal controller, multimodal network elements, and a cloud platform. The modal deployment and control system is connected to the multimodal controller and the multimodal network elements, which are deployed on the cloud platform. The modal deployment and control system is used for centralized control and modal distribution of the cloud platform, which has network interconnection with the modal deployment and control system. The multimodal controller is used to manage the local multimodal network elements in the cloud platform. The multimodal network elements are used for forwarding multimodal packets at the data layer within the multimodal network architecture. This technical solution solves the technical problem of traditional cloud service networks having a single network protocol and being unable to flexibly adapt to diverse network service needs. Furthermore, through centralized control by the modal deployment and control system, combined with the multimodal controller and multimodal network elements deployed on the cloud platform, efficient management of multimodal packets and cross-cloud platform data layer forwarding are achieved. Attached Figure Description

[0018] 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.

[0019] 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.

[0020] Figure 1 This is a structural block diagram of a multimodal network architecture according to an embodiment of this application;

[0021] Figure 2 This is a schematic diagram of a cloud-based multimodal network environment structure according to an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of a modal deployment control system device according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of a multimodal controller device according to an embodiment of this application;

[0024] Figure 5 This is a schematic diagram of a multimodal network element device structure according to an embodiment of this application;

[0025] Figure 6 This is a schematic diagram of an end-to-end multimodal data transmission process in the cloud according to an embodiment of this application;

[0026] Figure 7 This is a schematic diagram of a multimodal message carried by a VXLAN according to an embodiment of this application;

[0027] Figure 8 This is a schematic diagram of a VXLAN address matching table structure according to an embodiment of this application;

[0028] Figure 9 This is a hardware structure block diagram of a server device for a multimodal data transmission method according to an embodiment of this application;

[0029] Figure 10 This is a flowchart of a multimodal data transmission method according to an embodiment of this application. Detailed Implementation

[0030] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] According to one aspect of the embodiments of this application, a multimodal network architecture is provided. Figure 1 This is a structural block diagram of a multimodal network architecture according to an embodiment of this application, such as... Figure 1 As shown, it includes:

[0033] Modal deployment control system 12 is used for centralized control and modal distribution of cloud platforms that have network interconnection with the modal deployment control system;

[0034] The multimodal controller 14 is used to manage local multimodal network elements in the cloud platform.

[0035] Multimodal network element 16 is used for forwarding multimodal packets at the data layer in a multimodal network architecture.

[0036] In an exemplary embodiment, the modal deployment control system 12 includes: a modal deployment subsystem, a deployment control subsystem, and a data management subsystem; wherein, the modal deployment subsystem is used to convert the operational intent of the target object received by the modal deployment control system into a target program, and send the target program to the data management subsystem; the deployment control subsystem is used to collect full-domain information from the cloud platform connected to the modal deployment control system, and generate a configuration policy corresponding to the full-domain information, wherein the configuration policy is used to adjust the data transmission parameters on different cloud platforms, and the data transmission parameters include at least one of the following: network parameters, status parameters, and traffic parameters; the data management subsystem is used to store the target program generated by the modal deployment subsystem and adjust the timing of the target program's distribution.

[0037] In an exemplary embodiment, the modal deployment control system 12 further includes: a modal environment module and an intent translation module; wherein, the modal environment module is used to build a development environment for receiving operational intents and generate code text; the intent translation module is connected to the modal environment module and is used to receive the code text output by the modal environment module and compile the code text into a target program that allows multimodal network elements to run.

[0038] In an exemplary embodiment, the multimodal controller 14 includes: a modal deployment control system interface, a flow table management subsystem, a packet decision subsystem, a modal switching control subsystem, a topology management subsystem, a traffic scheduling subsystem, a load balancing subsystem, and a local resource management subsystem. The modal deployment control system interface is used to receive management commands and configuration policies issued by the modal deployment control system, wherein the management commands are determined by a target program. The flow table management subsystem is used to analyze and extract global flow table information corresponding to the cloud platform to obtain a target flow table. The traffic scheduling subsystem, connected to the flow table management subsystem, is used to generate first configuration information for adjusting the traffic of multimodal network element ports or specific modes on the cloud platform based on the target flow table when the target flow table is obtained. The packet decision subsystem is used to determine the type of special packets obtained by the multimodal network element ports, wherein the characteristic packets include at least one of the following: unknown packets, abnormal packets, and special packets are a portion of the multimodal packets to be transmitted. The system can handle all packets; the mode switching control subsystem, connected to the packet decision subsystem, marks the target mode corresponding to the characteristic packet as a mode that is not allowed to forward data when the type is discard; and marks the target mode corresponding to the characteristic packet as a mode that is allowed to forward data when the type is new; and generates second configuration information based on the marking results; the topology management subsystem is used to obtain traffic carrying information of other cloud platforms connected to the mode deployment and control system, and adjusts the number of links used by the cloud platform for data transmission based on the traffic carrying information; the load balancing subsystem, connected to the topology management subsystem, generates third configuration information to adjust the data transmission load on adjacent links when the number of links is fixed; and the local resource management subsystem, connected to the traffic scheduling subsystem, the mode switching control subsystem, and the load balancing subsystem, stores the first configuration information, the second configuration information, and the third configuration information, and transmits the configuration information to the multi-modal network element port for distribution to the multi-modal network element.

[0039] In an exemplary embodiment, the multimodal network element 16 includes: a multimodal controller interaction entry, a line-speed forwarding unit, and a modal network element application; the multimodal controller interaction entry is used to receive configuration information issued by the multimodal controller and synchronize the configuration information to the line-speed forwarding unit and the modal network element application for processing; the line-speed forwarding unit is used to forward the received multimodal packets to be transmitted according to the configuration information; the modal network element application is connected to the line-speed forwarding unit and is used to determine the processing method and forwarding port of the target multimodal packets whose modality is unknown in the multimodal packets to be transmitted.

[0040] The multimodal network architecture provided by the above-described device includes: a modal deployment and control system, a multimodal controller, multimodal network elements, and a cloud platform. The modal deployment and control system is connected to the multimodal controller and the multimodal network elements, which are deployed on the cloud platform. The modal deployment and control system is used for centralized control and modal distribution to the cloud platform with which it has network connectivity. The multimodal controller manages the local multimodal network elements on its cloud platform. The multimodal network elements are used for data-level forwarding of multimodal packets within the multimodal network architecture. This technical solution solves the technical problem of traditional cloud service networks having a single network protocol and being unable to flexibly adapt to diverse network service needs. Furthermore, through centralized control by the modal deployment and control system, combined with the multimodal controller and multimodal network elements deployed on the cloud platform, efficient management of multimodal packets and cross-cloud platform data-level forwarding are achieved.

