A flow scheduling method and device, electronic equipment and machine readable storage medium

Abstract

One or more embodiments of the present specification provide a traffic scheduling method and device, electronic equipment and machine readable storage medium. The method comprises: in response to a resource allocation event of a target service, determining a first target cloud computing resource to be scheduled from cloud computing resources carried by at least one edge cloud node, and creating a cloud computing instance for the target service based on the first target cloud computing resource; adding the cloud computing instance created for the target service on the same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and creating a logical service object for each cloud computing instance set of the target service; determining service capabilities that can be provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service; and scheduling service traffic of the target service by taking each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node.

Description

TECHNICAL FIELD

[0001] One or more embodiments of the present specification relate to the technical field of edge cloud computing, and in particular to a traffic scheduling method and device, an electronic device, and a machine-readable storage medium. BACKGROUND

[0002] A CDN (Content Delivery Network) is an overlay network built on the Internet infrastructure, which can provide static content acceleration, dynamic content acceleration, network security protection, and other functions for various Internet digital content (including but not limited to video, text, streaming media, and other content).

[0003] In the face of large traffic scenarios, traditional CDN vendors usually use resource hard isolation to scale resources for services. That is, different services are deployed on different physical devices, so that different services use independent machine resources, and a certain number of redundant machine resources are reserved for each service; when a certain service bursts, the redundant machine resources reserved for the service can be used for scaling.

[0004] In related technologies, if too many redundant machine resources are reserved for each service, there may be problems of low resource utilization, long construction period of the machine room, and high construction cost of the machine room; and if too few redundant machine resources are reserved for each service, resource scaling cannot be quickly realized when the traffic of a large traffic service suddenly increases, affecting the service experience of users. SUMMARY

[0005] The present application provides a traffic scheduling method, which is applied to a center control device in a distributed edge cloud computing system; the edge cloud computing system further includes at least one edge cloud node interconnected with the center control device; the method includes:

[0006] In response to a resource allocation event of a target service, determining a first target cloud computing resource to be scheduled from cloud computing resources carried by the at least one edge cloud node, and scheduling the first target cloud computing resource to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resource;

[0007] Adding cloud computing instances created for the target service on the same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and creating a logical service object for each cloud computing instance set of the target service;

[0008] determine service capabilities provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service;

[0009] schedule service traffic of the target service according to the service capabilities provided by each logical service object of the target service.

[0010] The application further provides a traffic scheduling device, which is applied to a center management and control equipment in a distributed edge cloud computing system; the edge cloud computing system further comprises at least one edge cloud node interconnected with the center management and control equipment; the device comprises:

[0011] a first scheduling unit configured to determine first target cloud computing resources to be scheduled from cloud computing resources carried by the at least one edge cloud node in response to a resource allocation event of a target service, and schedule the first target cloud computing resources to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resources;

[0012] a creating unit configured to add cloud computing instances created for the target service on a same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and create logical service objects for each cloud computing instance set of the target service;

[0013] a determining unit configured to determine service capabilities provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service;

[0014] a second scheduling unit configured to schedule service traffic of the target service according to the service capabilities provided by each logical service object of the target service, by taking each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node.

[0015] The application further provides an electronic device, which comprises a communication interface, a processor, a memory and a bus, the communication interface, the processor and the memory are interconnected through the bus;

[0016] The memory stores machine readable instructions, and the processor executes the above method by invoking the machine readable instructions.

[0017] The application further provides a machine readable storage medium, which stores machine readable instructions, and the machine readable instructions realize the above method when invoked and executed by a processor.

[0018] Through the above embodiments, in the case where expansion is needed for a target service, the first target cloud computing resource for expanding the target service can be determined from the cloud computing resources of the whole network, the cloud computing instance is created for the target service based on the first target cloud computing resource, and the several cloud computing instances created for the target service on the same edge cloud node can be created as a whole to create the corresponding logical service object. Subsequently, each logical service object of the target service can be taken as the traffic scheduling object corresponding to each edge cloud node, and the service traffic of the target service is scheduled according to the service capability provided by each logical service object of the target service. Through the above manner, in the scenario of large traffic service, the resource soft isolation manner is adopted to realize the resource scheduling between different services, allowing different services to run on the same machine resource, which can quickly perform elastic expansion on demand, guarantee better service performance and user experience, and also can improve the resource utilization rate of the existing machine resource, saving the construction cost of the edge infrastructure. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present specification, and other drawings can also be obtained according to these drawings without creative labor for those skilled in the art.

[0020] Figure 1 Fig. 1 is a system architecture schematic diagram of an edge cloud computing system according to an example embodiment;

[0021] Figure 2 Fig. 2 is a resource scheduling schematic diagram according to an example embodiment;

[0022] Figure 3 Fig. 3 is a flowchart of a traffic scheduling method according to an example embodiment;

[0023] Figure 4 Fig. 4 is another resource scheduling schematic diagram according to an example embodiment;

[0024] Figure 5 Fig. 5 is a structural schematic diagram of an electronic device in which a traffic scheduling apparatus according to an example embodiment is located;

[0025] Figure 6 Fig. 6 is a block diagram of a traffic scheduling apparatus according to an example embodiment. DETAILED DESCRIPTION

[0026] In order for those skilled in the art to better understand the technical solutions in the specification, the technical solutions in the specification will be clearly and completely described below in combination with the drawings in the specification. Obviously, the described embodiments are only some of the embodiments of the specification, not all. Based on the embodiments in the specification, all other embodiments obtained by those of ordinary skill in the art without creative labor should be within the scope of protection of the specification.

