WO2023186002A1 - 一种资源调度方法、装置及设备 - Google Patents

一种资源调度方法、装置及设备 Download PDF

Info

Publication number
WO2023186002A1
WO2023186002A1 PCT/CN2023/085043 CN2023085043W WO2023186002A1 WO 2023186002 A1 WO2023186002 A1 WO 2023186002A1 CN 2023085043 W CN2023085043 W CN 2023085043W WO 2023186002 A1 WO2023186002 A1 WO 2023186002A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
resource
scheduling
resources
node
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/085043
Other languages
English (en)
French (fr)
Inventor
姚惠娟
陆璐
付月霞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Mobile Communications Group Co Ltd
Research Institute of China Mobile Communication Co Ltd
Original Assignee
China Mobile Communications Group Co Ltd
Research Institute of China Mobile Communication Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China Mobile Communications Group Co Ltd, Research Institute of China Mobile Communication Co Ltd filed Critical China Mobile Communications Group Co Ltd
Publication of WO2023186002A1 publication Critical patent/WO2023186002A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/15—Flow control; Congestion control in relation to multipoint traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/78—Architectures of resource allocation
    • H04L47/783—Distributed allocation of resources, e.g. bandwidth brokers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/104—Peer-to-peer [P2P] networks
    • H04L67/1074—Peer-to-peer [P2P] networks for supporting data block transmission mechanisms

