CN112953993A

CN112953993A - Resource scheduling method, device, network system and storage medium

Info

Publication number: CN112953993A
Application number: CN201911266785.6A
Authority: CN
Inventors: 刘云剑
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2019-12-11
Filing date: 2019-12-11
Publication date: 2021-06-11
Anticipated expiration: 2039-12-11
Also published as: CN112953993B

Abstract

The embodiment of the application provides a resource scheduling method, equipment, a network system and a storage medium. In the embodiment of the application, the network system comprises at least one edge cloud node and a central control device which can perform resource scheduling on the at least one edge cloud node; the central control equipment supports at least one resource scheduling algorithm, can select a resource scheduling algorithm adaptive to the resource scheduling request under the condition of receiving the resource scheduling request, adopts the resource scheduling algorithm adaptive to the resource scheduling request to perform resource scheduling, is favorable for improving the rationality of resource scheduling, and can simultaneously give consideration to the efficiency and the accuracy of resource scheduling.

Description

Resource scheduling method, device, network system and storage medium

Technical Field

The present application relates to the field of network technologies, and in particular, to a resource scheduling method, device, network system, and storage medium.

Background

With the arrival of the age of 5G and the internet of things and the gradual increase of cloud computing applications, the requirements of a terminal on the performances of time delay, bandwidth and the like of cloud resources are higher and higher, and the traditional centralized cloud network cannot meet the increasingly high cloud resource requirements of the terminal.

With the advent of edge computing technology, the concept of edge clouds has been created. How to reasonably schedule resources of the edge cloud so as to improve the utilization rate of physical resources and ensure high availability of user services is an urgent problem to be solved in the development process of the edge cloud.

Disclosure of Invention

Aspects of the present application provide a resource scheduling method, device, network system, and storage medium, which are used to reasonably schedule resources in an edge cloud, improve a utilization rate of physical resources, and ensure high availability of user services.

An embodiment of the present application provides a network system, including: the system comprises a central management and control device and at least one edge cloud node; the central control equipment is used for receiving a resource scheduling request, and the resource scheduling request is used for requesting resource scheduling on at least one specification example; selecting a target resource scheduling algorithm which is adaptive to the resource scheduling request from at least one resource scheduling algorithm; adopting the target resource scheduling algorithm to perform resource scheduling on available resources in the schedulable edge cloud node for the at least one specification example; the schedulable edge cloud node is a part or all of the at least one edge cloud node.

The embodiment of the present application further provides a resource scheduling method, which is applicable to a central control device in a network system, and the method includes: receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on at least one specification instance; selecting a target resource scheduling algorithm adaptive to the resource scheduling request from at least one resource scheduling algorithm; and performing resource scheduling on available resources in the edge cloud node which can be scheduled at this time for the example of the at least one specification by adopting the target resource scheduling algorithm.

The embodiment of the present application further provides a resource scheduling method, which is applicable to a terminal device in a network system, and the method includes: displaying a human-computer interaction interface provided by a central control device; responding to the input operation on the human-computer interaction interface, acquiring a resource scheduling request, and requesting the central control equipment to perform resource scheduling on available resources in the edge-available nodes which can be scheduled at this time for at least one specification example; and receiving the resource scheduling result returned by the central control equipment, and outputting the resource scheduling result returned by the central control equipment.

An embodiment of the present application further provides a central management and control device, including: a memory and a processor; the memory for storing a computer program; when executed by the processor, the computer program causes the processor to implement the steps in the resource scheduling method that can be executed by the central control device according to the embodiment of the present application.

Embodiments of the present application further provide a computer-readable storage medium storing a computer program, which, when executed by one or more processors, causes the one or more processors to implement the steps in the resource scheduling method provided by the embodiments of the present application.

In the embodiment of the application, the network system comprises at least one edge cloud node and a central control device which can perform resource scheduling on the at least one edge cloud node; the central control equipment supports at least one resource scheduling algorithm, can select a resource scheduling algorithm adaptive to the resource scheduling request under the condition of receiving the resource scheduling request, adopts the resource scheduling algorithm adaptive to the resource scheduling request to perform resource scheduling, is favorable for improving the rationality of resource scheduling, and can simultaneously give consideration to the efficiency and the accuracy of resource scheduling.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic structural diagram of a network system according to an exemplary embodiment of the present application;

fig. 2a is a schematic diagram illustrating a result of resource scheduling according to an exemplary embodiment of the present application;

FIG. 2b is a diagram illustrating results of another resource scheduling provided by an exemplary embodiment of the present application;

fig. 3 is a schematic structural diagram of another network system provided in an exemplary embodiment of the present application;

fig. 4a is a flowchart illustrating a resource scheduling method according to an exemplary embodiment of the present application;

fig. 4b is a flowchart illustrating another resource scheduling method according to an exemplary embodiment of the present application;

fig. 4c is a schematic flowchart of another resource scheduling method according to an exemplary embodiment of the present application;

fig. 4d is a schematic flowchart of another resource scheduling method according to an exemplary embodiment of the present application;

fig. 5 is a flowchart illustrating a resource scheduling method described from a resource scheduling demander according to an exemplary embodiment of the present application;

fig. 6a is a schematic structural diagram of a central management and control device according to an exemplary embodiment of the present application;

fig. 6b is a schematic structural diagram of a terminal device according to an exemplary embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Aiming at the problem of resource scheduling faced by the edge cloud, in the embodiment of the application, the center control equipment is combined with the edge cloud node, the center control equipment supports at least one resource scheduling algorithm, and the resource scheduling algorithm adaptive to the resource scheduling request can be adopted to schedule resources for the edge cloud node according to the resource scheduling request, so that the rationality of resource scheduling is improved, and the efficiency and the accuracy of resource scheduling can be considered at the same time.

The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic structural diagram of a network system according to an exemplary embodiment of the present application. As shown in fig. 1, the network system 100 includes: a central management and control device 101 and at least one edge cloud node 102; at least one edge cloud node 102 is connected to the central management and control device 101 through a network.

The network system 100 of the present embodiment is a cloud computing platform constructed on an edge infrastructure based on cloud computing technology and edge computing capability, and is a cloud platform having computing, networking, storage, security, and other capabilities at an edge location.

The network system 100 of the present embodiment may be regarded as an edge cloud network system corresponding to a central cloud or a conventional cloud computing platform. The edge cloud is a relative concept, the edge cloud refers to a cloud computing platform relatively close to the terminal, or is different from a central cloud or a traditional cloud computing platform, the central cloud or the traditional cloud computing platform may include a data center with a large resource scale and a centralized position, a network range covered by the edge cloud nodes is wider, and therefore the edge cloud computing platform has the characteristic of being closer to the terminal, the resource scale of a single edge cloud node is smaller, but the number of the edge cloud nodes is large, and a plurality of edge cloud nodes form a component of the edge cloud in the embodiment. The terminal in this embodiment refers to a demand end of the cloud computing service, and may be, for example, a terminal or a user end in the internet, or a terminal or a user end in the internet of things. An edge cloud network is a network built based on infrastructure between a central cloud or traditional cloud computing system and terminals.

Therein, the network system 100 includes at least one edge cloud node 102, each edge cloud node 102 including a series of edge infrastructures including, but not limited to: a distributed Data Center (DC), a wireless room or cluster, an edge device such as a communication network of an operator, a core network device, a base station, an edge gateway, a home gateway, a computing device or a storage device, a corresponding network environment, and the like. It is noted that the location, capabilities, and infrastructure involved of different edge cloud nodes 102 may or may not be the same.

It should be noted that the network system 100 of this embodiment may be combined with a central network such as a central cloud or a conventional cloud computing platform, and further combined with a terminal, so as to form a "cloud-edge-end-three-body cooperation" network architecture, in the network architecture, tasks such as network forwarding, storage, computation, and intelligent data analysis may be placed in each edge cloud node 102 in the network system 100 for processing, and as each edge cloud node 102 is closer to the terminal, response delay may be reduced, pressure on the central cloud or the conventional cloud computing platform may be reduced, and bandwidth cost may be reduced. In addition, the network system 100 of the present embodiment may also be directly combined with a terminal, so as to form an "edge-to-end" network architecture.

Regardless of the network architecture, how to reasonably schedule resources in the edge cloud nodes and how to manage the edge cloud nodes to perform cloud computing services with correct and stable logic is an important challenge. In the network system 100 of this embodiment, a central control device 101 is deployed, and the central control device 101 uses the edge cloud node 102 as a control object to perform unified control on at least one edge cloud node 102 in the network system 100 in aspects of resource scheduling, mirror image management, instance control, operation and maintenance, network, security, and the like, so as to place cloud computing services into each edge cloud node 102 for processing. In terms of deployment implementation, the central management and control device 101 may be deployed in one or more cloud computing data centers, or may be deployed in one or more conventional data centers, and the central management and control device 101 may also form an edge cloud network together with at least one edge cloud node managed by the central management and control device, which is not limited in this embodiment.

For an edge cloud node 102, various resources, such as computing resources like CPUs and GPUs, storage resources like memories and hard disks, network resources like bandwidths, and the like, may be provided externally. In addition, respective instances may be deployed in the edge cloud nodes 102, which may provide various cloud computing services to the outside. The implementation form of the instance may be a Virtual Machine (VM), a container (Docker), a native application, or the like. The problem to be solved by resource scheduling is that an instance should be placed in which edge cloud node 102, and specifically on which physical machine in the edge cloud node 102.

Optionally, the central managing and controlling device 101 may provide a human-machine interaction interface to the outside, where the human-machine interaction interface may be a web page, an application page, a command window, or the like. The human-computer interaction interface is used for the resource scheduling demand to submit the resource scheduling request to the central control equipment 101. For the resource scheduling demander, a resource scheduling request can be submitted to the central control device 101 through the human-computer interaction interface. The resource scheduling request comprises: information related to resource scheduling, such as scale information of an instance requiring resources (referred to simply as an instance to be deployed), and the like. According to different service scenarios for initiating resource scheduling, information related to resource scheduling carried in the resource scheduling request may be different, which is not limited to this.

