Traffic distribution method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a traffic distribution method and device, computer equipment, a computer readable storage medium and a computer program product. Relates to the technical field of cloud. The method comprises the following steps: identifying a traffic source of traffic data in a service access layer; identifying gray traffic in the traffic data based on the traffic source, and endowing the gray traffic with a gray label; under the condition that the service processing layer executes the data processing service of the traffic data, identifying a data label carried by the traffic data; and calling a gray scale container of the service processing layer to process the flow data carrying the gray scale label. According to the invention, the traffic data is classified and the gray scale label is added in the service access layer, and then the special gray scale container is called to carry out isolation processing on the traffic data of the gray scale type according to the gray scale label in the service processing layer, so that the gray scale traffic does not relate to a full-amount container of the service, and the service efficiency is improved. The influence of gray flow on the system stability is reduced, and the system stability is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a traffic distribution method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] With the development of computer technology and applications, traffic distribution technology has emerged, which can enable various experiments during application version changes. New features of the application can be tested through traffic distribution. Taking the recommendation function as an example, the new version of the recommendation function can be published to the application system container through traffic distribution. However, due to the complexity of the recommendation system and the long execution chain, the experimental process may bring certain stability challenges to various modules in the application system chain.

[0003] The current traffic distribution process typically samples users and selects a portion of users for experimental traffic distribution. However, this method still requires all machines in the system to process the relevant traffic, so experimental changes under this method may pose a risk to system stability. Summary of the Invention

[0004] Therefore, it is necessary to provide a traffic distribution method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve system stability during the traffic distribution process, in order to address the aforementioned technical problems.

[0005] Firstly, this application provides a traffic distribution method, including:

[0006] Identify the traffic source of traffic data at the service access layer;

[0007] Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic;

[0008] When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data;

[0009] The grayscale container of the service processing layer is invoked to process traffic data carrying grayscale labels.

[0010] Secondly, this application also provides a traffic distribution device, comprising:

[0011] The traffic identification module is used to identify the source of traffic data at the service access layer;

[0012] The tag assignment module is used to identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale tags to the grayscale traffic;

[0013] The tag recognition module is used to identify the data tags carried by the traffic data when the data processing service of the traffic data is executed in the service processing layer;

[0014] The container invocation module is used to invoke the grayscale container of the service processing layer to process traffic data carrying grayscale labels.

[0015] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0016] Identify the traffic source of traffic data at the service access layer;

[0017] Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic;

[0018] When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data;

[0019] The grayscale container of the service processing layer is invoked to process traffic data carrying grayscale labels.

[0020] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0021] Identify the traffic source of traffic data at the service access layer;

[0022] Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic;

[0023] When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data;

[0024] The grayscale container of the service processing layer is invoked to process traffic data carrying grayscale labels.

[0025] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0026] Identify the traffic source of traffic data at the service access layer;

[0027] Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic;

[0028] When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data;

[0029] The grayscale container of the service processing layer is invoked to process traffic data carrying grayscale labels.

[0030] The aforementioned traffic distribution method, apparatus, computer equipment, computer-readable storage medium, and computer program product first identify the traffic source of traffic data at the service access layer; then, based on the traffic source, identify grayscale traffic within the traffic data and assign grayscale labels to the grayscale traffic. As traffic data flows into the service access layer, grayscale traffic types are identified based on the source of this traffic data. This traffic data is classified at the traffic entry point. Then, while the service processing layer is performing data processing services on the traffic data, the data labels carried by the traffic data are identified. That is, during the provision of data processing services, the grayscale type of traffic is identified by recognizing the data labels of the traffic data. Finally, the grayscale container in the service processing layer is invoked to process the traffic data carrying grayscale labels. In other words, the service processing layer uses a dedicated grayscale container to isolate and process this traffic data carrying grayscale labels, thereby preventing these grayscale traffic types from affecting system stability. This application classifies traffic data and adds grayscale tags at the service access layer, and then calls a dedicated grayscale container in the service processing layer based on the grayscale tags to isolate grayscale traffic data. This ensures that grayscale traffic does not involve the full container of the service, reducing the impact of grayscale traffic on system stability and improving system stability. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a diagram illustrating the application environment of a traffic distribution method in one embodiment;

[0033] Figure 2 This is a flowchart illustrating a traffic distribution method in one embodiment;

[0034] Figure 3 This is a schematic diagram of the architecture of a traffic distribution system in one embodiment;

[0035] Figure 4 This is a schematic diagram of the grayscale container selection process in one embodiment;

[0036] Figure 5This is a schematic diagram of the grayscale container deletion and insertion process in one embodiment;

[0037] Figure 6 This is a schematic diagram of the grayscale container state change process in one embodiment;

[0038] Figure 7 This is a schematic diagram of the grayscale container expansion process in one embodiment;

[0039] Figure 8 This is a schematic diagram of the grayscale container shrinkage process in one embodiment;

[0040] Figure 9 A schematic diagram illustrating the layered grayscale containers in one embodiment;

[0041] Figure 10 This is a schematic diagram of a small cluster traffic optimization process in one embodiment;

[0042] Figure 11 This is a flowchart illustrating the traffic distribution method in another embodiment;

[0043] Figure 12 This is a structural block diagram of a traffic distribution device in one embodiment;