Definitions

  • the present disclosure relates to the field of communication technology, and in particular, to a resource scheduling method, apparatus and equipment.
  • Computing Force Network (Computing Force Network) embeds computing resources and computing capabilities into the network through a high degree of collaboration between the network and computing, achieving efficient collaboration between cloud, network, edge, and terminal, and achieving a balanced distribution of services in the network and among computing resources.
  • Computing network is not a single technology, it is the integration and innovative application of a series of new network technologies.
  • the computing power network requires collaborative innovation in the network domain and computing domain.
  • the core routing protocol needs to be innovated to achieve fast, flexible, and on-demand business request scheduling.
  • the network orchestration management system needs to be innovated to achieve Union and collaborative orchestration of computing network resources.
  • computing power networks need to achieve unified measurement of heterogeneous computing power, establish a unified measurement mechanism, and further form a unified modeling mechanism for heterogeneous computing power.
  • the computing power network requires collaborative innovation based on the network domain and the computing domain to help operator networks transform from providing "connection" to providing "connection + computing” comprehensive information and communications technology (ICT) services.
  • ICT information and communications technology
  • edge computing, heterogeneous computing, and artificial intelligence computing power networks need to evolve collaboratively from aspects such as architecture, protocols, and metrics.
  • infrastructure as a service Infrastructure as a Service
  • IaaS Infrastructure as a Service
  • PaaS Platform as a Service
  • SaaS Software as a Service
  • NaaS Network as a Service
  • XaaS everything-as-a-service
  • the purpose of this disclosure is to provide a resource scheduling method, device and equipment to solve the problem in related technologies that the computing power network cannot realize the working mechanism of centralized orchestration and distributed scheduling collaboration.
  • embodiments of the present disclosure provide a resource scheduling method, which is applied to the first node, including:
  • target scheduling policy information perform resource scheduling for target business data corresponding to the target business requirements
  • the resources include: at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for the real-time status of resources;
  • the topology information is topology information for computing power, services and networks.
  • target scheduling policy information for target business requirements it also includes:
  • Obtaining target scheduling strategy information for target business requirements based on the overall scheduling strategy information includes:
  • the real-time status information of resources includes: real-time status information of resources for computing resources, At least one of resource real-time status information for service resources and resource real-time status information for network resources.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes:
  • Configure collection parameter information for resource real-time status information according to the topology information and configuration information in the overall scheduling policy information
  • the collection parameter information includes: at least one of collection method information and collection interval information.
  • Embodiments of the present disclosure also provide a resource scheduling method, applied to the second node, including:
  • the resource-related information refers to at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for the real-time status of resources;
  • the topology information is topology information for computing power, services and networks.
  • An embodiment of the present disclosure also provides a resource scheduling device, which is applied to the first node.
  • the resource scheduling device includes:
  • the first receiving module is used to receive the overall scheduling policy information sent by the second node;
  • a first acquisition module configured to acquire target scheduling strategy information for target business requirements according to the overall scheduling strategy information
  • the first scheduling module is configured to, according to the target scheduling policy information, target the target service needs. Find the corresponding target business data for resource scheduling;
  • the resources include: at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for the real-time status of resources;
  • the topology information is topology information for computing power, services and networks.
  • the resource scheduling device also includes:
  • the second acquisition module is configured to acquire real-time resource status information based on the topology information and configuration information in the overall scheduling policy information before acquiring target scheduling policy information for target business requirements according to the overall scheduling policy information;
  • Obtaining target scheduling strategy information for target business requirements based on the overall scheduling strategy information includes:
  • the real-time resource status information includes: at least one of real-time resource status information for computing resources, real-time resource status information for service resources, and real-time resource status information for network resources.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes:
  • Configure collection parameter information for resource real-time status information according to the topology information and configuration information in the overall scheduling policy information
  • the collection parameter information includes: at least one of collection method information and collection interval information.
  • An embodiment of the present disclosure also provides a resource scheduling device, which is applied to the second node.
  • the resource scheduling device includes:
  • the third acquisition module is used to obtain resource-related information of the computing power network
  • a first determination module configured to determine the overall scheduling policy information for each first node based on the resource-related information
  • a first sending module configured to send the overall scheduling policy information to the corresponding first node
  • the resource-related information refers to at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for the real-time status of resources;
  • the topology information is topology information for computing power, services and networks.
  • determining overall scheduling policy information for each first node based on the resource-related information includes:
  • Embodiments of the present disclosure also provide a resource scheduling device.
  • the resource scheduling device is a first node and includes: a processor and a transceiver;
  • the processor is configured to receive the overall scheduling policy information sent by the second node through the transceiver;
  • target scheduling policy information perform resource scheduling for target business data corresponding to the target business requirements
  • the resources include: at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for the real-time status of resources;
  • the topology information is topology information for computing power, services and networks.
  • the processor is also used to:
  • Obtaining target scheduling strategy information for target business requirements based on the overall scheduling strategy information includes:
  • the real-time resource status information includes: at least one of real-time resource status information for computing resources, real-time resource status information for service resources, and real-time resource status information for network resources.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes:
  • Configure collection parameter information for resource real-time status information according to the topology information and configuration information in the overall scheduling policy information
  • the collection parameter information includes: at least one of collection method information and collection interval information.
  • An embodiment of the present disclosure also provides a resource scheduling device, which is a second node and includes: a processor and a transceiver;
  • the processor is used to obtain resource-related information of the computing power network
  • the resource-related information refers to at least one of computing resources, service resources and network resources.
  • the topology information is topology information for computing power, services and networks.
  • determining overall scheduling policy information for each first node based on the resource-related information includes:
  • An embodiment of the present disclosure also provides a resource scheduling device, including a memory, a processor, and a program stored in the memory and executable on the processor; when the processor executes the program, the above-mentioned first step is implemented.
  • Resource scheduling method on the node side; or,
  • the resource scheduling method on the second node side is implemented.
  • Embodiments of the present disclosure also provide a readable storage medium on which a program is stored, and when the program is executed by a processor, the steps in the resource scheduling method on the first node side are implemented; or,
  • each first node obtains target scheduling policy information (that is, specific scheduling policy information) based on the overall scheduling policy information to perform resource scheduling and achieve distributed scheduling; thereby realizing centralized orchestration of the computing power network (such as overall resource A collaborative working mechanism between centralized orchestration by the second node) and distributed scheduling (for example, specific resources are scheduled by each first node); specifically, through the collaborative working mechanism between centralized orchestration and distributed scheduling, computing resources, service resources, Integrated scheduling with network resources provides users with rapid supply of computing power network services; it effectively solves the problem in related technologies that computing power networks cannot achieve centralized orchestration and distributed scheduling collaborative working mechanisms.
  • Figure 3 is a schematic diagram of the computing power network architecture of an embodiment of the present disclosure.
  • Figure 4 is a schematic diagram of the collaborative work of centralized orchestration and distributed scheduling according to an embodiment of the present disclosure
  • Figure 5 is a schematic diagram 1 of an example of collaborative work of centralized orchestration and distributed scheduling according to an embodiment of the present disclosure
  • Figure 6 is a schematic diagram 2 of an example of collaborative work of centralized orchestration and distributed scheduling according to an embodiment of the present disclosure
  • Figure 7 is a schematic structural diagram of a resource scheduling device according to an embodiment of the present disclosure.
  • Figure 8 is a schematic structural diagram 2 of a resource scheduling device according to an embodiment of the present disclosure.
  • Figure 9 is a schematic structural diagram of a resource scheduling device according to an embodiment of the present disclosure.
  • Figure 10 is a schematic structural diagram 2 of a resource scheduling device according to an embodiment of the present disclosure.
  • This disclosure aims at the problem in related technologies that the computing power network cannot realize the working mechanism of centralized orchestration and distributed scheduling collaboration, and provides a resource scheduling method, which is applied to the first node, as shown in Figure 1, including:
  • Step 11 Receive the overall scheduling policy information sent by the second node
  • Step 12 Obtain target scheduling strategy information for target business requirements based on the overall scheduling strategy information
  • Step 13 According to the target scheduling policy information, perform resource scheduling for the target business data corresponding to the target business demand; wherein the resources include: at least one of computing resources, service resources, and network resources.
  • the first node may include at least one of the computing power nodes, network nodes, computing network integration nodes and other nodes of the computing power network; the overall scheduling policy information is the resource related information of the second node according to the computing power network Obtained; wherein, the resource-related information is for at least one of computing resources, service resources and network resources; it can be understood that: the resource-related information includes: resource-related information for computing resources, service resources At least one of resource-related information and resource-related information for network resources. Resource-related information can also be understood as resource-related information for at least two first nodes in the computing power network, such as resource-related information used for resource allocation by at least two first nodes, but is not limited to this. Resource-related information may include: resource registration information.
  • the resource scheduling method receives the overall scheduling policy information sent by the second node; according to the overall scheduling policy information, obtains the target scheduling policy information for the target business requirements; according to the target scheduling policy information, Resource scheduling is performed for target business data corresponding to the target business requirements; wherein the resources include: at least one of computing resources, service resources, and network resources; capable of supporting centralized orchestration by the second node to obtain The overall scheduling strategy information, and each first node obtains the target scheduling strategy information (i.e., specific scheduling strategy information) based on the overall scheduling strategy information to perform resource scheduling and realize distributed scheduling; thereby realizing centralized orchestration of the computing power network (such as The overall resources are centrally orchestrated by the second node) and distributed scheduling (for example, specific resources are scheduled by each first node).
  • the target scheduling strategy information i.e., specific scheduling strategy information
  • computing resources and services can be realized through the working mechanism of centralized orchestration and distributed scheduling.
  • Integrated scheduling of resources and network resources provides users with rapid supply of computing network services; it effectively solves the problem in related technologies that computing networks cannot achieve centralized orchestration and distributed scheduling collaborative working mechanisms.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information, and configuration information collected for real-time status of resources; wherein the topology information is for computing power, services, and networks. topological information.
  • the policy execution node information can be used to determine possible controllers that execute the policy, or specific computing power routing nodes.
  • Topology information can be used to determine the computing power topology, network topology and service topology information of the entire network. If specific business request information (such as the above-mentioned target business requirements) is received later and a specific service is requested, the policy execution node can be determined based on the policy execution node information. The policy execution node will search for one or more deployed services based on the above topology information.
  • Computing power node specifically:
  • the business request type is computing power type
  • periodic collection can be carried out by extending the Border Gateway Protocol (BGP) or Interior Gateway Protocol (IGP).
  • BGP Border Gateway Protocol
  • IGP Interior Gateway Protocol
  • the required collection can be performed by extending in-band or out-of-band Operation Administration and Maintenance (OAM) information, but is not limited to this.
  • OAM Operation Administration and Maintenance
  • content not included in the overall scheduling policy information can be replaced by local configuration or locally stored relevant content, but is not limited to this.
  • target scheduling policy information for the target business requirements according to the overall scheduling policy information it also includes: obtaining the real-time status information of the resources according to the topology information and configuration information in the overall scheduling policy information;
  • Obtaining target scheduling strategy information for target business needs according to the overall scheduling strategy information includes: acquiring target scheduling strategy for target business needs based on scheduling strategy information in the overall scheduling strategy information and the real-time resource status information.
  • Information; wherein the resource real-time status information includes: at least one of resource real-time status information for computing resources, resource real-time status information for service resources, and resource real-time status information for network resources.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes: configuring real-time resource status based on the topology information and configuration information in the overall scheduling policy information.
  • Collection parameter information of status information obtain real-time status information of resources according to the collection parameter information; wherein the collection parameter information includes: at least one of collection method information and collection interval information.