[0027] It should be noted that the steps of the corresponding method are not necessarily performed in the order shown and described in the specification in other embodiments. In some other embodiments, the steps included in the method can be more or less than described in the specification. In addition, a single step described in the specification can be divided into multiple steps for description in other embodiments, and multiple steps described in the specification can be combined into a single step for description in other embodiments.

[0028] CDN (Content Delivery Network, content delivery network) is an overlay network built on the Internet infrastructure, which can provide static content acceleration, dynamic content acceleration, network security protection and other functions for various Internet digital content (including but not limited to video, text, streaming media and other content).

[0029] On the one hand, by deploying edge servers in a distributed manner, CDN can cache various Internet digital content from the Internet center to edge service devices close to the user side, thereby reducing the latency of user access to content and reducing the traffic through the backbone network, and thus improving the service quality.

[0030] On the other hand, in the face of DDoS (Distributed Denial of Service, Distributed Denial of Service attack) and other network attacks, CDN can avoid the attack on the content source station to be unable to normally provide content based on the widely distributed edge nodes and combined with attack traffic detection and traffic cleaning technology, thereby protecting the normal and smooth operation of various Internet services.

[0031] In the face of large traffic business scenarios, traditional CDN vendors usually use resource hard isolation to expand and shrink the resources for business. That is, different businesses are deployed on different physical devices, so that different businesses use independent machine resources, and a certain number of redundant machine resources are reserved for each business; when a certain business bursts, the redundant machine resources reserved for the business can be used for expansion.

[0032] For example, in scenarios involving high-traffic services such as live streaming and video-on-demand, separate dedicated data centers can be built for live streaming and video-on-demand services. Dedicated redundant machine resources can be reserved in the dedicated live streaming data center, and similarly, dedicated redundant machine resources can be reserved in the dedicated video-on-demand data center. If the traffic for live streaming suddenly increases, the reserved redundant machine resources in the dedicated live streaming data center can be used for capacity expansion.

[0033] Therefore, in the embodiments shown above, if too many redundant machine resources are reserved for each service, there may be problems such as low resource utilization, long data center construction cycle, and high data center construction cost; while if too few redundant machine resources are reserved for each service, it will be impossible to quickly expand resources when the business traffic of high-traffic services surges, affecting the user's business experience.

[0034] In view of this, this specification aims to propose a technical solution for resource scheduling and traffic scheduling based on edge cloud, so as to realize the dynamic co-location of different services, that is, to enable tasks of different services to be scheduled to the same machine resources.

[0035] Edge cloud is a cloud computing platform built on edge infrastructure, based on the core of cloud computing technology and the capabilities of edge computing. Edge cloud can also be called edge cloud computing system, edge cloud computing platform, etc., but this manual does not make a specific definition.

[0036] Edge cloud is a relative concept, distinct from central cloud or traditional cloud computing platforms. In practical applications, central cloud or traditional cloud computing platforms can include data centers with large-scale resources and centralized locations; while edge cloud nodes cover a wider network range, are closer to the terminals, and have a larger number of nodes, but each edge cloud node has a smaller resource scale.

[0037] Please see Figure 1 , Figure 1 This is a schematic diagram of the system architecture of an edge cloud computing system, illustrating an exemplary embodiment. (As shown...) Figure 1 As shown, the edge cloud computing system 100 may include a central control device 101, an edge cloud node 102, an edge cloud node 103, and an edge cloud node 104; the central control device 101 may be interconnected with the edge cloud node 102, the edge cloud node 103, and the edge cloud node 104 respectively.

[0038] It should be noted that, Figure 1The specification only illustrates three edge cloud nodes as an example, which does not imply any special limitation on this specification; in practical applications, the edge cloud computing system 100 may include at least one edge cloud node interconnected with the central management device 101. The central management device 101 may be deployed in one or more cloud computing data centers, or in one or more traditional data centers, and this specification does not impose any special limitations on this.

[0039] Each edge cloud node in the at least one edge cloud node may include a series of edge infrastructures; the edge infrastructure may specifically include, but is not limited to: distributed data centers (DCs), wireless equipment rooms or clusters, operator communication networks, core network equipment, base stations, edge gateways, home gateways, computing devices or storage devices, and other edge devices and their corresponding network environments, etc. It should be noted that the location, cloud computing capabilities, and included edge infrastructure of different edge cloud nodes may be the same or different, and this specification does not impose any special limitations on this.

[0040] Among them, such as Figure 1 The terminal shown may include the demand side of cloud computing services. For example, the terminal may include a user terminal in the Internet or the Internet of Things.

[0041] In such Figure 1 In the system architecture of the edge cloud computing system shown, the central cloud corresponding to the central control device, the edge cloud formed by the at least one edge cloud node, and the terminal can form an end-to-end technical architecture of "cloud-edge-device three-body collaboration". In the edge cloud computing system 100, the central control device 101 can treat each edge cloud node (such as edge cloud node 102, edge cloud node 103, and edge cloud node 104) as control objects, and perform unified control in terms of resource scheduling, image management, instance control, operation and maintenance, network, and security, thereby placing cloud computing services in each edge cloud node for processing, and providing cloud computing services to users by leveraging the various resources carried by the edge cloud nodes. Since the edge cloud nodes are closer to the user terminals, the response latency of cloud computing services can be reduced, the pressure on the central cloud or traditional cloud computing platform corresponding to the edge cloud nodes can be reduced, bandwidth costs can be reduced, and the resource requirements of terminals for large connections, low latency, and high bandwidth in cloud computing scenarios can be met.