[0044] Figure 13 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] The traffic distribution method provided in this application embodiment can be applied to, for example, Figure 1In the application environment shown, terminal 102 communicates with ingress server 104 via a network. Ingress server 104 subsequently includes multiple processing servers 106 for service processing. Both ingress server 104 and processing servers 106 are connected to traffic distribution server 108. Multiple container instances running on processing servers 106 provide data processing services. A data storage system can store data related to the container list that traffic distribution server 108 needs to process. The data storage system can be integrated on traffic distribution server 108 or placed on a cloud or other network server. When terminal 102 accesses ingress server 104 to request related service processing, ingress server 104 receives the requested traffic data. Traffic distribution server 108 identifies the source of the traffic data at the service access layer, identifies grayscale traffic within the traffic data based on the source, and assigns grayscale labels to the grayscale traffic. These traffic data are then transferred to the subsequent processing server 106. The traffic distribution server 108, when the current processing server 106 needs to perform data processing services on the traffic data, identifies the data tags carried by the traffic data entering the processing server 106 and calls the grayscale container in the processing server 106 to process the traffic data carrying grayscale tags. The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The entry server 104, processing server 106, and traffic distribution server 108 can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing cloud computing services.

[0047] In one exemplary embodiment, such as Figure 2 As shown, a traffic distribution method is provided, which can be applied to... Figure 1 Taking the traffic distribution server 108 as an example, the explanation includes the following steps 201 to 207. Wherein:

[0048] Step 201: Identify the traffic source of the traffic data at the service access layer.

[0049] The service access layer refers to the entry point for traffic data. When data processing is required, the traffic for data processing-related requests first converges at the service access layer, and then enters the subsequent service processing layer to execute the relevant data processing service operations. Traffic data refers to data transmitted through computer networks. For example, in a video recommendation system, a request submitted by a target user to obtain a list of recommended videos can be transmitted to the service access layer of the video recommendation service, where it is processed as traffic data for generating the recommendation list.

[0050] For example, this application is specifically applied to the distribution and processing of traffic data during the service process, so as to import this service-related traffic data into the corresponding container for data processing operations. When the target object initiates a data processing request through terminal 102, these requests will enter the service system through the network in the form of traffic data. The traffic data entering the service system will first converge at the access server 104 of the service access layer, and then sequentially enter the processing server 106, which provides the subsequent data processing server, where the container instance running on the processing server 106 will execute the specific data processing service. The traffic distribution server 108 is connected to the access server 104 and also to the processing server 106, and is used to control the distribution of traffic data. After the traffic data converges to the service access layer, the traffic distribution server 108 will first identify the source of the traffic data entering the service access layer. Specifically, this can be achieved by identifying the representation information attached to this traffic data. For example, traffic data carrying account identification information can be identified by the account identification to determine the data source. The data source is specifically used to identify the type of traffic data. In one embodiment, this application is specifically applied to the distribution of traffic data in a video recommendation system. After a video recommendation request is triggered, it typically goes through the access layer, engine layer, recall layer, fine ranking layer, and re-ranking layer, finally returning the highest-quality videos to the requesting object. The video recommendation request first enters the access layer of the recommendation system. At this point, the traffic control system connected to the access layer can read the request information and identify the requesting object by the account identifier carried in the request information, using it as the source of the traffic data.

[0051] Step 203: Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic.

[0052] In this context, grayscale traffic, or gray-scale data, refers to a state between black and white. For services, grayscale represents an intermediate state during version iterations and feature upgrades, a smooth transition from an older version to a newer one. Before upgrading to a new version, grayscale can be used—selecting a portion of traffic data as grayscale traffic—to verify the performance of the new version. Grayscale labels are used to identify whether the traffic data belongs to grayscale or the normal mains data.

[0053] For example, after determining the source of traffic data, these traffic volumes can be classified according to their source to identify gray-scale traffic used for experiments and normal processing traffic. For the identified gray-scale traffic, a gray-scale label can be assigned, thereby marking these gray-scale traffic volumes at the traffic entry point of the service access layer. In one embodiment, this application is specifically applied to traffic data distribution in a video recommendation system. Upgrade prompt messages can be pushed to selected objects in the video application. When an object chooses to receive the upgrade prompt message and upgrade, that object can be identified as a gray-scale object, and the request submitted by the gray-scale object is gray-scale traffic. At the service access layer, when the traffic source is identified as a gray-scale object, the traffic data can be identified as gray-scale traffic, and a gray-scale label can be assigned to these gray-scale traffic volumes.

[0054] Step 205: When performing data processing services for traffic data at the service processing layer, identify the data tags carried by the traffic data.

[0055] Step 207: Call the grayscale container of the service processing layer to process the traffic data carrying grayscale labels.

[0056] The service processing layer is responsible for processing traffic data. After being aggregated by the service access layer, the traffic data continues to be processed by subsequent service processing layers. For example, in a video recommendation system, after a requester submits a video recommendation request, requests from different users enter the recommendation system through the service access layer. Then, the system goes through the engine layer, recall layer, fine ranking layer, and re-ranking layer, finally returning the best videos to the user. The engine layer, recall layer, fine ranking layer, and re-ranking layer constitute the service processing layer of the recommendation system, responsible for performing data processing services related to the recommendation and list generation process. The service processing layer contains container instances that execute services. These container instances can be divided into main containers and gray-scale containers. Gray-scale containers are those that perform data processing related to gray-scale traffic. Gray-scale containers isolate gray-scale traffic from the main traffic data, preventing it from entering normal container instances and thus preventing this gray-scale traffic from affecting the overall system.