  • Embodiments of the present disclosure also provide a resource scheduling method, applied to the second node, as shown in Figure 2, including:
  • Step 21 Obtain resource-related information of the computing power network
  • Step 22 Determine the overall scheduling strategy information for each first node (in the computing power network) based on the resource-related information
  • Step 23 Send the overall scheduling policy information to the corresponding first node; wherein the resource-related information is for at least one of computing resources, service resources and network resources (understandable It is: the resource-related information includes: at least one of resource-related information for computing resources, resource-related information for service resources, and resource-related information for network resources).
  • the resource-related information may include: resource registration information; the overall scheduling policy information of each first node is the same or different, which is not limited here.
  • Resource-related information can also be understood as resource-related information for at least two first nodes in the computing power network, such as resource-related information used for resource allocation by at least two first nodes, but is not limited to this.
  • the resource scheduling method obtains resource-related information of the computing power network; determines the overall scheduling strategy information for each first node based on the resource-related information; and sends the overall scheduling strategy information to the corresponding The first node of One node obtains the target scheduling strategy information (i.e., specific scheduling strategy information) based on the overall scheduling strategy information to perform resource scheduling and achieve distributed scheduling; thereby achieving centralized orchestration of the computing power network (for example, the overall resources are centrally orchestrated by the second node)
  • a working mechanism that coordinates with distributed scheduling for example, specific resources are scheduled by each first node); specifically, the integrated scheduling of computing resources, service resources, and network resources can be achieved through a working mechanism that coordinates centralized orchestration and distributed scheduling. , providing users with the rapid supply of computing power network services; well solving the problem in related technologies that the computing power network cannot realize the working mechanism of centralized orchestration and distributed scheduling coordination.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information and configuration information collected for real-time status of resources; wherein the topology information is topology information for computing power, services and networks.
  • determining the overall scheduling policy information for each first node based on the resource-related information includes: obtaining topology information based on the resource-related information; and obtaining topology information based on historical business requirements and the topology information. At least one item of determining the overall scheduling policy information for each first node.
  • the method further includes: receiving the failure notification information sent by the first node; performing at least one of the following operations according to the failure notification information: (1) Returning to execute the "Obtain resources related to the computing power network" again. Information; determine the overall scheduling strategy information for each first node based on the resource-related information; send the overall scheduling strategy information to the corresponding first node"; (2) Directly return to re-execute "according to the resource-related information, determine the overall scheduling strategy information for each first node; send the overall scheduling strategy information to the corresponding first node" (3) According to the resource-related information and/ or overall scheduling policy information, to schedule resources for the service data corresponding to the failure notification information.
  • the computing power network may also be called a computing network fusion system, a computing power aware network, a computing power endogenous network or a computing network integrated network, which is not limited here.
  • the following is an example of the resource scheduling method provided by the embodiment of the present disclosure.
  • a resource scheduling method which can be specifically implemented as a collaborative method of centralized orchestration and distributed scheduling (for a computing power network). More specifically, it can be: computing-oriented
  • the future data communication network of network integration also known as computing power network or computing power aware network or computing power endogenous network, provides a working mechanism of centralized orchestration and distributed scheduling collaboration, and defines new functional modules and interfaces Support collaborative work.
  • this solution can be implemented based on a new network architecture - Computing-aware Networking (CAN).
  • CAN can interconnect dynamically distributed computing resources based on ubiquitous network connections, through network, storage, computing power
  • the unified and coordinated scheduling of multi-dimensional resources enables a large number of applications to call computing resources in different places on demand in real time, achieve global optimization of connections and computing power in the network, and provide a consistent user experience.
  • this solution provides a computing and network collaborative orchestration and scheduling mechanism based on CAN, which can achieve integrated scheduling of computing resources and network resources through centralized orchestration and distributed scheduling collaborative working mechanisms, and provide users with computing power. Rapid provisioning of network services.
  • FIG. 3 a computing network architecture that supports collaborative orchestration of computing networks.
  • VR in the figure represents virtual reality.
  • AR stands for Augmented Reality
  • V2X stands for Vehicle to everything
  • APP stands for application
  • API Gateway stands for application interface gateway
  • CPU stands for Central Processing Unit
  • GPU stands for Graphics Processing Unit
  • NPU Network Processing Unit
  • Stroage stands for memory
  • FPGA Field Programmable Gate Array.
  • ASICs stands for Application Specific Integrated Circuits (Application Specific Integrated Circuits)
  • OAM stands for operation management and maintenance.
  • the computing power-aware network architecture system can be divided into a computing power service layer, a computing network management and orchestration layer, and a computing power resources logically.
  • a computing power service layer There are five functional modules: layer, computing power routing layer and network resource layer.
  • the computing power routing layer includes the control plane and the forwarding plane.
  • Computing power service layer carries various services and applications of ubiquitous computing, supports distributed microservice architecture, and supports application decomposition into atomic functional components and algorithm libraries, which are applied by the application interface gateway (API Gateway)
  • API Gateway application interface gateway
  • Unified scheduling realizes functions such as service decomposition and service scheduling.
  • Computing power routing layer includes the control plane and the forwarding plane; based on the abstracted computing network resource discovery and comprehensive consideration of network conditions and computing resource conditions, services can be flexibly scheduled to different computing resource nodes on demand.
  • Computing network management and orchestration layer It needs to support the registration, update and deregistration of computing nodes, network nodes and service information of the entire network, that is, it supports the computing power resource layer, network resource layer and computing power service layer.
  • the scheduling and orchestration center (corresponding to the computing network management and orchestration layer) registers and generates computing power, services and network topology information (corresponding to the above topology information).
  • Computing resource layer Utilize existing computing infrastructure to provide computing resources.
  • the computing infrastructure includes: from single-core CPU (central processing unit) to multi-core CPU, to CPU+GPU+FPGA (field programmable gate array) and other combinations of various computing capabilities; in order to meet the diverse computing needs in the edge computing field and for different applications, on the basis of physical computing resources, computing power models, computing power application programming interfaces (APIs), and computing networks are provided Resource identification and other functions.
  • Network resource layer Utilize existing network infrastructure to provide ubiquitous network connections to every corner of the network.
  • Network infrastructure includes access networks, metropolitan area networks, and backbone networks.
  • the computing resource layer and the network resource layer are the infrastructure layers of the computing network.
  • the computing network management and orchestration layer and the computing routing layer are the two core functional modules that realize the computing power awareness functional system. Users and applications Access the network through the computing power routing layer.
  • the computing power-aware network architecture is based on the five defined functional modules to realize the perception, control and scheduling of computing network resources.
  • this solution provides collaboration through centralized orchestration + distributed scheduling, as shown in Figure 4 (a collaborative working diagram that supports computing network collaborative orchestration and distributed scheduling), realizing computing power resources and network Integrated scheduling of resources provides users and applications with rapid supply of computing power network services.
  • IoT in the figure stands for Internet of Things
  • MEP stands for edge computing platform
  • DCGW stands for Data Center-Gateway.
  • Figure 4 can be understood as the topology diagram corresponding to Figure 3. in:
  • Computing network orchestration management center (corresponding to the computing network management and orchestration layer): completes the unified management of computing resources and network resources, including the perception, measurement and OAM (operation and maintenance management) management of computing resources and network resources, etc. ; Construct global topology information of computing power, network and services (corresponding to the above topology information) to achieve unified operation of the computing network. Based on the received computing power, network, and service information, and according to business needs, the computing network orchestration management center further coordinates the scheduling strategy.
  • the management plane corresponding to the computing network orchestration management center
  • the control plane Corresponding to the controller, it can be integrated into the computing network orchestration management center) and/or the data plane (corresponding to the computing power routing layer) to execute the scheduling strategy.
  • the computing network orchestration management center needs to support the management of registration, update and cancellation of computing power nodes, network nodes and service information of the entire network. For example, after a computing power node goes online, it notifies its computing power enabling information to the computing power management platform, including the computing power node identification or the identification of computing power resources, including equipment type, chip type, storage type, computing type and other information (corresponding to information related to the above resources). It may further include but is not limited to the computing network management layer (corresponding to the computing network management and orchestration layer) that needs to implement method configuration and management including awareness of services, networks and computing resources, mainly including but not limited to:
  • the template includes perception parameters, such as service ID (and/or function ID), at least one of the parameters such as available computing resources, available bandwidth, and/or the number of available connections for the service. and other information.
  • the computing network orchestration management layer (i.e., the computing network management orchestration layer) can adaptively order the collected sensing and measurement parameters, feedback frequency, etc.
  • the computing network management layer can use the received multi-dimensional resources to
  • the sensing and measurement results are updated to form the initial topology of computing power, services and network (corresponding to the above topology information), which can then realize network + computing power programmability and automatic business adaptation.
  • the computing network management and orchestration layer, controller (not shown in Figure 3) and computing power routing nodes (corresponding to the computing power routing layer) are the core functional modules to realize collaborative functional work:
  • Computing network orchestration management center responsible for maintaining global static computing power and network topology information (corresponding to the above-mentioned topology information).
  • the overall scheduling execution strategy (corresponding to the above-mentioned overall scheduling strategy information) is generated according to the network deployment or (historical) business requirements.
  • the specific scheduling (corresponding to the above-mentioned target scheduling strategy information) can be done at the control level (corresponding to the controller) or the data level (corresponding to the controller). Corresponding to the computing power routing node) do.
  • the controller or computing power routing node (corresponding to the above-mentioned first node): is responsible for maintaining the status information of the computing power and network (corresponding to the real-time status of the above-mentioned resources) according to the received scheduling instruction message (corresponding to the above-mentioned overall scheduling policy information) information), combined with business needs (corresponding to the above target business needs), schedule the business to the appropriate node.
  • the computing network orchestration management center receives the computing power nodes, network nodes, computing network integration nodes, service information registration and other information (corresponding to the above resource-related information) of the entire computing network (corresponding to the computing power network).
  • the way to obtain information can be through the registration of various computing power network nodes, or through various information obtained from network management platforms or cloud management platforms.
  • the computing network orchestration management center can update the initial topology information of computing power, services and networks (corresponding to the above topology information).
  • the computing network orchestration management center generates an overall scheduling strategy (corresponding to the above overall scheduling strategy information) based on business requirements (corresponding to the above historical business requirements) or network deployment conditions (corresponding to the above topology information), specifically including: scheduling executors (controller or computing power routing node, corresponding to the above-mentioned policy execution node information), topology information indication (corresponding to the above-mentioned topology information), real-time resource status information configuration (corresponding to the above-mentioned configuration information) and other information, and provide information to the controller or The computing power routing node sends the scheduling strategy (corresponding to the above overall scheduling strategy information).
  • scheduling executors controller or computing power routing node, corresponding to the above-mentioned policy execution node information
  • topology information indication corresponding to the above-mentioned topology information
  • real-time resource status information configuration corresponding to the above-mentioned configuration information
  • the controller or computing power routing node configures the real-time computing power information (corresponding to the real-time status information of the above resources) based on the topology information and real-time resource information configuration (corresponding to the above-mentioned configuration information) in the received overall scheduling policy information.
  • the collection method corresponding to the above collection method information
  • interval corresponding to the above collection interval information
  • other information corresponding to the above collection parameter information
  • Nodes such as computing resource nodes and network resource nodes collect computing resource information and network information.
  • the collection method it can be BGP or IGP extension, or it can be OAM mechanism in-band or out-of-band OAM.
  • the collection method can be updated proactively or triggered by the business. For detailed collection methods, please refer to the current solution and will not be repeated here.
  • PE represents the operator edge router (Provider Edge, corresponding to the routing node)
  • AS represents the autonomous domain.
  • FIG. 6 An example of collaborative orchestration and management of computing network (without controller collaboration) can be seen in Figure 6.
