Deployment method and deployment apparatus for virtual network function

By introducing load group objects and group scheduling methods, multiple virtual deployment units are used as a whole for resource scheduling, which solves the problem of unreasonable resource allocation in the existing technology and improves the deployment success rate and efficiency of virtual network functions.

WO2024239710A1PCT designated stage expired Publication Date: 2024-11-28HUAWEI TECH CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2024/075147
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-19
Filing Date
2024-02-01
Publication Date
2024-11-28

AI Technical Summary

Technical Problem

In network functional virtualization (NFV), when the prior art uses a single virtual deployment unit (VDU) to perform resource scheduling as granularity, it leads to unreasonable resource allocation, resulting in VNF deployment failure or resource fragmentation unavailability, reducing network deployment efficiency.

Method used

The load group object is introduced as a custom resource, and multiple VDUs are dispatched as a whole through the group scheduling method until the entire group of VDUs obtain resources or fails, improving the rationality of resource allocation and VNF deployment success rate.

Benefits of technology

It improves the deployment success rate and efficiency of virtual network functions, reduces unnecessary operations during the deployment process, optimizes resource utilization, and improves the scalability of the network system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2024075147_28112024_PF_FP_ABST
    Figure CN2024075147_28112024_PF_FP_ABST
Patent Text Reader

Abstract

The present application provides a deployment method and deployment apparatus for a virtual network function. When allocating resources to VDUs required by a VNF to be deployed, the deployment method provided by the present application uses all of the required VDUs as an overall group, and uses the group as a granularity in one computing cycle so as to compute and schedule the resources required by all of the VDUs, until the whole group of VDUs obtains the corresponding scheduled resources or the whole group of VDUs fails in scheduling. When the group is used as the granularity so as to compute the deployment positions of the VDUs of the VNF to be deployed, recomputing and adjusting the allocation mode which is not conducive to overall deployment increases the success rate of VNF deployment.
Need to check novelty before this filing date? Find Prior Art

Description

Virtual network function deployment method and deployment device

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on May 19, 2023, with application number 202310576405.9 and application name “Deployment Method and Deployment Device for Virtual Network Functions”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the technical field of network function virtualization, and in particular to a method and device for deploying virtual network functions. Background Art

[0003] Network function virtualization (NFV) technology decouples upper-layer software from underlying general-purpose hardware by virtualizing computing, storage, and network resources. This allows software-defined network functions to be flexibly deployed on general-purpose servers, switches, storage devices, and other devices, promoting efficient and cost-effective network operations. In the NFV framework, software-implemented network functions are referred to as virtualized network functions (VNFs). To ensure the proper operation of VNFs and the interoperability of different VNFs, the cloud operating system (OS) must schedule resources and deploy VNFs on general-purpose hardware platforms.

[0004] Summary of the Invention

[0005] The present application provides a method and device for deploying virtual network functions, aiming to improve network deployment efficiency.

[0006] In a first aspect, the present application provides a method for deploying a virtual network function, which is applied to an infrastructure service manager. The method includes: receiving a first request message sent from a virtual network function manager VNFM, where the first request message is used to request deployment of a virtual deployment unit VDU for a first virtual network function VNF; creating M VDUs for the first VNF, where the M VDUs are associated as VDUs in the same group, where M is a positive integer; and deploying the VDUs in the same group using a group scheduling method, where the group scheduling method allocates resources to each VDU in the same group within a calculation cycle until each VDU in the same group obtains corresponding resources or the allocation of resources to the entire group of VDUs in the same group fails, and the calculation cycle is used to calculate the corresponding resources for each VDU in the same group.

[0007] In this method, the group scheduling method uses multiple VDUs as the granularity and performs unified scheduling calculations on the resources required by the entire VDU, thereby improving the rationality of resource allocation and increasing the possibility of successful VDU deployment.

[0008] In one possible implementation, the first request information carries first information, and the first information indicates that the load group object identifier associated with the VDU required by the first VNF ​​is the first load group object identifier, and the load group object identifier is used to indicate the load group object, and the load group object includes the following attribute information: the maximum number of VDUs that can be associated with the load group object identifier, the scheduling policy when deploying the VDU associated with the load group object identifier using the group scheduling method, the scheduling priority when deploying the VDU associated with the load group object identifier using the group scheduling method, and the scheduling timeout time when deploying the VDU associated with the load group object identifier using the group scheduling method.

[0009] The infrastructure service manager creates M VDUs for the first VNF, including: the infrastructure service manager creates M VDUs, and each of the M VDUs is associated with a first load group object identifier.

[0010] Deploying M VDUs using a group scheduling method includes: deploying the VDUs associated with the first load group object identifier according to attribute information of the load group object indicated by the first load group object identifier using the group scheduling method.

[0011] The load group object is introduced as a custom resource. The load group object acts as a group. Multiple VDUs required for the VNF to be deployed are associated with the load group object through tags and become members of the load group object. Furthermore, the multiple attributes of the load group object facilitate subsequent deployment using group scheduling methods.

[0012] In some implementations, deploying a VDU associated with a first load group object identifier using a group scheduling method according to attribute information of a load group object indicated by the first load group object identifier includes:

[0013] Obtain an undeployed VDU; when the load group object identifier associated with the undeployed VDU is a first load group object identifier, add the undeployed VDU to the scheduling queue associated with the obtained first load group object identifier; use the group scheduling method to deploy the VDU in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier.

[0014] According to the load group object identifier associated with the undeployed VDU, ordinary VDUs and VDUs that need to be deployed using the group scheduling method can be added to different scheduling queues to facilitate subsequent deployment of different VDUs.

[0015] In some implementations, using a group scheduling method, deploying the VDUs in the scheduling queue according to attribute information of the load group object indicated by the first load group object identifier includes:

[0016] When the number of VDUs in the scheduling queue is equal to the maximum number in the load group object indicated by the first load group object identifier, the group scheduling method is used to deploy the VDUs in the scheduling queue according to the scheduling policy and scheduling priority in the load group object indicated by the first load group object identifier until each VDU in the scheduling queue is successfully deployed or the deployment duration exceeds the scheduling timeout duration in the load group object indicated by the first load group object identifier.

[0017] The number of members of the load group object is used as the dequeue rule of the scheduling queue to ensure that the VDUs scheduled out of the queue are associated with the same load group object, thereby realizing group scheduling method for multiple VDUs at the group granularity.

[0018] In one possible implementation, the method further includes:

[0019] Second information is received from the VNFM, where the second information is used to instruct creation of a load group object identified as the first load group object identifier.