[0042] In addition, in such Figure 1In the system architecture of the edge cloud computing system shown, in addition to the center management device 101, the edge cloud computing system 100 can also include one or more edge management devices. The edge management device can be used to assist and cooperate with the center management device to manage and schedule the at least one edge cloud node, so as to improve the management and scheduling of the edge cloud node by the center management device, and thus achieve the purpose of fully utilizing edge resources. The edge management device can be deployed in one or more edge cloud nodes; specifically, each edge cloud node can include one or more resource devices, and the edge management device can be deployed on one or more resource devices; or, in addition to the resource devices, each edge cloud node can also include one or more special devices different from the resource devices, which are specially used to deploy edge management devices, and the edge management device can be deployed on one or more special devices. The edge management device can also be deployed together with the center management device, which is not specially limited in the present specification.

[0043] At present, more and more edge businesses are considering the upper edge cloud, and more local and suitable center applications running in edge cloud nodes are sinking to the edge, and the demand for edge cloud is increasing.

[0044] The scenario of large flow business usually has the characteristics of strong burst, regional distribution difference, and business flow this disappears the other. For example, in the process of sports event live broadcast, electronic mall promotion activity, online course live broadcast, and sudden hot event, the business flow of related business may surge. For another example, in the area with more online users, the business flow may be more, and in the area with less online users, the business flow is less, and the business flow of the same business is not uniform in regional distribution. For another example, in the process of accessing live broadcast business, users may not access on-demand business, and vice versa, so there may be a relationship between the needs of this disappearing the other between multiple businesses.

[0045] Therefore, the widely distributed edge infrastructure can be used to schedule the cloud computing resources carried by the edge cloud node, create cloud computing instances for each business on the edge cloud node, and the several cloud computing instances created for the same business on the same edge cloud node can be regarded as a whole to provide cloud computing service capability to the outside. It should be noted that through resource soft isolation, when the business flow of a single business surges, resource scheduling and flow scheduling can be quickly performed in the edge cloud computing system to realize on-demand dynamic expansion and contraction, thereby improving resource utilization under the premise of ensuring service quality.

[0046] In some possible implementation manners, the central management device in the edge cloud computing system can determine, in response to a resource allocation event of a target service, a first target cloud computing resource to be scheduled from cloud computing resources carried by at least one edge cloud node interconnected with the central device, and schedule the first target cloud computing resource to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resource; further, the central management device can add cloud computing instances created for the target service on the same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and create a logical service object for each cloud computing instance set of the target service; further, the central management device can determine service capabilities that can be provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service; further, the central management device can take each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node, and schedule service traffic of the target service according to the service capabilities that can be provided by each logical service object of the target service.

[0047] The target service can be any service that needs to be expanded in resource or service capability. The cloud computing resource can be a physical resource or a logical resource obtained by abstracting a physical resource in the edge cloud computing system, and the present specification does not limit the cloud computing resource. The cloud computing instance can include, but is not limited to, a container instance, a virtual machine instance, an application instance, and the like, and the present specification does not limit the cloud computing instance.

[0048] For example, please refer to Figure 2 , Figure 2 is a schematic diagram of resource scheduling according to an exemplary embodiment. As Figure 2 shown, the edge cloud node 102 runs cloud computing instances podA1 and podA2 created for a live service, and cloud computing instances podB1 and podB2 created for an on-demand service; the logical service object podgroupA.node102 is created for the cloud computing instance set {podA1, podA2} of the live service corresponding to the edge cloud node 102, and the logical service object podgroupB.node102 is created for the cloud computing instance set {podB1, podB2} of the on-demand service corresponding to the edge cloud node 102.

[0049] In response to a resource allocation event of the live service, the central management device 101 can determine that the live service needs to be scaled out, determine the first target cloud computing resource to be scheduled from the cloud computing resources carried by the edge cloud node 102, and schedule the first target cloud computing resource to create a cloud computing instance podA3 for the live service on the edge cloud node 102. Further, the cloud computing instance podA3 can be added to the cloud computing instance set of the live service corresponding to the edge cloud node 102, and a logical service object podgroupA.node102 of the cloud computing instance set {podA1, podA2, podA3} of the live service is created, that is, the cloud computing instance set of the live service corresponding to the logical service object podgroupA.node102 can be updated to {podA1, podA2, podA3}; further, the service capability that can be provided by the logical service object podgroupA.node102 of the live service can be determined according to the cloud computing resources allocated to the cloud computing instance podA1, the cloud computing instance podA2 and the cloud computing instance podA3 of the live service. Based on a similar process, the service capability that can be provided by the logical service object podgroupB.node102 of the on-demand service can also be determined according to the cloud computing resources allocated to the cloud computing instance podB1 and the cloud computing instance podB2 of the on-demand service.

[0050] In addition, based on a similar process, the service capability that can be provided by other logical service objects of the live service on other edge cloud nodes and the service capability that can be provided by other logical service objects of the on-demand service on other edge cloud nodes can also be determined, which will not be described one by one here. According to one or more embodiments exemplified by the edge cloud node 102 in the present specification, those skilled in the art can know the implementation manner of the technical solutions in the present specification on other edge cloud nodes without creative labor.

[0051] Further, the central management device 101 can schedule the service traffic of the live service according to the service capability that can be provided by each logical service object of the live service on each edge cloud node, taking each logical service object of the live service as a traffic scheduling object corresponding to each edge cloud node.

[0052] As can be seen, after the live streaming service cloud computing instance podA3 is added to the live streaming service cloud computing instance set corresponding to the edge cloud node 102, more cloud computing resources are allocated to the live streaming service on the edge cloud node 102, and thus the service capability provided by the logical service object podgroupA.node102 of the live streaming service is improved, and more service traffic can be accessed by the logical service object podgroupA.node102 of the live streaming service in the process of scheduling the service traffic of the live streaming service, and the on-demand online expansion of the live streaming service is realized.