  • the computing power routing node performs collaborative orchestration and scheduling of computing power and network, that is, the computing network management and orchestration layer will perform static computing power and network coordination.
  • Topology information specific scheduling is implemented by the data plane of the computing power routing node, and distributed scheduling is done through the computing power routing node.
  • the solution provided by the embodiments of the present disclosure relates to a collaborative method and system for centralized orchestration and distributed scheduling, wherein:
  • the second node performs centralized orchestration: completes the registration, update and deregistration management of computing power, network and service information, generates computing power and network topology (corresponding to the above topology information), and needs to complete multi-dimensional resource awareness configuration and management, while generating an overall scheduling strategy. It may further include: completing the delivery of multi-dimensional resource sensing templates to the computing power service layer, computing power resource layer and network resource layer, so that the computing power service layer, computing power resource layer and network resource layer can perform resource sensing and measurement according to the template. .
  • the second node generates an overall scheduling execution strategy based on network deployment, and notifies the computing power routing node to execute the scheduling strategy.
  • the computing power routing node maintains real-time status resource information of computing power resources and network resources. After receiving a business request, it can comprehensively consider the resource status information (to generate specific policies) and schedule the business to Suitable node.
  • the computing network orchestration management center receives information such as registration of computing power nodes, network nodes, computing network integration nodes, and service information of the entire computing network. Based on this information, it can update the topology information of computing power, services, and networks.
  • the computing network orchestration management center generates an overall scheduling strategy based on business needs and topology information, and sends the scheduling strategy to the controller or computing power routing node.
  • the controller or computing power routing node maintains real-time status resource information of computing power resources and network resources. After receiving a business request, it can comprehensively consider the resource status information and schedule the business to the appropriate node.
  • this solution provides a method and system that supports collaborative management and orchestration of computing networks and can:
  • Embodiments of the present disclosure also provide a resource scheduling device, applied to the first node, as shown in Figure 7, including:
  • the first receiving module 71 is used to receive the overall scheduling policy information sent by the second node;
  • the first acquisition module 72 is configured to acquire target scheduling policy information for target business requirements according to the overall scheduling policy information
  • the first scheduling module 73 is configured to perform resource scheduling for target business data corresponding to the target business requirements according to the target scheduling policy information;
  • the resources include: at least one of computing resources, service resources and network resources.
  • the resource scheduling device receives the overall scheduling strategy information sent by the second node; according to the overall scheduling strategy information, obtains the target scheduling strategy information for the target business requirements; according to the target scheduling strategy information, Resource scheduling is performed for target business data corresponding to the target business requirements; wherein the resources include: at least one of computing resources, service resources, and network resources; capable of supporting centralized orchestration by the second node to obtain Overall scheduling strategy information, and each first node obtains the target scheduling strategy based on the overall scheduling strategy information.
  • One-node scheduling collaborative working mechanism; specifically, through the working mechanism of centralized orchestration and distributed scheduling collaboration, integrated scheduling of computing resources, service resources, and network resources can be achieved to provide users with rapid supply of computing network services. ; It well solves the problem in related technologies that the computing power network cannot realize the working mechanism of centralized orchestration and distributed scheduling coordination.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information and configuration information collected for real-time status of resources; wherein the topology information is topology information for computing power, services and networks.
  • the resource scheduling device further includes: a second acquisition module, configured to obtain target scheduling policy information for target business requirements according to the overall scheduling policy information according to the overall scheduling policy information. Topology information and configuration information, and obtaining real-time status information of resources; obtaining target scheduling strategy information for target business requirements according to the overall scheduling strategy information, including: according to the scheduling strategy information in the overall scheduling strategy information and the Resource real-time status information is used to obtain target scheduling policy information for target business requirements; wherein, the resource real-time status information includes: resource real-time status information for computing resources, resource real-time status information for service resources, and network resources. At least one item of real-time status information of the resource.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes: configuring real-time resource status based on the topology information and configuration information in the overall scheduling policy information.
  • Collection parameter information of status information obtain real-time status information of resources according to the collection parameter information; wherein the collection parameter information includes: at least one of collection method information and collection interval information.
  • the above-mentioned implementation embodiments of the resource scheduling method on the first node side are applicable to the embodiments of the resource scheduling device, and can also achieve the same technical effect.
  • Embodiments of the present disclosure also provide a resource scheduling device, applied to the second node, as shown in Figure 8, including:
  • the third acquisition module 81 is used to acquire resource-related information of the computing power network
  • the first determination module 82 is configured to determine the overall scheduling policy information for each first node based on the resource-related information
  • the first sending module 83 is used to send the overall scheduling policy information to the corresponding first node
  • the resource-related information refers to at least one of computing resources, service resources and network resources.
  • the resource scheduling device obtains resource-related information of the computing power network; determines the overall scheduling strategy information for each first node based on the resource-related information; and sends the overall scheduling strategy information to the corresponding The first node of One node obtains the target scheduling strategy information (i.e., specific scheduling strategy information) based on the overall scheduling strategy information to perform resource scheduling and achieve distributed scheduling; thereby achieving centralized orchestration of the computing power network (for example, the overall resources are centrally orchestrated by the second node)
  • a working mechanism that coordinates with distributed scheduling for example, specific resources are scheduled by each first node); specifically, the integrated scheduling of computing resources, service resources, and network resources can be achieved through a working mechanism that coordinates centralized orchestration and distributed scheduling. , providing users with the rapid supply of computing power network services; well solving the problem in related technologies that the computing power network cannot realize the working mechanism of centralized orchestration and distributed scheduling coordination.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information and configuration information collected for real-time status of resources; wherein the topology information is topology information for computing power, services and networks.
  • determining the overall scheduling policy information for each first node based on the resource-related information includes: obtaining topology information based on the resource-related information; and obtaining topology information based on historical business requirements and the topology information. At least one item of determining the overall scheduling policy information for each first node.
  • the above-mentioned implementation embodiments of the resource scheduling method on the second node side are applicable to the embodiments of the resource scheduling device, and can also achieve the same technical effect.
  • Embodiments of the present disclosure also provide a resource scheduling device.
  • the resource scheduling device is a first node. As shown in Figure 9, it includes: a processor 91 and a transceiver 92;
  • the processor 91 is configured to receive the overall scheduling policy information sent by the second node through the transceiver 92;
  • target scheduling strategy information for target business requirements is obtained. interest;
  • target scheduling policy information perform resource scheduling for target business data corresponding to the target business requirements
  • the resources include: at least one of computing resources, service resources and network resources.
  • the resource scheduling device receives the overall scheduling strategy information sent by the second node; according to the overall scheduling strategy information, obtains the target scheduling strategy information for the target business requirements; according to the target scheduling strategy information, Resource scheduling is performed for target business data corresponding to the target business requirements; wherein the resources include: at least one of computing resources, service resources, and network resources; capable of supporting centralized orchestration by the second node to obtain The overall scheduling strategy information, and each first node obtains the target scheduling strategy information (i.e., specific scheduling strategy information) based on the overall scheduling strategy information to perform resource scheduling and realize distributed scheduling; thereby realizing centralized orchestration of the computing power network (such as The overall resources are centrally orchestrated by the second node) and distributed scheduling (for example, specific resources are scheduled by each first node).
  • the target scheduling strategy information i.e., specific scheduling strategy information
  • computing resources and services can be realized through the working mechanism of centralized orchestration and distributed scheduling.
  • Integrated scheduling of resources and network resources provides users with rapid supply of computing network services; it effectively solves the problem in related technologies that computing networks cannot achieve centralized orchestration and distributed scheduling collaborative working mechanisms.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information and configuration information collected for real-time status of resources; wherein the topology information is topology information for computing power, services and networks.
  • the processor is further configured to: before obtaining target scheduling policy information for target business requirements according to the overall scheduling policy information, obtain real-time resource allocation based on the topology information and configuration information in the overall scheduling policy information.
  • Status information obtaining target scheduling policy information for target business requirements according to the overall scheduling policy information includes: obtaining target scheduling policy information for the target business according to the scheduling policy information in the overall scheduling policy information and the real-time status information of the resources.
  • Required target scheduling policy information wherein the real-time resource status information includes: at least one of resource real-time status information for computing resources, resource real-time status information for service resources, and resource real-time status information for network resources.
  • obtaining real-time resource status information based on the topology information and configuration information in the overall scheduling policy information includes:
  • the above-mentioned implementation embodiments of the resource scheduling method on the first node side are applicable to the embodiments of the resource scheduling device, and can also achieve the same technical effect.
  • the embodiment of the present disclosure also provides a resource scheduling device.
  • the resource scheduling device is a second node. As shown in Figure 10, it includes: a processor 101 and a transceiver 102;
  • the processor 101 is used to obtain resource-related information of the computing power network
  • the resource-related information refers to at least one of computing resources, service resources and network resources.
  • the overall scheduling policy information includes: at least one of policy execution node information, topology information and configuration information collected for real-time status of resources; wherein the topology information is topology information for computing power, services and networks.
  • the resource-related information for each first node is determined according to the resource-related information.
  • the overall scheduling strategy information includes: obtaining topology information based on the resource-related information; and determining overall scheduling strategy information for each first node based on at least one of historical business requirements and the topology information.
  • the above-mentioned implementation embodiments of the resource scheduling method on the second node side are applicable to the embodiments of the resource scheduling device, and can also achieve the same technical effect.
  • An embodiment of the present disclosure also provides a resource scheduling device, including a memory, a processor, and a program stored in the memory and executable on the processor; when the processor executes the program, the above-mentioned first step is implemented.
  • the implementation embodiments of the resource scheduling method on the first node side or the second node side are applicable to the embodiments of the resource scheduling device, and can also achieve the same technical effect.
  • Embodiments of the present disclosure also provide a readable storage medium on which a program is stored.
  • the program is executed by a processor, the steps in the resource scheduling method on the first node side are implemented; or, when the program is executed by the processor, Implement the steps in the resource scheduling method on the second node side.
  • the above-mentioned implementation embodiments of the resource scheduling method on the first node side or the second node side are applicable to the embodiments of the readable storage medium, and can also achieve the same technical effect.
  • modules may be implemented in software so as to be executed by various types of processors.
  • an identified module of executable code may include one or more physical or logical blocks of computer instructions, which may be structured, for example, as an object, procedure, or function. Nonetheless, the executable code of an identified module need not be physically located together, but may include different instructions stored on different bits that, when logically combined, constitute the module and implement the provisions of the module Purpose.
  • an executable code module can be a single instruction or many instructions, and can even be distributed over multiple different code segments, distributed among different programs, and distributed across multiple memory devices.
  • operational data may be identified within modules and may be implemented in any suitable form and organized within any suitable type of data structure. The operational data may be collected as a single data set or may be distributed across different locations (including on different storage devices). on) and may exist, at least in part, solely as electronic signals on a system or network.
  • the module can be implemented using software, taking into account the level of existing hardware technology, those skilled in the art can build corresponding hardware circuits to implement the corresponding functions without considering the cost.
  • the hardware circuits include conventional Very Large Scale Integration Circuit (VLSI) circuits or gate arrays, as well as existing semiconductors such as logic chips, transistors, or other discrete components.
  • VLSI Very Large Scale Integration Circuit
  • Modules can also be implemented using programmable hardware devices, such as field programmable gate arrays, programmable array logic, programmable logic devices, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本公开提供了一种资源调度方法、装置及设备，其中，资源调度方法包括：接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。