[0020] In some implementations, the method further includes:

[0021] Sending a load group object list to the VNFM; wherein receiving the second information from the VNFM includes: when the load group object list does not include a load group object identified as the first load group object identifier, receiving the second information from the VNFM.

[0022] In some implementations, the method further includes:

[0023] Receive third information from the VNFM, where the third information is used to instruct to update attribute information in the load group object identified by the first load group object identifier.

[0024] In some implementations, the method further includes:

[0025] Sending a load group object list to the VNFM; wherein receiving the third information from the VNFM includes: when the load group object list includes a load group object identified as the first load group object identifier, receiving the third information from the VNFM.

[0026] In some implementations, before sending the load group object list to the VNFM, the method further includes:

[0027] A second request message is received from the VNFM, where the second request message is used to request a load group object list.

[0028] The deployment method of the virtual network function proposed in this application can not only independently perform add, delete, modify and query operations on load group objects, but also realize group scheduling verification during the deployment of VDU.

[0029] In one possible implementation, a group scheduling method is used to deploy the VDUs in the scheduling queue according to attribute information of the load group object indicated by the first load group object identifier, including:

[0030] The first scheduler is used to associate a group scheduling label with the first node, the first node includes an alternative deployment node for the VDU in the scheduling queue, and the first scheduler is used to allocate resources to the VDU; the second scheduler uses a group scheduling method to deploy the VDU in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier, and the second scheduler is a copy of the first scheduler.

[0031] In some implementations, using a group scheduling method, deploying the VDUs in the scheduling queue according to attribute information of the load group object indicated by the first load group object identifier includes:

[0032] A group scheduling label is associated with the first node through the first scheduler, the first node includes an alternative deployment node for the VDU in the scheduling queue, and the first scheduler is used to allocate resources to the VDU; a group scheduling method is used by at least one scheduling solver to deploy the VDU in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier, and the scheduling solver is an independent third-party scheduler.

[0033] The computational load can be alleviated by using a second scheduler or a group scheduling method in the scheduling solver.

[0034] In a second aspect, the present application provides a method for deploying a virtual network function, which is applied to a virtual network function manager (VNFM). The method includes:

[0035] Receive third request information, where the third request information is used to trigger deployment of a virtual deployment unit VDU for the first virtual network function VNF; and send first request information to the infrastructure service manager, where the first request information is used to request deployment of the VDU for the first VNF.

[0036] In one possible implementation, the third request information and the first request information carry first information, the first information indicates that the load group object identifier associated with the VDU required by the first VNF ​​is the first load group object identifier, the load group object identifier is used to indicate the load group object, and the load group object includes the following attribute information: the maximum number of VDUs that can be associated with the load group object identifier, the scheduling policy when deploying the VDU associated with the load group object identifier using the group scheduling method, the scheduling priority when deploying the VDU associated with the load group object identifier using the group scheduling method, and the scheduling timeout time when deploying the VDU associated with the load group object identifier using the group scheduling method.

[0037] In one possible implementation, the method further includes:

[0038] Second information is sent to the infrastructure service manager, where the second information is used to instruct creation of a load group object identified as the first load group object identifier.

[0039] In some implementations, the method further includes:

[0040] Receiving a load group object list from an infrastructure service manager; wherein, sending second information to the infrastructure service manager includes: when the load group object list does not contain a load group object identified as a first load group object identifier, sending the second information to the infrastructure service manager.

[0041] In some implementations, the method further includes:

[0042] Send third information to the infrastructure service manager, where the third information is used to instruct to update the attribute information in the load group object identified as the first load group object identifier.

[0043] In some implementations, the method further includes:

[0044] Receiving a load group object list from an infrastructure service manager; wherein, sending third information to the infrastructure service manager includes: when the load group object list includes a load group object identified as a first load group object identifier, sending the third information to the infrastructure service manager.

[0045] In some implementations, before receiving the load group object list from the infrastructure service manager, the method further includes:

[0046] A second request message is sent to the infrastructure service manager, where the second request message is used to request a load group object list.

[0047] In some implementations, the third request information is specifically used to request instantiation of the first VNF, or the third request information is specifically used to request expansion of the VDU for the first VNF.

[0048] Whether the third request information is used to request the instantiation of the first VNF ​​or to request capacity expansion for the first VNF, scheduling resources need to be allocated to the VDU required by the first VNF.

[0049] In some implementations, the request information received by the VNFM is input by an administrator on the VNFM, or the request information received by the VNFM is received by the VNFM from the NFVO.

[0050] In the NFV system architecture, both NFVO and VNFM can manage VNFs.

[0051] In a third aspect, the present application provides a virtual network function deployment device, which includes various functional modules for implementing any of the deployment methods mentioned in the above implementations. In some implementations, each module can be implemented by software and / or hardware.

[0052] In a fourth aspect, the present application provides a virtual network function deployment device, comprising: a processor, and a memory communicatively connected to the processor; the memory storing computer-executable instructions. The processor executes the computer-executable instructions stored in the memory to implement the virtual network function deployment method provided in the first aspect.

[0053] In a fifth aspect, the present application provides a virtual network function deployment device, comprising: a processor, and a memory communicatively connected to the processor; the memory storing computer-executable instructions. The processor executes the computer-executable instructions stored in the memory to implement the virtual network function deployment method provided in the second aspect.

[0054] In a sixth aspect, the present application provides a virtual network function deployment system, which includes the device in the fourth aspect and the device in the fifth aspect.

[0055] In a seventh aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method in the first aspect, the second aspect, or any possible implementation thereof.