In application scenario 1, an operator or an operating system of the network system 100 may need to query how many instances of a certain specification may be put down in one or more edge cloud nodes 102 according to an operation requirement, for example, how many 2-core 4G VMs may be put down in one or more edge cloud nodes 102. In this application scenario, an operator or an operation system of the network system 100 may serve as a resource scheduling demander, and submit a resource scheduling request to the central control device 101 through a human-computer interaction interface provided by the central control device 101, where the resource scheduling request at least includes: information pointing to one or more edge cloud nodes involved in the query, specification information of instances involved in the query, and the like. The information pointing to the edge cloud nodes involved in the query may be identification information of an area where the edge cloud nodes are located, or may also be identification information of the edge cloud nodes themselves.

In application scenario 2, a customer, such as an individual or a business, of network system 100 may need to query its own instance requirements in a certain region (e.g., northwest region), for example, may need to query which edge cloud nodes 102 in the northwest region may have 10 VMs with 4 cores and 8 VMs with 4 cores and 4G, and at most several sets of such VM combinations. In the application scenario, a customer as a resource scheduling demander may submit a resource scheduling request to the central control device 101 through a human-computer interaction interface provided by the central control device 101, where the resource scheduling request at least includes: information pointing to the edge cloud nodes involved in the query, information such as specification and number of VMs involved in the query, and the like. The information pointing to the edge cloud nodes involved in the query may be identification information of an area where the edge cloud nodes are located, or may also be identification information of the edge cloud nodes themselves.

In application scenario 3, as a client of network system 100, for example, an individual or a business, a request to purchase an instance is made to network system 100. In this application scenario, a client may submit a resource scheduling request to the central control device 101 through a human-computer interaction interface provided by the central control device 101, where the resource scheduling request at least includes: information on the specification, number, etc. of the purchased instances is required. For example, a customer may request to purchase 10 VMs that are 4-core 8G in size. Alternatively, the customer may request to purchase 10 VMs of 4-core 8G specification and 8 VMs of 4-core 4G specification.

The implementation form of the human-computer interaction interface is not limited in this embodiment. In an optional embodiment, the man-machine interface provided by the central control apparatus 101 includes a query page and a purchase page. For the application scenarios 1 and 2, an operator or a customer of the network system 100 may enter the query page provided by the central control device 101, provide information of the edge cloud nodes involved in the query and information such as specification and number of instances involved in the query through the query page, and carry the information in the resource scheduling request to the central control device 101, so that the central control device 101 performs resource scheduling accordingly. For the application scenario 3, a customer of the network system 100 may enter a purchase page provided by the central control apparatus 101, provide information such as specifications and number of instances to be purchased through the purchase page, and provide the information to the central control apparatus 101 in a resource scheduling request, so that the central control apparatus 101 performs resource scheduling accordingly. Further alternatively, the query page and the purchase page may be the same page, and the query page and the purchase page are provided with a query control and a purchase control, so that an operator or a customer of the network system 100 can initiate a resource scheduling request according to a requirement.

It should be noted that, besides the above man-machine interface manner, the central control device 101 may also obtain the resource scheduling request in other manners. For example, the resource scheduling demander may also embed a resource scheduling task in the central control device 101 in advance, and the central control device 101 may execute the built-in resource scheduling task to obtain the resource scheduling request. Or, the resource scheduling demander may transmit the resource scheduling request to the central control device 101 through other devices that can communicate with the central control device 101, such as a terminal device or a configuration device, in a wired or wireless communication manner, so that the central control device 101 may receive the resource scheduling request transmitted by the other devices.

No matter how the central management and control device 101 obtains the resource scheduling request, the resource scheduling may be performed on the edge cloud node 102 in the network system 100 according to the received resource scheduling request.

In this embodiment, the resource scheduling request is used to request the central control apparatus 101 to perform resource scheduling on the instance of the at least one specification. The example here runs depending on available resources in the edge cloud node, and its implementation form may be a Virtual Machine (VM), a container (Docker), or a native application, etc. The central control device 101 may perform resource scheduling for the instance with at least one specification, and the specification of the instance is not limited, and may perform resource scheduling for the instance with at least one specification online, so as to ensure real-time performance of resource scheduling.

In order to more reasonably schedule resources for at least one specification instance, the central control device 101 supports at least one resource scheduling algorithm, and algorithm logics, calculation amount and time consumption of different resource scheduling algorithms and accuracy and precision of given resource scheduling results are different, so that the method can adapt to different resource scheduling requirements. After receiving the resource scheduling request, the central control device 101 may select a resource scheduling algorithm adapted to the resource scheduling request from at least one resource scheduling algorithm. For convenience of description, the selected resource scheduling algorithm adapted to the resource scheduling request is referred to as a target resource scheduling algorithm. Then, the central control device 101 may perform resource scheduling on available resources in the schedulable edge cloud node for an example with at least one specification by using a target resource scheduling algorithm. The schedulable edge cloud nodes are part or all of at least one edge cloud node. The available resources in the schedulable edge cloud node mainly refer to the resources (or referred to as remaining resources) in the edge cloud node that are not used yet. These resources include, but are not limited to: CPU, internal memory, hard disk and/or network card, etc.

In this embodiment, a resource scheduling algorithm adapted to the resource scheduling request is adopted to perform resource scheduling for at least one specification instance, which is beneficial to improving the rationality of resource scheduling, meeting the requirement of resource scheduling, and considering both the efficiency and the accuracy of resource scheduling.

In some optional embodiments, the central management and control device 101 provides a human-computer interaction interface to the resource scheduling demander, so that the resource scheduling demander can initiate a resource scheduling request. For the resource scheduling demander, a resource scheduling request can be provided to the central control device 101 through the human-computer interaction interface, the resource scheduling request carries information related to resource scheduling, and the information can reflect that the resource scheduling demander needs the central control device 101 to perform resource scheduling for an instance of at least one specification. For the central control device 101, a resource scheduling request provided by a resource scheduling demander through a human-computer interaction interface may be received, and it is determined that resource scheduling needs to be performed for an instance of at least one specification according to the resource scheduling request, so that a target resource scheduling algorithm adapted to the resource scheduling request is selected from the at least one resource scheduling algorithm, and a target resource scheduling algorithm is adopted to perform resource scheduling for an instance of at least one specification on available resources in an edge cloud node that can be scheduled this time.

The central control apparatus 10 may succeed or fail in scheduling the resource for the instance with at least one specification. For the central control device 101, the resource scheduling result of this time, that is, the result of the successful resource scheduling or the failed resource scheduling of this time, may also be output to the resource scheduling demander. It should be noted that the resource scheduling result herein refers to an overall result of performing resource scheduling for an instance with at least one specification, in other words, if the current resource scheduling result is successful, it means that the resource scheduling corresponding to all instances in the instance with at least one specification is successful; if the resource scheduling result is failure, it means that the resource scheduling corresponding to at least one of the instances under at least one specification is failed.

Further, in some resource scheduling scenarios, the central management and control device 101 may further generate a resource scheduling scheme corresponding to an instance of at least one specification under the condition that the current resource scheduling is successful. The resource scheduling scheme is another product of resource scheduling, and comprises the following steps: examples of at least one specification can be placed in which edge cloud nodes, and information of available resources in which the examples can be placed, information of VMs specifically placed on the available resources, and the like. For the central control device 101, the resource scheduling result of this time may be returned to the resource scheduling demander, and further, the resource scheduling scheme of this time resource scheduling may be returned to the resource scheduling demander under the condition that this time resource scheduling is successful.

In the present embodiment, the whole resource scheduling process is exemplarily described in connection with an application scenario. Assume that an enterprise customer needs to query whether an edge cloud node of network system 100 in a specified area (e.g., northwest) can meet its VM needs. Then, the enterprise customer may enter a query page provided by the central control device 101, enter the VM specification and number to be queried on the query page, for example, 10 VMs with specification of 4-core and 8G and 8 VMs with specification of 4-core and 4G, and enter its query requirement on the query page, that is, on which physical machines of which edge cloud nodes of a specified area a query can be put down, and at most, several groups of such VM combinations can be put down; after the information input is completed, a determination button on the query page may be clicked, and a resource scheduling request is sent to the central control device 101. The central control device 101 selects a proper resource scheduling algorithm, namely a target resource scheduling algorithm, for enterprise customers, and performs resource scheduling on edge cloud nodes in a specified area by using the target resource scheduling algorithm, wherein the target resource scheduling algorithm comprises 10 VMs with the specification of 4 cores and 8 VMs with the specification of 4 cores and 4G, so as to give a resource scheduling result. The resource scheduling result includes: whether result information of 10 VMs with the specification of 4 cores and 8 VMs with the specification of 4 cores and 4G can be put down in the edge cloud node of the specified area or not; if 10 VMs with the specification of 4-core and 8-core and 4-core and 8 VMs with the specification of 4-core and 4-core can be put down on available resources in the edge cloud node of the designated area, the resource scheduling result is successful, and the result of the successful resource scheduling can be output to enterprise customers through a human-computer interaction interface; if 10 VMs with the specification of 4-core 8G and 8 VMs with the specification of 4-core 4G cannot be put down on available resources in the edge cloud node of the specified area, the resource scheduling result is failure, and the result of the failure of the resource scheduling can be output to enterprise customers through a human-computer interaction interface.