[0057] For example, when traffic data enters the system from the service access layer, it sequentially enters each service processing layer to perform corresponding service processing operations. Therefore, when data processing services are executed via RPC (Remote Procedure Call) or other methods, the traffic distribution server 108 can identify the data tags carried by the traffic data through the processing server 106 in the service processing layer. In one embodiment, routing plugins can be pre-deployed in each service processing layer, and then the traffic distribution server 108 can identify grayscale tags on the traffic through the routing plugins, thereby determining which traffic entering the service processing layer belongs to grayscale traffic and which belongs to non-grayscale disk traffic. Based on the identification of grayscale traffic, a grayscale container in the service processing layer can be called to process the traffic data carrying grayscale tags. At the same time, if it is disk traffic carrying grayscale tags, an instance is selected from the disk instances for calling. Through this mechanism, this application achieves physical isolation of grayscale in the system. Grayscale traffic will only request grayscale containers in each service processing layer, thereby eliminating the risk to disk instances caused by experiments, upgrades, and other events. This can effectively improve the stability of the system. In one embodiment, this application is specifically applied to traffic data distribution in a video recommendation system, and the system architecture diagram can be referred to in this case. Figure 3 As shown, traffic first enters the system through the service entry point of the service access layer. The responsible traffic distribution server can tag the traffic entering the system at the service entry point, assigning gray-scale labels to the gray-scale traffic. Then, these tagged gray-scale traffic and untagged normal traffic will enter the subsequent service access layers A, B, C, D, etc. in sequence. The traffic distribution server can identify the data labels carried by the traffic data at each service access layer. The service access layer contains two types of container instances: gray-scale container instances and normal large-scale container instances. As shown by the arrows in the figure, the traffic distribution server can call the gray-scale container instances of the service access layer to process gray-scale traffic and call the large-scale container instances to process normal traffic, thereby achieving physical isolation processing of gray-scale traffic.

[0058] The aforementioned traffic distribution method first identifies the traffic source of traffic data at the service access layer; then, based on the traffic source, it identifies gray-scale traffic within the traffic data and assigns gray-scale labels to it. As traffic data flows into the service access layer, gray-scale traffic is identified based on its source. This traffic data is categorized at the traffic entry point. Then, during data processing at the service processing layer, the data labels carried by the traffic data are identified. In other words, during the data processing service, the gray-scale type of traffic is identified by recognizing its data labels. Finally, the gray-scale container in the service processing layer is called to process the traffic data carrying gray-scale labels. This means that the service processing layer uses a dedicated gray-scale container to isolate and process this gray-scale traffic data, thereby preventing it from affecting system stability. This application, by classifying traffic data and adding gray-scale labels at the service access layer, and then using the gray-scale labels in the service processing layer to call a dedicated gray-scale container to isolate and process gray-scale traffic data, ensures that this gray-scale traffic does not involve the full service container, reducing the impact of gray-scale traffic on system stability and improving system stability.

[0059] In an exemplary embodiment, the method further includes: obtaining container address data of data processing services through service discovery; generating a container list of different data processing services based on the container address data; selecting a corresponding proportion of large-scale containers from each container list as grayscale containers, and marking the grayscale containers in the container list.

[0060] Service discovery refers to using a registry center to record information about all services in a distributed system, enabling other services to quickly find these registered services. For example, a service discovery platform can be used to process service discovery and obtain the container address data for different data processing services. Container address data refers to the container's IP (Internet Protocol) address. When creating a container, a corresponding IP address needs to be assigned to it, and the container address data allows for locating the container on the network. As for the container list, a corresponding container list can be created for each data processing service. The list contains all container instances executing that service, and its contents are modifiable, meaning they can be modified according to changes in the service's container instances. Initially, all deployed containers can serve as the main container pool. Later, for each data processing service, a portion of these main containers can be selected as canary containers specifically for handling corresponding canary traffic.

[0061] For example, before performing traffic distribution processing related to grayscale traffic, it is necessary to elect a portion of container instances from the container instances executing data processing services as grayscale containers. First, it is necessary to collect container-related information for different data processing services. This can be done through service discovery, collecting container address data of container instances under each data processing service from the system. The names of these container instances are randomly generated, and the order of the returned list is also random. Then, a corresponding proportion of large-scale containers can be selected from each container list as grayscale containers, and these grayscale containers are marked in the container list. To ensure the selection of a fixed batch of grayscale containers, all container lists can be sorted, for example, based on container names, and then a fixed proportion of containers at the top of the sequence are selected. In one embodiment, such as... Figure 4 As shown, the marking of gray-scale containers can be implemented through a container management platform. After identifying the container address data of the service and generating a corresponding container list, 5% of the large-disk containers can be selected as gray-scale containers. These selected gray-scale containers are then marked in the container list, thus categorizing containers into two types: ordinary large-disk containers and gray-scale containers. Furthermore, a copy of the container list with gray-scale container markings can be generated, and the container list and its copy are stored based on a strong consistency mechanism. To maintain the accuracy of the gray-scale container identification process and ensure high availability of traffic distribution, multiple identical list copy modules can be stored within the system to provide the same traffic distribution service capabilities, reducing the probability of failures due to single points of failure. Simultaneously, the etcd strong consistency mechanism is used internally to ensure the consistency of the stored gray-scale container list. In this embodiment, generating the container list by obtaining container address data, then selecting gray-scale containers from it and marking the selected gray-scale containers effectively ensures the accuracy of gray-scale container selection, thereby guaranteeing the effectiveness of traffic distribution.

[0062] In an exemplary embodiment, the method further includes: obtaining a state change request for a data processing service; identifying the state change type of the state change request; updating the grayscale container markers of the container list and list copies when the state change type is a grayscale container change type; and changing the grayscale container status of the container list and list copies to an invalid state when the state change type is a service change type.