[0056] In an eighth aspect, the present application provides a computer-readable medium storing a program code for execution by a device, wherein the program code includes a method for executing the method in the first aspect, the second aspect, or any possible implementation thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0058] Figure 1 is a schematic diagram of the standardized NFV system architecture proposed by the European Telecommunications Standards Institute;

[0059] FIG2 is a conceptual diagram of a load group provided in one embodiment of the present application;

[0060] FIG3 is a flow chart of a method for deploying a virtual network function according to an embodiment of the present application;

[0061] FIG4 is a flow chart of a method for deploying a virtual network function according to an embodiment of the present application;

[0062] FIG5 is a schematic diagram of a process for implementing addition, deletion, modification, and querying of a load group object list according to an embodiment of the present application;

[0063] FIG6 is a schematic diagram of a VDU associated load group object provided by an embodiment of the present application;

[0064] FIG7 is a schematic diagram of a component architecture of K8S extended group scheduling provided in an embodiment of the present application;

[0065] FIG8 is a schematic diagram of a process for implementing group scheduling in K8S according to an embodiment of the present application;

[0066] FIG9 is a schematic diagram of a framework of a K8S extended coordinator supporting group scheduling calculations provided by an embodiment of the present application;

[0067] FIG10 is a flow chart of group scheduling implemented by K8S through a master coordinator according to an embodiment of the present application;

[0068] FIG11 is a schematic diagram of a framework of a K8S extended scheduling solver supporting group scheduling calculations provided by an embodiment of the present application;

[0069] FIG12 is a flow diagram of group scheduling implemented by K8S through a master scheduler and a scheduling solver according to an embodiment of the present application;

[0070] FIG13 is a schematic diagram of the structure of a device for deploying a virtual network function according to an embodiment of the present application;

[0071] FIG14 is a schematic diagram of the structure of a device for deploying a virtual network function according to another embodiment of the present application;

[0072] FIG15 is a schematic structural diagram of a virtual network function deployment device provided in yet another embodiment of the present application.

[0073] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0074] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0075] In this application, "at least one" means one or more, and "more" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In the text description of this application, the character " / " generally indicates that the previous and next associated objects are in an "or" relationship. Traditional telecommunications network systems are composed of various specialized hardware devices, and different dedicated hardware devices run corresponding functions. With the growth of network scale, telecommunications network systems are becoming more and more complex, and scalability is also declining. NFV technology uses virtualization technology to separate the software functions of the equipment from the underlying hardware, realize the software and hardware decoupling of telecommunications network function nodes, and is the core technology of telecommunications-grade service cloudification. Based on NFV technology, VNFs with different functions can be deployed on a common hardware platform, thereby reducing the cost of the telecommunications network system and providing better scalability.

[0076] Different manufacturers may develop VNFs with different functions. Service providers can flexibly select VNFs based on their own needs and combine different VNFs. To ensure that all selected VNFs can work together normally and run on a common hardware platform, a standardized architecture is required to constrain VNF development. Figure 1 is a schematic diagram of the standardized architecture of the NFV system proposed by the European Telecommunications Standards Institute (ETSI). As shown in Figure 1, the NFV system includes the NFV management and orchestration (NFV-MAMO) module, the NFV foundation layer (network function virtualization infrastructure, NFVI), VNFs, the element management (EM) module, and the operation support system / business support system (OSS / BSS).

[0077] Among them, the NFV-MAMO module includes four functional modules: NFV orchestrator (NFVO), VNF manager (VNFM), container infrastructure service manager (CISM), and virtualized infrastructure manager (VIM).

[0078] NFVO has a global view of virtual resources and is primarily responsible for handling the lifecycle management of virtualized services, the allocation and scheduling of virtual resources in virtual infrastructure and NFVI. NFVO can also execute resource-related requests from VNFM, send configuration information to VNFM, and collect VNF ​​status information.

[0079] A VNFM is primarily responsible for the lifecycle management of one or more VNFs. By communicating with the VNFs, the VNFM can perform operations such as instantiation, update, query, elastic scaling, and termination. An NFV system can have multiple VNFMs, each responsible for the lifecycle management of different types of VNFs.

[0080] VIM performs resource management functions, such as managing infrastructure resource allocation, resource discovery, and troubleshooting. VIM and VNFM can communicate with each other to allocate resources and exchange configuration and status information for virtualized hardware resources. CISM functions similarly to VIM, primarily managing resources for container objects. Both VIM and CISM are infrastructure service managers.

[0081] NFVI is mainly used to provide hardware resources for upper-layer VNFs. The underlying physical hardware resources in NFVI are abstracted and converted into virtualized hardware resources such as virtual machines (VMs) and containers through the virtualization layer. The upper-layer VNFs run in a virtualized resource environment, achieving complete decoupling of software and hardware.

[0082] The EM module performs traditional fault, user and configuration, performance, and security management functions for VNFs.

[0083] OSS / BSS supports various end-to-end telecommunications services. OSS supports management functions such as network configuration, service provision, and fault management, while BSS supports product management, order management, revenue management, and customer management.

[0084] Under the NFV architecture, cloud OS can realize comprehensive virtualization of computing resources, storage resources and network resources, and can uniformly manage, monitor and optimize physical hardware virtualization resources.

[0085] As the core technology for cloud computing, NFV technology introduces three levels of services to achieve cloudification of telecom-grade services: infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS).

[0086] Among them, IaaS focuses on providing basic resources for users, which are usually provided in the form of VMs in actual applications. Therefore, when implementing network function virtualization under the IaaS architecture, corresponding to the virtual machine scenario, one or more VMs are called to deploy VNFs with corresponding functions.

[0087] PaaS is a business model that provides a server platform as a service to the outside world. In the PaaS architecture, interface services are provided to the outside world, and other business developers can use these service interface capabilities to develop businesses or applications.

[0088] Container as a Service (CaaS) provides services to the outside world in the form of containers and can be considered a specific type of PaaS service. Containers, as an operating system-level virtualization technology, isolate different processes through operating system isolation technologies such as CGroup and Namespace under Linux. There is no operating system inside the container, only processes, so containers are more lightweight than VMs, and containers can be deployed not only on bare metal but also in virtual machines. Containerized systems deployed on bare metal are called bare metal containers, and containerized systems deployed on virtual machines are called virtual machine containers. When implementing network function virtualization under the CaaS architecture, corresponding to the container scenario, one or more containers are called to deploy the VNF of the corresponding function.

[0089] Under the CaaS architecture, the most widely used container management and orchestration system currently is the Kubernetes (K8S) container cluster management system based on the open source platform. K8S can provide a complete set of functions such as resource scheduling, deployment and operation, service discovery, and scaling for containerized applications. In essence, it can be regarded as a mini-CaaS manager based on container technology.

[0090] Based on the above scenarios, deploying a VNF requires one or more smaller-granularity virtualized deployment units (VDUs). In the virtual machine (VM) scenario, a VDU object refers to a VM. In the container scenario, Kubernetes uses the smallest deployment unit, a "Pod," to manage containers. Each "Pod" contains one or more closely related containers, so a VDU object refers to a single "Pod." When CISM / VIM schedules resources to deploy VNFs, it actually allocates resources for the VDUs contained within the VNF to be deployed. Deployment of the VNF is considered successful only when all VDUs for the VNF to be deployed are bound to the corresponding resource nodes.