[0041] In an exemplary embodiment, the modal deployment control system includes: a modal deployment subsystem, a deployment control subsystem, and a data management subsystem; wherein, the modal deployment subsystem is used to convert the operational intent of the target object received by the modal deployment control system into a target program, and send the target program to the data management subsystem; the deployment control subsystem is used to collect full-domain information from the cloud platform connected to the modal deployment control system, and generate a configuration policy corresponding to the full-domain information, wherein the configuration policy is used to adjust the data transmission parameters on different cloud platforms, and the data transmission parameters include at least one of the following: network parameters, status parameters, and traffic parameters; the data management subsystem is used to store the target program generated by the modal deployment subsystem and adjust the timing of the target program's distribution.

[0042] Optionally, the modal deployment management system adopts a centralized deployment and can be deployed in any public cloud on the Internet, as long as it is reachable from the IP address of the managed cloud platform. It is responsible for developing and deploying application modalities, collecting necessary information about network operation, and indirectly configuring and managing multimodal network elements in various cloud platforms. The modal deployment management system mainly performs centralized management and control tasks, and has the following two main functions:

[0043] (1) Modal deployment: Based on the user's intent, complete the entire process of modal program development, compilation and deployment, generate a modal compiler that can be run directly, and distribute it to the multimodal controller in the cloud.

[0044] (2) Operation and control: By collecting information from multimodal network elements in each cloud multimodal controller, the operation and control system can grasp the equipment, status and traffic information of all cloud controllers and multimodal network elements. It is the only device that gathers global information and can perform resource management, status monitoring and modal control on the multimodal network during operation, and configure the target multimodal network element.

[0045] Optionally, a control and management subsystem is deployed. This subsystem obtains network, status, and traffic information of multimodal network elements from each cloud controller via the controller's interaction interface and uploads it to the data management subsystem.

[0046] Optional, a data management subsystem. This includes a modal program repository, a network cloud database, and a status database, which stores and manages key programs and configuration information. Compiled executable programs are stored in the modal program repository and then distributed via a program / configuration distribution agent. Data is uploaded to the deployment and control subsystem via the data management subsystem. This system collects information across the entire domain, analyzes network status, and generates configuration policies, playing a crucial role in the stable transmission of multimodal data in the cloud.

[0047] In an exemplary embodiment, the modal deployment subsystem comprises a modal environment module and an intent translation module; wherein the modal environment module is used to build a development environment for receiving operational intents and to generate code text; the intent translation module is connected to the modal environment module and is used to receive the code text output by the modal environment module and compile the code text into a target program that allows multimodal network elements to run.

[0048] Optionally, a modal deployment subsystem. The front-end interaction interface of the modal deployment management system is presented in a graphical interface. Users first input their intent, i.e., a modal program, through the front-end interaction interface to guide the multimodal network elements in processing multimodal data. This program is written in a high-level language. Then, the modal program is distributed to the modal deployment subsystem, which sets up the development environment for the modal program. Through the intent translation module, the modal program is compiled to generate a JSON program that can be directly run on the network elements and sent to the data management subsystem.

[0049] In an exemplary embodiment, the multimodal controller includes a modal deployment control system interface, a flow table management subsystem, a packet decision subsystem, a modal switching control subsystem, a topology management subsystem, a traffic scheduling subsystem, a load balancing subsystem, and a local resource management subsystem. The modal deployment control system interface receives management commands and configuration policies issued by the modal deployment control system, wherein the management commands are determined by a target program. The flow table management subsystem analyzes and extracts global flow table information corresponding to the cloud platform to obtain a target flow table. The traffic scheduling subsystem, connected to the flow table management subsystem, generates first configuration information for adjusting the traffic of multimodal network element ports or specific modes on the cloud platform based on the target flow table, provided the target flow table is obtained. The packet decision subsystem determines the type of special packets obtained by the multimodal network element ports, wherein the characteristic packets include at least one of the following: unknown packets, abnormal packets, and special packets are part or all of the multimodal packets to be transmitted. The system comprises three subsystems: a message dispatch control subsystem (connected to the message decision subsystem), a topology management subsystem (used to mark the target mode corresponding to a characteristic message as a mode that is not allowed to forward data when the message type is discarded, and a message type that is allowed to forward data when the message type is added), and a topology management subsystem (used to obtain traffic carrying information of other cloud platforms connected to the topology management subsystem, and adjust the number of links used by the cloud platform for data transmission based on the traffic carrying information), a load balancing subsystem (connected to the topology management subsystem, used to generate third configuration information to adjust the data transmission load on adjacent links when the number of links is fixed), and a local resource management subsystem (connected to the traffic scheduling subsystem, the topology dispatch control subsystem, and the load balancing subsystem, used to store the first configuration information, the second configuration information, and the third configuration information, and to transmit the configuration information to the multi-modal network element port for distribution to the multi-modal network element).

[0050] Optionally, the flow table management subsystem analyzes and extracts management commands and configuration information obtained from the modal deployment control system interface to form dedicated flow tables for corresponding devices and modes within the cloud. Then, the traffic scheduling subsystem adjusts the multimodal network element port traffic or specific mode traffic in real time, primarily determining the amount of specific mode traffic and the corresponding matching traffic. When a multimodal network element receives an unknown or abnormal packet, it uploads it to the packet decision subsystem for decision-making. If it decides to discard the packet type, the mode will be marked in the mode switching control system, and all subsequent packets of that mode will be discarded. If the packet type is a newly added mode, the flow table management subsystem issues the corresponding flow table and mode program, and then the mode switching subsystem notifies the multimodal network element to execute the forwarding action for that mode. The topology management subsystem obtains the traffic carrying capacity of other multimodal switches connected to its cloud from the modal deployment control system. When multiple links exist, the load balancing subsystem dynamically selects and establishes transmission links based on the load size to achieve load balancing on adjacent links. Because each cloud platform deploys a multimodal controller, end-to-end load optimization is achieved. Finally, the local resource management subsystem stores important configuration information, such as fixed flow tables and global information, that is distributed to multimodal network elements.

[0051] In an exemplary embodiment, the multimodal network element includes: a multimodal controller interaction entry, a line-speed forwarding unit, and a modal network element application; the multimodal controller interaction entry is used to receive configuration information issued by the multimodal controller and synchronize the configuration information to the line-speed forwarding unit and the modal network element application for processing; the line-speed forwarding unit is used to forward the received multimodal packets to be transmitted according to the configuration information; the modal network element application, connected to the line-speed forwarding unit, is used to determine the processing method and forwarding port of the target multimodal packets whose modality is unknown in the multimodal packets to be transmitted.

[0052] Optional multimodal network elements, using custom-developed programmable switches, are capable of parsing, encapsulating, and forwarding different modal packets at line speed. They include two main functions: line-speed forwarding and modal loading, primarily handling data plane-related tasks. A multimodal network element consists of the following two parts:

[0053] (1) Line-speed forwarding unit. The line-speed forwarding unit is the main hardware device of the multi-mode network element. It contains multiple forwarding channels, which can classify modes and complete line-speed forwarding according to priority.