Description

一种资源调度方法、装置及设备
相关申请的交叉引用
本申请主张在2022年4月1日在中国提交的中国专利申请号No.202210347746.4的优先权，其全部内容通过引用包含于此。
技术领域
本公开涉及通信技术领域，尤其涉及一种资源调度方法、装置及设备。
背景技术
结合未来计算形态云-边-端泛在分布的趋势，计算与网络的融合将会更加紧密，为了提升“联接+计算”的能力，需要计算和网络两大产业的有机协同，相互配合，亟需探索云-边-端多级计算资源和服务能力如何通过网络实现智能调度和高效分配。算力网络(Computing Force Network)通过网络和计算高度协同，将计算资源和计算能力嵌入网络，实现云、网、边、端的高效协同，实现业务在网络中、在计算资源间的均衡分布。
算力网络不是一个单项技术，它是一系列网络新技术的集成融合和创新应用。算力网络需要网络域、计算域协同创新，在网络域，需要对核心路由协议进行创新，从而实现快速、灵活、按需的业务请求调度；此外，需要对网络编排管理系统进行创新，实现对算网资源的联合、协同编排。在计算域，随着异构计算的发展，算力网络需要实现对异构算力的统一度量，建立统一度量机制，进一步形成异构算力的统一建模机制。算力网络需要基于网络域和计算域的协同创新，助力运营商网络从提供“连接”转变为提供“连接+计算”的综合信息与通信技术(Information and Communications Technology，ICT)服务。
在架构层面上看，面对边缘计算、异构计算、人工智能算力网络需要从架构、协议、度量等方面协同演进，构建面向算网一体化的新型基础网络，需要在基础设施即服务(IaaS)资源层编排的基础上，研究向平台即服务(PaaS)、软件即服务(SaaS)、网络即服务(NaaS)等一系列上层算法、业 务、函数和/或能力的编排演进，并协同管理、控制和数据面进一步探索实现编排系统与网络调度系统的协作，实现一切即服务(XaaS)能力按需灵活部署。
但是，目前并没有具有集中式编排和分布式调度协同的工作机制的算力网络。
发明内容
本公开的目的在于提供一种资源调度方法、装置及设备，以解决相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
为了解决上述技术问题，本公开实施例提供一种资源调度方法，应用于第一节点，包括：
接收第二节点发送的总体调度策略信息；
根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，还包括：
根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；
所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：
根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；
其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、 针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
可选的，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；
根据所述采集参数信息，获取资源实时状态信息；
其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
本公开实施例还提供了一种资源调度方法，应用于第二节点，包括：
获取算力网络的资源相关信息；
根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
将所述总体调度策略信息发送给对应的第一节点；
其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：
根据所述资源相关信息，得到拓扑信息；
根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
本公开实施例还提供了一种资源调度装置，应用于第一节点，所述资源调度装置包括：
第一接收模块，用于接收第二节点发送的总体调度策略信息；
第一获取模块，用于根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
第一调度模块，用于根据所述目标调度策略信息，针对所述目标业务需 求对应的目标业务数据进行资源调度；
其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，所述资源调度装置还包括：
第二获取模块，用于在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；
所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：
根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；
其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
可选的，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；
根据所述采集参数信息，获取资源实时状态信息；
其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
本公开实施例还提供了一种资源调度装置，应用于第二节点，所述资源调度装置包括：
第三获取模块，用于获取算力网络的资源相关信息；
第一确定模块，用于根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
第一发送模块，用于将所述总体调度策略信息发送给对应的第一节点；
其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：
根据所述资源相关信息，得到拓扑信息；
根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
本公开实施例还提供了一种资源调度设备，所述资源调度设备为第一节点，包括：处理器和收发机；
所述处理器，用于通过所述收发机接收第二节点发送的总体调度策略信息；
根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，所述处理器还用于：
在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；
所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：
根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息， 获取针对目标业务需求的目标调度策略信息；
其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
可选的，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；
根据所述采集参数信息，获取资源实时状态信息；
其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
本公开实施例还提供了一种资源调度设备，所述资源调度设备为第二节点，包括：处理器和收发机；
所述处理器，用于获取算力网络的资源相关信息；
根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
通过所述收发机将所述总体调度策略信息发送给对应的第一节点；
其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
可选的，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
可选的，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：
根据所述资源相关信息，得到拓扑信息；
根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
本公开实施例还提供了一种资源调度设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的程序；所述处理器执行所述程序时实现上述第一节点侧的资源调度方法；或者，
所述处理器执行所述程序时实现上述第二节点侧的资源调度方法。
本公开实施例还提供了一种可读存储介质，其上存储有程序，该程序被处理器执行时实现上述第一节点侧的资源调度方法中的步骤；或者，
该程序被处理器执行时实现上述第二节点侧的资源调度方法中的步骤。
本公开的上述技术方案的有益效果如下：
上述方案中，所述资源调度方法通过接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
附图说明
图1为本公开实施例的资源调度方法流程示意图一；
图2为本公开实施例的资源调度方法流程示意图二；
图3为本公开实施例的算力网络体系结构示意图；
图4为本公开实施例的集中式编排和分布式调度协同工作示意图；
图5为本公开实施例的集中式编排和分布式调度协同工作示例示意图一；
图6为本公开实施例的集中式编排和分布式调度协同工作示例示意图二；
图7为本公开实施例的资源调度装置结构示意图一；
图8为本公开实施例的资源调度装置结构示意图二；
图9为本公开实施例的资源调度设备结构示意图一；
图10为本公开实施例的资源调度设备结构示意图二。
具体实施方式
为使本公开要解决的技术问题、技术方案和优点更加清楚，下面将结合附图及具体实施例进行详细描述。
本公开针对相关的技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题，提供一种资源调度方法，应用于第一节点，如图1所示，包括：
步骤11：接收第二节点发送的总体调度策略信息；
步骤12：根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
步骤13：根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
其中，第一节点可包括算力网络的算力节点、网络节点、算网融合节点等节点中的至少一项；所述总体调度策略信息是所述第二节点根据算力网络的资源相关信息得到的；其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项；可理解为：所述资源相关信息包括：针对算力资源的资源相关信息、针对服务资源的资源相关信息以及针对网络资源的资源相关信息中的至少一项。资源相关信息也可理解为针对算力网络中的至少两个第一节点的资源相关信息，比如用于至少两个第一节点进行资源调配的资源相关信息，但并不以此为限。资源相关信息可包括：资源注册信息。第一节点“进行资源调度”中的资源是与总体调度策略信息相关的，比如包含于总体调度策略信息对应的资源中，但并不以此为限。其中的目标业务需求来自用户(终端)，用户可以单独发送业务需求(作为目标业务需求)，也可以在数据中捎带发送，在此不作限定。
步骤13具体可包括：根据所述目标调度策略信息，获取算力资源、服务资源以及网络资源信息，根据所述资源信息，针对所述目标业务需求对应的目标业务数据进行资源调度。此外，“针对所述目标业务需求对应的目标业务数据进行资源调度”，可包括：针对所述目标业务需求对应的目标业务数据进 行资源调度和业务调度。
本公开实施例提供的所述资源调度方法通过接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
本公开实施例中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