[0053] As can be seen, in the technical solutions in the present specification, in the case where expansion is needed for a target service, the first target cloud computing resource for expanding the target service can be determined from the cloud computing resources of the whole network, the cloud computing instance for the target service is created based on the first target cloud computing resource, and the several cloud computing instances created for the target service on the same edge cloud node can be taken as a whole to create a corresponding logical service object, and the logical service objects of the target service can be taken as the traffic scheduling objects corresponding to the edge cloud nodes in the subsequent process, and the service traffic of the target service is scheduled according to the service capability provided by the logical service objects of the target service. In the above manner, in the scenario of a large-traffic service, the resource soft isolation is adopted to realize the resource scheduling between different services, allowing different services to run on the same machine resources, which can quickly and flexibly expand on demand, guarantee good service performance and user experience, and improve the resource utilization rate of the existing machine resources and save the construction cost of the edge infrastructure.

[0054] The present application will be described below through specific embodiments and in combination with specific application scenarios.

[0055] Please refer to Figure 3 , Figure 3 is a flowchart of a traffic scheduling method according to an exemplary embodiment. The traffic scheduling method can be applied to a center management and control device in a distributed edge cloud computing system; the edge cloud computing system can also include at least one edge cloud node interconnected with the center management and control device.

[0056] For example, the method can be applied to the center management and control device 101 in the edge cloud computing system 100 as shown in Figure 1 . Specifically, the scheduling system can be run on the center management and control device 101, and the method can be executed by the scheduling system.

[0057] As shown in Figure 3 , the traffic scheduling method can perform the following steps:

[0058] Step 302: in response to the resource allocation event of the target service, determining the first target cloud computing resource to be scheduled from the cloud computing resources carried by the at least one edge cloud node, and scheduling the first target cloud computing resource to create at least one cloud computing instance for the target service on the edge cloud node carrying the first target cloud computing resource.

[0059] For example, the target service can be a live streaming service. As shown in Figure 2 , in response to the resource allocation event of the live streaming service, the central management device 101 can determine that the live streaming service needs to be scaled out, determine the first target cloud computing resource to be scheduled from the cloud computing resources carried by the edge cloud node 102, and schedule the first target cloud computing resource to create a cloud computing instance podA3 for the live streaming service on the edge cloud node 102.

[0060] It should be noted that, in the embodiment as shown in Figure 2 , the first target cloud computing resource can be a cloud computing resource in an idle state (i.e., not allocated to any cloud computing instance) on the edge cloud node 102. Alternatively, the first target cloud computing resource can be a cloud computing resource on the edge cloud node 102 that has been allocated to a cloud computing instance of another service. Please refer to Figure 4 , Figure 4 is another schematic diagram of resource scheduling shown in an exemplary embodiment.

[0061] For another example, as shown in Figure 4 , the target service can be a live streaming service, in response to the resource allocation event of the live streaming service, the central management device 101 can determine that the live streaming service needs to be scaled out, determine the first target cloud computing resource to be scheduled from the cloud computing resources carried by the edge cloud node 102, and schedule the first target cloud computing resource to create a cloud computing instance podA3 for the live streaming service on the edge cloud node 102.

[0062] It should be noted that, in the embodiment as shown in Figure 2 , the first target cloud computing resource can be a cloud computing resource in an idle state (i.e., not allocated to any cloud computing instance) on the edge cloud node 102. Alternatively, the first target cloud computing resource can be a cloud computing resource on the edge cloud node 102 that has been allocated to a cloud computing instance of another service. Please refer to Figure 4

[0063] ​In the step 302, the target service can be any service that needs to be expanded in resource or service capability. Figure 2 and Figure 4 In the embodiments shown in the figures, only live services and on-demand services are exemplarily shown, which does not represent a special limitation to the present specification. In the present specification, one or more services can be simultaneously run on the same edge cloud node, which can specifically include but is not limited to live services, on-demand services, load balancing services, transcoding services, dynamic acceleration services, etc. Among them, the cloud computing instances of different services on the same edge cloud node, and the cloud computing instances created for different tenants of the same service, can be isolated through namespaces.

[0064] In the step 302, the resource allocation event of the target service can be understood as an event indicating that the target service needs to be expanded. For example, in response to receiving a user-initiated expansion instruction for the target service, or in response to detecting that the cloud computing resources allocated to the target service have an exception (machine failure, etc.), or monitoring that the resource usage of the cloud computing resources allocated to the target service is continuously high, it can be considered that the resource allocation event of the target service is detected.

[0065] In one embodiment shown, the central management device can combine resource portraits and service portraits to match resources and services, and then arrange services on suitable edge cloud nodes. In this case, determining the first target cloud computing resource to be scheduled from the cloud computing resources carried by the at least one edge cloud node can specifically include: obtaining the resource portrait of the edge cloud computing system and the service portrait of the target service; from the cloud computing resources carried by the at least one edge cloud node, determining the cloud computing resource whose resource features match the service features of the target service as the first target cloud computing resource to be scheduled.

[0066] The resource portrait can be used to describe the resource features of the cloud computing resources carried by each edge cloud node in the edge cloud computing system, such as static hardware configuration data such as CPU, memory, etc., and real-time data such as network coverage of the area where the cloud computing resources are located, etc. The service portrait can be used to describe the service features of the target service, such as the resource consumption of a single service request, the sensitivity of service quality to different network indicators, etc.

[0067] In some possible embodiments, the obtaining of the resource profile of the edge cloud computing system and the service profile of the target service can specifically include: collecting resource information of cloud computing resources carried by the at least one edge cloud node, and collecting service information of each cloud computing instance created on each edge cloud node for the target service; generating the resource profile of the edge cloud computing system according to the collected resource information, and generating the service profile of the target service according to the collected service information.