[0054] (2) Modal APP Subsystem. The modal APP subsystem identifies and parses control or management messages of different modes, matches them according to the modal protocol, and performs corresponding encapsulation and decapsulation operations. The processed messages are forwarded by the line-speed forwarding unit.

[0055] Optionally, the line-rate forwarding unit includes: an input processing engine, a modality classifier, a switching matrix, and an output scheduler; the modality APP subsystem includes: a control plane proxy, a policy execution engine, a modality protocol stack, and custom modules. The specific execution flow is as follows: First, the input processing engine in the line-rate forwarding unit is mainly used to receive input multimodal data, which is then passed to the modality classifier for classification. The modality classifier performs Layer 2 parsing on the received data packets to identify the modality type and distinguish between known and unknown modalities. For known modalities, they are directly sent to the switching matrix and forwarded according to the flow table. During the simultaneous forwarding of multiple types of modal packets, the output scheduler can forward them on demand according to queue priority or weight. For unknown modalities, the line-rate forwarding unit will access the policy execution engine in the modality APP subsystem. The policy execution engine receives the modality program and corresponding flow table through the control plane proxy, thereby guiding the processing method and forwarding port of the unknown modality packets. The modal protocol stack stores standard modal protocol stack data, used to encapsulate data packets in a fixed format; the custom module stores custom fields, allowing for custom encapsulation of packet content and format.

[0056] To better understand the process of the above multimodal data transmission method, the following description of the multimodal data transmission method flow is further illustrated with reference to optional embodiments, but is not intended to limit the technical solutions of the embodiments of this application.

[0057] This embodiment provides a method for constructing a multimodal network in the cloud. This method builds a network environment supporting multimodal data processing and transmission by incrementally deploying multimodal network elements, a multimodal controller, and a modal deployment management system in the cloud. The modal deployment management system is responsible for centralized management and configuration, and the development and deployment of modal programs. The multimodal controller performs flow table management and packet decision-making for multimodal network elements within the cloud. The multimodal network elements have line-rate forwarding and modal loading capabilities, and can identify and process various network modalities. Furthermore, this method provides a method for transmitting multimodal data in the cloud, specifically a VXLAN-based multimodal packet encapsulation and transmission method. It utilizes reserved fields in the VXLAN header as modal identifiers to ensure transparent transmission of non-IP packets in the IP network, achieving stable end-to-end transmission of multimodal data.

[0058] Optional, Figure 2 This is a schematic diagram of a cloud-based multimodal network environment structure according to an embodiment of this application, such as... Figure 2 As shown, Figure 2This demonstrates the interconnection between user and operator networks (Internet) and cloud platforms (provided by Internet companies / operators) in a multimodal network environment. Users access the Internet through operator access networks. The cloud egress points of each provincial cloud platform connect to the Internet via egress devices, i.e., firewalls, enabling communication with user traffic. The firewalls, Spine devices, and Leaf devices within the cloud are existing cloud resources.

[0059] Optional, Figure 2 The multimodal network environment constructed in the system includes at least: a modal deployment and control system 22, a multimodal controller 24, and a multimodal network element 26.

[0060] Optional, Modal Deployment Management System 22. The Modal Deployment Management System adopts a centralized deployment and can be deployed in any public cloud on the Internet, as long as it is reachable from the managed cloud platform's IP address. It is responsible for developing and deploying application modalities, collecting necessary information about network operation, and indirectly configuring and managing multimodal network elements in various cloud platforms. The Modal Deployment Management System mainly performs centralized management and control tasks, and has the following two main functions:

[0061] (1) Modal deployment: Based on the user's intent, complete the entire process of modal program development, compilation and deployment, generate a modal compiler that can be run directly, and distribute it to the multimodal controller in the cloud.

[0062] (2) Operation and control: By collecting information from multimodal network elements in each cloud multimodal controller, the operation and control system can grasp the equipment, status and traffic information of all cloud controllers and multimodal network elements. It is the only device that gathers global information and can perform resource management, status monitoring and modal control on the multimodal network during operation, and configure the target multimodal network element.

[0063] Optional, Figure 3 This is a schematic diagram of a modal deployment control system device according to an embodiment of this application. The system includes at least: a modal deployment subsystem 32, a deployment control subsystem 34, and a data management subsystem 36.

[0064] Optionally, there is a modal deployment subsystem 32. The front-end interaction entry is presented in a graphical interface. Users first input their intent, i.e., a modal program, through the front-end interaction entry to guide the multimodal network elements in processing multimodal data. This program is written in a high-level language. Then, the modal program is distributed to the modal deployment subsystem 32, which sets up the development environment for the modal program. The modal program is compiled through the intent translation module to generate a JSON program that can be directly run on the network element and sent to the data management subsystem 36.

[0065] Optionally, deploy a control and management subsystem 34. The control and management subsystem 34 obtains network information, status information, and traffic information of multimodal network elements from each cloud controller through the controller interaction interface, and uploads them to the data management subsystem 36.

[0066] Optionally, a data management subsystem 36 is included. This subsystem comprises a modal program repository, a network cloud database, and a status database, capable of storing and managing key programs and configuration information. Compiled executable programs are stored in the modal program repository and then distributed via a program / configuration distribution agent. Data is uploaded via the data management subsystem 36 to the deployment and control subsystem 34. This system collects information across the entire domain, analyzes network status, and generates configuration policies, playing a crucial role in the stable transmission of multimodal data in the cloud.

[0067] Optional, a multimodal controller 24. The multimodal controller 24 is deployed in the cloud, forming a two-tier control architecture with the modal deployment management system 22. It primarily performs control plane-related tasks, as detailed below:

[0068] (1) The multimodal controller 24 receives configuration and control information from the modal deployment and control system 22 and completes functions such as flow table distribution and topology definition for the multimodal network element 26.

[0069] (2) Receive Packet_in messages from multimodal network element 26, make decisions on the handling of unknown or abnormal packets, and also perform traffic scheduling by mode or port according to different service requirements.