[0091] In the existing technology, the CISM / VIM scheduler calculates the deployment location based on the granularity of a single VDU. When hardware resources are limited but VNF ​​deployment requirements are met, some VDUs are deployed to the corresponding hardware locations first, and the remaining VDUs may not be deployed due to insufficient remaining resources, resulting in VNF deployment failure. For example:

[0092] In the container scenario, the virtual deployment unit (VDU) corresponds to a pod. The currently deployed VNF requires one pod each: A-Pod {Cpu(6v) / Mem(40G)}, B-Pod {Cpu(2v) / Mem(20G)}, C-Pod {Cpu(4v) / Mem(40G)}, and D-Pod {Cpu(4v) / Mem(30G)}. Available resources include the numa0 {FreeCpu(8v) / Mem(100G)} and numa1 {FreeCpu(8v) / Mem(100G)} boards in frame 0, slot 0. CISM calculates the deployment location for each pod based on the order in which B, C, D, and A are dequeued. The calculation shows that B-Pod and C-Pod are deployed on numa0, and D-Pod is deployed on numa1. However, when calculating the deployment location for A-Pod, resource allocation fails because the remaining resources on numa0 and numa1 do not meet A-Pod's requirements. Because existing technologies perform resource calculations at the granularity of a single "Pod", when an A-Pod fails to be deployed, CISM can only repeatedly calculate the deployment location of the A-Pod within the deployment time limit until it times out, resulting in VNF deployment failure.

[0093] On the other hand, when the deployment location is calculated at the granularity of a single VDU, when multiple VNFs share a host aggregate (HA) resource pool, the resource requirements of the VDU in the VNF are discrete. The remaining resources of multiple boards in the HA are insufficient to deploy other VDUs, resulting in a large amount of resource fragmentation that cannot be utilized.

[0094] To solve the above problems, the present application provides a method for deploying virtual network functions, aiming to improve network deployment efficiency.

[0095] The technical concept of this application is: when allocating resources for the VDUs required by the VNF to be deployed, all the required VDUs are associated as a whole group, and the resources required by the full VDUs are calculated and scheduled at the group granularity within a calculation cycle. When the scheduling of any VDU fails, the available resources are recalculated for the VDUs in the group that have been successfully allocated resources until the corresponding resources are scheduled for the entire group of VDUs or the scheduling of the entire group of VDUs fails.

[0096] When calculating the deployment location of the VDU of the VNF to be deployed at the group granularity, the allocation method that is not conducive to the overall deployment is recalculated and adjusted, thereby improving the success rate of VNF deployment.

[0097] FIG2 is a conceptual diagram of a workload group provided in an embodiment of the present application. The present application introduces the concept of workload group to assist the infrastructure service manager in group scheduling of the VDUs required for the deployment of VNFs. For the infrastructure service manager, a workload group is a custom resource, and a workload group object is an instantiation of this type of resource. As shown in FIG2 , workload group 1 (WorkloadGroup1) is a workload group object, which has a series of attributes, including: workload group name identifier, number of group members, group scheduling policy, group scheduling priority, and group scheduling timeout.

[0098] The three VDUs (workload1, workload2, and workload3) are the VDUs required for the VNF to be deployed. They are associated with WorkloadGroup1's name through a label and become members of WorkloadGroup1. Workload4 is a common VDU for comparison. The group scheduling policy is the scheduling policy used by the CISM / VIM when deploying WorkloadGroup1 members (workload1-3). Based on resource requirements, it can be categorized as Least Host, Least Allocate, or Most Allocate. The group scheduling priority distinguishes the scheduling priorities of different workload group objects. A lower group scheduling priority indicates a higher priority. It should be noted that VDU priorities are manually set, and members of the same workload group object have the same priority. The group scheduling timeout is the time limit for the CISM / VIM to allocate resources to workload1-3. If any VDU of workload1-3 does not receive the required scheduling resources within the group scheduling timeout, the deployment is considered a failure.

[0099] Figure 3 is an exemplary interactive flow chart of a method for deploying a virtual network function according to an embodiment of the present application. This method can be applied to the system shown in Figure 1. The application of this deployment method in the NFV system shown in Figure 1 is described in detail below in conjunction with Figure 3. The infrastructure service manager in this embodiment of the present application can be a CISM or a VIM.

[0100] S301, the VNFM receives third request information, where the third request information is used to trigger deployment of a virtual deployment unit VDU for a first virtual network function VNF.

[0101] In this step, the third request information is used to trigger the deployment of the VDU for the first VNF. As an example, the third request information is a request information for instantiating the first VNF. As the VNF management module, the VNFM allows the administrator to trigger the first VNF ​​instantiation request through the VNFM deployment interface and send the third request information to the VNFM.

[0102] It should be noted that before triggering an instantiation request through the VNFM deployment interface, the deployment package and virtualized network function module descriptor (VNFD) for the first VNF ​​must be uploaded. The deployment package contains the software image used to run the VNF, which is distributed to the Infrastructure Service Manager. The VNFD is a descriptive template for deploying the VNF, from which the VNFM extracts the input parameters required to deploy the first VNF. The deployment package and VNFD template can be uploaded through the VNF Portal.

[0103] In some implementations, the third request message also carries first information, where the first information indicates that the load group object identifier associated with the VDU required by the first VNF ​​is the first load group object identifier, and thus the VDU required by the first VNF ​​becomes a group member of the first load group object. It is understood that the load group object identifier is the load group name identifier in the load group object attributes, and the various attribute information of the first load group object is used to assist the infrastructure service manager in group scheduling the VDU required by the first VNF.

[0104] It should be noted that the load group object identifier associated with the VDU required by the first VNF ​​indicated by the first information and the attribute information contained in the first load group object are obtained based on the VNFD. In other words, when the load group concept is introduced to assist the CISM / VIM in group scheduling of the VDUs required for the VNF to be deployed, the VNFD template also needs to include the group scheduling tag of the VDU required by the VNF to be deployed and the attribute parameters of the associated load group object. After obtaining this information based on the VNFD, the administrator can make further modifications as needed.

[0105] In this embodiment, when deploying the VDU associated with the load group object according to the attribute information of the load group object, as shown in FIG4 , the method further includes S301 - 1 , S301 - 2 and S301 - 3 .

[0106] S301-1, the VNFM sends a second request message to the infrastructure service manager to obtain a load group object list and information.

[0107] The infrastructure service manager provides the VNFM with an interface for adding, deleting, modifying, and querying load group objects. The VNFM sends a second request message to the infrastructure service manager and can query the identifiers and attribute information of the load group objects contained in the load group object list.