[0068] The resource information can specifically include but is not limited to resource configuration information of the cloud computing resources themselves, historical usage data of the cloud computing resources, and the like. The service data can specifically include but is not limited to resource distribution of the cloud computing resources allocated for the service, resource usage rate of the allocated cloud computing resources by the service, and the like.

[0069] For example, the monitoring data of each edge cloud node, such as the CPU water level, the memory water level, and the network state of each machine, can be collected periodically, feature extraction is performed on the collected detection data, a model is trained by means of machine learning in combination with data features and service requirements, and the service profile and the resource profile output by the model are obtained.

[0070] In an embodiment shown, CPU dynamic core binding can be supported, so as to avoid scheduling of service threads between different CPU cores, reduce overhead caused by thread scheduling, and ensure service performance of the target service. In this case, the target service can be a large-flow service; the cloud computing instance of the target service runs multiple threads of the target service; and the cloud computing resources can include CPU resources. The method can further include: performing core binding operation on the multiple threads of the target service running in the cloud computing instance of the target service and the CPU resources allocated for the cloud computing instance of the target service.

[0071] Step 304: adding the cloud computing instance created on the same edge cloud node for the target service to the cloud computing instance set of the target service corresponding to the edge cloud node, and creating a logical service object for each cloud computing instance set of the target service.

[0072] Step 306: determining service capabilities that can be provided by each logical service object of the target service according to the cloud computing resources allocated for each cloud computing instance of the target service.

[0073] For example, as shown in FIG. 4, the service capabilities of the target service can include service performance, service stability, service availability, and the like. Figure 2 or Figure 4As shown, after the cloud computing instance podA3 is created based on the first target cloud computing resource on the edge cloud node 102 for the live streaming service, the cloud computing instance podA3 can be added to the cloud computing instance set of the live streaming service corresponding to the edge cloud node 102, and a logical service object podgroupA.node102 of the cloud computing instance set {podA1, podA2, podA3} of the live streaming service is created, that is, the cloud computing instance set of the live streaming service corresponding to the logical service object podgroupA.node102 can be updated to {podA1, podA2, podA3}; further, the service capability that can be provided by the logical service object podgroupA.node102 of the live streaming service can be determined according to the cloud computing resources allocated to the cloud computing instance podA1, the cloud computing instance podA2 and the cloud computing instance podA3 of the live streaming service.

[0074] It can be seen that after the cloud computing instance podA3 of the live streaming service is added to the cloud computing instance set of the live streaming service corresponding to the edge cloud node 102, more cloud computing resources are allocated to the live streaming service on the edge cloud node 102, so that the service capability that can be provided by the logical service object podgroupA.node102 of the live streaming service is also improved, and further, more service traffic can be accessed by the logical service object podgroupA.node102 of the live streaming service in the process of scheduling the service traffic of the live streaming service, and the on-line expansion of the live streaming service on demand is realized.

[0075] Step 308: regarding each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node, scheduling the service traffic of the target service according to the service capability that can be provided by each logical service object of the target service.

[0076] For example, the center management and control device 101 can regard each logical service object of the live streaming service as a traffic scheduling object corresponding to each edge cloud node according to the service capability that can be provided by each logical service object of the live streaming service on each edge cloud node, and schedule the service traffic of the live streaming service.

[0077] In an embodiment shown, the scheduling of the service traffic of the target service according to the service capabilities provided by the respective logical service objects of the target service can specifically include: in response to a user-initiated access request for the target service, determining a target logical service object for providing services to the user from the respective logical service objects of the target service according to the service capabilities provided by the respective logical service objects of the target service; and delivering the access request to a target edge cloud node corresponding to the target logical service object, so that the cloud computing instance of the target service created on the target edge cloud node provides services to the user.

[0078] The object information of the logical service object can include, in addition to the service identifier, the node identifier of the edge cloud node where the logical service object is located, the object identifier of the logical service object, the instance identifier of the cloud computing instance corresponding to the logical service object, the oversellable ratio, the available resource amount, and the like, and can further include the region to which the logical service object belongs and the VIP group (Virtual IP address group) corresponding to the logical service object (which includes the pre-allocated VIP for each cloud computing instance).

[0079] For example, on the edge cloud node 102, the edge cloud node 103, and the edge cloud node 104, the logical service objects of the live streaming service created are podgroupA.node102, podgroupA.node103, and podgroupA.node104, respectively. In response to a user-initiated access request for the live streaming service, the central management device 101 can determine the target logical service object closest to the user as podgroupA.node102 from the respective logical service objects of the live streaming service according to the region to which the respective logical service objects of the live streaming service belong, so as to reduce the access latency; further, according to the VIP group corresponding to the target logical service object podgroupA.node102, the user-initiated access request for the live streaming service is delivered to the target edge cloud node 102 corresponding to the target logical service object podgroupA.node102, so that one or more cloud computing instances of the live streaming service created on the target edge cloud node 102 provide content services to the user.

[0080] In an embodiment shown, before scheduling the service traffic of the target service according to the service capabilities available to each logical service object of the target service, the method can further include: in response to the creation of each logical service object of the target service, marking each logical service object of the target service as a stop-service state, and performing service availability verification on each logical service object of the target service; and if each logical service object of the target service passes the service availability verification, updating each logical service object of the target service to an allow-service state.

[0081] Here, the creation of the logical service object can be understood as that the cloud computing instances corresponding to the logical service object are all ready for production, and each cloud computing instance has been assigned a VIP.