[0070] Optional, Figure 4 This is a schematic diagram of a multimodal controller device according to an embodiment of this application. The device includes at least: a flow table management subsystem 41, a packet decision subsystem 42, a topology management subsystem 43, a traffic scheduling subsystem 44, a mode switching subsystem 44, and a load balancing subsystem 46. The specific execution flow is as follows:

[0071] First, the flow table management subsystem 41 analyzes and extracts management commands and configuration information obtained from the modal deployment control system interface to form dedicated flow tables for corresponding devices and modalities within the cloud. Then, the traffic scheduling subsystem 44 adjusts the multimodal network element port traffic or specific modal traffic in real time, mainly determining the amount of specific modal traffic and the corresponding matching traffic. When a multimodal network element receives an unknown or abnormal packet, it uploads it to the packet decision subsystem 42 for decision-making. If it decides to discard the packet of this type, the modality will be marked in the modality switching control system, and all subsequent packets of this modality will be discarded. If the packet type is a newly added modality, the flow table management subsystem 41 issues the corresponding flow table and modality program, and then the modality switching subsystem 44 notifies the multimodal network element to execute the forwarding action for this modality. The topology management subsystem 43 obtains the traffic load information of other multimodal switches connected to its own cloud from the modal deployment control system. When multiple links exist, the load balancing subsystem 46 dynamically selects the transmission link based on the load size to achieve load balancing on adjacent links. Since each cloud platform deploys a multimodal controller, end-to-end load optimization is achieved. Finally, local resource management is used to store important configuration information distributed to multimodal network elements, such as fixed flow tables and global information.

[0072] Optional, multimodal network element 26. Using a custom-developed programmable switch, it can parse, encapsulate, and forward different modal packets at line speed, including two main functions: line-speed forwarding and modal loading. It primarily performs data plane related tasks. The multimodal network element consists of the following two parts:

[0073] (1) Line-speed forwarding unit. The line-speed forwarding unit is the main hardware device of the multi-mode network element. It contains multiple forwarding channels, which can classify modes and complete line-speed forwarding according to priority.

[0074] (2) Modal APP Subsystem. The modal APP subsystem identifies and parses control or management messages of different modes, matches them according to the modal protocol, and performs corresponding encapsulation and decapsulation operations. The processed messages are forwarded by the line-speed forwarding unit.

[0075] Optional, Figure 5 This is a schematic diagram of a multimodal network element device according to an embodiment of this application. The device includes at least: a line-speed forwarding unit 52 and a modal application subsystem 54. The line-speed forwarding unit 52 includes: an input processing engine 522, a modality classifier 524, a switching matrix 526, and an output scheduler 528. The modal application subsystem 54 includes: a control plane proxy 542, a policy execution engine 544, a modal protocol stack 546, and a custom module 548. The specific execution flow is as follows:

[0076] First, the input processing engine 522 in the line-speed forwarding unit 52 is mainly used to receive input multimodal data, which is then passed to the modality classifier 524 for classification. The modality classifier 524 performs Layer 2 parsing on the received data packets to identify the modality type and distinguish between known and unknown modalities. For known modalities, the packets are directly sent to the switching matrix 526 and forwarded according to the flow table. During the simultaneous forwarding of multiple types of modal packets, the output scheduler 528 can forward them on demand according to queue priority or weight. For unknown modalities, the line-speed forwarding unit 52 accesses the policy execution engine 544 in the modality APP subsystem 54. The policy execution engine 544 receives the modality program and corresponding flow table through the control plane proxy 542, thereby guiding the processing method and forwarding port of the unknown modality packets. The modality protocol stack 546 stores the standard modality protocol stack data, which is used to encapsulate data packets in a fixed format; the custom module 548 is used to store custom fields, which can customize the encapsulation of the content and format of the packets.

[0077] Optionally, in a distributed deployment scenario, a single service can be broken down into multiple sub-services, each deployed in different locations. Similarly, after migrating the service to the cloud, it can be distributed across different clouds. If the number of clouds exceeds one, inter-cloud data transmission becomes unavoidable. To address this, this application presents a method for operating a multimodal cloud network in a typical scenario where two users in different regions access their respective cloud platforms to process cloud services and then communicate between them. This method solves the end-to-end data transmission problem after distributively deploying multimodal services in the cloud.

[0078] Optionally, this embodiment also provides a method for operating a multimodal network in the cloud. This method mainly utilizes VXLAN tunneling technology, modifying its header reserved field as a modal identifier to achieve transparent transmission of non-IP modal data in the IP network. During the data inbound phase, the user terminal upgrades its software protocol stack to support VXLAN encapsulation of multimodal packets, ensuring that data can successfully reach the cloud through the operator's network. Multimodal network elements within the cloud make resolution and forwarding decisions based on the identifier. During the inter-cloud transfer phase, the key parameters at both ends of the tunnel are precisely configured using a VXLAN address matching table to ensure stable transmission of multimodal packets between clouds. The VXLAN tunnel, constructed using virtualization technology, demonstrates good flexibility and scalability. Finally, the data outbound process is the reverse of the inbound process, completing the entire transmission of multimodal data. The entire method, through unified configuration and management of the modal deployment and control system, achieves seamless transmission of multimodal data in the cloud environment, breaking through the protocol limitations of traditional cloud networks and providing strong support for diverse network services.

[0079] Optionally, VXLAN tunneling technology can be used to modify the reserved field between the "Flag" and "VNI" fields in the VXLAN header into a modal identifier, used to fill in the payload type. In other words, multimodal network elements can clearly identify the type of the encapsulated multimodal packet during the decapsulation of VXLAN packets. If the multimodal network element contains a flow table for that modality, it will forward the packet according to the table; if it does not contain the relevant flow table, a Packet_in message will be triggered, uploaded to the controller, and await the controller's instruction (either discard the packet or the controller issues a new flow table and forwards it accordingly). Furthermore, upgrading the software protocol stack at the user terminal can enable VXLAN transport of multimodal packets on the user side and ensure stable transmission in the operator's existing network. Data is transmitted through the operator's network to the cloud, where it is then decapsulated and processed by multimodal network elements; the process of exiting the cloud is the reverse.

[0080] Optional, Figure 6 This is a schematic diagram of an end-to-end multimodal data transmission process in the cloud according to an embodiment of this application, such as... Figure 6 As shown, the service is distributed across Cloud A and Cloud B. User A accesses Cloud A, while User B accesses Cloud B. User A first sends a multimodal packet encapsulated in IP, which is then forwarded to Cloud A via the Internet. The multimodal network elements within Cloud A parse and match the multimodal packet and, through inter-cloud routing, send it to Cloud B, which is near the receiving user. Finally, the packet is re-encapsulated in a tunnel by the multimodal network elements and sent to User B. The complete packet forwarding process can be divided into the following three parts:

[0081] Part 1: Data Transfer to Cloud Process Design. User data transfer to the cloud requires traversing the carrier's network. The specific transmission methods during the data transfer process are as follows:

[0082] like Figure 6 As shown, user terminal A first accesses the operator's optical access network (BRAS, Broadband Remote Access Server) through devices such as switches and firewalls within its local area. The BRAS is mainly used for access control and traffic statistics. The BRAS connects to the MB (Metropolitan Area Backbone), and the MB then connects to the PB (Provincial Area Backbone), interconnecting with cloud resources via the Internet. The optical access network, BRAS, MB, and PB are all operator equipment.