[0108] S301 - 2 , the VNFM determines whether to create or update the first load group object, generates and sends configuration information to the infrastructure service manager.

[0109] In this step, the VNFM may query, based on the first information, whether the load group object list includes a load group object identified as the first load group object identifier.

[0110] If the load group object list does not include the load group object identified as the first load group object identifier, the VNFM determines that the first load group object needs to be created. Accordingly, the VNFM generates second information for instructing the creation of the load group object identified as the first load group object.

[0111] If the load group object list includes a load group object identified by the first load group object identifier, the load group object is recorded as the initial first load group object. The VNFM further determines whether the attribute information of the initial first load group object is the same as the attribute information of the first load group object indicated by the first information. If the attribute information of the initial first load group object is different from the attribute information of the first load group object indicated by the first information, the VNFM determines that the attribute information of the initial first load group object in the load group object list needs to be updated. Accordingly, the VNFM generates third information, which is used to indicate that the attribute information of the load group object identified by the first load group object identifier needs to be updated.

[0112] The second information or the third information indicates the configuration information required by the first load group object, and the VNFM sends it to the infrastructure service manager

[0113] S301 - 3 : The infrastructure service manager creates or updates a first load group object in response to the configuration information.

[0114] The second information or the third information includes the identification and attribute information of the first load group object. According to the configuration information, the infrastructure service manager creates the first load group object, or updates the attribute information of the initial first load group object in the load group object list to be consistent with the attribute information of the first load group object included in the configuration information.

[0115] It can be understood that the above three steps S301-1, S301-2 and S301-3 serve as a group scheduling verification function, which is to prevent the identification and other attribute information of the first load group object indicated by the first information from conflicting with the existing load group object information in the load group object list, thereby avoiding the failure of the first VNF ​​deployment. It is a non-essential step in the first VNF ​​deployment process.

[0116] S302: The VNFM sends a first request message to the infrastructure service manager, where the first request message is used to request deployment of a VDU for a first VNF.

[0117] In this step, the first request information is used to request the infrastructure service manager to create a VDU for the first VNF, wherein the number of VDU instances required by the first VNF ​​is determined according to the third request information, and thus the first request information is generated based on the third request information.

[0118] In one possible implementation, when the third request information carries first information for indicating the load group object identifier associated with the VDU required by the first VNF, in order to pass the content indicated in the first information to the infrastructure service manager, the first request information also carries the first information.

[0119] S303: The infrastructure service manager creates M VDUs for the first VNF. The M VDUs are associated as VDUs in the same group, where M is a positive integer.

[0120] Among them, according to the first request information, it can be known that M VDUs are required to deploy the first VNF. The infrastructure service manager creates the M VDUs required for the first VNF, and the M VDUs are associated as VDUs in the same group.

[0121] In a possible implementation, when the first request information carries the first information, the M VDUs created by the infrastructure service manager are also associated with the first load group object identifier through a label. At this time, the M VDUs become group members of the first load group object.

[0122] S304, the infrastructure service manager deploys VDUs in the same group using a group scheduling method. The group scheduling method allocates resources to each VDU in the same group within a calculation cycle until each VDU in the same group obtains the corresponding resources or the allocation of resources to the entire group of VDUs in the same group fails. The calculation cycle is used to calculate the corresponding resources for each VDU in the same group.

[0123] When M VDUs are associated as a group, the Infrastructure Service Manager can use group scheduling methods to calculate corresponding deployment locations for all group members. If any group member does not have a corresponding mapped location in the schedulable resources after calculation, the group member VDU that has obtained the corresponding mapped location will be backtracked and the schedulable resources will be adjusted until all M VDUs have calculated the corresponding mapped location or the calculation time exceeds one calculation cycle. If the calculation time exceeds one calculation cycle, the entire group of VDUs fails to be scheduled. After all M VDUs have obtained the corresponding scheduling location, the Infrastructure Service Manager will bind each VDU to the corresponding resource node, and the first VNF ​​is considered to be successfully deployed.

[0124] In this embodiment, VNFs are deployed using a group scheduling approach, which recalculates and adjusts allocation methods that are unfavorable to overall deployment, thereby improving the success rate of VNF deployment. Furthermore, the concept of load groups is introduced, and by using a calculation-before-binding approach, unnecessary operations during deployment are reduced, thereby improving VNF deployment efficiency.

[0125] In the above embodiment, the group scheduling verification function is implemented through VNFM during the VNF instantiation process. In some implementations, the load group object can also be manually managed through VNFM before instantiation. The method of manually managing load group objects is described below with reference to Figure 5 of the embodiment. As shown in Figure 5, in the VNF deployment method provided in the embodiment of the present application, the deployment planning of the load group object can also be implemented before instantiation, specifically including:

[0126] S501: The VNFM receives fourth request information, where the fourth request information is used to query a load group object list.

[0127] Before instantiating a VNF, you can query the status of existing load group objects in the load group object list through VNFM to facilitate advance planning based on the deployment requirements of the VNF to be deployed.

[0128] S502: The VNFM sends a second request message to the infrastructure service manager to obtain a load group object list.

[0129] The load group object is an instantiation of a custom resource and is essentially also a resource that can be managed by the infrastructure service manager. The second request information is used to obtain a list of load group objects from the infrastructure service manager.

[0130] S503: VNFM outputs a load group object list.

[0131] The VNFM feeds back the obtained load group object list to the inquirer. Based on the deployment requirements of the VNF to be deployed, the inquirer determines whether to create a new load group object or update the attribute information of the existing load group object in the load group object list.

[0132] For this embodiment, deploying the first VNF ​​requires attribute information of the first load group object, and querying whether the load group object list contains a load group object identified as the first load group object identifier.

[0133] If the load group object list does not include a load group object identified as the first load group object identifier, it is determined that the first load group object needs to be created.

[0134] If the load group object list contains a load group object identified as the first load group object identifier, the load group object is recorded as the initial first load group object. If the attribute information of the initial first load group object is inconsistent with the attribute information of the first load group object required for the first VNF ​​deployment, it is determined that the attribute information of the initial first load group object in the load group object list needs to be updated.

[0135] S504: The VNFM receives fifth request information, where the fifth request information is used to request creation or update of a load group object whose load group object identifier is the first load group object identifier.

[0136] In this step, the inquirer sends a fifth request message to the VNFM based on the judgment result in S503. If the judgment result is that the first load group object needs to be created, the fifth request message is used to request the creation of a load group object whose load group object identifier is the first load group object identifier. If the judgment result is that the attribute information of a load group object in the load group object list needs to be updated, the fifth request message is used to request the update of the load group object whose load group object identifier is the first load group object identifier.