[0063] A state change request refers to a request used to change the state of a canary container. The canary container's state is divided into an active state and an inactive state. In an active state, it can receive service requests from other upstream layers, while in an inactive state, it cannot accept service requests from other layers. Generally, situations such as normal service image deployment, routine restarts, container migration, container deletion, and service crashes due to internal anomalies require setting it to an inactive state to prevent unresponsiveness for normal requests. Therefore, in these situations, a state change request can be used to change the state of the service's canary container.

[0064] For example, the state of a grayscale container can be set to an active or inactive state, and state change requests can be used to control the state change of the grayscale container. There are two types of state change: one is the grayscale container change type, such as when a container is deleted or migrated. In these cases, the container needs to be completely removed from the grayscale list, and then a new active, available container is added. The grayscale container markers in the container list and its copy are updated. After deleting the currently marked grayscale container, a new large-scale container is selected and set as the grayscale container. Figure 5 As shown, the process of a container being deleted or migrated can be achieved by deleting the corresponding grayscale container from the container list and then adding a new grayscale container.

[0065] For service change types, such as service image releases, routine restarts, and service crashes, these can all be considered service change types. In these cases, canary containers cannot be removed from the list; instead, traffic to these containers needs to be temporarily removed. After the service recovers and the status is reset to valid, these nodes will still serve as valid canary containers. The canary container status in the container list and its replicas will be changed to invalid. For example... Figure 6 As shown, the process of changing the status of image releases, daily restarts, and service crashes can be achieved by modifying the status (valid status value) of the grayscale containers in the container list.

[0066] In practical applications, container resources are hosted on a big data platform for unified resource scheduling. When traffic changes significantly, container scaling up or down frequently occurs. Additionally, existing container resources may be migrated due to platform resource integration. Generally, these changes are imperceptible to the business. However, in this application, the server responsible for traffic distribution manages the entire process of gray-scale containers from creation to destruction. If a container that is automatically changed is a gray-scale machine, it can be dynamically taken offline and replaced with a valid node to supplement the gray-scale list, ensuring that only valid gray-scale containers are used online. The server scaling process is as follows: Figure 7 As shown, the grayscale machine will scale up proportionally according to the expansion ratio, while the scaling down process is as follows: Figure 8 As shown, the capacity is reduced proportionally according to the same reduction ratio. This ensures consistency in the processing.

[0067] In an exemplary embodiment, selecting a container of a corresponding proportion from each container list as a grayscale container and marking the grayscale container in the container list includes: finding the grayscale proportion of different data processing services; selecting a container of a corresponding proportion from each container list as a grayscale container based on the grayscale proportion, and marking the grayscale container in the container list.

[0068] For example, since the system call chain involves numerous services at each module layer, and each service consumes different amounts of resources, it's not feasible to simply set a fixed value or ratio for each service's grayscale ratio. The grayscale ratio must match the traffic of the experimental users. Therefore, this application requires a dynamic approach to pre-set the grayscale ratios of different data processing services, thereby reducing manual maintenance workload. After setting the corresponding grayscale ratio for each data processing service, when selecting grayscale containers, containers with the corresponding ratio can be selected from each container list based on the grayscale ratio, and these grayscale containers are marked in the container list. In this embodiment, selecting appropriate grayscale containers according to the grayscale ratio can effectively select a number of grayscale containers that match the input traffic, thereby effectively balancing system stability and processing performance.

[0069] In an exemplary embodiment, step 207 includes: identifying a first type of grayscale container and a second type of grayscale container in the grayscale container of the service processing layer; processing traffic data carrying grayscale labels through the first type of grayscale container to obtain the traffic data processing result of the first type of grayscale container; and, if the traffic data processing result indicates that there are no abnormalities in the processing, processing the traffic data carrying grayscale labels through the second type of grayscale container.

[0070] The first type of grayscale container refers to a container instance that performs initial processing of grayscale traffic, while the second type of grayscale container is a container instance that performs full processing of grayscale traffic. Both types of grayscale containers are suitable for performing staged processing of grayscale traffic.

[0071] For example, since a grayscale container is a division of all containers under a service according to a certain ratio, such as setting a 10% grayscale ratio for a service with 100 container instances, the container list would have 10 grayscale containers. To further reduce the risk of user experience degradation due to program crashes during the grayscale experiment, the grayscale containers in the container list can be further divided into two stages: a first-class grayscale container and a second-class grayscale container. When a grayscale container enters and starts processing traffic data, the traffic data carrying grayscale tags can first be processed by the first-class grayscale container, and the processing results of the first-class grayscale container can be collected. If the processing results are normal, indicating that the data processing service is not abnormal, the second-class grayscale container in the second stage can then process the traffic data carrying grayscale tags. In one embodiment, such as... Figure 9 As shown, at the start of the experiment, a 5-minute period was set as the first phase. During this phase, the gray-scale traffic entering the system was only distributed to the first type of gray-scale containers. After 5 minutes of verification, the gray-scale traffic was then distributed to the second type of gray-scale containers. To ensure service availability, the first type of gray-scale containers consisted of two containers, allowing for daily image deployment and other operations with two replicas. The second type of gray-scale containers comprised all five gray-scale containers, including the first type. In this embodiment, the phased processing of gray-scale traffic using two types of gray-scale containers further improves the stability of the gray-scale traffic processing.

[0072] In one exemplary embodiment, step 207 includes: finding the latest container list of the data processing service; selecting a target grayscale container from the container list; and calling the target grayscale container to process the traffic data carrying grayscale labels.