[0083] Optionally, current operator optical access networks and BRAS (Brake-Based Access Server) equipment are designed for the IP protocol and only support the forwarding of IPv4 and IPv6 packets. To achieve non-IP mode pass-through in the Internet, a new packet forwarding scheme needs to be designed. Considering the characteristics of the access network, the solution adopts a network mode bearer method based on IP tunnels, that is, multimodal packets are encapsulated in IP on the user side, such as through a VXLAN (Virtual Extensible LAN) tunnel. Specifically, a tunnel is established between the user terminal and the multimodal network element equipment in the cloud, and intermediate network devices forward packets according to the IP protocol. Details are as follows:

[0084] (1) Upgrade of user terminal functions.

[0085] First, a multimodal software protocol stack is installed on the user terminal to enable the transmission of non-IP protocol packets. The software protocol stack is part of the end-side function development during modal application development; it generates multimodal packets. The three-layer destination identifier (similar to a destination IP address, but since non-IP modes do not rely on IP for addressing, this identifier is used uniformly here) of the multimodal packet is the identifier of the receiving terminal (user terminal B), and the source identifier is the identifier of the sending terminal. To establish a tunnel, the terminal needs to implement IP tunnel encapsulation and decapsulation functions. Considering the potential NAT traversal problem during data transmission, a multimodal packet transmission method based on VXLAN is proposed, using VXLAN technology to encapsulate multimodal packets. Therefore, this application deploys a VXLAN encapsulation program on the user terminal, which mainly implements the following two functions:

[0086] Function 1: Perform VXLAN encapsulation on the generated multimodal packets and decapsulate the VXLAN-encapsulated packets.

[0087] Function 2: Record network information of the VXLAN tunnel start and end points, such as IP and MAC addresses, and fill in the corresponding fields.

[0088] Optionally, in the encapsulated VXLAN tunnel, the sending VTEP is in software form (deployed on the user terminal), and the receiving VTEP is a multimodal network element at cloud A. Multimodal packets sent by the user terminal are encapsulated in software before being transmitted (e.g., ...). Figure 6 In process 1), the field filling content of multimodal packets carried by VXLAN is as follows: Figure 7 As shown. Figure 7This is a schematic diagram of a multimodal packet carried by VXLAN according to an embodiment of this application. The reserved field between the "Flag" and "VNI" fields in the VXLAN header is modified to a modality identifier, used to fill in the payload type. Its main function is that the multimodal network element can clearly identify the type of the encapsulated multimodal packet during the decapsulation of the VXLAN packet. If the multimodal network element contains a flow table for that modality, it will forward the packet according to the table matching. If the multimodal network element does not contain a relevant flow table, a Packet_in message will be triggered, uploaded to the controller, and await the controller's instruction (either discard the packet or the controller issues a new flow table and forwards it accordingly).

[0089] Optionally, the key fields should be filled in as follows: the inner destination MAC address is the MAC address of the service server within Cloud A, and the inner source MAC address is the MAC address of the sending terminal; the outer destination IP address is the IP address of the receiving VTEP, i.e., the IP address of the multimodal network element at Cloud A, which can be obtained through the VXLAN encapsulation program deployed on the user terminal. The outer source IP address is the IP address of the sending VTEP device; the outer destination MAC address is the MAC address of the next-hop device, configured through the VXLAN encapsulation program local to the user terminal. The outer source MAC address is the MAC address of the sending VTEP device.

[0090] Optionally, after receiving the data packet, the multimodal network element decapsulates the VXLAN tunnel message (e.g., Figure 6 In process 2), multimodal packets based on Ethernet bearers are obtained. Then, the modality type and destination address fields in the multimodal packets are parsed, and actions are matched according to the pipeline configuration or flow table from the controller.

[0091] (2) Configuration of Spine switch in cloud.

[0092] Cloud-based switches mainly refer to existing carrier equipment (such as Spine-Leaf switches), while multimodal network elements are new equipment added to the project. To enable packet forwarding to multimodal network elements, the cloud-based Spine switches need to be configured in advance. Specifically, flow tables are configured on the Spine switches to point to the IP addresses of the multimodal network elements, ensuring that all traffic packets destined for those IP addresses are routed to the multimodal network element.

[0093] Part Two: Cloud-to-Cloud Transmission Design. Considering compatibility with the existing network architecture and consistency in cloud-to-cloud and cloud-to-cloud encapsulation methods, the packet encapsulation method for cloud-to-cloud transmission adopts the VXLAN-based transmission method for data entering and leaving the cloud. This method is universally applicable and can be used for data transmission between multiple clouds. A VXLAN tunnel is established between multimodal network elements in two communicating clouds. VXLAN tunnel endpoints (VTEPs) are created on the two multimodal network elements, and VXLAN packets are configured for encapsulation.

[0094] Optional, such as Figure 7 The inner and outer source and destination MAC addresses shown can be obtained from the local ARP table of the multimodal network element. However, the destination IP address and VNI number of the outer packet cannot be directly obtained locally. To accurately fill the key fields of the VXLAN packet, this application uses a modal deployment management system to distribute configurations and designs a VXLAN address matching table within the multimodal network element to match the correct peer address. The modal deployment management system (containing global information) distributes the configuration information required for communication with the destination cloud to the cloud controller. The controller then distributes the information to the multimodal network element in the form of a flow table and stores it locally to obtain and configure the outer destination IP address and the VNI number in the VXLAN header.

[0095] Optional, Figure 8 This is a schematic diagram of a VXLAN address matching table structure according to an embodiment of this application. This address matching table can accurately match and fill specific fields, such as... Figure 8 As shown, this matching table stores configuration information issued by the modal deployment management system, including protocol type, destination identifier, and corresponding outer IP and VNI numbers. The matching table uses "modal type + destination identifier" as the matching entry to query the corresponding outer IP address and VNI number. When multiple paths exist, the modal deployment management system can assign priorities to each matching table entry, meaning that when a data packet matches multiple entries, the one with the higher priority will be applied first. The VXLAN matching table is implemented as a structure, and control information can be added before and after the priority level as needed.

[0096] Optionally, the sender encapsulates the original message into a VXLAN message format (e.g., Figure 6 In process 3), the packet then passes through the Spine switch and is forwarded to the Internet. The source IP address of the VXLAN packet is the IP address of the sending CGMN network element tunnel point, and the destination IP address is the IP address of the receiving multimode network element tunnel point, obtained by querying the VXLAN address matching table. After receiving the data packet, the cloud switch at the receiving end forwards it to the designated multimode network element according to the destination IP address. The multimode network element decapsulates the VXLAN packet to obtain a generated multimode packet (such as...). Figure 6 Process 4 in the middle.