Further, under the condition that 10 VMs with the specification of 4 core and 8G and 8 VMs with the specification of 4 core and 4G can be put down on available resources in an edge cloud node in a specified area, a resource scheduling scheme of this resource scheduling can be output to enterprise customers through a human-computer interaction interface, where the resource scheduling scheme includes: information of the physical machines that can drop these VMs, information of VMs that can be placed on each physical machine, and the number of groups of VM combinations that can be dropped at most, and so on. For example, an example of a resource scheduling scheme includes: placing 10 VMs with the specification of 4 cores and 8G into the same edge cloud node, and respectively placing 10 VMs on different physical machines in the edge cloud node; 8 VMs with the specification of 4 cores and 4G are placed in another edge cloud node, and the 8 VMs are respectively placed on different physical machines in the edge cloud node. As another example, another example of a resource scheduling scheme includes: 10 VMs with the specification of 4 cores and 8 VMs with the specification of 4 cores and 4G are placed in the same edge cloud node, and 18 VMs are respectively placed on different physical machines in the edge cloud node. As another example, another example of a resource scheduling scheme includes: 6 VMs with the specification of 4 cores and 8G and 4 VMs with the specification of 4 cores and 4G are placed in the same edge cloud node, the 10 VMs are respectively placed on different physical machines, the other 4 VMs with the specification of 4 cores and 8G and the other 4 VMs with the specification of 4 cores and 4G are placed in the other edge cloud node, and every 2 VMs with the same specification are placed on the same physical machine. In the above example of the resource scheduling scheme, the placement of the VM is only an exemplary illustration, and the placement scheme of the VM is related to the adopted target resource scheduling algorithm, which is not limited to this.

In some application scenarios, for example, in the above query scenario, after the resource scheduling result is obtained, the central management and control device 101 only needs to return the resource scheduling result to the resource scheduling demand party. In other application scenarios, for example, in the scenario of purchasing a VM, after the resource scheduling result is obtained, the central control device 101 not only needs to return the resource scheduling result and the resource scheduling scheme when the resource scheduling is successful to the resource scheduling demander, but also needs to deploy an instance of at least one specification on the available resource scheduled in the edge cloud node that can be scheduled this time according to the resource scheduling scheme. Continuing with the above example, the central management and control device 101 further needs to deploy 10 VMs with specification of 4 core and 8G and 8 VMs with specification of 4 core and 4G to corresponding physical machines according to the resource scheduling scheme.

In the embodiments of the present application, the central control apparatus 101 needs to select a target resource scheduling algorithm adapted to the resource scheduling request from at least one resource scheduling algorithm. The selection mode used by the central control device 101 may be various, and may be flexibly set according to the application requirement. In the following exemplary embodiments of the present application, several specific options will be enumerated.

In some embodiments, the resource scheduling request may carry information related to resource scheduling. Based on this, after receiving the resource scheduling request, the central control device may select a target resource scheduling algorithm adapted to the resource scheduling request from at least one resource scheduling algorithm according to information carried by the resource scheduling request. The information carried by the resource scheduling request comprises at least one of source address information and example scale information of the resource scheduling request.

The source address information of the resource scheduling request may reflect a source of the resource scheduling request, such as an IP address and a service interface for initiating the resource scheduling request, and these pieces of information may reflect a service scenario of resource scheduling. Example size information typically includes, among other things: the specification of the instances that need to be deployed, the number of specifications, the number of instances under each specification, etc.

In an exemplary embodiment a1, the resource scheduling algorithm is associated with a service scenario, and service scenarios corresponding to different resource scheduling algorithms are maintained in advance, where the different resource scheduling algorithms correspond to different service scenarios. Based on this, the central control device 101 determines a service scenario for initiating the resource scheduling by analyzing the source address information of the resource scheduling request; and selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to the service scene initiating the resource scheduling. For example, the central management and control device 101 may analyze a service interface initiating the resource scheduling in the source address information of the resource scheduling request, such as a query interface, a preview interface, or an interface for purchasing a VM; determining a service scene for initiating the resource scheduling according to a service interface for initiating the resource scheduling; for example, the query interface corresponds to a query scene, the preview interface corresponds to a preview scene, and the interface for purchasing the VM corresponds to a purchase scene; and then, according to the service scene initiating the resource scheduling, inquiring the corresponding relation between different resource scheduling algorithms and different service scenes, and determining the target resource scheduling algorithm required to be used by the resource scheduling.

In the present embodiment, at least one resource scheduling algorithm supported by the center controlling device 101 is not limited. Optionally, the resource scheduling algorithm supported by the central control apparatus 101 includes a deterministic heuristic algorithm and a non-deterministic heuristic algorithm. Deterministic heuristics may include, but are not limited to: a resource scheduling algorithm based on NF (Next Fit), a resource scheduling algorithm based on BF (best Fit), a resource scheduling algorithm based on FF (first Fit), and the like; non-deterministic heuristics may include, but are not limited to: and (3) a resource scheduling algorithm based on a genetic algorithm. Compared with a non-deterministic heuristic algorithm, the deterministic heuristic algorithm has relatively simple algorithm logic, less calculation amount and high calculation speed, but the calculation result is not necessarily optimal, and if the result of the algorithm is adopted for example deployment, the resource utilization rate of the edge cloud node is possibly low; the logic of the non-deterministic heuristic algorithm is relatively complex, the calculation amount is relatively large, the time consumption is relatively long, the calculation result is relatively excellent, and if the result of the algorithm is adopted for example deployment, the resource utilization rate of the edge cloud node is favorably improved.

In some embodiments, the business scenario may be divided into a query scenario and a non-query scenario. The non-query scenario may include a scenario of previewing VM deployment and a scenario of purchasing a VM, and both scenarios require the central management and control device 101 to provide a relatively accurate resource scheduling scheme, compared with the query scenario; more in the query scene, whether the instance of the query specified specification can be put down in the specified edge cloud node or not is judged, the result is relatively rough, and the requirement on the accuracy of the resource scheduling scheme is low. In view of this, a deterministic heuristic may be associated with a query scenario and a non-deterministic heuristic may be associated with a non-query scenario. In these embodiments, the central management and control device 101 may determine whether the service scenario initiating the resource scheduling belongs to an inquiry scenario; if the service scene initiating the resource scheduling is an inquiry scene, selecting a deterministic heuristic algorithm as a target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene, selecting a nondeterministic heuristic algorithm as a target resource scheduling algorithm.

In an exemplary embodiment a2, the resource scheduling algorithm is associated with instance size information and instance size ranges corresponding to different resource scheduling algorithms are maintained, wherein different resource scheduling algorithms correspond to different instance size ranges. Based on this, when selecting the target resource scheduling algorithm, the central control device 101 may specifically select the target resource scheduling algorithm from at least one resource scheduling algorithm according to the instance scale information carried by the resource scheduling request. For example, the central management and control device 101 may compare the instance scale information carried by the resource scheduling request with different instance scale ranges corresponding to different resource scheduling algorithms maintained in advance, and determine in which instance scale range the instance scale information carried by the resource scheduling request falls; and taking the resource scheduling algorithm corresponding to the example scale range as a target resource scheduling algorithm.

In the present embodiment, at least one resource scheduling algorithm supported by the center controlling device 101 is not limited. Optionally, the at least one resource scheduling algorithm supported by the central control device 101 includes: deterministic heuristics and non-deterministic heuristics. For the related description of the deterministic heuristic algorithm and the non-deterministic heuristic algorithm, reference may be made to the foregoing embodiments, which are not described herein again.

Based on the above, when the central management and control device 101 selects the target resource scheduling algorithm, it is specifically configured to: comparing example scale information carried by the resource scheduling request with a set scale threshold; if the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as a target resource scheduling algorithm; and if the example scale information carried by the resource scheduling request is greater than or equal to the set scale threshold, selecting a non-deterministic heuristic algorithm as a target resource scheduling algorithm.

Wherein, according to different content of example scale information, the threshold value contained in the scale threshold value is different. If the instance size information refers to the total number of instances (e.g., to be queried or to be purchased) involved in the resource scheduling, the size threshold is an instance number threshold; if the example scale information refers to the specification quantity of the example (to be inquired or purchased) related to the resource scheduling, the scale threshold is the specification quantity threshold; if the instance size information includes both the total number of instances involved in resource scheduling and the specification number, the size threshold may include: a specification quantity threshold and an instance quantity threshold.

In some application scenarios, the instance size information carried in the resource scheduling request refers to the total number of instances involved in resource scheduling, and accordingly, the size threshold is an instance number threshold. For example, assuming that an enterprise customer needs to inquire whether 10 instances of the first specification and 8 instances of the second specification can be dropped in a specified edge cloud node, a resource scheduling request may be submitted to the central management and control device 101, where the resource scheduling request carries the two specifications of the instances and the number of the instances in each specification. After receiving the resource scheduling request, the central management and control device 101 may parse out the total number of VMs from the resource scheduling request, that is, 10+8 is 18, and then compare the total number of instances 18 with a set threshold of the number of instances, for example, 10; since the total number of instances 18 is greater than the threshold number of instances 10, the central managing and controlling device 101 selects a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

In other application scenarios, the instance size information carried in the resource scheduling request refers to the specification number of the instances involved in resource scheduling, and accordingly, the size threshold is the specification number threshold. For example, assuming that an enterprise customer needs to inquire whether 10 instances of the first specification and 8 instances of the second specification can be dropped in a specified edge cloud node, a resource scheduling request may be submitted to the central management and control device 101, where the resource scheduling request carries the two specifications of the instances and the number of the instances in each specification. After receiving the resource scheduling request, the central management and control device 101 may parse out that the specification number of the instance is 2 from the resource scheduling request; this instance number of specifications 2 is then compared to a set number of specifications threshold, e.g. 3; since the instance number of metrics 2 is smaller than the threshold number of metrics 3, the central management and control device 101 selects a deterministic heuristic algorithm as the target resource scheduling algorithm.