[0073] For example, when it is determined that the traffic data carries a grayscale label, the corresponding grayscale container can be selected to perform the corresponding data processing operation. The container list changes continuously as the container status changes. To ensure the accuracy of data processing, the latest container list can be read first, and then grayscale containers with valid status can be filtered from the grayscale containers marked in this list. One of these grayscale containers can be selected as the target grayscale container, and the target grayscale container can be called to process the traffic data carrying the grayscale label. In one embodiment, for the selection process of the target grayscale container, the real-time load status table of grayscale containers can be read, and then the container with the lowest load can be selected as the target grayscale container. In this embodiment, by searching the latest container list to select the target grayscale container to perform the corresponding traffic data processing operation, the efficiency and accuracy of traffic data processing can be effectively guaranteed, and the stability of the system can be improved.

[0074] In an exemplary embodiment, the method further includes: if no grayscale traffic is detected in the traffic data within a preset time interval, randomly selecting a corresponding proportion of pseudo-grayscale traffic from the traffic data and assigning pseudo-grayscale labels to the pseudo-grayscale traffic; if the service processing layer performs data processing services on the traffic data, identifying the data labels carried by the traffic data; identifying the container load status of the grayscale container in the service processing layer; and if the container load status indicates that there is a grayscale container in an unoverloaded state, calling the unoverloaded grayscale container to process the traffic data carrying pseudo-grayscale labels.

[0075] For example, during experimental releases, configuration modifications, or deployments, grayscale traffic continuously requests grayscale instances at each layer. However, under normal circumstances, when there is no grayscale traffic, grayscale instances remain idle, potentially wasting computing resources. Therefore, a pseudo-grayscale mechanism can be used to prevent this waste. Specifically, a preset time interval can be used to detect whether the system is in a grayscale experimental phase. This preset time interval can be set according to the experimental time interval of the specific version update phase of the service system. For example, during the system's experimental phase, if a portion of traffic is selected as grayscale traffic every 30 minutes, the preset time interval can be set to 40 minutes. If no grayscale traffic is detected within the preset time interval, it indicates that the system is not currently in an experimental phase, and the grayscale container can be used for processing normal system traffic. Therefore, a corresponding proportion of pseudo-grayscale traffic can be randomly selected from the traffic data and given a pseudo-grayscale label. Similar to grayscale traffic, pseudo-grayscale traffic enters the grayscale container to provide corresponding data processing services, also continuously requesting grayscale instances at each layer. However, during the processing of pseudo-grayscale traffic, it doesn't 100% permeate the grayscale container instance. If the grayscale container instance is overloaded, some pseudo-grayscale traffic will be diverted to the main disk instance. When the container load status indicates that all grayscale containers are overloaded, the main disk container in the service processing layer is called to process the traffic data carrying pseudo-grayscale tags. Because this pseudo-grayscale traffic is essentially still normal traffic from the main disk, importing this pseudo-grayscale traffic into the main disk container does not pose an online risk. In this embodiment, by setting pseudo-grayscale tags, the utilization rate of container computing resources in the system can be effectively improved.

[0076] In an exemplary embodiment, the method further includes: finding a list of containers for the next-layer data processing service of the data processing service; determining the proportion of grayscale containers in the next-layer data processing service based on the container list; if the proportion of grayscale containers is lower than a proportion threshold, processing the disk traffic (excluding pseudo-grayscale traffic) in the traffic data through a disk container in the next-layer data processing service; if the proportion of grayscale containers is higher than or equal to the proportion threshold, processing the disk traffic (excluding pseudo-grayscale traffic) in the traffic data through a disk container and grayscale containers in the next-layer data processing service.

[0077] For example, the solution in this application also provides methods for optimizing specific scenarios during the service process. For instance, in a general scenario, large-scale traffic will continuously request large-scale instances at each layer. However, this is an exception for small cluster scenarios. To ensure service stability, each service generally requires at least two gray-scale container instances. However, some small cluster scenarios only have two or three container instances, so there may only be one large-scale container instance or even none at all. Gray-scale traffic and pseudo-gray-scale traffic only account for a small portion, while the majority is large-scale traffic. Therefore, the processing of normal traffic can be discussed in categories. When processing normal large-scale traffic, the container list of the next layer of data processing services is searched; the proportion of gray-scale containers in the next layer of data processing services is determined based on the container list. Since the proportion of gray-scale containers in small clusters will be much higher than in normal cases, the proportion of gray-scale containers in the next layer of data processing services is first obtained through routing plugins, and then the proportion of gray-scale containers is compared with the proportion threshold to determine whether to import the normal traffic of the large-scale traffic into the gray-scale containers for processing. When the proportion of grayscale containers is below the threshold, the next-layer data processing service uses large-scale containers to process the large-scale traffic data, excluding pseudo-grayscale traffic. When the proportion of grayscale containers is above or equal to the threshold, the next-layer data processing service uses both large-scale containers and grayscale containers to process the large-scale traffic data, excluding pseudo-grayscale traffic. This allows normal large-scale traffic to run on grayscale containers, preventing a shortage of regular containers. See details in [link to relevant documentation]. Figure 10 As shown, a threshold of 30% is set. Services A and C are non-small cluster scenarios, where the proportion of gray-scale containers is less than 30%. Pseudo-gray-scale traffic and normal traffic are respectively imported into gray-scale containers and large-scale containers for processing. Service B is a small cluster scenario, where the proportion of gray-scale containers is 50%, higher than 30%. Large-scale traffic output from service A needs to be processed jointly by large-scale containers and gray-scale containers. In addition to optimizations for small cluster scenarios, this application also provides optimization solutions for abnormal gray-scale container scenarios. For example, for daily migrations and restarts within gray-scale containers, the proportion of unavailable gray-scale containers can be automatically monitored. When the proportion of unavailable gray-scale machines in the system exceeds an abnormal threshold (e.g., an abnormal threshold can be set to 30%, calculated as: unavailable gray-scale machines / all gray-scale machines), pseudo-gray-scale traffic will be automatically adjusted to ordinary large-scale containers, thus preventing abnormal pseudo-gray-scale traffic. It is particularly important to note that the optimizations for the aforementioned special scenarios only affect the service processing layer where an anomaly occurs. Figure 10The optimization methods for small cluster scenarios only involve component B, which belongs to the small cluster scenario. For normal components A and C, the pseudo-grayscale traffic / mainstream traffic distribution process is performed according to the traffic distribution method described above. In this embodiment, the proportion of grayscale containers is used to manage the small cluster scenario, thereby achieving the allocation and processing of normal mainstream traffic. This can effectively improve the utilization efficiency of computing resources and enhance system stability.