[0097] Part Three: Data Outbound Process Design. The process of user data outbound from the cloud is the reverse of the inbound process. The specific process is as follows:

[0098] First, multimodal network elements need to encapsulate multimodal packets using VXLAN (e.g., ...). Figure 6In process 5), the encapsulated message is transmitted via the Internet to the receiving user terminal B. Terminal B decapsulates the VXLAN message using a VXLAN encapsulation program, restoring the multimodal message. In the transmission tunnel, the sending VTEP is the CGMN network element at cloud B, and the receiving VTEP is third-party software deployed on terminal B. During message filling, a VXLAN address matching table is also used. The fields are: inner destination MAC address is the MAC address of the receiving terminal, i.e., user terminal B; inner source MAC address is the MAC address of the multimodal network element at cloud B; outer destination IP address is the IP address of the receiving terminal VTEP, i.e., the IP address of user terminal B; outer source IP address is the IP address of the sending VTEP, i.e., the IP address of the multimodal network element at cloud B; outer destination MAC address is the MAC address of the next-hop device.

[0099] Optionally, after VXLAN encapsulation, the packet is forwarded to the cloud via a Spine switch. Data is then forwarded to user B via intermediate devices such as PB, MB, and BRAS. User B then decapsulates the VXLAN tunnel (e.g., ...). Figure 6 In process 6), the data is restored to the original multimodal message and sent to the relevant business application of the terminal to complete the data reception.

[0100] In summary, this application establishes a multi-layered, intelligent network management system by incrementally deploying a multi-modal network environment in the cloud, including multi-modal network elements and controllers, supplemented by a centralized modal deployment and control system. Particularly in data transmission, it creatively employs a packet encapsulation method based on VXLAN technology. By modifying reserved fields in the VXLAN header to identify different network modes, it enables transparent transmission of multi-modal data within IP networks. This not only ensures stable end-to-end data transmission from user terminals to the cloud platform and across cloud platforms but also incorporates a VXLAN address matching table to achieve accurate filling and matching of key packet fields, significantly improving the accuracy and efficiency of cross-cloud data transmission.

[0101] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a server to be installed (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0102] According to another embodiment of this application, a multimodal data transmission method is provided, applied to any of the above-mentioned multimodal network architectures, comprising: when it is determined that a first terminal needs to perform multimodal data transmission with a second terminal, determining a first cloud platform connected to the first terminal and a second cloud platform connected to the second terminal, and establishing a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform; when using the communication tunnels for multimodal data transmission, converting the multimodal data to be transmitted by the first terminal into multimodal packets; uploading the multimodal packets to the first cloud platform and forwarding them to the second cloud platform through the first cloud platform; and the second cloud platform sending the received forwarded packets to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

[0103] As an optional implementation, the method embodiments provided in this application can be executed on a server device or a similar computing device. Taking running on a server device as an example, Figure 9 This is a hardware structure block diagram of a server device according to an embodiment of a multimodal data transmission method of this application. For example... Figure 9 As shown, the server device may include one or more ( Figure 9 Only one is shown in the diagram. A processor 902 (which may include, but is not limited to, a microprocessor MPU or a programmable logic device FPGA, etc.) and a memory 904 for storing data are also shown. The server device may further include a transmission device 906 for communication functions and an input / output device 908. Those skilled in the art will understand that... Figure 9 The structure shown is for illustrative purposes only and does not limit the structure of the server equipment described above. For example, the server equipment may also include components that are more... Figure 9 The more or fewer components shown, or having the same Figure 9 The different configurations shown.

[0104] The memory 904 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the method for executing early warning events in this embodiment. The processor 902 executes various functional applications and data processing by running the computer program stored in the memory 904, thus implementing the aforementioned method. The memory 904 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 904 may further include memory remotely located relative to the processor 902, and these remote memories can be connected to server devices via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0105] Transmission device 906 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the server equipment. In one example, transmission device 906 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, transmission device 906 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0106] This embodiment provides a multimodal data transmission method applied to the aforementioned server device. Figure 10 This is a flowchart of a multimodal data transmission method according to an embodiment of this application, such as... Figure 10 As shown, the process includes the following steps:

[0107] Step S1002: When it is determined that the first terminal needs to perform multimodal data transmission with the second terminal, the first cloud platform connected to the first terminal and the second cloud platform connected to the second terminal are determined, and a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform are established respectively.

[0108] Step S1004: When using a communication tunnel for multimodal data transmission, convert the multimodal data to be transmitted by the first terminal into a multimodal message.

[0109] Step S1006: Upload the multimodal message to the first cloud platform, and forward it to the second cloud platform through the first cloud platform;

[0110] In step S1008, the second cloud platform sends the received forwarded message to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

[0111] Through the above steps, when it is determined that the first terminal needs to perform multimodal data transmission with the second terminal, the first cloud platform to which the first terminal connects and the second cloud platform to which the second terminal connects are identified. A first communication tunnel is established between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform. When using these communication tunnels for multimodal data transmission, the multimodal data to be transmitted by the first terminal is converted into multimodal packets. These multimodal packets are uploaded to the first cloud platform and forwarded to the second cloud platform. The second cloud platform then sends the received forwarded packets to the second terminal for decapsulation processing, thus completing the multimodal data transmission between the first and second terminals. This technical solution solves the technical problem of traditional cloud service networks having a single network protocol and being unable to flexibly adapt to diverse network service needs. Furthermore, through centralized control by a modal deployment management system, combined with multimodal controllers and multimodal network elements deployed on the cloud platform, efficient management of multimodal packets and cross-cloud platform data-level forwarding are achieved.

[0112] In an exemplary embodiment, converting multimodal data to be transmitted by a first terminal into a multimodal message includes: determining a message template supported by the communication tunnel, wherein the message template includes at least one of the following: an outer Ethernet header byte, an outer IP header byte, a tunnel header byte, and an inner Ethernet frame byte; filling the inner Ethernet frame byte in the message template according to the first MAC address corresponding to the service server in the first cloud platform, the second MAC address corresponding to the first terminal, and the multimodal data; filling the outer IP header byte in the message template according to the first IP address corresponding to the multimodal network element in the first cloud platform and the second IP address corresponding to the first terminal; filling the outer Ethernet header byte in the message template according to the second MAC address corresponding to the first terminal and the third MAC address corresponding to the second terminal; filling the tunnel header byte in the message template according to the mode type corresponding to the multimodal data; and obtaining a multimodal message corresponding to the multimodal data after filling is completed.