[0137] According to the judgment result, the fifth request information carries the second information or the third information, and the second information or the third information includes the attribute information that the load group object whose load group object identifier is the first load group object identifier should include or should be updated.

[0138] S505: The VNFM sends a fifth request message to the infrastructure service manager.

[0139] The Infrastructure Service Manager provides VNFM with an interface for adding, deleting, modifying, and querying load group objects.

[0140] S506: The infrastructure service manager creates or updates a load group object whose load group object identifier is the first load group object identifier based on the fifth request information.

[0141] If the fifth request message carries the second information, the infrastructure service manager creates a first load group object; if the fifth request message carries the third information, the infrastructure service manager updates the attribute information of the initial first load group object in the load group object list to be consistent with the attribute information of the first load group object contained in the third information.

[0142] The above steps are mainly for the deployment planning of the first VNF. For the deployment of multiple VNFs, the load group objects can also be manually managed through VNFM. For example:

[0143] As shown in Figure 6, the VNF-1 to be deployed contains three pods: Control Pod 1, Compute Pod 1, and Compute Pod 2. The VNF-2 to be deployed contains three pods: Control Pod 2, Compute Pod 3, and Compute Pod 4. VNF-1 and VNF-2 need to be co-deployed. To ensure compact deployment of VNF-1 and VNF-2, a load group object, WG1, is defined. WG1's attributes are as follows: Load group name: WG1; Number of group members: 6; Group scheduling policy: Minimum number of used boards; Group scheduling priority: 1; Group scheduling timeout: 30s. Each pod in VNF-1 and VNF-2 is associated with the name of WG1 through a label.

[0144] After querying the load group object list, no load group object with the name identifier WG1 exists, so it is determined that the WG1 object needs to be created. According to the defined attribute parameters of the WG1 object, the fifth request message and the included second information are configured, and the infrastructure service manager creates the WG1 object based on the fifth request message.

[0145] In this embodiment, by manually managing load group objects through VNFM, it is possible to add, delete, modify and query load group objects, and it can also avoid conflicts between the attributes of the load group objects associated with the VDU required for the VNF to be deployed and the existing load group objects in the load group object list.

[0146] In this application, the Infrastructure Service Manager uses a group scheduling method to allocate resources to M VDUs associated with a first load group object. For the Infrastructure Service Manager, a load group is a new custom resource, so it is necessary to distinguish between VDUs associated with a load group object and ordinary load VDUs. The following describes a scheduling method for VDUs associated with a load group object using a load group name identifier, using a bare metal container scenario as an example.

[0147] Figure 7 is a schematic diagram of the component architecture of the K8S extended group scheduling provided in an embodiment of the present application. As shown in Figure 7, "kubectl" is used as a command line tool to control and manage the K8S cluster. IF1, as an extended interface, provides operations for adding, deleting, modifying, and querying load group objects. User-defined resources (custom resource definition, CRD) are built-in resource types of K8S. Through CRD resources, custom resources such as load group WG can be converted into resource types on K8S. K8S mainly includes four major components: API server (API Server), controller manager (Controller-Manager), "Etcd" and scheduler (Scheduler).

[0148] Among them, the API server is the unified access entrance for all resources in the K8S cluster, and exists as the core hub for data interaction between various components in the cluster; the controller manager is the core component of K8S for automatic management of various resources. Through automatic management, it can monitor the changes of various states in the cluster in real time; "Etcd" is responsible for storing information on various resource objects and status data in the cluster; the scheduler is responsible for cluster resource scheduling and scheduling "Pod" to the appropriate node.

[0149] The K8S scheduler expands the group scheduling queue (GroupPriorityQueue) and group scheduling process (ScheduleGroup) based on the original priority queue (PriorityQueue), scheduling process (Scheduleone) and scheduler cache (Cache). Among them, the group scheduling queue is used to store the load group members to be scheduled, and the group scheduling process provides a group scheduling algorithm to complete the group scheduling calculation for the load group members that have left the queue. It can be understood that the group scheduling queue and the group scheduling process can distinguish between the VDU of the associated load group object and the ordinary load VDU, ensuring that there is no conflict between the ordinary load VDU and the VDU of the associated load group object during the scheduling process.

[0150] K8S can essentially be seen as a mini-CaaS manager, so K8S can be used as an example of an infrastructure service manager. Figure 8 is a schematic diagram of the process of implementing group scheduling in K8S according to an embodiment of the present application. As shown in Figure 8, the group scheduling process may include S801 to S807.

[0151] S801: The API server receives the first request information and creates the "Pod" required by the first VNF.

[0152] In this step, a first request message is used to request the deployment of a "Pod" for the first VNF. Based on the first request message, it is known that M "Pods" are required to deploy the first VNF. The first request message carries first information, indicating that the load group object identifier associated with the "Pod" required by the first VNF ​​is the first load group object identifier WG1. M "Pods" associated with the first load group object identifier WG1 are created through the reset interface (REST API) of the API server.

[0153] S802: The API server stores the data of the "Pod" required by the first VNF ​​into "Etcd".

[0154] Since "Etcd" is equivalent to the K8S database, the API server stores the data of the M created "Pods" in "Etcd".

[0155] It can be understood that steps S801 and S802 are actually equivalent to step S303 in the embodiment shown in Figure 3. K8S creates M VDUs through the API server and "Etcd".

[0156] It should be noted that if the load group object list does not contain a load group object identified as the first load group object identifier WG1, the first load group object WG1 can also be created through the API server and "Etcd", and a load group queue WG1 corresponding to the first load group object WG1 can be created in the group scheduling queue.

[0157] S803, the monitor (Informer) in the scheduler monitors the status change of "Etcd" and adds it to the load group queue corresponding to the group scheduling queue according to the load group object identifier associated with the "Pod" to be scheduled.

[0158] The monitor in the scheduler uses the "list-Watch" mechanism to detect the addition of an unscheduled "Pod" in "Etcd". The group scheduling queue contains a load group queue WG1 corresponding to the first load group object WG1. The scheduler adds the "Pod" with the associated load group object identified as WG1 to the load group queue WG1. If the "Pod" does not have an associated load group object, it is added to the priority queue. Since the API server serves as the hub for data interaction between components in the K8S framework, when the monitor monitors changes in the "Etcd" status, it needs to be transferred through the API server.