In still other application scenarios, the instance size information carried in the resource scheduling request refers to the specification number and the total number of instances involved in resource scheduling, and accordingly, the size threshold is the specification number threshold and the instance number threshold. For example, assuming that an enterprise customer needs to inquire whether 10 instances of the first specification and 8 instances of the second specification can be dropped in a specified edge cloud node, a resource scheduling request may be submitted to the central management and control device 101, where the resource scheduling request carries the two specifications of the instances and the number of the instances in each specification. After receiving the resource scheduling request, the central management and control device 101 may parse out, from the resource scheduling request, that the specification number of the instances is 2, and the total number of the instances is 10+8 — 18; this instance number of specifications 2 is then compared to a set number of specifications threshold, e.g., 3, and the total number of instances 18 is compared to a set number of instances threshold, e.g., 20; since the instance number of metrics 2 is smaller than the threshold number of metrics 3 and the instance total number 18 is smaller than the threshold number of instances, the central management and control device 101 selects a deterministic heuristic algorithm as the target resource scheduling algorithm. On the contrary, if the specification number of the instances is greater than the specification number threshold, or if the total number of the instances is greater than the instance number threshold, the central management and control device 101 may select a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

In an exemplary embodiment a3, a resource scheduling algorithm is associated with the traffic scenario and instance size information. In the exemplary embodiment a3, the service scenarios may be divided into two categories, i.e., a first service scenario and a second service scenario; for the first service scene, a target resource scheduling algorithm adapted to the first service scene can be directly determined; and for the second service scene, determining a target resource scheduling algorithm adapted to the second service scene by combining the example scale information. Based on this, when selecting the target resource scheduling algorithm, the central control device 101 is specifically configured to: determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request; judging whether the service scene initiating the resource scheduling is a first service scene or not; if so, selecting a resource scheduling algorithm corresponding to the first service scene from at least one resource scheduling algorithm as a target resource scheduling algorithm; if not, the service scene initiating the resource scheduling is the second service scene, and then a resource scheduling algorithm adaptive to the example scale information can be selected from at least one resource scheduling algorithm as a target resource scheduling algorithm according to the example scale information carried by the resource scheduling request.

Optionally, the first service scenario is a query scenario, and the second service scenario is a non-query scenario. The non-query scenarios may include scenarios of previewing VM deployment and scenarios of purchasing VMs. More in the query scene, whether the instance of the query specified specification can be put down in the specified edge cloud node or not is judged, the result is relatively rough, and the requirement on the accuracy of the resource scheduling scheme is low; compared with the query scenario, the non-query scenario requires the central control device 101 to provide a relatively accurate resource scheduling scheme, and the complexity of the whole scheduling process is greatly related to the example scale information related to resource scheduling. In view of this, deterministic heuristics can be directly associated with the query scenario; regarding the non-query scenario, the non-query scenario with small instance scale information may be associated with a deterministic heuristic algorithm, and the non-query scenario with relatively large instance scale information may be associated with a non-deterministic heuristic algorithm. The central control device 101 may determine whether the service scenario initiating the resource scheduling belongs to the query scenario in the process of selecting the target resource scheduling algorithm; if the service scene initiating the resource scheduling is an inquiry scene, which indicates that the resource scheduling process is simpler, a deterministic heuristic algorithm is selected as a target resource scheduling algorithm; if the service scene initiating the resource scheduling is a non-query scene, further comparing example scale information carried in the resource scheduling request with a set scale threshold; if the service scene initiating the resource scheduling is a non-query scene and the example scale information carried by the resource scheduling request is smaller than the set scale threshold, the resource scheduling process is simple, and a deterministic heuristic algorithm is selected as a target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene, and the example scale information carried by the resource scheduling request is greater than or equal to the set scale threshold, which indicates that the resource scheduling process is complex, selecting a non-deterministic heuristic algorithm as a target resource scheduling algorithm.

In the above exemplary embodiments a1-A3, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics are exemplified. The deterministic heuristic algorithm can be a resource scheduling algorithm based on NF, a resource scheduling algorithm based on BF, a resource scheduling algorithm based on FF, or the like; the non-deterministic heuristic algorithm may be, but is not limited to, a genetic algorithm based resource scheduling algorithm. In the following embodiments, taking an example that at least one resource scheduling algorithm includes a BF-based resource scheduling algorithm and a genetic algorithm-based resource scheduling algorithm, how to select a target scheduling resource algorithm is exemplified in combination with the following 3 application scenarios. The BF-based resource scheduling algorithm corresponds to a query scene, a non-query scene with the specification number smaller than 3 and the total number of instances smaller than 15; other scenarios correspond to a genetic algorithm based resource scheduling algorithm.

In application scenario 1, an operator of the network system 100 needs to query how many VMs with the specification of 4 core 4G can be put down in one or more edge cloud nodes 102 according to an operation requirement, and then submits a resource scheduling request to the central management and control device 101, where the resource scheduling request carries the identifier of the edge cloud node and the specification of the VM of 4 core 4G. After receiving the resource scheduling request, the central management and control device 101 determines that the first service scenario, that is, the query scenario, is the scenario initiating the resource scheduling request; a BF-based resource scheduling algorithm may be used as a target resource scheduling algorithm; then, resource scheduling is performed for the VM with the specification of 4 core 4G by using a BF-based resource scheduling algorithm, and a resource scheduling result is returned to an operator of the network system 100.

In application scenario 2, an enterprise customer of the network system 100 needs to query the network system 100 for its VM requirements in edge cloud nodes in the northwest region, for example, query which edge cloud nodes 102 in the northwest region can put down 10 VMs with the specification of 4 core 8G and 8 VMs with the specification of 4 core 4G, and at most several groups of such VM combinations, and then submit a resource scheduling request to the central management and control device 101, where the resource scheduling request carries the identifier of the edge cloud node, the specifications of the VMs, 4 core 4G and 4 core 8G, and the number of instances under each specification, that is, 10 and 8. After receiving the resource scheduling request, the central control device 101 determines that the first service scenario, i.e., the query scenario, initiating the resource scheduling is the first service scenario according to the source address information of the resource scheduling request; a BF-based resource scheduling algorithm may be used as a target resource scheduling algorithm; and then, performing resource scheduling on the VMs with the specifications of 4-core 4G and 4-core 8G by using a BF-based resource scheduling algorithm, and returning a resource scheduling result to the enterprise client.

In application scenario 3, an enterprise customer of the network system 100 needs to purchase 10 VMs of 4 cores and 8 VMs of 4 cores and 4G from the network system 100, and then submit a resource scheduling request to the central management and control device 101, where the resource scheduling request carries specifications of the VMs, i.e., 4 cores and 4 cores, and 8, and the number of instances under each specification, i.e., 10 and 8. After receiving the resource scheduling request, the central control device 101 determines, according to the source address information of the resource scheduling request, that it is a second service scenario, that is, a non-query scenario, to initiate the resource scheduling this time. In a non-query scenario, the central management and control device 101 further compares the specification number and the total number of instances carried in the resource scheduling request with a set specification number threshold and an set instance number threshold, respectively; in this embodiment, the threshold of the number of specifications is 3, and the threshold of the number of instances is 10, then the number of instances 2 involved in the current resource scheduling is less than the threshold of the number of specifications 3, but the total number of instances 10+8 involved in the current resource scheduling is greater than the threshold of the number of instances 10, which does not meet the requirement of the BF-based resource scheduling algorithm, so the resource scheduling algorithm based on the genetic algorithm is selected as the target resource scheduling algorithm.

No matter which way the target resource scheduling algorithm is selected, after the target resource scheduling algorithm is selected, the central control device 101 may perform resource scheduling on available resources in the current schedulable edge cloud node for an example of at least one specification by using the target resource scheduling algorithm. The schedulable edge cloud nodes are part or all of at least one edge cloud node.

Optionally, the resource scheduling demander may specify an edge cloud node to be scheduled, and carry information pointing to the edge cloud node to be scheduled in the resource scheduling request to provide the resource scheduling request to the central management and control device 101. Based on this, the central control device 101 may further analyze the edge cloud node specified by the resource scheduling demander from the resource scheduling request, and use the edge cloud node as the edge cloud node that can be scheduled this time. Or, the resource scheduling demander may not specify the edge cloud node to be scheduled, but the central management and control device 101 determines the schedulable edge cloud node according to a certain policy. For example, the center control device 101 may use one or more edge cloud nodes closest to the resource scheduling demander as the edge cloud nodes schedulable this time. For another example, the center management and control device 101 may use one or more edge cloud nodes with the most sufficient resources as the edge cloud nodes that can be scheduled this time. For another example, the central management and control device 101 may also use all edge cloud nodes in the entire network system 100 as the edge cloud nodes that can be scheduled this time.

After determining the schedulable edge cloud node, the central control device 101 may perform resource scheduling on the available resources in the schedulable edge cloud node for the example with at least one specification by using a target resource scheduling algorithm. The process of scheduling the available resources in the schedulable edge cloud node at this time by using the target resource scheduling algorithm as the example of at least one specification mainly refers to a process of placing the example of at least one specification on the available resources in the schedulable edge cloud node at this time according to the target resource scheduling algorithm. The scheduling process needs to maximize resource utilization or improve resource scheduling efficiency and improve competitiveness of the network system 100 on the premise of ensuring the customer resource demand and stability of service. Of course, according to different target resource scheduling algorithms, the effect of maximizing resource utilization or improving resource scheduling efficiency brought by resource scheduling may be different.

Further, in order to improve the rationality of resource scheduling, the central control device 101 may perform resource scheduling on available resources in the edge cloud node that can be scheduled this time by using a target resource scheduling algorithm for an example of at least one specification in combination with the resource constraint condition required by the resource scheduling this time. The resource constraint condition is a condition for constraining or limiting resources where instances can be placed, and aims to improve the stability of a resource scheduling result and ensure the quality of service of a client.

Further optionally, the resource constraints include, but are not limited to: affinity resource constraints and exclusivity resource constraints. Among these, resource constraints for affinity are some that define instances that need to be deployed on the same physical machine. For example, VMs with strong dependencies need to be placed on the same physical machine, so that the VMs cooperate with each other to improve the quality of service and efficiency. Exclusive resource constraints are conditions that define instances that need to be deployed scattered across different physical machines. For example, different VMs for the same client or the same service are distributed on different physical machines as much as possible, so as to avoid single point of failure and improve service security. For another example, when placing a VM, the VM is not placed on a physical machine that has been down more than 2 times in the last 7 days.