其中，策略执行节点信息可用于确定执行策略的可能控制器，或特定的算力路由节点。拓扑信息可用于确定全网的算力拓扑、网络拓扑和服务拓扑信息。后续如果接收的具体业务请求信息(比如上述目标业务需求)，请求特定服务，则可以根据策略执行节点信息确定策略执行节点，由策略执行节点根据上述拓扑信息，寻找已部署服务的一个或者多个算力节点；具体的：
如果业务请求类型是算力类型，可以根据算力拓扑信息，首先确定满足业务需求的一个或者多个算力节点；之后，根据资源实时状态采集的配置信息，支持周期性实时采集也可以按需进行资源实时状态信息的采集；然后根据获取的资源状态信息，从多个算力节点和多个网络路径中，确定最佳服务节点和合适的网络路径，从而保证业务体验的同时，提高网络和算力资源的利用率。
其中，关于周期性采集，可以通过扩展边界网关协议(Border Gateway Protocol，BGP)或内部网关协议(Interior Gateway Protocol，IGP)进行，按 需采集可以通过扩展带内或者带外操作管理维护(Operation Administration and Maintenance，OAM)信息进行，但并不以此为限。
此外，在总体调度策略信息中不包含的内容，可采用本地配置或者本地已存储的相关内容来替代，但并不以此为限。
进一步的，在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，还包括：根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
本公开实施例中，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；根据所述采集参数信息，获取资源实时状态信息；其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
进一步的，所述方法还包括：在进行资源调度失败的情况下，向第二节点发送失败通知信息；返回再次执行“接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度”。
本公开实施例还提供了一种资源调度方法，应用于第二节点，如图2所示，包括：
步骤21：获取算力网络的资源相关信息；
步骤22：根据所述资源相关信息，确定针对(算力网络中)各第一节点的总体调度策略信息；
步骤23：将所述总体调度策略信息发送给对应的第一节点；其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项(可理解 为：所述资源相关信息包括：针对算力资源的资源相关信息、针对服务资源的资源相关信息以及针对网络资源的资源相关信息中的至少一项)。
其中，资源相关信息可包括：资源注册信息；各第一节点的总体调度策略信息相同或不同，在此不作限定。资源相关信息也可理解为针对算力网络中的至少两个第一节点的资源相关信息，比如用于至少两个第一节点进行资源调配的资源相关信息，但并不以此为限。
本公开实施例提供的所述资源调度方法通过获取算力网络的资源相关信息；根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；将所述总体调度策略信息发送给对应的第一节点；其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
本公开实施例中，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：根据所述资源相关信息，得到拓扑信息；根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
进一步的，所述方法还包括：接收所述第一节点发送的失败通知信息；根据所述失败通知信息执行以下至少一项操作：(1)返回再次执行所述“获取算力网络的资源相关信息；根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；将所述总体调度策略信息发送给对应的第一节点”；(2) 直接返回重新执行“根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；将所述总体调度策略信息发送给对应的第一节点”(3)根据所述资源相关信息和/或总体调度策略信息，给所述失败通知信息对应的业务数据调度资源。
在此说明，本公开实施例中，算力网络也可称为计算网络融合系统、算力感知网络、算力内生网络或者算网一体化网络，在此不作限定。
下面对本公开实施例提供的资源调度方法进行举例说明。
针对上述技术问题，本公开实施例提供了一种资源调度方法，具体可实现为一种(对算力网络进行)集中式编排和分布式调度的协同方法，更具体的，可以是：面向计算网络融合的未来数据通信网络，或者称为算力网络或者算力感知网络或者算力内生网络，提供一种集中式编排和分布式调度协同的工作机制，并定义了新的功能模块和接口支持协同工作。其中，本方案可以是基于新型网络架构-算力感知网络(Computing-aware Networking，CAN)来实现，CAN可以基于无处不在的网络连接将动态分布的计算资源互联，通过网络、存储、算力等多维度资源的统一协同调度，使海量的应用能够按需、实时调用不同地方的计算资源，实现连接和算力在网络的全局优化，提供一致的用户体验。
由上，本方案基于CAN提供了一种支持算网协同编排调度机制，可通过集中式的编排和分布式调度协同工作机制，实现算力资源和网络资源的一体化调度，为用户提供算力网络服务的快速供给。
一、关于具体实现本方案的网络架构(支持集中式编排和分布式调度的协同的新型网络架构)；
关于上述新型网络架构-算力网络(Computing-aware Networking，CAN)，其基本理念是通过面向服务的网络重构，计算感知的网络路由、算网协同编排管理，构建运营商全新的ICT基础设施，成为智能社会的基础设施。基于此，本方案提供一种可实现上述方法的网络体系架构示例，可参见图3(一种支持算网协同编排的算力网络体系结构)，图中的VR表示虚拟现实(Virtual Reality)，AR表示增强现实(Augmented Reality)，V2X表示车联网(Vehicle to everything)，APP表示应用程序，API Gateway表示应用接口网关；CPU 表示中央处理单元(Central Processing Unit)，GPU表示图形处理器(Graphics Processing Unit)，NPU表示网络处理器(Network Processing Unit)，Strorage表示存储器，FPGA表示现场可编程门阵列(Field Programmable Gate Array)，ASICs表示专用集成电路(Application Specific Integrated Circuits)，OAM表示操作管理维护。
为了实现对泛在的计算和服务的感知、互联和协同调度，如图3所示，算力感知网络架构体系从逻辑功能上可划分为算力服务层、算网管理编排层、算力资源层、算力路由层和网络资源层五大功能模块，其中，算力路由层包含控制面和转发面。
基于网络无处不在的算力资源，具体如下：
1.算力服务层：承载泛在计算的各类服务及应用，支持分布式微服务架构，算力服务层支持应用分解构成原子化功能组件并组成算法库，由应用接口网关(API Gateway)应用统一调度，实现服务分解、服务调度等功能。
2.算力路由层：包含控制面和转发面；基于抽象后的算网资源发现，综合考虑网络状况和计算资源状况，将业务灵活按需调度到不同的计算资源节点中。
3.算网管理编排层：需要支持对全网的算力节点、网络节点和服务信息的注册、更新和注销等管理，即支持算力资源层、网络资源层和算力服务层向算力调度编排中心(对应于算网管理编排层)注册，生成算力、服务和网络拓扑信息(对应于上述拓扑信息)。
4.算力资源层：利用现有的计算基础设施提供算力资源，计算基础设施包括：从单核CPU(中央处理单元)到多核CPU，到CPU+GPU+FPGA(现场可编程门阵列)等多种计算能力的组合；为满足边缘计算领域多样性计算需求，面向不同应用，在物理计算资源基础上，提供算力模型、算力应用程序编程接口(Application Programming Interface，API)、算网资源标识等功能。
5.网络资源层：利用现有的网络基础设施为网络中的各个角落提供无处不在的网络连接，网络基础设施包括接入网、城域网和骨干网。
其中，算力资源层和网络资源层是算力网络的基础设施层，算网管理编排层和算力路由层是实现算力感知功能体系的两大核心功能模块，用户及应 用通过算力路由层接入网络中。算力感知网络体系架构基于所定义的五大功能模块，实现了对算网资源的感知、控制和调度。
二、关于支持集中式编排和分布式调度的协同的工作流；
基于上述算力网络架构，本方案提供通过集中式编排+分布式调度的协同，如图4所示(一种支持算网协同编排与分布式调度的协同工作图)，实现算力资源和网络资源的一体化调度，为用户和应用提供算力网络服务的快速供给。图中的IoT表示物联网(Internet of Things)，MEP表示边缘计算平台，DCGW表示数据中心网关(Data Center-Gateway)。图4可理解为图3对应的拓扑图。其中：
1.算网编排管理中心(对应于算网管理编排层)：完成对算力资源和网络资源的统一管理，包括对算力资源和网络资源的感知、度量和OAM(操作维护管理)管理等；构建算力、网络和服务全局拓扑信息(对应于上述拓扑信息)，实现对算网统一运营。基于接收到的算力、网络、服务信息，根据业务需求，由算网编排管理中心进一步协同调度策略，根据网络部署状态，可以选择在管理面(对应于算网编排管理中心)、控制面(对应于控制器，可集成在算网编排管理中心)和/或数据面(对应于算力路由层)执行调度策略。
2.进一步算网编排管理中心需要支持对全网的算力节点、网络节点和服务信息的注册、更新和注销等管理。比如，算力节点上线后，向算力管理平台通告其算力使能信息，包括算力节点标识或者算力资源的标识，包含设备类型、芯片类型、存储类型、计算类型等信息(对应于上述资源相关信息)。进一步可包括不限于算网管理层(对应于算网管理编排层)需要实现包括对服务、网络和算力资源的感知的方法配置和管理，主要包括不限于：
(1)向算力服务层、算力资源层和/或网络资源层下发多维资源感知的模版，用于算力服务层、算力资源层和/或网络资源层按照模版进行感知和度量。其中，该模板包括感知参数，比如服务标识(service ID)(和/或功能标识(fucntion ID))、算力可用资源、可用带宽等参数中的至少一项，和/或，服务可用连接数等信息。
(2)同时，算网编排管理层(即算网管理编排层)可以自适应订购采集的感知和测量的参数、反馈的频率等。算网管理层可以根据收到的多维资源 的感知和测量的结果，更新形成算力、服务和网络的初始化拓扑(对应于上述拓扑信息)，进而可以实现网络+算力可编程，以及业务自动适配。
其中，算网管理编排层、控制器(图3中未体现)和算力路由节点(对应于算力路由层)是实现协同功能工作的核心功能模块：
1.算网编排管理中心(对应于上述第二节点)：负责维护全局静态的算力和网络拓扑信息(对应于上述拓扑信息)。根据网络部署或者(历史)业务需求生成总体调度执行策略(对应于上述总体调度策略信息)，具体调度(对应于上述目标调度策略信息)可以在控制层面(对应于控制器)做或者数据层面(对应于算力路由节点)做。
2.控制器或者算力路由节点(对应于上述第一节点)：根据接收的调度指示消息(对应于上述总体调度策略信息)，负责维护算力和网络的状态信息(对应于上述资源实时状态信息)，结合业务需求(对应于上述目标业务需求)，将业务调度到合适的节点。
下面对本公开实施例提供的方案进行具体实现流程举例：
1.算网编排管理中心接收到算全网(对应于算力网络)的算力节点、网络节点、算网融合节点、服务信息注册等信息(对应于上述资源相关信息)。其中获取信息的方式可以是通过各种算力网络节点的注册，也可以通过网管平台或云管平台的各种信息获取。算网编排管理中心根据这些信息可以更新形成算力、服务和网络的初始化拓扑信息(对应于上述拓扑信息)。