[0082] For example, in response to the creation of the logical service object podgroupA.node102 of the live broadcast service on the edge cloud node 102, the logical service object podgroupA.node102 can be marked as an Unready state, and cannot provide services to the outside, and a probe service can be triggered to perform service availability verification on the logical service object podgroupA.node102; if the logical service object podgroupA.node102 passes the service availability verification, the logical service object podgroupA.node102 can be updated to a Ready state, and can provide services to the outside. Based on a similar process, service availability verification can also be performed on other logical service objects of the target service, which will not be described here.

[0083] In this specification, in addition to performing elastic expansion of the target service on demand, the target service can also be elastically contracted on demand, so as to release cloud computing resources in a timely manner and improve resource utilization.

[0084] In this case, the method can further include: in response to a resource recycling event of the target service, determining a second target cloud computing resource to be scheduled from the cloud computing resources allocated to the cloud computing instances of the target service, and determining the reduced service capabilities of each logical service object of the target service according to the second target cloud computing resource; scheduling the service traffic of the target service according to the reduced service capabilities of each logical service object of the target service; and in response to the service traffic accessing the cloud computing instances created based on the second target cloud computing resource being less than a preset first threshold, recycling the second target cloud computing resource.

[0085] For example, as shown in FIG. 6, in response to the service traffic accessing the cloud computing instances created based on the second target cloud computing resource being less than the preset first threshold, the second target cloud computing resource can be recycled. Figure 4As shown, the target service can be an on-demand service, in response to a resource recovery event of the on-demand service, the central management device 101 can determine that it is necessary to scale down the on-demand service, determine the second target cloud computing resource to be scheduled from the cloud computing resources allocated to the on-demand service carried by the edge cloud node 102, and determine the service capability of the logical service object podgroupB.node102 of the on-demand service after reduction according to the second target cloud computing resource. According to the service capability of the logical service object podgroupB.node102 of the on-demand service after reduction, the central management device 101 can schedule the service traffic of the on-demand service; in response to the service traffic of the cloud computing instance podB2 being less than a preset first threshold, the instance specification of the cloud computing instance podB2 can be reduced, and the second target cloud computing resource can be recovered.

[0086] The resource recovery event of the target service can be understood as an event indicating that the target service needs to be scaled down. For example, in response to detecting a user-initiated scaling instruction for the target service, or in response to detecting a reduction configuration instruction for a specified logical service object of the target service, or in response to detecting a resource allocation event of another service, it can be considered that the resource recovery event of the target service is detected.

[0087] In an embodiment shown, if it is determined that a certain logical service object of the target service is no longer used as a traffic scheduling object, the logical service object can be deleted to completely release the resources occupied by the cloud computing instance corresponding to the logical service object.

[0088] In this case, the method can further include: in response to a logical service object deletion event of the target service, updating the target logical service object to be deleted from the allowed service state to the stopped service state, and switching the service traffic accessing the target logical service object to other logical service objects of the target service; in response to the cloud computing resource usage rate of the target logical service object being less than a preset second threshold, deleting a target cloud computing instance set corresponding to the target logical service object, recovering the cloud computing resources allocated to the cloud computing instances included in the target cloud computing instance set, and deleting the target logical service object.

[0089] For example, the target service can be an on-demand service, in response to a logical service object deletion event for the on-demand service, the target logical service object podgroupB.node102 to be deleted can be updated from the Ready state to the Unready state, and the service traffic accessing the target logical service object podgroupB.node102 can be switched to other logical service objects of the on-demand service on other edge cloud nodes; in response to the CPU water level of the target logical service object podgroupB.node102 being less than 5%, the cloud computing instance podB1 and the cloud computing instance podB2 corresponding to the target logical service object podgroupB.node102 can be deleted first, the cloud computing resources allocated to the cloud computing instance podB1 and the cloud computing instance podB2 are recovered, and then the target logical service object podgroupB.node102 is deleted.

[0090] Through the above technical solutions, in the case of needing to expand the target service, the first target cloud computing resource for expanding the target service can be determined from the cloud computing resources of the whole network, the cloud computing instance for the target service can be created based on the first target cloud computing resource, and the several cloud computing instances created for the target service on the same edge cloud node can be created as a whole to create the corresponding logical service object. Subsequently, each logical service object of the target service can be used as a traffic scheduling object corresponding to each edge cloud node, and the service traffic of the target service can be scheduled according to the service capability provided by each logical service object of the target service. Through the above method, in the scenario of a large amount of traffic, the resource soft isolation method is used to realize the resource scheduling between different services, allowing different services to run on the same machine resources. This can quickly and flexibly expand on demand, ensure good service performance and user experience, and also can improve the resource utilization rate of the existing machine resources and save the construction cost of the edge infrastructure.

[0091] Corresponding to the above-mentioned embodiments of the traffic scheduling method, the present specification also provides an embodiment of a traffic scheduling device.

[0092] Please refer to Figure 5 , Figure 5Fig. 1 is a hardware structure diagram of an electronic device in which a traffic scheduling apparatus shown in an exemplary embodiment is located. At the hardware level, the device includes a processor 502, an internal bus 504, a network interface 506, a memory 508, and a non-volatile memory 510, and can also include other required hardware. One or more embodiments of the present specification can be implemented in a software manner, such as reading a corresponding computer program from the non-volatile memory 510 into the memory 508 by the processor 502 and then running. Of course, in addition to the software implementation, one or more embodiments of the present specification do not exclude other implementation manners, such as a logic device or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or a logic device.