[0078] In an exemplary embodiment, the traffic data includes recommendation system traffic data. Step 201 includes: identifying the object identifier of the recommendation system traffic data at the service access layer; selecting a target grayscale experiment object based on the object identifier, and adding a grayscale object label to the target grayscale experiment object. Step 203 includes: identifying grayscale traffic carrying grayscale object labels in the traffic data based on the traffic source, and assigning grayscale labels to the grayscale traffic.

[0079] For example, this application can be specifically applied to the traffic distribution process of a recommendation system. Taking a video recommendation system as an example, a user submits a request on the terminal to obtain a list of recommended videos. This request generally needs to go through the access layer, engine layer, recall layer, fine ranking layer, and re-ranking layer before finally returning the user with the best videos in this recommendation. Therefore, in order to improve the user's viewing experience, continuous training is required to continuously improve the metrics. For example, improving the click-through rate during the fine ranking process will be experimented with initiating an experiment at the fine ranking layer. After the service access layer detects the change of initiating the experiment, it will mark a portion of the imported traffic as target gray-scale experimental objects, and then add gray-scale object tags to these target gray-scale experimental objects. During the traffic marking process, gray-scale traffic carrying gray-scale object tags can be identified based on the traffic source, and gray-scale tags can be assigned to gray-scale traffic.

[0080] This application also provides an application scenario, which is illustrated by taking the above-mentioned traffic distribution method as an example. The traffic distribution method specifically includes:

[0081] In the field of video recommendation systems, when an object requests a list of videos, the request typically goes through an access layer, an engine layer, a recall layer, a fine-ranking layer, and a re-ranking layer, finally returning the requesting object with the highest-quality videos in this recommendation. To improve the user experience, continuous training is needed to continuously improve metrics. This can be achieved through experiments to improve and adjust the recommendation system. The experiment process is divided into two phases: In the first phase, requests from these users are prioritized and routed to a portion of containers in each layer for processing. This phase lasts for 30 minutes or several hours as needed. Once the traffic in some containers has been verified to be normal, the first phase ends. The second phase starts automatically, at which point the traffic from these users is fully distributed (here, "fully" means that requests from some users are received by all containers, not all users' requests). Finally, these user requests are distributed to all machines in the recommendation system for processing, and the process runs for several days for final verification. However, even a small percentage of traffic requires processing from all machines in the system, which cannot guarantee system stability. Therefore, the traffic control method of this application can be used to distribute traffic in the recommendation system content. The traffic distribution method of this application is applied to the scheduling process between backend services. Each backend service relies on the grayscale container list provided by this application to schedule and route traffic, thereby isolating grayscale traffic from normal mains traffic and improving system stability.

[0082] First, this application requires constructing a container list. This can be achieved by obtaining container address data for data processing services through service discovery; generating container lists for different data processing services based on the container address data; determining the grayscale ratios for different data processing services; selecting containers from each container list according to the grayscale ratio as grayscale containers and marking them in the container list; then generating a list copy of the container list with the grayscale container markings; storing the container list and list copy based on a strong consistency mechanism to reduce the probability of failures caused by single points of failure. Furthermore, a node validity mechanism can be used to monitor container status by obtaining state change requests for data processing services; identifying the state change type of the state change request; updating the grayscale container markings in the container list and list copy when the state change type is a grayscale container change; and changing the grayscale container status in the container list and list copy to an invalid state when the state change type is a service change. The system stability can be further improved by using a layered mechanism for grayscale containers. This involves dividing grayscale containers into a first type and a second type. When processing traffic data, the first and second types of grayscale containers in the service processing layer are first identified. Traffic data carrying grayscale labels is processed through the first type of grayscale container to obtain the traffic data processing result for the first type of grayscale container. If the traffic data processing result indicates that there are no abnormalities, the traffic data carrying grayscale labels is processed through the second type of grayscale container. Furthermore, to improve the utilization of computing resources, a portion of traffic data can be extracted and pseudo-grayscale labeled during non-experimental phases. If no grayscale traffic is detected within a preset time interval, a corresponding proportion of pseudo-grayscale traffic is randomly selected from the traffic data and assigned a pseudo-grayscale label. When the service processing layer executes data processing services for the traffic data, the data labels carried by the traffic data are identified. The container load status of the grayscale containers in the service processing layer is also identified. If the container load status indicates that there are grayscale containers in an unloaded state, the unloaded grayscale containers are called to process the traffic data carrying pseudo-grayscale labels. If the container load status indicates that all grayscale containers are overloaded, the large-scale container in the service processing layer is called to process the traffic data carrying pseudo-grayscale labels. Finally, for small cluster services, since the number of containers for this type of service is relatively small, traffic optimization can be performed by identifying the proportion of grayscale containers. If the proportion of grayscale containers is lower than a threshold, it indicates that the next layer does not belong to a small cluster scenario. Therefore, in the next layer data processing service, the large-scale container directly processes the large-scale traffic data excluding pseudo-grayscale traffic. For small cluster scenarios where the proportion of grayscale containers is higher than or equal to the threshold, the large-scale traffic, excluding pseudo-grayscale traffic, can be processed in the next layer of data processing service through large-scale containers and grayscale containers.