The resource constraint condition required by the current resource scheduling can be specified by a resource scheduling demander; alternatively, the setting may also be set by the central management and control device 101; or part of the resource constraint conditions are from the specification of the resource scheduling demander and part of the resource constraint conditions are from the setting of the central control device 101. In an optional embodiment, before resource scheduling is performed on available resources in the edge cloud node that can be scheduled this time in combination with the resource constraint condition required by the resource scheduling this time, the central control device 101 may analyze the resource constraint condition required by the resource scheduling request this time from the resource scheduling request, or acquire a locally preset resource constraint condition as the resource constraint condition required by the resource scheduling this time; or, analyzing part of resource constraint conditions from the resource scheduling request, and acquiring resource constraint conditions different from the resource constraint conditions analyzed from the resource scheduling request from locally preset resource constraint conditions to form the resource constraint conditions required by the current resource scheduling request.

In the embodiments of the present application, it is not limited that the center management and control device 101 performs a resource scheduling process on available resources in the schedulable edge cloud node of this time by using a target resource scheduling algorithm. This process is related to the target resource scheduling algorithm. The resource scheduling process is exemplified by using FF or BF-based resource scheduling algorithm as an example.

For example, if the target resource scheduling algorithm is a deterministic heuristic algorithm, and specifically is a resource scheduling algorithm based on FF, in the process of performing resource scheduling for an instance of at least one specification, for each instance of at least one specification, first checking all non-empty physical machines in the edge cloud node which can be scheduled this time, and if a first non-empty physical machine which can put down the instance is found and the non-empty physical machine meets a resource constraint condition required by the resource scheduling this time, determining to place the instance on the first found non-empty physical machine which meets the resource constraint condition; if the virtual machine is not found, further judging whether an empty physical machine still exists in the schedulable edge cloud node, if so, determining to place the instance on a new empty physical machine; if not, determining that the resource scheduling fails. The non-empty physical machine refers to a physical machine where an instance has been placed, and correspondingly, the empty physical machine refers to a physical machine where no instance has been placed.

Wherein, assuming that 5 specifications of VMs need to be placed on 4 physical machines, and the number of VMs in each specification is 1, a resource scheduling result using the FF-based resource scheduling algorithm is shown in fig. 2 a. In fig. 2a, the out-of-number numbers behind the physical machine indicate the sequence in which the physical machine is used, i.e. the first number of the physical machine is used; the numbers in parentheses behind the physical machines indicate the resource specifications that the physical machines have. Similarly, the number outside the number behind the VM indicates the order in which the VM is placed, i.e., the VM is placed the second time; the number in parentheses after a VM indicates the resource specification that the VM needs. It should be noted that the resource specification is a quantized resource specification, which is convenient for calculation.

For another example, if the target resource scheduling algorithm is a deterministic heuristic algorithm, and is specifically a BF-based resource scheduling algorithm, in the process of performing resource scheduling for an instance of at least one specification, for each instance of at least one specification, first checking all non-empty physical machines in the edge cloud node that can be scheduled this time, and if a non-empty physical machine that is most suitable for the instance is found and the non-empty physical machine satisfies a resource constraint condition required for the resource scheduling this time, determining to place the instance on the found non-empty physical machine that is most suitable for the instance and satisfies the resource constraint condition; if the virtual machine is not found, further judging whether an empty physical machine still exists in the schedulable edge cloud node, if so, determining to place the instance on a new empty physical machine; if not, determining that the resource scheduling fails. And for a certain example, placing the example on a non-empty physical machine, wherein the resource waste of the non-empty physical machine is minimum or no resource waste, and the non-empty physical machine is most suitable for the example.

Similarly, assuming that 5 specifications of VMs need to be placed on 4 physical machines, and the number of VMs in each specification is 1, a resource scheduling result using a BF-based resource scheduling algorithm is shown in fig. 2 b. Comparing fig. 2a and fig. 2b, it can be known that the resource scheduling schemes obtained by using different resource scheduling algorithms are different.

In some exemplary embodiments of the present application, in a case that the target resource scheduling algorithm is a deterministic heuristic algorithm, the central control device 101 may further monitor, for at least one specification example, a result of performing resource scheduling on available resources in the current schedulable edge cloud node according to the deterministic heuristic algorithm, that is, monitoring the current resource scheduling result; if the resource scheduling result is failure under the condition of adopting a deterministic heuristic algorithm, judging whether the difference between the resources required by the example with at least one specification and the available resources in the schedulable edge cloud node is smaller than a set difference threshold value or not; if the judgment result is yes, the available resources in the schedulable edge cloud node are possibly examples capable of placing at least one next specification, and the placement is failed only because the placement sequence given by the adopted deterministic heuristic algorithm is not optimized enough, so that the non-deterministic heuristic algorithm can be further adopted to re-schedule the resources for the examples of at least one specification, and the final result of the resource scheduling is given; if the judgment result is negative, the failure of the resource scheduling can be directly determined.

For example, under the condition that the scheduling result of the BF-based resource scheduling algorithm is failed, a resource scheduling algorithm based on a genetic algorithm may be further adopted, and as an example of at least one specification, resource scheduling is performed once again on available resources in the edge cloud node that can be scheduled this time, so that the accuracy of resource scheduling is improved in a controllable range as much as possible, and the balance between the accuracy and the efficiency of resource scheduling is realized.

Regarding the resource scheduling algorithm based on the genetic algorithm: genetic Algorithm (Genetic Algorithm) is a computational model of the biological evolution process that simulates the natural selection and Genetic mechanism of darwinian biological evolution theory, and is a method for searching for an optimal solution by simulating the natural evolution process. Genetic algorithms start with a population (population) representing a possible potential solution set to the problem, and a population consists of a certain number of individuals (individual) encoded by genes (gene). Each individual is actually an entity with a characteristic of the chromosome (chromosome). In a resource scheduling algorithm based on a genetic algorithm, a scheduling mode of placing an instance of at least one specification on available resources in a current schedulable edge cloud node is regarded as an individual in the genetic algorithm; the method comprises the following steps that a plurality of scheduling modes for placing at least one specification on available resources in the schedulable edge cloud node can form a population in a genetic algorithm, namely a combination of a plurality of individuals; an example of one resource available in a scheduling approach corresponding to it is considered a gene segment in a genetic algorithm. Furthermore, by combining the principles of population generation, gene intersection, gene variation, fitness calculation and the like in the genetic algorithm, the resource scheduling is performed on the available resources in the schedulable edge cloud node, and a better resource scheduling result can be obtained. In the embodiment of the present application, a detailed process of resource scheduling implemented according to the above principle is not limited, and all resource scheduling schemes implemented based on a genetic algorithm are applicable to the embodiment of the present application.

It should be noted that, in the network system 100, the central management apparatus 101 may directly manage and schedule the at least one edge cloud node 102, but is not limited thereto. As shown in fig. 3, in the network system 100, in addition to the center management apparatus 101 and the at least one edge cloud node 102, an edge management apparatus 103 is further included. The number of the edge control devices 103 may be one or multiple. In addition, the edge policing device 103 may be deployed in one or more edge cloud nodes 102. In an optional embodiment, as shown in fig. 3, an edge management and control device 103 is respectively deployed in each edge cloud node 102. Further, each edge cloud node includes one or more resource devices, and optionally, the edge management and control device 103 may be centrally deployed on one resource device, or may be dispersedly deployed on multiple resource devices. In addition, each edge cloud node may include one or more proprietary devices in addition to the resource device, where the edge management and control device 103 may be centrally deployed on one proprietary device or dispersedly deployed on multiple proprietary devices. The proprietary device refers to a physical device used to deploy the edge controlling device 103, and is different from the resource device. Furthermore, the edge management device 103 may also be deployed together with the central management device 101, and is not limited herein.

In this embodiment, the edge management and control device 103 may assist and cooperate with the central management and control device 101 to manage and schedule at least one edge cloud node 102. With the assistance of the edge management and control device 103, the central management and control device 101 can more conveniently and efficiently manage and control and schedule at least one edge cloud node 102, so as to achieve the purpose of fully utilizing edge resources.

A secure and encrypted communication channel may be established between the central control device 101 and the edge control device 103, and interaction is performed based on the communication channel. The edge management and control device 103 may monitor the state of each physical machine in the edge cloud node 102, the resource usage amount, the remaining amount, which instances are occupied by, and the like, and report monitoring information to the central management and control device 101. In addition, the edge management and control device 103 may also monitor states (for example, health conditions) of instances in the edge cloud node 102, and report corresponding monitoring information to the central management and control device 101.

The central control device 101 may receive various monitoring information provided by the edge control device 103, and obtain information required for resource scheduling from the monitoring information, for example, the usage states of the physical machines in the edge cloud node, which physical machines are idle, which physical machines are not idle, which VMs are occupied, how many resources the VMs already occupy the physical machines, and the like. Based on the information, the central control device 101 may select a target resource scheduling algorithm adapted to the resource scheduling request from the at least one supported resource scheduling algorithm after receiving the resource scheduling request requesting to perform resource scheduling on the instance of the at least one specification; and then, a target resource scheduling algorithm is adopted, and for at least one specification example, resource scheduling is carried out on available resources in the edge cloud nodes which can be scheduled at this time. For a detailed description of the resource scheduling process, reference may be made to the foregoing embodiments, which are not described herein again.

Further optionally, the central management and control device 101 may deploy, according to a resource scheduling scheme, an instance of at least one specification on an available resource scheduled in the current schedulable edge cloud node. In this case, the edge management and control device 103 may also cooperate with the center management and control device 101 to deploy the instance of at least one specification on the available resource scheduled in the edge cloud node that is schedulable this time. For example, for each instance under at least one specification, the central management and control device 101 may provide the image file of the instance and the resource information to which the instance needs to be deployed to the edge management and control device 103; the edge management and control device 103 provides the image file of the instance to the resource device, such as a physical machine, to which the instance needs to be deployed, so that the corresponding resource device creates the instance by using the image file.