[0159] It's important to note that as a custom resource, load group objects store information in "Etcd." Before group scheduling begins, the scheduler monitors resource changes in "Etcd" and creates a corresponding load group queue in the group scheduling queue for each load group object in the load group object list.

[0160] At the same time, in some implementations, the scheduler's monitor is also responsible for updating the scheduler cache, which caches the information of the "Pod" to be scheduled and the resource node in the "Etcd".

[0161] S804: The group scheduling queue dequeues the load group member "Pod" that meets the dequeue rules.

[0162] To implement group scheduling calculations for load group member "Pods", load group member "Pods" in the group scheduling queue can be dequeued and subsequent group scheduling operations performed only if the dequeuing rules are met.

[0163] As an example, the dequeue rules are as follows:

[0164] (1) Check the length of each load group queue in the group scheduling queue. When the load group queue length is greater than or equal to the upper limit of the number of group members of the load group, refer to rule (2). The load group queue length represents the number of "Pods" in the load group queue.

[0165] (2) Check the group scheduling priority of the load group object associated with each load group queue. The load group queue with the smallest value has the highest scheduling priority and is dequeued the earliest. If there are load group queues with the same group scheduling priority, refer to rule (3);

[0166] (3) Check the creation time of the load group queues that meet the conditions of rule (2). Compare the creation time of the load group queues with the same group scheduling priority. The load group queue created earliest will be dequeued the earliest.

[0167] S805 , the group scheduling process in the scheduler calls the group scheduling algorithm to complete the group scheduling calculation according to the group scheduling policy of the dequeued load group object.

[0168] The group scheduling process provides a group scheduling algorithm. The group scheduling policy in the WG1 attribute of the outgoing load group object corresponds to the group scheduling algorithm one-to-one. The group scheduling process calls the corresponding group scheduling algorithm based on the group scheduling policy of the outgoing load group object to calculate the nodes to be scheduled for the M "Pods" in the outgoing load group queue.

[0169] In this step, the group scheduling calculation ends until each "Pod" in the outgoing load group queue obtains the corresponding node or the time consumed exceeds the group scheduling timeout indicated in the load group object attribute information.

[0170] S806: The scheduler performs a scheduling operation on the load group member "Pod".

[0171] In this step, the scheduler modifies the value of the "nodeName" field of the "Pod" object to the name of the "Node" scheduled by the "Pod" object. At the same time, the scheduler updates the "Pod" information and "Node" information in the scheduler cache.

[0172] S807: The scheduler initiates an update request to the API server. The update request is used to update the Pod data in "Etcd".

[0173] The scheduler initiates a request to update the "Pod" data in "Etcd" through the API server. Only at this step is the "Pod" bound to the corresponding node and resources are allocated to the "Pod".

[0174] In this embodiment, according to the dequeueing rules of the load group queue, the entire group of "Pods" in the load group queue is dequeued, so that the group scheduling algorithm can be called to calculate the corresponding resource nodes for the entire group of "Pods". Resource allocation is achieved through calculation and binding, thereby improving the success rate of "Pod" deployment and avoiding the complex operations of repeated binding and unbinding.

[0175] It should be noted that after completing the group scheduling deployment, VNFM can check whether the full amount of "Pod" required for the first VNF ​​to be deployed can be successfully started. If the full amount of "Pod" required for the first VNF ​​is successfully started, it is judged that the first VNF ​​is instantiated successfully and the first VNF ​​is successfully deployed.

[0176] The group scheduling algorithm proposed in the above embodiment requires calculation for the entire group of "Pods". Compared with the method of calculating at the granularity of a single Pod, the computing resources consumed are greatly increased. Therefore, the following introduces an extended group scheduling method that can alleviate the pressure of computing resource consumption.

[0177] Figure 9 is a schematic diagram of the framework of the K8S extended coordinator supporting group scheduling calculations provided by an embodiment of the present application. As shown in Figure 9, based on Figure 7, the K8s extended coordinator (Co-Scheduler) is a replica of the manually created main scheduler. Therefore, the coordinator also includes a group scheduling queue and a group scheduling process.

[0178] Figure 10 is a schematic diagram of a process flow for implementing group scheduling in K8S through a master coordinator according to an embodiment of the present application. The specific process for implementing group scheduling in the K8S architecture when a coordinator is included in Figure 9 is described below in conjunction with Figure 10. As shown in Figure 10, the group scheduling process implemented by the master coordinator may include S1001 to S1008.

[0179] S1001: The coordinator creates a load group queue.

[0180] Before group scheduling begins, the master scheduler monitors resource changes in "Etcd" through the monitor and has created a corresponding load group queue for each load group object in the load group object list in the group scheduling queue.

[0181] In this embodiment, the coordinator, as a replica of the master scheduler, also needs to create a corresponding load group queue for each load group object in the load group object list in the group scheduling queue within the coordinator. It should be noted that there is no order relationship between the master scheduler and the coordinator when they create the load group queues.

[0182] S1002: The API server receives the first request information and creates the "Pod" required by the first VNF.

[0183] S1003: The API server stores the data of the "Pod" required by the first VNF ​​into "Etcd".

[0184] The processes of steps S1002 and S1003 are consistent with steps S801 and S802 in the embodiment shown in FIG8 .

[0185] S1004: The main scheduler pre-selects a node for the "Pod" required by the first VNF ​​and locks the candidate node.

[0186] In this step, the master scheduler first adds the Pod required by the first VNF ​​to the associated load group queue according to steps S803 and S804 in the embodiment of Figure 8 . It then dequeues the Pod from the load group queue after the dequeue rules are met. Then, based on a pre-selected policy, the master scheduler uses the API server to filter nodes from the "Etcd" table that do not meet the policy, and locks the candidate nodes in the scheduler cache.

[0187] S1005: The master scheduler updates the "Node" information, "Pod" information, and the information of the load group object associated with the "Pod" in the "Etcd" through the API server.

[0188] The main scheduler uses the API server to label the candidate node in "Etcd" with the label of the first load group object to prevent "Pod" that is not a member of the first load group from calling the candidate node. At the same time, it updates the information of the dequeued "Pod" in "Etcd" and the preliminary scheduling information of the first load group object to facilitate subsequent coordinators to perform group scheduling operations.

[0189] S1006: The coordinator calls the group scheduling algorithm to allocate a resource node for the load group member "Pod" that has been dequeued from the candidate nodes.

[0190] The coordinator monitors resource changes in "Etcd" through a monitor. Based on the preliminary scheduling information of the first load group object, the scheduler within the coordinator caches the nodes with the first load group object's tag and the information of the dequeued "Pods". The above process of the monitor monitoring "Etcd" status changes is relayed through the API server.