[0093] Please refer to Figure 6 , Figure 6 Fig. 2 is a block diagram of a traffic scheduling apparatus shown in an exemplary embodiment. The traffic scheduling apparatus can be applied to an electronic device as shown in Figure 5 , so as to implement the technical solutions of the present specification. The traffic scheduling apparatus can include:

[0094] A first scheduling unit 602 is configured to determine a first target cloud computing resource to be scheduled from cloud computing resources carried by at least one edge cloud node in response to a resource allocation event of a target service, and schedule the first target cloud computing resource to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resource;

[0095] A creating unit 604 is configured to add cloud computing instances created for the target service on the same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and create a logical service object for each cloud computing instance set of the target service;

[0096] A determining unit 606 is configured to determine service capabilities that can be provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service;

[0097] A second scheduling unit 608 is configured to schedule service traffic of the target service according to service capabilities that can be provided by each logical service object of the target service, by taking each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node.

[0098] In the present embodiment, the first scheduling unit 602 is specifically configured to:

[0099] obtain a resource portrait of the edge cloud computing system and a service portrait of the target service; the resource portrait is used to describe resource features of cloud computing resources carried by each edge cloud node in the edge cloud computing system; the service portrait is used to describe service features of the target service;

[0100] From the cloud computing resources carried by the at least one edge cloud node, a cloud computing resource with a resource feature matched with a service feature of the target service is determined as the first target cloud computing resource to be scheduled.

[0101] In this embodiment, the first scheduling unit 602 is specifically configured to:

[0102] collect resource information of the cloud computing resources carried by the at least one edge cloud node, and collect service information of each cloud computing instance created on each edge cloud node for the target service;

[0103] generate the resource portrait of the edge cloud computing system according to the collected resource information, and generate the service portrait of the target service according to the collected service information.

[0104] In this embodiment, the creating unit 604 is further configured to:

[0105] in response to completion of creation of each logical service object of the target service, mark each logical service object of the target service as a service stop state, and perform service availability verification on each logical service object of the target service.

[0106] If each logical service object of the target service passes the service availability verification, update each logical service object of the target service to a service allowed state.

[0107] In this embodiment, the second scheduling unit 608 is specifically configured to:

[0108] in response to an access request initiated by a user for the target service, determine a target logical service object for providing services to the user from each logical service object of the target service according to service capabilities that can be provided by each logical service object of the target service.

[0109] send the access request to a target edge cloud node corresponding to the target logical service object, so that a cloud computing instance of the target service created on the target edge cloud node provides services to the user.

[0110] In this embodiment, the apparatus further includes:

[0111] a third scheduling unit, configured to determine a second target cloud computing resource to be scheduled from cloud computing resources allocated to a cloud computing instance of the target service in response to a resource recycling event of the target service, and determine a reduced service capability of each logical service object of the target service according to the second target cloud computing resource;

[0112] a fourth scheduling unit, configured to schedule service traffic of the target service according to the reduced service capability of each logical service object of the target service;

[0113] a first recycling unit, configured to recycle the second target cloud computing resource in response to service traffic accessing the cloud computing instance created based on the second target cloud computing resource being less than a preset first threshold.

[0114] In this embodiment, the apparatus further includes:

[0115] a fifth scheduling unit, configured to update a target logical service object to be deleted from an allowed service state to a service-stopped state in response to a logical service object deletion event of the target service, and switch service traffic accessing the target logical service object to other logical service objects of the target service;

[0116] a second recycling unit, configured to delete a target cloud computing instance set corresponding to the target logical service object in response to a cloud computing resource usage rate of the target logical service object being less than a preset second threshold, recycle cloud computing resources allocated to cloud computing instances included in the target cloud computing instance set, and delete the target logical service object.

[0117] In this embodiment, the target service is a large-flow service; the cloud computing instance of the target service runs a plurality of threads of the target service; and the cloud computing resources include CPU resources.

[0118] The apparatus further includes:

[0119] a binding unit, configured to perform a binding operation on the plurality of threads run by the cloud computing instance of the target service and the CPU resources allocated to the cloud computing instance of the target service.

[0120] The functions and effects of each unit of the apparatus are achieved in the implementation process of the corresponding steps in the above method, which will not be described here again.

[0121] For apparatus embodiments, since they basically correspond to the method embodiments, relevant parts are explained by referring to the parts of the method embodiments. The apparatus embodiments described above are only illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purposes of the solutions in the specification according to actual needs. Those skilled in the art can understand and implement it without creative effort.

[0122] The systems, apparatuses, modules or units illustrated by the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, and the specific form of the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an e-mail device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0123] In a typical configuration, a computer includes one or more processors (CPUs), input / output interfaces, network interfaces, and memories.

[0124] The memory can include non-persistent memory in computer readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or Flash memory. The memory is an example of computer readable media.

[0125] The computer readable media includes non-transitory and transitory, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile discs (DVDs) or other optical storage, magnetic cassettes, magnetic disks storage, quantum memory, graphene-based storage media, or other magnetic storage devices, or any other non-transitory medium that can be used to store information accessible to computing devices, according to the definition in this paper. According to the definition in this paper, computer readable media does not include transitory computer readable media, such as modulated data signals and carriers.

[0126] The user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation portal for user to choose authorization or refusal.

[0127] It should also be noted that the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the identified element.

[0128] The above describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different than the order in which they are recited in the embodiments and still achieve the desired result. In addition, the processes depicted in the figures do not necessarily require the particular order shown or sequential order to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous or possible.

[0129] The terms used in one or more embodiments of the present specification are merely for the purpose of describing specific embodiments and are not intended to limit one or more embodiments of the present specification. The singular forms "a", "an" and "the" used in one or more embodiments of the present specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0130] It should be understood that although the terms first, second, third, etc. can be employed in one or more embodiments of the present specification to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another type of information. For example, without departing from the scope of one or more embodiments of the present specification, first information can also be referred to as second information, and similarly, second information can also be referred to as first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "upon determination" or "in response to a determination".