The image file of the instance is a basic file required for creating the instance in the edge cloud node, and may be, for example, an image file such as an operating system, an application, or an operation configuration required for providing a cloud computing service for a user, and the image file may be a file that meets computing deployment requirements of the edge cloud node and is manufactured according to a certain format according to a specific series of files. In addition, the implementation form of the image file is also various corresponding to the instance, for example, the image file may be a Virtual Machine (VM) image file, a container (Docker) image file, or an application package file of each type, and the form of the image file may be related to the virtualization technology used by the instance, which is not limited in this embodiment.

Besides the above system embodiments, the present application also provides some resource scheduling method embodiments, which are mainly described from the perspective of a central management and control device and a resource scheduling demand side. The following describes embodiments of a resource scheduling method provided in the present application.

Fig. 4a is a flowchart illustrating a resource scheduling method according to an exemplary embodiment of the present application. The method is described from the perspective of a central management device, as shown in fig. 4a, and comprises:

41a, receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on the at least one specification instance.

42a, selecting a target resource scheduling algorithm which is matched with the resource scheduling request from at least one resource scheduling algorithm.

43a, adopting a target resource scheduling algorithm, and performing resource scheduling on available resources in the schedulable edge cloud node for at least one specification example.

In some optional embodiments, receiving a resource scheduling request comprises: providing a human-computer interaction interface facing to a resource scheduling demander for the resource scheduling demander to initiate a resource scheduling request; and receiving a resource scheduling request provided by a resource scheduling demander through a human-computer interaction interface.

Further optionally, the method further comprises: and outputting the result of the current resource scheduling to the resource scheduling demander.

Further optionally, the method further comprises: and under the condition that the current resource scheduling is successful, generating a resource scheduling scheme corresponding to the instance with at least one specification.

Still further, the method further comprises: deploying at least one specification of instances on the scheduled available resources in the schedulable edge cloud node according to a resource scheduling scheme; and/or outputting the resource scheduling scheme.

In an optional embodiment, selecting a target resource scheduling algorithm adapted to the resource scheduling request from at least one resource scheduling algorithm includes: selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried by the resource scheduling request; wherein the information includes at least one of source address information and instance size information of the resource scheduling request.

In an exemplary embodiment, selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to information carried by a resource scheduling request includes: determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request; selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to a service scene for initiating the resource scheduling; wherein, different resource scheduling algorithms correspond to different service scenarios.

Optionally, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics. A deterministic heuristic algorithm is applied to the query scenario; a non-deterministic heuristic is applicable to non-query scenarios. As shown in fig. 4b, another resource scheduling method includes:

41b, receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on the instances of at least one specification.

And 42b, determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request.

43b, judging whether the service scene initiating the resource scheduling is an inquiry scene, if so, executing the step 44 b; if the determination result is negative, go to step 45 b.

44b, a deterministic heuristic is selected as the target resource scheduling algorithm, and step 46b is performed.

45b, a non-deterministic heuristic is selected as the target resource scheduling algorithm, and step 46b is performed.

And 46b, adopting a target resource scheduling algorithm, and performing resource scheduling on available resources in the schedulable edge cloud node for at least one specification example.

In the embodiment, an adaptive resource scheduling algorithm is selected in combination with a service scene, and a relatively complex and excellent non-deterministic heuristic algorithm can be selected in the service scene needing an accurate resource scheduling scheme, so that the accuracy requirement on the scheduling result is met; in a service scene without a precise resource scheduling scheme, a relatively simple and quick deterministic heuristic algorithm can be selected, a resource scheduling result is quickly given, and the requirement on scheduling efficiency is met.

In another exemplary embodiment, selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to information carried by a resource scheduling request includes: selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to example scale information carried by the resource scheduling request; different resource scheduling algorithms correspond to different example size ranges.

Optionally, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics. The deterministic heuristic algorithm is suitable for the resource scheduling condition with smaller example scale information; the non-deterministic heuristic algorithm is suitable for the resource scheduling situation with larger example scale information. As shown in fig. 4c, another resource scheduling method includes:

41c, receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on the at least one specification instance.

42c, judging whether the example scale information carried by the resource scheduling request is smaller than a set scale threshold value; if yes, go to step 43 c; if the determination result is negative, go to step 44 c.

43c, a deterministic heuristic is selected as the target resource scheduling algorithm and step 45c is performed.

44c, selecting a non-deterministic heuristic algorithm as the target resource scheduling algorithm, and executing step 45 c.

And 45c, performing resource scheduling on available resources in the schedulable edge cloud node by adopting a target resource scheduling algorithm as an example of at least one specification.

In the embodiment, an adaptive resource scheduling algorithm is selected by combining the example scale information, and a relatively complex and excellent non-deterministic heuristic algorithm can be selected for the case of large or complex example scale, so that the precision requirement on the scheduling result is met; for the case that the example scale is small or simple, a heuristic algorithm with relatively simple and rapid determinacy can be selected, the resource scheduling result is rapidly given, and the requirement on the scheduling efficiency is met.

In another exemplary embodiment, selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to information carried in a resource scheduling request includes: determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request; under the condition that the service scene initiating the resource scheduling is a first service scene, selecting a resource scheduling algorithm corresponding to the first service scene from at least one resource scheduling algorithm as a target resource scheduling algorithm; under the condition that the service scene initiating the resource scheduling is a second service scene, selecting a resource scheduling algorithm adaptive to the example scale information from at least one resource scheduling algorithm as a target resource scheduling algorithm according to the example scale information carried by the resource scheduling request; wherein, different resource scheduling algorithms correspond to different service scenes or example scale ranges.

Optionally, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics. The first service scenario is a query scenario, and the second service scenario is a non-query scenario. The deterministic heuristic algorithm is suitable for query scenes and non-query scenes with smaller instance scale information; the non-deterministic heuristic algorithm is suitable for the non-query scene with larger instance scale information. As shown in fig. 4d, another resource scheduling method includes:

41d, receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on the at least one specification instance.

And 42d, determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request.

43d, judging whether the service scene initiating the resource scheduling is a query scene, and if so, executing the step 44 d; if the determination result is negative, go to step 45 d.

44d, a deterministic heuristic is selected as the target resource scheduling algorithm and step 47d is performed.

45d, judging whether the example scale information carried by the resource scheduling request is smaller than a set scale threshold value; if yes, go to step 44 d; if the determination result is negative, go to step 46 d.

46d, a non-deterministic heuristic is selected as the target resource scheduling algorithm, and step 47d is performed.

And 47d, performing resource scheduling on available resources in the schedulable edge cloud node by adopting a target resource scheduling algorithm, which is an example of at least one specification.

In the embodiment, the adaptive resource scheduling algorithm is selected by combining the service scene and the example scale information, and a relatively complex and relatively excellent non-deterministic heuristic algorithm can be selected in the service scene which needs an accurate resource scheduling scheme and has a large or relatively complex example scale, so that the accuracy requirement on the scheduling result is met; in a service scene without a precise resource scheduling scheme, or in a service scene with a precise resource scheduling scheme but a small or simple example scale, a relatively simple and quick deterministic heuristic algorithm can be selected to quickly provide a resource scheduling result and meet the requirement on scheduling efficiency.

In some embodiments, a target resource scheduling algorithm is adopted, and as an example of at least one specification, performing resource scheduling on available resources in the edge cloud node that can be scheduled this time includes: and performing resource scheduling on available resources in the schedulable edge cloud node for at least one specification example by adopting a target resource scheduling algorithm according to the resource constraint condition required by the current resource scheduling.

Further, if the target resource scheduling algorithm is a deterministic heuristic algorithm and the current resource scheduling using the deterministic heuristic algorithm fails, the method further comprises: judging whether the difference between the resources required for placing the instances with at least one specification and the available resources in the schedulable edge cloud node is smaller than a set difference threshold value or not; if the judgment result is yes, adopting a non-deterministic heuristic algorithm, and performing resource scheduling on available resources in the schedulable edge cloud node for the at least one specification instance again.

For each step or operation in the foregoing method embodiment, the detailed implementation or process may refer to the corresponding description in the foregoing system embodiment, and is not described again in this embodiment.

In the resource scheduling provided by this embodiment, the central control device supports at least one resource scheduling algorithm, and when receiving a resource scheduling request, can select a resource scheduling algorithm adapted to the resource scheduling request for the resource scheduling request, and perform resource scheduling by using the resource scheduling algorithm adapted to the resource scheduling request, which is beneficial to improving the rationality of resource scheduling and can give consideration to both the efficiency and the accuracy of resource scheduling.

Fig. 5 is a flowchart illustrating another resource scheduling method according to an exemplary embodiment of the present application. The method is described from the description of the resource scheduling demander, as shown in fig. 5, and includes:

51. and displaying a human-computer interaction interface provided by the central control equipment.

52. And responding to the input operation on the human-computer interaction interface, and generating a resource scheduling request for requesting resource scheduling for the instance with at least one specification.

53. And sending the resource scheduling request to the central control equipment so that the central control equipment can perform resource scheduling on available resources in the edge nodes which can be scheduled at this time for at least one specification instance.

In an optional embodiment, the method further comprises: receiving a resource scheduling result of the current time returned by the central control equipment; and outputting the resource scheduling result returned by the central control equipment. The output mode may be display on a display screen, or output in a voice mode, etc.

In this embodiment, the resource scheduling demander and the central control device cooperate with each other, the central control device supports at least one resource scheduling algorithm, and when receiving the resource scheduling request, the resource scheduling algorithm adapted to the resource scheduling request can be selected for the resource scheduling request, and the resource scheduling algorithm adapted to the resource scheduling request is adopted to perform resource scheduling, which is beneficial to improving the rationality of resource scheduling, and can simultaneously consider the efficiency and accuracy of resource scheduling.

It should be noted that the execution subjects of the steps of the methods provided in the above embodiments may be the same device, or different devices may be used as the execution subjects of the methods. For example, the execution subjects of steps 51 to 52 may be device a; for another example, the execution subject of step 51 may be device a, and the execution subject of step 52 may be device B; and so on.