[0113] Optionally, if User A (the first terminal) needs to transmit data based on the RoCE (RDMA over Converged Ethernet, or RoCE for short) protocol to User B (the second terminal), and User A and User B are located in Cloud A and Cloud B respectively, and Cloud A and Cloud B are connected via the Internet, and User A's software protocol stack has been updated through the modal deployment management system, enabling it to generate multimodal data using the RoCE protocol, then first select a VXLAN tunnel message template that supports the RoCE protocol. Then, the key fields are filled in: the destination MAC address of the inner Ethernet frame is set to the MAC address of a service server in cloud A (used to receive and process RoCE data), the source MAC address is set to the MAC address of user A, the Type field is identified as RoCE protocol, followed by the source identifier (user A's ID), the destination identifier (user B's ID), and the RoCE protocol data; the destination IP address of the outer IP header is set to the IP address of the multimodal network element (VTEP) in cloud A, and the source IP address is set to the IP address of user A; the destination MAC address of the outer Ethernet header is set to the MAC address of the multimodal network element in cloud A, and the source MAC address is set to the MAC address of user A; the tunnel header bytes are filled in, and the reserved field in the VXLAN tunnel header is modified to the modality identifier and filled with "RoCE" to identify the multimodal data type being transmitted. After User A's software protocol stack completes the encapsulation of the aforementioned information, it generates a VXLAN multimodal packet carrying RoCE data. This packet can successfully pass through the network devices of Cloud A, be transmitted via the Internet to Cloud B, and then be decapsulated and identified as RoCE type by the multimodal network elements within Cloud B. Finally, it is processed according to the correct protocol and forwarded to User B. Through the above steps, cross-cloud transmission of RoCE protocol data in the cloud environment is realized, demonstrating the comprehensive advantages of this application's solution in terms of compatibility, efficiency, security, and flexibility.

[0114] In an exemplary embodiment, after uploading the multimodal message to the first cloud platform, the method further includes: instructing the first cloud platform to parse the data packet corresponding to the multimodal message, so as to determine the modality type and destination address field when the multimodal message is forwarded from the first cloud platform to the second cloud platform based on the parsing result; and, if the parsing is completed, requesting the first cloud platform to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and destination address field.

[0115] Optionally, if user A generates multimodal data based on the InfiniBand (IB) protocol in cloud A, it needs to be sent to user B located in cloud B. An IP network connection has been established between cloud A and cloud B, but the IB protocol is not IP-based, therefore encapsulation and transmission require tunneling technology. The IB protocol multimodal data generated by user A is first encapsulated into a VXLAN multimodal packet and then uploaded to cloud A. After receiving the packet, the multimodal controller in cloud A identifies that the packet carries IB protocol data based on the modality identifier field in the tunnel header. Subsequently, the controller parses the destination identifier (user B's identifier) ​​from the inner Ethernet frame. After parsing, the multimodal controller queries the VXLAN address matching table based on the parsing result (modality type: IB, destination identifier: user B) to obtain the outer IP address and VNI number of the multimodal network element in cloud B corresponding to user B. The controller then issues a flow table to the multimodal network element in cloud A, instructing it to establish a VXLAN tunnel transmission with the multimodal network element in cloud B. The above embodiments enable secure, efficient, and intelligent data transmission.

[0116] In an exemplary embodiment, after requesting the first cloud platform to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and destination address fields, the method further includes: determining that the modality deployment control system sends configuration information required for communication with the second cloud platform to the first multimodal controller within the first cloud platform; determining a tunnel address matching table based on the configuration information, wherein each entry in the tunnel address matching table includes a matching field, an outer IP address and tunnel number field, and a priority field, and the matching field consists of a modality type byte and a destination identifier byte; and controlling the multimodal data transmission between the first cloud platform and the second cloud platform based on the tunnel address matching table.

[0117] Optionally, if user A initiates a compute-intensive task based on RDMA in cloud A, it needs to exchange data with user B in cloud B. First, a tunnel transmission foundation is established between cloud A and cloud B; that is, the first cloud platform establishes a tunnel based on the modality type (RDMA) and the destination address (the location of user B). Next, the modality deployment management system sends the configuration information required for communication with user B to the first multimodal controller in cloud A, including user B's specific identifier, the outer IP address of the multimodal network element in cloud B, the VNI number of the tunnel (used for VXLAN tunnels), and the priority of the data flow. After receiving the configuration information, the first multimodal controller generates a tunnel address matching table. Assume the table contains the following entries: Matching field: "RDMA protocol + user B identifier"; Outer IP address and tunnel number: the IP address and corresponding VNI number of the multimodal network element in cloud B; Priority: "High". When the RDMA multimodal data packet generated by user A arrives at the multimodal network element in cloud A, the network element will refer to the above tunnel address matching table to find a matching entry. Since the modality type and destination identifier of the data packet match the fields in the table, the network element will send the data packet to the corresponding multimodal network element in Cloud B using a high-priority tunnel, as configured. During this process, the data packet is encapsulated through a VXLAN tunnel, which includes the IP address and VNI number of user B, ensuring direct and efficient transmission from Cloud A to Cloud B. Upon arrival at Cloud B, the multimodal network element decapsulates the data packet, restores the RDMA data, and delivers it to user B, thereby achieving seamless data exchange for RDMA-based computing tasks in a multi-cloud environment. This embodiment, through automated configuration and intelligent flow control, significantly optimizes the communication process across cloud platforms, improving user experience and business execution efficiency.

[0118] Embodiments of this application also provide a storage medium including a stored program, wherein the program executes any of the methods described above when it is run.

[0119] Optionally, in this embodiment, the storage medium may be configured to store program code for performing the following steps:

[0120] S1, when it is determined that the first terminal needs to perform multimodal data transmission with the second terminal, the first cloud platform connected to the first terminal and the second cloud platform connected to the second terminal are determined, and a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform are established respectively.

[0121] S2, when using a communication tunnel for multimodal data transmission, convert the multimodal data to be transmitted by the first terminal into a multimodal message;

[0122] S3, upload the multimodal message to the first cloud platform, and forward it to the second cloud platform through the first cloud platform;

[0123] S4, the second cloud platform sends the received forwarded message to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

[0124] Embodiments of this application also provide an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0125] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0126] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0127] S1, when it is determined that the first terminal needs to perform multimodal data transmission with the second terminal, the first cloud platform connected to the first terminal and the second cloud platform connected to the second terminal are determined, and a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform are established respectively.

[0128] S2, when using a communication tunnel for multimodal data transmission, convert the multimodal data to be transmitted by the first terminal into a multimodal message;

[0129] S3, upload the multimodal message to the first cloud platform, and forward it to the second cloud platform through the first cloud platform;

[0130] S4, the second cloud platform sends the received forwarded message to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

[0131] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0132] Optionally, embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0133] Optionally, embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0134] Optionally, embodiments of this application also provide a computer program that includes computer instructions stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in any of the above method embodiments.