2.算网编排管理中心根据业务需求(对应于上述历史业务需求)或者网络部署情况(对应于上述拓扑信息)，生成总体调度策略(对应于上述总体调度策略信息)，具体包括：调度执行者(控制器或者算力路由节点，对应于上述策略执行节点信息)，拓扑信息指示(对应于上述拓扑信息)，实时资源状态信息的配置(对应于上述配置信息)等信息，并向控制器或者算力路由节点发送调度策略(对应于上述总体调度策略信息)。
3.控制器或者算力路由节点，根据收到的总体调度策略信息中的拓扑信息和实时资源信息配置(对应于上述配置信息)等信息，配置算力实时信息(对应于上述资源实时状态信息)的采集方法(对应于上述采集方式信息)和间隔(对应于上述采集间隔信息)等信息(对应于上述采集参数信息)，向 算力资源节点和网络资源节点等节点，执行算力资源信息和网络的信息的收集。关于采集方法，可以是BGP或IGP扩展，也可以是OAM机制带内或者带外OAM等方法。采集的方法可以主动更新，或者由业务触发性更新，采集的详细方法可参见目前方案，在此不再赘述。
4.控制器或者算力路由节点收到业务需求(对应于上述目标业务需求)后，综合考虑网络和算力的状态信息(对应于上述资源实时状态信息)，生成具体业务的调度策略(对应于上述目标调度策略信息)，将业务数据包调度到合适的节点继续进行处理(对应于上述根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息)。此外，在资源调度失败的情况下，还可告知算网编排管理中心(对应于上述在进行资源调度失败的情况下，向第二节点发送失败通知信息)，以进行调配。
其中，关于算网编排管理+控制器协同的示例可参见图5，其中的PE表示运营商边缘路由器(Provider Edge，对应于路由节点)，AS表示自治域。
关于算网协同编排管理(无控制器协同)的示例可参见图6，其中，由算力路由节点做算力和网络的协同编排调度，即算网管理编排层会做静态的算力和网络拓扑信息，具体调度由算力路由节点数据面实现，通过算力路由节点做分布式调度。
由上，本公开实施例提供的方案涉及一种集中式编排和分布式调度的协同方法和系统，其中：
(1)第二节点执行集中式编排：完成算力、网络和服务信息的注册、更新和注销等管理，生成算力和网络拓扑(对应于上述拓扑信息)，需要完成多维资源感知的配置和管理，同时生成总体调度策略。进一步可包括：完成向算力服务层、算力资源层以及网络资源层下发多维资源感知的模版，用于算力服务层、算力资源层以及网络资源层按照该模版进行资源感知和度量。
具体，第二节点根据网络部署生成总体调度执行策略，通知算力路由节点执行调度策略。
(2)算力路由节点维护算力资源和网络资源的实时状态资源信息，接收到业务请求后，可综合考虑资源状态信息(以生成具体策略)，将业务调度到 合适的节点。
结合以上，本公开实施例提供的方案主要涉及：
1)算网编排管理中心接收到算全网的算力节点、网络节点、算网融合节点、服务信息的注册等信息，根据这些信息可以更新形成算力、服务和网络的拓扑信息。
2)算网编排管理中心根据业务需求和拓扑信息，生成总体调度策略，并向控制器或者算力路由节点发送调度策略。
3)控制器或者算力路由节点，维护算力资源和网络资源的实时状态资源信息，接收到业务请求后，可综合考虑资源状态信息，将业务调度到合适的节点。
综上，本方案提供一种支持算网协同管理编排的方法和系统，能够：
(1)通过算网协同编排管理，实现对算力和网络资源的统一管控和管理，实现算力和网络资源的全局优化配置；
(2)提供管理面、控制面和数据面协同的算网协同编排管理的多种方案，利于分阶段在现网的部署，有助于推动算力网络的实践进展。
本公开实施例还提供了一种资源调度装置，应用于第一节点，如图7所示，包括：
第一接收模块71，用于接收第二节点发送的总体调度策略信息；
第一获取模块72，用于根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
第一调度模块73，用于根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
本公开实施例提供的所述资源调度装置通过接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略 信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
进一步的，所述的资源调度装置，还包括：第二获取模块，用于在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
本公开实施例中，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；根据所述采集参数信息，获取资源实时状态信息；其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
其中，上述第一节点侧的资源调度方法的所述实现实施例均适用于该资源调度装置的实施例中，也能达到相同的技术效果。
本公开实施例还提供了一种资源调度装置，应用于第二节点，如图8所示，包括：
第三获取模块81，用于获取算力网络的资源相关信息；
第一确定模块82，用于根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
第一发送模块83，用于将所述总体调度策略信息发送给对应的第一节点；
其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
本公开实施例提供的所述资源调度装置通过获取算力网络的资源相关信息；根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；将所述总体调度策略信息发送给对应的第一节点；其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
本公开实施例中，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：根据所述资源相关信息，得到拓扑信息；根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
其中，上述第二节点侧的资源调度方法的所述实现实施例均适用于该资源调度装置的实施例中，也能达到相同的技术效果。
本公开实施例还提供了一种资源调度设备，所述资源调度设备为第一节点，如图9所示，包括：处理器91和收发机92；
所述处理器91，用于通过所述收发机92接收第二节点发送的总体调度策略信息；
根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信 息；
根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
本公开实施例提供的所述资源调度设备通过接收第二节点发送的总体调度策略信息；根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
进一步的，所述处理器还用于：在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
本公开实施例中，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；根据所述采集参数信息，获取资源实时状态信息；其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
其中，上述第一节点侧的资源调度方法的所述实现实施例均适用于该资源调度设备的实施例中，也能达到相同的技术效果。
本公开实施例还提供了一种资源调度设备，所述资源调度设备为第二节点，如图10所示，包括：处理器101和收发机102；
所述处理器101，用于获取算力网络的资源相关信息；
根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
通过所述收发机102将所述总体调度策略信息发送给对应的第一节点；
其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
本公开实施例提供的所述资源调度设备通过获取算力网络的资源相关信息；根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；将所述总体调度策略信息发送给对应的第一节点；其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项；能够支撑实现由第二节点进行集中式编排以得到总体调度策略信息，并由各第一节点基于总体调度策略信息得到目标调度策略信息(即具体的调度策略信息)以进行资源调度，实现分布式调度；从而实现算力网络的集中式编排(如总体资源由第二节点集中编排)和分布式调度(如具体资源由各第一节点调度)协同的工作机制；具体可通过集中式编排和分布式调度协同的工作机制，实现算力资源、服务资源、和网络资源的一体化调度，为用户提供算力网络服务的快速供给；很好的解决了相关技术中算力网络无法实现集中式编排和分布式调度协同的工作机制的问题。
其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
本公开实施例中，所述根据所述资源相关信息，确定针对各第一节点的 总体调度策略信息，包括：根据所述资源相关信息，得到拓扑信息；根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
其中，上述第二节点侧的资源调度方法的所述实现实施例均适用于该资源调度设备的实施例中，也能达到相同的技术效果。
本公开实施例还提供了一种资源调度设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的程序；所述处理器执行所述程序时实现上述第一节点侧的资源调度方法；或者，所述处理器执行所述程序时实现上述第二节点侧的资源调度方法。
其中，上述第一节点侧或第二节点侧的资源调度方法的所述实现实施例均适用于该资源调度设备的实施例中，也能达到相同的技术效果。
本公开实施例还提供了一种可读存储介质，其上存储有程序，该程序被处理器执行时实现上述第一节点侧的资源调度方法中的步骤；或者，该程序被处理器执行时实现上述第二节点侧的资源调度方法中的步骤。
其中，上述第一节点侧或第二节点侧的资源调度方法的所述实现实施例均适用于该可读存储介质的实施例中，也能达到相同的技术效果。
需要说明的是，此说明书中所描述的许多功能部件都被称为模块，以便更加特别地强调其实现方式的独立性。
本公开实施例中，模块可以用软件实现，以便由各种类型的处理器执行。举例来说，一个标识的可执行代码模块可以包括计算机指令的一个或多个物理或者逻辑块，举例来说，其可以被构建为对象、过程或函数。尽管如此，所标识模块的可执行代码无需物理地位于一起，而是可以包括存储在不同位里上的不同的指令，当这些指令逻辑上结合在一起时，其构成模块并且实现该模块的规定目的。
实际上，可执行代码模块可以是单条指令或者是许多条指令，并且甚至可以分布在多个不同的代码段上，分布在不同程序当中，以及跨越多个存储器设备分布。同样地，操作数据可以在模块内被识别，并且可以依照任何适当的形式实现并且被组织在任何适当类型的数据结构内。所述操作数据可以作为单个数据集被收集，或者可以分布在不同位置上(包括在不同存储设备 上)，并且至少部分地可以仅作为电子信号存在于系统或网络上。
在模块可以利用软件实现时，考虑到现有硬件工艺的水平，所以可以以软件实现的模块，在不考虑成本的情况下，本领域技术人员都可以搭建对应的硬件电路来实现对应的功能，所述硬件电路包括常规的超大规模集成(Very Large Scale Integration Circuit，VLSI)电路或者门阵列以及诸如逻辑芯片、晶体管之类的现有半导体或者是其它分立的元件。模块还可以用可编程硬件设备，诸如现场可编程门阵列、可编程阵列逻辑、可编程逻辑设备等实现。
以上所述的是本公开的优选实施方式，应当指出对于本技术领域的普通人员来说，在不脱离本公开所述原理前提下，还可以作出若干改进和润饰，这些改进和润饰也应视为本公开的保护范围。