In addition, in some of the flows described in the above embodiments and the drawings, a plurality of operations are included in a specific order, but it should be clearly understood that the operations may be executed out of the order presented herein or in parallel, and the sequence numbers of the operations, such as 41a, 42a, etc., are merely used for distinguishing various operations, and the sequence numbers themselves do not represent any execution order. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor limit the types of "first" and "second" to be different.

Fig. 6a is a schematic structural diagram of a central management and control device according to an exemplary embodiment of the present application. As shown in fig. 6a, the center regulating apparatus includes: a memory 61a and a processor 62 a.

A memory 61a for storing a computer program and may be configured to store other various data to support operations on the central administration device. Examples of such data include instructions, messages, pictures, videos, etc. for any application or method operating on the central governing device.

A processor 62a, coupled to the memory 61a, for executing computer programs in the memory 61a for: receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on at least one specification instance; selecting a target resource scheduling algorithm adaptive to the resource scheduling request from at least one resource scheduling algorithm; and performing resource scheduling on available resources in the schedulable edge cloud node for at least one specification example by adopting a target resource scheduling algorithm.

In an optional embodiment, when receiving the resource scheduling request, the processor 62a is specifically configured to: providing a human-computer interaction interface facing a resource scheduling demander for the resource scheduling demander to initiate a resource scheduling request; and receiving a resource scheduling request provided by the resource scheduling demander through the human-computer interaction interface.

Further, the processor 62a is further configured to: and outputting the result of the current resource scheduling to the resource scheduling demander.

In an alternative embodiment, the processor 62a is further configured to: and under the condition that the current resource scheduling is successful, generating a resource scheduling scheme corresponding to the instance with at least one specification.

Further, the processor 62a is also configured to perform at least one of the following operations:

deploying the instance of the at least one specification on the scheduled available resource in the current schedulable edge cloud node according to the resource scheduling scheme;

and outputting the resource scheduling scheme.

In an alternative embodiment, the processor 62a, when selecting the target resource scheduling algorithm, is specifically configured to: selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried by the resource scheduling request; wherein the information comprises at least one of source address information and instance size information of the resource scheduling request.

Further optionally, when the processor 62a selects the target resource scheduling algorithm according to the information carried in the resource scheduling request, it is specifically configured to: determining a service scene for initiating the resource scheduling according to source address information carried in the resource scheduling request; selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the service scene initiating the resource scheduling; wherein, different resource scheduling algorithms correspond to different service scenarios.

Further, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics. Based on this, when the processor 62a selects the target resource scheduling algorithm according to the service scenario initiating the resource scheduling of this time, the processor is specifically configured to: if the service scene initiating the resource scheduling is an inquiry scene, selecting a deterministic heuristic algorithm as a target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene, selecting a nondeterministic heuristic algorithm as a target resource scheduling algorithm.

Further optionally, when the processor 62a selects the target resource scheduling algorithm according to the information carried in the resource scheduling request, it is specifically configured to: selecting a target resource scheduling algorithm from at least one resource scheduling algorithm according to example scale information carried by the resource scheduling request; different resource scheduling algorithms correspond to different example size ranges.

Further, the at least one resource scheduling algorithm comprises: deterministic heuristics and non-deterministic heuristics. Based on this, when the processor 62a selects the target resource scheduling algorithm according to the example scale information carried by the resource scheduling request, it is specifically configured to: if the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as a target resource scheduling algorithm; and if the example scale information carried by the resource scheduling request is greater than or equal to the set scale threshold, selecting a non-deterministic heuristic algorithm as a target resource scheduling algorithm.

Further optionally, when the processor 62a selects the target resource scheduling algorithm according to the information carried in the resource scheduling request, it is specifically configured to: determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request; under the condition that the service scene initiating the resource scheduling is a first service scene, selecting a resource scheduling algorithm corresponding to the first service scene from at least one resource scheduling algorithm as a target resource scheduling algorithm; and under the condition that the service scene initiating the resource scheduling is the second service scene, selecting a resource scheduling algorithm adaptive to the example scale information from at least one resource scheduling algorithm as a target resource scheduling algorithm according to the example scale information carried by the resource scheduling request. Wherein, different resource scheduling algorithms correspond to different service scenes or example scale ranges.

Further, the at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic; correspondingly, the first service scenario is a query scenario, and the second service scenario is a non-query scenario. Based on this, when selecting the target resource scheduling algorithm, the processor 62a is specifically configured to: if the service scene initiating the resource scheduling is an inquiry scene or the example scale information carried by the resource scheduling request is smaller than a set scale threshold, selecting a deterministic heuristic algorithm as a target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene and the example scale information carried by the resource scheduling request is greater than or equal to the set scale threshold, selecting a non-deterministic heuristic algorithm as a target resource scheduling algorithm.

In an optional embodiment, when performing resource scheduling on available resources in the edge cloud node that can be scheduled this time, the processor 55 is specifically configured to: and performing resource scheduling on available resources in the schedulable edge cloud node for at least one specification example by adopting a target resource scheduling algorithm according to the resource constraint condition required by the current resource scheduling.

Further optionally, the processor 62a is further configured to: before using the resource constraint condition required by the current resource scheduling, resolving the resource constraint condition required by the current resource scheduling from the resource scheduling request; or acquiring a locally preset resource constraint condition as a resource constraint condition required by the current resource scheduling.

Optionally, the resource constraint condition required by the current resource scheduling includes an affinity resource constraint condition and an exclusivity resource constraint condition; wherein the resource constraints of the affinity define conditions that require instances to be deployed on the same physical machine; exclusive resource constraints define conditions that require instances deployed on different physical machines to be decentralized.

Further optionally, the processor 62a is further configured to: before resource scheduling is carried out on available resources in the edge cloud nodes which can be scheduled at this time, analyzing specified edge cloud nodes from the resource scheduling request to serve as the edge cloud nodes which can be scheduled at this time; or at least one edge cloud node in a network system to which the central control equipment belongs is used as the schedulable edge cloud node.

In an optional embodiment, the target resource scheduling algorithm is a FF-based resource scheduling algorithm, and based on this, when performing resource scheduling on available resources in the edge cloud node that can be scheduled this time, the processor 62a is specifically configured to: for each example, checking all non-empty physical machines in the edge cloud node which can be scheduled at this time, and if a non-empty physical machine which is most suitable for the example is found and meets the resource constraint condition of the resource scheduling requirement at this time, determining to place the example on the found non-empty physical machine; if not, the instance is determined to be placed on a new empty physical machine.

In an optional embodiment, the target resource scheduling algorithm is a BF-based resource scheduling algorithm, and based on this, the processor 62a is specifically configured to, when performing resource scheduling on available resources in the edge cloud node that can be scheduled this time: for each example, checking all non-empty physical machines in the edge cloud node which can be scheduled at this time, and if a non-empty physical machine which is most suitable for the example is found and the non-empty physical machine meets the resource constraint condition required by the resource scheduling at this time, determining to place the example on the found non-empty physical machine; if not, the instance is determined to be placed on a new empty physical machine.

In an alternative embodiment, the processor 62a is further configured to: when the target resource scheduling algorithm is a deterministic heuristic algorithm and the current resource scheduling adopting the deterministic heuristic algorithm fails, judging whether the difference between the resources required by placing the instances with at least one specification and the available resources in the current schedulable edge cloud node is smaller than a set difference threshold value; if the judgment result is yes, adopting a non-deterministic heuristic algorithm, and performing resource scheduling on available resources in the schedulable edge cloud node for the at least one specification instance again.

Further, as shown in fig. 6a, the center managing and controlling apparatus further includes: communication components 63a, display 64a, power components 65a, audio components 66a, and the like. Only some of the components are schematically shown in fig. 6a, and it is not meant that the central managing device comprises only the components shown in fig. 6 a. In addition, the components within the dashed box in fig. 6a are optional components, not necessary components, and may be determined according to the product form of the central control device. The central control device of this embodiment may be implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, or an IOT device, or may be a server device such as a conventional server, a cloud server, or a server array. If the central control device of this embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., the central control device may include components within a dashed line frame in fig. 6 a; if the central management and control device of this embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, the components in the dashed box in fig. 6a may not be included.

Accordingly, the present application further provides a computer-readable storage medium storing a computer program, where the computer program is capable of implementing the steps that can be executed by the central control device in the foregoing method embodiments when executed.

Fig. 6b is a schematic structural diagram of a terminal device according to an exemplary embodiment of the present application. As shown in fig. 6b, the terminal device includes: a memory 61b, a processor 62b and a communication component 63 b.

The memory 61b is used for storing a computer program and may be configured to store other various data to support operations on the terminal device. Examples of such data include instructions for any application or method operating on the terminal device, contact data, phonebook data, messages, pictures, videos, etc.

A processor 62b, coupled to the memory 61b, for executing computer programs in the memory 61b for: displaying a human-computer interaction interface provided by a central control device; responding to input operation on a human-computer interaction interface, and generating a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling for an instance with at least one specification; and sending the resource scheduling request to the central control equipment so that the central control equipment can perform resource scheduling on available resources in the edge nodes which can be scheduled at this time for at least one specification instance.

Further, the processor 62b is also configured to: receiving a resource scheduling result of the current time returned by the central control equipment; and outputting the resource scheduling result returned by the central control equipment.

Further, as shown in fig. 6b, the terminal device further includes: communication components 63b, display 64b, power components 65b, audio components 66b, and the like. The terminal device of this embodiment may be implemented as a desktop computer, a notebook computer, a smart phone, or an IOT device.

Accordingly, the present application further provides a computer readable storage medium storing a computer program, where the computer program is capable of implementing the steps that can be executed by the resource scheduling demander in the foregoing method embodiments when executed.

The communication components of fig. 6a and 6b described above are configured to facilitate communication between the device in which the communication component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a wireless network based on a communication standard, such as a WiFi, a 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component may further include a Near Field Communication (NFC) module, Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and the like.