[0135] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0136] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0137] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A multimodal network architecture, characterized in that, include: A modal deployment control system includes a multimodal controller, multimodal network elements, and a cloud platform. The modal control module is connected to the multimodal controller and the multimodal network elements, respectively. The multimodal controller and the multimodal network elements are deployed on the cloud platform. The modal deployment management system is used for centralized control and modal distribution of cloud platforms that are interconnected with the modal deployment management system via the network; The multimodal controller is used to manage local multimodal network elements in the cloud platform. The multimodal network element is used to forward multimodal packets at the data layer in a multimodal network architecture.

2. The multimodal network architecture according to claim 1, characterized in that, The modal deployment control system includes: a modal deployment subsystem, a deployment control subsystem, and a data management subsystem; The modal deployment subsystem is used to convert the operational intent of the target object received by the modal deployment control system into a target program, and send the target program to the data management subsystem. The deployment control subsystem is used to collect full-domain information from the cloud platform connected to the modal deployment control system and generate configuration strategies corresponding to the full-domain information. The configuration strategies are used to adjust data transmission parameters on different cloud platforms. The data transmission parameters include at least one of the following: network parameters, status parameters, and traffic parameters. The data management subsystem is used to store the target program generated by the modal deployment subsystem and adjust the timing of the target program's distribution.

3. The multimodal network architecture according to claim 2, characterized in that, The modal deployment subsystem includes: a modal environment module and an intent translation module. The modal environment module is used to build a development environment for receiving the operation intent and generate code text; the intent translation module is connected to the modal environment module and is used to receive the code text output by the modal environment module and compile the code text into a target program that allows multimodal network elements to run.

4. The multimodal network architecture according to claim 1, characterized in that, The multimodal controller includes a modal deployment control system interface, a flow table management subsystem, a packet decision subsystem, a modal switching control subsystem, a topology management subsystem, a traffic scheduling subsystem, a load balancing subsystem, and a local resource management subsystem. The modal deployment management system interface is used to receive management commands and configuration policies issued by the modal deployment management system, wherein the management commands are determined by the target program; The flow table management subsystem is used to analyze and extract global flow table information corresponding to the cloud platform to obtain the target flow table; The traffic scheduling subsystem is connected to the flow table management subsystem and is used to generate first configuration information for adjusting the traffic of multimodal network element ports or specific modes on the cloud platform based on the target flow table when the target flow table is obtained. The message decision subsystem is used to determine the type of special message obtained by the multimodal network element port, wherein the special message includes at least one of the following: unknown message, abnormal message, and the special message is part or all of the multimodal message to be transmitted; The mode switching control subsystem, connected to the packet decision subsystem, is used to mark the target mode corresponding to the feature packet as a mode that does not allow data forwarding when the type is discard; and to mark the target mode corresponding to the feature packet as a mode that allows data forwarding when the type is new; and to generate second configuration information based on the marking results. The topology management subsystem is used to obtain traffic carrying information of other cloud platforms that are connected to the modal deployment control system, and adjust the number of links used by the cloud platform for data transmission based on the traffic carrying information; The load balancing subsystem is connected to the topology management subsystem and is used to generate third configuration information to adjust the data transmission load on adjacent links when the number of links is fixed. The local resource management subsystem is connected to the traffic scheduling subsystem, the mode switching control subsystem, and the load balancing subsystem. It is used to store the first configuration information, the second configuration information, and the third configuration information, and to transmit the configuration information to the multi-modal network element port for distribution to the multi-modal network element.

5. The multimodal network architecture according to claim 1, characterized in that, The multimodal network element includes: a multimodal controller interaction entry point, a line-speed forwarding unit, and a modal network element application; The multimodal controller interaction interface is used to receive configuration information issued by the multimodal controller and synchronize the configuration information to the line-speed forwarding unit and the modal network element application for processing; The line-speed forwarding unit is used to forward the received multimodal packets to be transmitted according to the configuration information. The modal network element application, connected to the line-speed forwarding unit, is used to determine the processing method and forwarding port for target multimodal packets with unknown modalities in the multimodal packets to be transmitted.

6. A multimodal data transmission method applied to the multimodal network architecture described in any one of claims 1 to 5, characterized in that, include: When it is determined that the first terminal needs to perform multimodal data transmission with the second terminal, the first cloud platform connected to the first terminal and the second cloud platform connected to the second terminal are determined, and a first communication tunnel between the first terminal and the first cloud platform, a second communication tunnel between the second terminal and the second cloud platform, and a third communication tunnel between the first cloud platform and the second cloud platform are established respectively. When using a communication tunnel for multimodal data transmission, the multimodal data to be transmitted by the first terminal is converted into a multimodal message; The multimodal message is uploaded to the first cloud platform and then forwarded to the second cloud platform through the first cloud platform; The second cloud platform sends the received forwarded message to the second terminal for decapsulation processing to complete the multimodal data transmission between the first terminal and the second terminal.

7. The multimodal data transmission method according to claim 6, characterized in that, Converting the multimodal data to be transmitted from the first terminal into a multimodal message includes: The message template supported by the communication tunnel is determined, wherein the message template includes at least one of the following: outer Ethernet header bytes, outer IP header bytes, tunnel header bytes, and inner Ethernet frame bytes; The inner Ethernet frame bytes in the message template are filled according to the first MAC address corresponding to the business server in the first cloud platform, the second MAC address corresponding to the first terminal, and the multimodal data. The outer IP header bytes in the message template are filled according to the first IP address corresponding to the multimodal network element in the first cloud platform and the second IP address corresponding to the first terminal; The outer Ethernet header bytes in the message template are filled according to the second MAC address corresponding to the first terminal and the third MAC address corresponding to the second terminal. The tunnel header bytes in the message template are filled according to the modality type corresponding to the multimodal data; Once the filling is complete, the multimodal message corresponding to the multimodal data is obtained.

8. The multimodal data transmission method according to claim 6, characterized in that, After uploading the multimodal message to the first cloud platform, the method further includes: The first cloud platform is instructed to parse the data packet corresponding to the multimodal message, so as to determine the modality type and destination address field when the multimodal message is forwarded from the first cloud platform to the second cloud platform based on the parsing result; If the parsing is completed, the first cloud platform is requested to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and the destination address field.

9. The multimodal data transmission method according to claim 8, characterized in that, After requesting the first cloud platform to establish a tunnel transmission between the first cloud platform and the second cloud platform based on the modality type and the destination address field, the method further includes: The modal deployment control system determines that it sends the configuration information required for communication with the second cloud platform to the first multimodal controller within the first cloud platform; A tunnel address matching table is determined based on the configuration information. Each entry in the tunnel address matching table includes a matching field, an outer IP address and tunnel number field, and a priority field. The matching field consists of modality type bytes and destination identifier bytes; The multimodal data transmission between the first cloud platform and the second cloud platform is controlled based on the tunnel address matching table.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method described in any one of claims 6 to 9.

11. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 6 to 9 through the computer program.

Images