[0131] The above description is only the preferred embodiment of one or more embodiments of the specification, and is not used to limit one or more embodiments of the specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of the specification should be included in the protection range of one or more embodiments of the specification.

Claims

1. A traffic scheduling method, applied to a central management device in a distributed edge cloud computing system. The edge cloud computing system further comprises at least one edge cloud node interconnected with the central management device; the method comprises: In response to a resource allocation event of a target service, determining a first target cloud computing resource to be scheduled from cloud computing resources carried by the at least one edge cloud node, and scheduling the first target cloud computing resource to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resource; adding cloud computing instances created for the target service on the same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and creating a logical service object for each cloud computing instance set of the target service, wherein the cloud computing instance set of an edge node comprises cloud computing instances created for the target service on the edge node and original cloud computing instances of the target service on the edge node; determining service capabilities that can be provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service; scheduling service traffic of the target service according to service capabilities that can be provided by each logical service object of the target service, by taking each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node.

2. The method of claim 1, wherein determining a first target cloud computing resource to be scheduled from cloud computing resources carried by the at least one edge cloud node comprises: obtaining a resource profile of the edge cloud computing system and a service profile of the target service; wherein the resource profile is used to describe resource characteristics of cloud computing resources carried by each edge cloud node in the edge cloud computing system; and the service profile is used to describe service characteristics of the target service; from cloud computing resources carried by the at least one edge cloud node, determining a cloud computing resource with resource characteristics matching the service characteristics of the target service as the first target cloud computing resource to be scheduled.

3. The method of claim 2, wherein obtaining the resource profile of the edge cloud computing system and the service profile of the target service comprises: collecting resource information of cloud computing resources carried by the at least one edge cloud node, and collecting service information of each cloud computing instance created for the target service on each edge cloud node; generating the resource profile of the edge cloud computing system according to the collected resource information, and generating the service profile of the target service according to the collected service information.

4. The method of claim 1, wherein before scheduling service traffic of the target service according to service capabilities that can be provided by each logical service object of the target service, the method further comprises: in response to completion of creation of each logical service object of the target service, marking each logical service object of the target service as a stop-service state, and performing service availability verification on each logical service object of the target service. If each logical service object of the target service passes the service availability check, each logical service object of the target service is updated to an allowed service state.

5. The method of claim 1, wherein the scheduling of the service traffic of the target service according to the service capability provided by each logical service object of the target service comprises: in response to a user-initiated access request for the target service, determining a target logical service object for providing service to the user from each logical service object of the target service according to the service capability provided by each logical service object of the target service; downloading the access request to a target edge cloud node corresponding to the target logical service object, so that the cloud computing instance of the target service created on the target edge cloud node provides service to the user.

6. The method of claim 1, further comprising: in response to a resource recycling event of the target service, determining a second target cloud computing resource to be scheduled from the cloud computing resources allocated to the cloud computing instance of the target service, and determining the reduced service capability of each logical service object of the target service according to the second target cloud computing resource; scheduling the service traffic of the target service according to the reduced service capability of each logical service object of the target service; in response to the service traffic accessing the cloud computing instance created based on the second target cloud computing resource being less than a preset first threshold, recycling the second target cloud computing resource.

7. The method of claim 1, further comprising: in response to a logical service object deletion event of the target service, updating a target logical service object to be deleted from an allowed service state to a service-stopped state, and switching the service traffic accessing the target logical service object to other logical service objects of the target service; in response to the cloud computing resource usage of the target logical service object being less than a preset second threshold, deleting a target cloud computing instance set corresponding to the target logical service object, recycling the cloud computing resources allocated to the cloud computing instances included in the target cloud computing instance set, and deleting the target logical service object.

8. The method of claim 1, wherein the target service is a large-flow service; the cloud computing instance of the target service runs a plurality of threads of the target service; and the cloud computing resources include CPU resources. The method further comprises: binding the plurality of threads run by the cloud computing instance of the target service to the CPU resources allocated to the cloud computing instance of the target service. 9.A flow scheduling apparatus, the apparatus being applied to a central management device in a distributed edge cloud computing system. The edge cloud computing system further comprises at least one edge cloud node interconnected with the central management device; and the apparatus comprises: a first scheduling unit, configured to determine a first target cloud computing resource to be scheduled from cloud computing resources carried by the at least one edge cloud node in response to a resource allocation event of a target service, and schedule the first target cloud computing resource to create at least one cloud computing instance for the target service on an edge cloud node carrying the first target cloud computing resource; a creating unit, configured to add a cloud computing instance created for the target service on a same edge cloud node to a cloud computing instance set of the target service corresponding to the edge cloud node, and create a logical service object for each cloud computing instance set of the target service, wherein the cloud computing instance set of an edge node includes a cloud computing instance created for the target service on the edge node and an original cloud computing instance of the target service on the edge node; a determining unit, configured to determine service capabilities that can be provided by each logical service object of the target service according to cloud computing resources allocated to each cloud computing instance of the target service; a second scheduling unit, configured to schedule service traffic of the target service according to the service capabilities that can be provided by each logical service object of the target service, by taking each logical service object of the target service as a traffic scheduling object corresponding to each edge cloud node. 10.An electronic device, comprising a communication interface, a processor, a memory and a bus, the communication interface, the processor and the memory are connected with each other through the bus; the memory stores machine readable instructions, and the processor executes the method of any one of claims 1 to 8 by invoking the machine readable instructions. 11.A machine readable storage medium, storing machine readable instructions, which when invoked and executed by a processor, implement the method of any one of claims 1 to 8.

Images