Claims (18)

  1. 一种资源调度方法，应用于第一节点，所述方法包括：
    接收第二节点发送的总体调度策略信息；
    根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
    根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
    其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
  2. 根据权利要求1所述的资源调度方法，其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
    其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
  3. 根据权利要求2所述的资源调度方法，其中，在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，还包括：
    根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；
    所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：
    根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；
    其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
  4. 根据权利要求3所述的资源调度方法，其中，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
    根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；
    根据所述采集参数信息，获取资源实时状态信息；
    其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
  5. 一种资源调度方法，应用于第二节点，所述方法包括：
    获取算力网络的资源相关信息；
    根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
    将所述总体调度策略信息发送给对应的第一节点；
    其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
  6. 根据权利要求5所述的资源调度方法，其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
    其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
  7. 根据权利要求5或6所述的资源调度方法，其中，所述根据所述资源相关信息，确定针对各第一节点的总体调度策略信息，包括：
    根据所述资源相关信息，得到拓扑信息；
    根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
  8. 一种资源调度装置，应用于第一节点，所述资源调度装置包括：
    第一接收模块，用于接收第二节点发送的总体调度策略信息；
    第一获取模块，用于根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
    第一调度模块，用于根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
    其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
  9. 根据权利要求8所述的资源调度装置，其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
    其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
  10. 根据权利要求9所述的资源调度装置，还包括：
    第二获取模块，用于在根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息之前，根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息；
    所述根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息，包括：
    根据所述总体调度策略信息中的调度策略信息和所述资源实时状态信息，获取针对目标业务需求的目标调度策略信息；
    其中，所述资源实时状态信息包括：针对算力资源的资源实时状态信息、针对服务资源的资源实时状态信息，以及针对网络资源的资源实时状态信息中的至少一项。
  11. 根据权利要求10所述的资源调度装置，其中，所述根据所述总体调度策略信息中的拓扑信息和配置信息，获取资源实时状态信息，包括：
    根据所述总体调度策略信息中的拓扑信息和配置信息，配置针对资源实时状态信息的采集参数信息；
    根据所述采集参数信息，获取资源实时状态信息；
    其中，所述采集参数信息包括：采集方式信息和采集间隔信息中的至少一项。
  12. 一种资源调度装置，应用于第二节点，所述资源调度装置包括：
    第三获取模块，用于获取算力网络的资源相关信息；
    第一确定模块，用于根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
    第一发送模块，用于将所述总体调度策略信息发送给对应的第一节点；
    其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
  13. 根据权利要求12所述的资源调度装置，其中，所述总体调度策略信息包括：策略执行节点信息、拓扑信息和针对资源实时状态采集的配置信息中的至少一项；
    其中，所述拓扑信息为针对算力、服务和网络的拓扑信息。
  14. 根据权利要求12或13所述的资源调度装置，其中，所述根据所述 资源相关信息，确定针对各第一节点的总体调度策略信息，包括：
    根据所述资源相关信息，得到拓扑信息；
    根据历史业务需求和所述拓扑信息中的至少一项，确定针对各第一节点的总体调度策略信息。
  15. 一种资源调度设备，所述资源调度设备为第一节点，所述资源调度设备包括：处理器和收发机；
    所述处理器，用于通过所述收发机接收第二节点发送的总体调度策略信息；
    根据所述总体调度策略信息，获取针对目标业务需求的目标调度策略信息；
    根据所述目标调度策略信息，针对所述目标业务需求对应的目标业务数据进行资源调度；
    其中，所述资源包括：算力资源、服务资源以及网络资源中的至少一项。
  16. 一种资源调度设备，所述资源调度设备为第二节点，所述资源调度设备包括：处理器和收发机；
    所述处理器，用于获取算力网络的资源相关信息；
    根据所述资源相关信息，确定针对各第一节点的总体调度策略信息；
    通过所述收发机将所述总体调度策略信息发送给对应的第一节点；
    其中，所述资源相关信息针对算力资源、服务资源以及网络资源中的至少一项。
  17. 一种资源调度设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的程序；其中，所述处理器执行所述程序时实现如权利要求1至4中任一项所述的资源调度方法；或者，
    所述处理器执行所述程序时实现如权利要求5至7中任一项所述的资源调度方法。
  18. 一种可读存储介质，其上存储有程序，其中，该程序被处理器执行时实现如权利要求1至4中任一项所述的资源调度方法中的步骤；或者，
    该程序被处理器执行时实现如权利要求5至7中任一项所述的资源调度方法中的步骤。
PCT/CN2023/085043 2022-04-01 2023-03-30 一种资源调度方法、装置及设备 Ceased WO2023186002A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210347746.4A CN116938837A (zh) 2022-04-01 2022-04-01 一种资源调度方法、装置及设备
CN202210347746.4 2022-04-01

Publications (1)

Publication Number Publication Date
WO2023186002A1 true WO2023186002A1 (zh) 2023-10-05

Family

ID=88199415

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/085043 Ceased WO2023186002A1 (zh) 2022-04-01 2023-03-30 一种资源调度方法、装置及设备

Country Status (2)

Country Link
CN (1) CN116938837A (zh)
WO (1) WO2023186002A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117216758A (zh) * 2023-11-08 2023-12-12 新华三网络信息安全软件有限公司 应用安全检测系统及方法
US20240146735A1 (en) * 2022-10-31 2024-05-02 Beijing Youzhuju Network Technology Co., Ltd. Method and apparatus for processing multi-cloud service, electronic device and storage medium
CN118484410A (zh) * 2024-04-10 2024-08-13 湖南大唐先一科技有限公司 一种用于发电企业基于fpga的物联网边缘算力调度物理卡
CN118827789A (zh) * 2024-01-11 2024-10-22 中国移动通信有限公司研究院 一种算网调度方法、算网编排系统及存储介质
CN119443684A (zh) * 2024-11-04 2025-02-14 杭州万澜科技有限公司 基于动态调度的项目资源分配方法和分配系统
WO2025232417A1 (zh) * 2024-05-08 2025-11-13 大唐移动通信设备有限公司 一种通信网络架构

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118250243B (zh) * 2024-02-26 2025-11-18 浪潮通信信息系统有限公司 算网一体化资源调度系统、方法、电子设备及存储介质
CN120583007A (zh) * 2024-03-01 2025-09-02 中兴通讯股份有限公司 一种算网操作维护管理系统、方法和存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108366020A (zh) * 2018-02-02 2018-08-03 网宿科技股份有限公司 一种发送数据资源的获取请求的方法和系统
CN111327076A (zh) * 2020-03-16 2020-06-23 明阳智慧能源集团股份公司 一种基于分布式记账的储能型风机调度响应方法
WO2020142897A1 (zh) * 2019-01-08 2020-07-16 华为技术有限公司 调度方法、装置、存储介质及通信系统
CN112925634A (zh) * 2019-12-06 2021-06-08 中国电信股份有限公司 异构资源调度方法和系统
CN113051053A (zh) * 2021-03-24 2021-06-29 依瞳科技(深圳)有限公司 异构资源调度方法、装置、设备和计算机可读存储介质
CN113810438A (zh) * 2020-06-12 2021-12-17 中国移动通信有限公司研究院 服务算力资源的调度、请求方法、节点设备及终端

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108366020A (zh) * 2018-02-02 2018-08-03 网宿科技股份有限公司 一种发送数据资源的获取请求的方法和系统
WO2020142897A1 (zh) * 2019-01-08 2020-07-16 华为技术有限公司 调度方法、装置、存储介质及通信系统
CN112925634A (zh) * 2019-12-06 2021-06-08 中国电信股份有限公司 异构资源调度方法和系统
CN111327076A (zh) * 2020-03-16 2020-06-23 明阳智慧能源集团股份公司 一种基于分布式记账的储能型风机调度响应方法
CN113810438A (zh) * 2020-06-12 2021-12-17 中国移动通信有限公司研究院 服务算力资源的调度、请求方法、节点设备及终端
CN113051053A (zh) * 2021-03-24 2021-06-29 依瞳科技(深圳)有限公司 异构资源调度方法、装置、设备和计算机可读存储介质

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20240146735A1 (en) * 2022-10-31 2024-05-02 Beijing Youzhuju Network Technology Co., Ltd. Method and apparatus for processing multi-cloud service, electronic device and storage medium
US12549551B2 (en) * 2022-10-31 2026-02-10 Beijing Youzhuju Network Technology Co., Ltd. Method and apparatus for processing multi-cloud service, electronic device and storage medium
CN117216758A (zh) * 2023-11-08 2023-12-12 新华三网络信息安全软件有限公司 应用安全检测系统及方法
CN117216758B (zh) * 2023-11-08 2024-02-23 新华三网络信息安全软件有限公司 应用安全检测系统及方法
CN118827789A (zh) * 2024-01-11 2024-10-22 中国移动通信有限公司研究院 一种算网调度方法、算网编排系统及存储介质
CN118484410A (zh) * 2024-04-10 2024-08-13 湖南大唐先一科技有限公司 一种用于发电企业基于fpga的物联网边缘算力调度物理卡
WO2025232417A1 (zh) * 2024-05-08 2025-11-13 大唐移动通信设备有限公司 一种通信网络架构
CN119443684A (zh) * 2024-11-04 2025-02-14 杭州万澜科技有限公司 基于动态调度的项目资源分配方法和分配系统

Also Published As

Publication number Publication date
CN116938837A (zh) 2023-10-24

Similar Documents

Publication Publication Date Title
WO2023186002A1 (zh) 一种资源调度方法、装置及设备
WO2021190482A1 (zh) 算力处理的网络系统及算力处理方法
CN115622904A (zh) 管理和调度方法、装置、节点及存储介质
Giang et al. Developing iot applications in the fog: A distributed dataflow approach
CN104468688B (zh) 用于网络虚拟化的方法和设备
Nguyen et al. Placement and chaining for run-time IoT service deployment in edge-cloud
EP3143511B1 (en) Method and apparatus for affinity-based network configuration
US9461877B1 (en) Aggregating network resource allocation information and network resource configuration information
WO2022184094A1 (zh) 算力处理的网络系统、业务处理方法及算力网元节点
CN114095577A (zh) 资源请求方法、装置、算力网元节点及算力应用设备
US20120269053A1 (en) Co-Scheduling of Network Resource Provisioning and Host-to-Host Bandwidth Reservation on High-Performance Network and Storage Systems
CN113542371B (zh) 基于边缘网关的资源调度方法及系统
Ermolenko et al. Internet of Things services orchestration framework based on Kubernetes and edge computing
CN103747107A (zh) 一种兼容式云操作平台及其实现方法
CN111654541A (zh) 面向边缘计算业务的服务功能链编排方法、系统及编排器
Nguyen et al. Studying and developing a resource allocation algorithm in Fog computing
AbdelBaky et al. Computing in the continuum: Combining pervasive devices and services to support data-driven applications
CN107924332B (zh) Ict服务供应的方法和系统
Ungureanu et al. Collaborative cloud-edge: A declarative api orchestration model for the nextgen 5g core
Al-Kasassbeh et al. Analysis of mobile agents in network fault management
CN110601897A (zh) 一种网络资源配置方法及装置
CN112491574A (zh) 一种数据处理方法及装置
Guo et al. Service orchestration for integrating edge computing and 5g network: State of the art and challenges
CN103067476B (zh) 一种基于虚拟机的动态网络重构方法
JP6564839B2 (ja) 組み込み型オペレーティングシステムに基づくｍｐｉ実現システムおよび方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23778367

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 27/01/2025)

122 Ep: pct application non-entry in european phase

Ref document number: 23778367

Country of ref document: EP

Kind code of ref document: A1