The displays in fig. 6a and 6b described above include screens, which may include Liquid Crystal Displays (LCDs) and Touch Panels (TPs). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

The power supply components of fig. 6a and 6b described above provide power to the various components of the device in which the power supply components are located. The power components may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device in which the power component is located.

The audio components of fig. 6a and 6b described above may be configured to output and/or input audio signals. For example, the audio component includes a Microphone (MIC) configured to receive an external audio signal when the device in which the audio component is located is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, 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 (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation 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 may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A network system, comprising: the system comprises a central management and control device and at least one edge cloud node;

the central management and control equipment is used for: receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on at least one specification instance; selecting a target resource scheduling algorithm which is adaptive to the resource scheduling request from at least one resource scheduling algorithm; adopting the target resource scheduling algorithm to perform resource scheduling on available resources in the schedulable edge cloud node for the at least one specification example; the schedulable edge cloud node is a part or all of the at least one edge cloud node.

2. The system according to claim 1, characterized in that the central management and control device is specifically configured to:

providing a human-computer interaction interface facing a resource scheduling demander for the resource scheduling demander to initiate a resource scheduling request; and

and receiving a resource scheduling request provided by the resource scheduling demander through the human-computer interaction interface.

3. The system of claim 2, wherein the central management device is further configured to:

and outputting the result of the current resource scheduling to the resource scheduling demander.

4. The system of claim 1, wherein the central management device is further configured to:

and under the condition that the current resource scheduling is successful, generating a resource scheduling scheme corresponding to the instance of the at least one specification.

5. The system of claim 4, wherein the central management facility is further configured to perform at least one of:

and outputting the resource scheduling scheme.

6. The system according to any one of claims 1 to 5, wherein the central management and control device, when selecting the target resource scheduling algorithm, is specifically configured to:

selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried by the resource scheduling request; wherein the information comprises at least one of source address information and instance size information of the resource scheduling request.

7. The system according to claim 6, characterized in that the central management and control device is specifically configured to:

determining a service scene for initiating the resource scheduling according to the source address information of the resource scheduling request;

selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the service scene initiating the resource scheduling; wherein, different resource scheduling algorithms correspond to different service scenarios.

8. The system according to claim 7, wherein said at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic;

the central management and control device is specifically configured to: if the service scene initiating the resource scheduling is an inquiry scene, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene, selecting a nondeterministic heuristic algorithm as the target resource scheduling algorithm.

9. The system according to claim 6, characterized in that the central management and control device is specifically configured to:

selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the example scale information carried by the resource scheduling request; different resource scheduling algorithms correspond to different example size ranges.

10. The system according to claim 9, wherein said at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic;

the central management and control device is specifically configured to: if the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm; and if the example scale information carried by the resource scheduling request is greater than or equal to a set scale threshold value, selecting a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

11. The system according to claim 6, characterized in that the central management and control device is specifically configured to:

under the condition that the service scene initiating the resource scheduling is a first service scene, selecting a resource scheduling algorithm corresponding to the first service scene from at least one resource scheduling algorithm as the target resource scheduling algorithm;

under the condition that the service scene initiating the current resource scheduling is a second service scene, selecting a resource scheduling algorithm adaptive to the example scale information from at least one resource scheduling algorithm as the target resource scheduling algorithm according to the example scale information carried by the resource scheduling request;

wherein, different resource scheduling algorithms correspond to different service scenes or example scale ranges.

12. The system according to claim 11, wherein said at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic; the first service scene is a query scene, and the second service scene is a non-query scene;

the central management and control device is specifically configured to: if the service scene initiating the resource scheduling is an inquiry scene or the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm; and if the service scene initiating the resource scheduling is a non-query scene and the example scale information carried by the resource scheduling request is greater than or equal to a set scale threshold, selecting a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

13. The system according to any one of claims 1 to 5, wherein the central management and control device is specifically configured to, when performing resource scheduling:

and performing resource scheduling on available resources in the schedulable edge cloud node for the example of the at least one specification by adopting the target resource scheduling algorithm in combination with the resource constraint condition required by the current resource scheduling.

14. The system of claim 13, wherein the central management device is further configured to:

analyzing the resource constraint condition required by the current resource scheduling from the resource scheduling request; or

And acquiring a locally preset resource constraint condition as a resource constraint condition required by the current resource scheduling.

15. The system according to claim 13, wherein the resource constraint condition of the current resource scheduling requirement includes an affinity resource constraint condition and an exclusivity resource constraint condition;

wherein the resource constraints of the affinity define conditions that require instances to be deployed on the same physical machine; the exclusive resource constraints define conditions that require instances deployed on different physical machines to be decentralized.

16. The system of claim 13, wherein the central management device is further configured to:

analyzing the appointed edge cloud node from the resource scheduling request to serve as the schedulable edge cloud node;

or

And taking the at least one edge cloud node as the schedulable edge cloud node.

17. The system according to claim 13, wherein the target resource scheduling algorithm is a FF-based resource scheduling algorithm; the central management and control device is specifically configured to: for each example, checking all non-empty physical machines in the edge cloud node which can be scheduled at this time, and if a first non-empty physical machine which can put down the example is found and the non-empty physical machine meets the resource constraint condition of the resource scheduling requirement at this time, determining to place the example on the found non-empty physical machine; if not, the instance is determined to be placed on a new empty physical machine.

18. The system of claim 13, wherein the target resource scheduling algorithm is a BF-based resource scheduling algorithm; the central management and control device is specifically configured to: for each example, checking all non-empty physical machines in the edge cloud node which can be scheduled at this time, and if a non-empty physical machine which is most suitable for the example is found and meets the resource constraint condition of the resource scheduling requirement at this time, determining to place the example on the found non-empty physical machine; if not, the instance is determined to be placed on a new empty physical machine.

19. The system according to any one of claims 1-5, wherein the central management device is further configured to:

when the target resource scheduling algorithm is a deterministic heuristic algorithm and the current resource scheduling adopting the deterministic heuristic algorithm fails, judging whether the difference between the resources required for placing the at least one specification example and the available resources in the current schedulable edge cloud node is smaller than a set difference threshold value;

and if so, performing resource scheduling on available resources in the schedulable edge cloud node for the at least one specification instance again by adopting a non-deterministic heuristic algorithm.

20. A method for scheduling resources, comprising:

receiving a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling on at least one specification instance;

selecting a target resource scheduling algorithm adaptive to the resource scheduling request from at least one resource scheduling algorithm;

and performing resource scheduling on available resources in the edge cloud node which can be scheduled at this time for the example of the at least one specification by adopting the target resource scheduling algorithm.

21. The method of claim 20, wherein selecting a target resource scheduling algorithm adapted to the resource scheduling request from at least one resource scheduling algorithm comprises:

selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried by the resource scheduling request;

wherein the information comprises at least one of source address information and instance size information of the resource scheduling request.

22. The method of claim 21, wherein selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried in the resource scheduling request comprises:

23. The method of claim 22, wherein the at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic;

the selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the service scene initiating the resource scheduling comprises:

if the service scene initiating the resource scheduling is an inquiry scene, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm;

and if the service scene initiating the resource scheduling is a non-query scene, selecting a nondeterministic heuristic algorithm as the target resource scheduling algorithm.

24. The method of claim 21, wherein selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried in the resource scheduling request comprises:

25. The method of claim 24, wherein the at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic;

the selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the example scale information carried by the resource scheduling request includes:

if the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm;

and if the example scale information carried by the resource scheduling request is greater than or equal to a set scale threshold value, selecting a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

26. The method of claim 21, wherein selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the information carried in the resource scheduling request comprises:

27. The method of claim 26, wherein the at least one resource scheduling algorithm comprises: a deterministic heuristic and a non-deterministic heuristic; the first service scene is a query scene, and the second service scene is a non-query scene;

under the condition that the service scene initiating the resource scheduling is an inquiry scene, selecting a resource scheduling algorithm corresponding to the first service scene from at least one resource scheduling algorithm as the target resource scheduling algorithm, wherein the resource scheduling algorithm comprises the following steps: selecting a deterministic heuristic algorithm as the target resource scheduling algorithm;

under the condition that the service scene initiating the resource scheduling is a non-query scene, selecting the target resource scheduling algorithm from at least one resource scheduling algorithm according to the example scale information carried by the resource scheduling request, wherein the method comprises the following steps: if the example scale information carried by the resource scheduling request is smaller than a set scale threshold value, selecting a deterministic heuristic algorithm as the target resource scheduling algorithm; and if the example scale information carried by the resource scheduling request is greater than or equal to a set scale threshold value, selecting a non-deterministic heuristic algorithm as the target resource scheduling algorithm.

28. The system according to any one of claims 20 to 27, wherein the resource scheduling for the available resources in the schedulable edge cloud node for the example of the at least one specification using the target resource scheduling algorithm comprises:

29. The system according to any of claims 20-27, wherein if said target resource scheduling algorithm is a deterministic heuristic algorithm and the current resource scheduling using said deterministic heuristic algorithm fails, said method further comprises:

judging whether the difference between the resources required for placing the at least one specification example and the available resources in the schedulable edge cloud node is smaller than a set difference threshold value or not;

30. A method for scheduling resources, comprising:

displaying a human-computer interaction interface provided by a central control device;

responding to the input operation on the human-computer interaction interface, and generating a resource scheduling request, wherein the resource scheduling request is used for requesting resource scheduling for at least one specification instance;

and sending the resource scheduling request to the central control equipment so that the central control equipment can perform resource scheduling on available resources in the edge nodes which can be scheduled at this time for the at least one specification instance.

31. The method of claim 30, further comprising:

receiving a resource scheduling result of the current time returned by the central control equipment;

and outputting the resource scheduling result returned by the central control equipment.

32. A central management and control device, comprising: a memory and a processor;

the memory for storing a computer program; the computer program, when executed by the processor, causes the processor to carry out the steps of the method of any one of claims 20-29.

33. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by one or more processors, causes the one or more processors to implement the steps of the method of any one of claims 20-31.