[0083] In one embodiment, the traffic distribution process of the traffic distribution method of this application is specifically as follows: Figure 11 As shown, it includes:

[0084] Step 1102: Obtain container address data for data processing services through service discovery. Step 1104: Generate a container list for different data processing services based on the container address data. Step 1106: Find the grayscale ratio for different data processing services. Step 1108: Based on the grayscale ratio, select containers with the corresponding ratio from each container list as grayscale containers and mark them in the container list. Step 1110: Identify the traffic source of traffic data at the service access layer. Step 1112: Identify grayscale traffic in the traffic data based on the traffic source and assign grayscale labels to the grayscale traffic. Step 1114: When executing data processing services for traffic data at the service processing layer, identify the data labels carried by the traffic data. Step 1116: Identify the first type and second type of grayscale containers in the grayscale containers of the service processing layer. Step 1118: Process the traffic data carrying grayscale labels through the first type of grayscale containers to obtain the traffic data processing result for the first type of grayscale containers. Step 1120: If there are no abnormalities in the characterization of the traffic data processing results, process the traffic data carrying grayscale labels through the second type of grayscale container.

[0085] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0086] Based on the same inventive concept, this application also provides a traffic distribution device for implementing the traffic distribution method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more traffic distribution device embodiments provided below can be found in the limitations of the traffic distribution method described above, and will not be repeated here.

[0087] In one exemplary embodiment, such as Figure 12 As shown, a traffic distribution device is provided, comprising:

[0088] Traffic identification module 1201 is used to identify the traffic source of traffic data at the service access layer.

[0089] The label assignment module 1203 is used to identify grayscale traffic in traffic data based on traffic source and assign grayscale labels to grayscale traffic.

[0090] The tag recognition module 1205 is used to identify the data tags carried by traffic data when performing data processing services on traffic data in the service processing layer.

[0091] Container call module 1207 is used to call the grayscale container of the service processing layer to process traffic data carrying grayscale labels.

[0092] In one embodiment, a list generation module is further included, configured to: obtain container address data of data processing services through service discovery; generate a container list of different data processing services based on the container address data; select a corresponding proportion of large-scale containers from each container list as grayscale containers, and mark the grayscale containers in the container list.

[0093] In one embodiment, the system further includes a list storage module, configured to: generate a list copy of the container list after adding grayscale container tags; and store the container list and the list copy based on a strong consistency mechanism.

[0094] In one embodiment, a state update module is further included, configured to: obtain a state change request for the data processing service; identify the state change type of the state change request; update the grayscale container markers of the container list and list copy when the state change type is a grayscale container change type; and change the grayscale container status of the container list and list copy to an invalid state when the state change type is a service change type.

[0095] In one embodiment, the list generation module is further configured to: find the grayscale ratio of different data processing services; select containers with the corresponding ratio from each container list as grayscale containers based on the grayscale ratio, and mark the grayscale containers in the container list.

[0096] In one embodiment, the container invocation module 1207 is specifically used to: identify a first type of grayscale container and a second type of grayscale container in the grayscale container of the service processing layer; process traffic data carrying grayscale labels through the first type of grayscale container to obtain the traffic data processing result of the first type of grayscale container; and, if the traffic data processing result indicates that there are no abnormalities in the processing, process the traffic data carrying grayscale labels through the second type of grayscale container.

[0097] In one embodiment, the container invocation module 1207 is specifically used to: find the latest container list of the data processing service; select the target grayscale container in the container list, and invoke the target grayscale container to process the traffic data carrying grayscale labels.

[0098] In one embodiment, a pseudo-grayscale processing module is further included, configured to: randomly select a corresponding proportion of pseudo-grayscale traffic from the traffic data and assign pseudo-grayscale labels to the pseudo-grayscale traffic when no grayscale traffic is detected in the traffic data within a preset time interval; identify the data labels carried by the traffic data when the data processing service of the traffic data is executed in the service processing layer; identify the container load status of the grayscale container in the service processing layer; and when the container load status indicates that there is a grayscale container in an unoverloaded state, call the unoverloaded grayscale container to process the traffic data carrying pseudo-grayscale labels.

[0099] In one embodiment, the pseudo-grayscale processing module is further configured to: when the container load status indicates that all grayscale containers are in an overloaded state, invoke the large-scale container of the service processing layer to process the traffic data carrying pseudo-grayscale labels.

[0100] In one embodiment, the container invocation module 1207 is further configured to: find the container list of the next-layer data processing service of the data processing service; determine the proportion of grayscale containers in the next-layer data processing service based on the container list; if the proportion of grayscale containers is lower than the proportion threshold, process the disk traffic (excluding pseudo-grayscale traffic) in the traffic data through the disk container in the next-layer data processing service; if the proportion of grayscale containers is higher than or equal to the proportion threshold, process the disk traffic (excluding pseudo-grayscale traffic) in the traffic data through the disk container and grayscale containers in the next-layer data processing service.

[0101] In one embodiment, the traffic data includes recommendation system traffic data; the traffic identification module 1201 is specifically used to: identify the object identifier of the recommendation system traffic data at the service access layer; select target grayscale experimental objects based on the object identifier, and add grayscale object tags to the target grayscale experimental objects. The tag assignment module 1203 is specifically used to: identify grayscale traffic carrying grayscale object tags in the traffic data based on the traffic source, and assign grayscale tags to the grayscale traffic.

[0102] Each module in the aforementioned traffic distribution device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0103] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 13 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores data such as container lists required for the traffic distribution process. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a traffic distribution method.

[0104] Those skilled in the art will understand that Figure 13 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0105] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0106] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0107] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including 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 the above method embodiments.

[0108] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0109] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0110] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0111] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A traffic distribution method, characterized in that, The method includes: Identify the traffic source of traffic data at the service access layer; Identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale labels to the grayscale traffic; When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data; The grayscale container of the service processing layer is invoked to process traffic data carrying grayscale labels.

2. The method according to claim 1, characterized in that, The method further includes: Obtain the container address data of the data processing service through service discovery; A list of containers for different data processing services is generated based on the container address data. Select a large container of the corresponding proportion from each of the container lists as a grayscale container, and mark the grayscale container in the container list.

3. The method according to claim 2, characterized in that, The method further includes: Generate a list copy of the container list after adding grayscale container markers; The container list and its copy are stored based on a strong consistency mechanism.

4. The method according to claim 3, characterized in that, The method further includes: Get the status change request for the data processing service; Identify the type of state change requested in the state change request; If the state change type is a grayscale container change type, update the grayscale container flags of the container list and the list copy. If the state change type is a service change type, change the grayscale container status of the container list and the list copy to an invalid state.

5. The method according to claim 3, characterized in that, The step of selecting a large container of a corresponding proportion from each of the container lists as a grayscale container and marking the grayscale container in the container list includes: Find the grayscale ratio of different data processing services; Based on the grayscale ratio, a container with the corresponding ratio is selected from each of the container lists as a grayscale container, and the grayscale container is marked in the container list.

6. The method according to claim 1, characterized in that, The process of calling the grayscale container in the service processing layer to process traffic data carrying grayscale tags includes: Identify the first type of grayscale container and the second type of grayscale container in the grayscale container of the service processing layer; The traffic data carrying grayscale labels is processed by the first type of grayscale container to obtain the traffic data processing result of the first type of grayscale container; If the traffic data processing results indicate no abnormalities, the traffic data carrying grayscale labels is processed using the second type of grayscale container.

7. The method according to claim 1, characterized in that, The process of calling the grayscale container in the service processing layer to process traffic data carrying grayscale tags includes: Find the latest container list for the data processing service; Select the target grayscale container from the container list, and call the target grayscale container to process the traffic data carrying grayscale labels.

8. The method according to claim 1, characterized in that, The method further includes: If no grayscale traffic is detected in the traffic data within a preset time interval, a corresponding proportion of pseudo-grayscale traffic is randomly selected from the traffic data, and a pseudo-grayscale label is assigned to the pseudo-grayscale traffic. When the service processing layer performs data processing services on the traffic data, it identifies the data tags carried by the traffic data; Identify the container load status of the grayscale container in the service processing layer; When the container load status indicates that there is a grayscale container in an unloaded state, the unloaded grayscale container is invoked to process the traffic data carrying pseudo-grayscale labels.

9. The method according to claim 8, characterized in that, The method further includes: When the container load status indicates that all grayscale containers are overloaded, the large container of the service processing layer is invoked to process the traffic data carrying pseudo-grayscale labels.

10. The method according to claim 9, characterized in that, The method further includes: Find the list of containers for the next layer of data processing services of the aforementioned data processing service; The proportion of grayscale containers for the next-level data processing service is determined based on the container list. When the proportion of the grayscale container is lower than the proportion threshold, in the next layer of data processing service, the large disk traffic in the traffic data, excluding the pseudo grayscale traffic, is processed through the large disk container. When the proportion of the grayscale container is higher than or equal to the proportion threshold, in the next layer of data processing service, the large disk traffic in the traffic data, excluding the pseudo grayscale traffic, is processed through the large disk container and the grayscale container.

11. The method according to any one of claims 1 to 10, characterized in that, The traffic data includes recommendation system traffic data; The traffic sources identified at the service access layer include: Identify the object identifier of the recommendation system traffic data at the service access layer; Select a target grayscale experimental object based on the object identifier, and add a grayscale object label to the target grayscale experimental object; The step of identifying grayscale traffic in the traffic data based on the traffic source and assigning grayscale labels to the grayscale traffic includes: Based on the traffic source, identify grayscale traffic carrying grayscale object labels in the traffic data, and assign grayscale labels to the grayscale traffic.

12. A flow distribution device, characterized in that, The device includes: The traffic identification module is used to identify the source of traffic data at the service access layer; The tag assignment module is used to identify grayscale traffic in the traffic data based on the traffic source, and assign grayscale tags to the grayscale traffic; The tag recognition module is used to identify the data tags carried by the traffic data when the data processing service of the traffic data is executed in the service processing layer; The container invocation module is used to invoke the grayscale container of the service processing layer to process traffic data carrying grayscale labels.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

Images