[0191] The outgoing load group member "Pod" is associated with the first load group object. At the low computing load node, the coordinator calls the corresponding group scheduling algorithm according to the group scheduling strategy of the first load group object, and calculates the node to be scheduled for the outgoing "Pod" within the range of nodes with the first load group object label. The group scheduling calculation ends only when each "Pod" in the outgoing load group queue obtains the corresponding node or the time consumed exceeds the group scheduling timeout indicated in the attribute information of the first load group object.

[0192] After the group scheduling calculation is completed, the coordinator modifies the value of the "nodeName" field of the "Pod" object and updates the "Pod" information and "Node" information in the scheduler cache within the coordinator.

[0193] S1007: The coordinator updates the “Node” information, “Pod” information, and the information of the load group object associated with the “Pod” in the “Etcd” through the API server.

[0194] In this step, the coordinator updates the "Pod" information in "Etcd" through the API server, binds the "Pod" to the corresponding node, and clears the load group object identifier of the alternative node. The preliminary scheduling information of the first load group object is updated to the scheduled information.

[0195] S1008: The main scheduler updates the cache information.

[0196] The master scheduler unlocks the candidate node in the scheduler cache and clears the candidate node's load group object identifier. This step is not required for deploying the Pod required for the first VNF; it primarily prevents deployment failures of other load group member Pods due to inability to access the candidate node. Because the master scheduler updates cached information based on the Pod information in the Etcd, this step requires intermediary processing via the API server.

[0197] In this embodiment, a low computing load node dynamic start-stop coordinator is used to complete the group scheduling calculation of a load group object with a huge number of group members, which more efficiently utilizes computing resources and alleviates computing overhead pressure.

[0198] In some implementations, the scheduler's computational overhead can be shared by extending the schedule solver. Figure 11 is a schematic diagram of the framework of the K8S extended schedule solver supporting group scheduling calculations, provided in an embodiment of the present application. The schedule solver is a third-party scheduler independent of the cloud OS.

[0199] Figure 12 is a schematic diagram of a process flow for implementing group scheduling in K8S through a master scheduler and a scheduling solver according to an embodiment of the present application. The specific process for implementing group scheduling in the K8S architecture when a scheduling solver is included in Figure 11 will be described below in conjunction with Figure 12. As shown in Figure 12, the group scheduling process implemented by the master scheduler and the scheduling solver may include S1201 to S1208.

[0200] S1201: The main scheduler pulls up the scheduling solver and registers it for use.

[0201] Multiple scheduling solver instances are distributed and deployed on low computing load nodes. Since the scheduler solver is an independent third-party scheduler, the main scheduler needs to predict the computing resources consumed by group scheduling based on the number of group members of the load group object and pull up one or more scheduling solvers on demand.

[0202] S1202: The API server receives the first request information and creates the "Pod" required by the first VNF.

[0203] S1203: The API server stores the data of the "Pod" required by the first VNF ​​into "Etcd".

[0204] S1204: The main scheduler pre-selects a node for the "Pod" required by the first VNF ​​and locks the candidate node.

[0205] The above steps S1202 to S1204 are consistent with steps S1002 to S1004 in the embodiment shown in FIG. 10 .

[0206] S1205: The main scheduler selects a scheduling solver instance and pushes candidate nodes, out-of-queue load group member "Pod" information, and load group information to it.

[0207] In this step, the master scheduler can select one or more scheduling solver instances based on the load balancing strategy and push the candidate nodes, the "Pod" information of the dequeued load group member, and the attribute information of the first load group object associated with the "Pod" to the selected scheduling solver instance. It should be noted that if multiple scheduling solver instances are selected, the information to be pushed will be sent to the multiple scheduling solver instances in batches, and the multiple scheduling solver instances will operate in parallel.

[0208] S1206 , the scheduling solver calls the group scheduling algorithm to allocate a resource node for the load group member “Pod” that has been dequeued from the candidate nodes.

[0209] The outgoing load group member "Pod" is associated with the first load group object. The scheduling solver calls the corresponding group scheduling algorithm according to the group scheduling strategy of the first load group object, and calculates the node to be scheduled for the outgoing "Pod" within the range of nodes with the first load group object label. The group scheduling calculation ends only when each "Pod" in the outgoing load group queue obtains the corresponding node or the time consumed exceeds the group scheduling timeout indicated in the attribute information of the first load group object.

[0210] S1207, the main scheduler receives the group scheduling calculation result sent by the scheduling solver and updates the main scheduler cache information.

[0211] In this step, the master scheduler modifies the value of the nodeName field of the Pod object based on the group scheduling calculation result and updates the Pod and Node information in the scheduler cache within the master scheduler.

[0212] S1208: The master scheduler updates the "Node" information, "Pod" information, and the information of the load group object associated with the "Pod" in the "Etcd" through the API server.

[0213] This step is substantially the same as S1007 in the embodiment of Figure 10, except that in step S1007 the coordinator initiates the update request, while in step S1208 the master scheduler initiates the update request. S1209: The master scheduler updates the cache information.

[0214] This step is consistent with S1008 in the embodiment Figure 10. It is also a non-essential step for deploying the "Pod" required for the first VNF. Its main function is to prevent the member "Pod" of other load group objects from failing to deploy due to the inability to call alternative nodes.

[0215] In this embodiment, group scheduling calculation is implemented by deploying a scheduling solver on a low computing load node, which can more efficiently utilize computing resources and alleviate computing overhead pressure.

[0216] In the above embodiments, the first request information is used to request the instantiation of the first VNF. In some implementations, the first request information can also be used to request the expansion of the VDU for the first VNF. It should be noted that when the first request information is used to request the expansion of the VDU for the first VNF, the first information carried by the first request information indicates that the load group object identifier associated with the expanded VDU of the first VNF ​​is the first load group object identifier. Expanding the VDU for the first VNF ​​means that the current group member VDU that becomes the first load group object through label association has reached the upper limit of the group members of the first load group, so the first request information also carries second information, and the second information indicates the new attribute information of the first load group object. The new attribute information of the first load group object modifies the upper limit of the number of group members of the original first load group to the upper limit of the number of group members after expansion.

[0217] It can be understood that only when the first VNF ​​expansion VDU is fully started, the first VNF ​​is considered to have been successfully expanded. Therefore, the first VNF ​​expansion VDU needs to be deployed to the corresponding resource node subsequently. The deployment method of the expansion VDU is consistent with the deployment method of the VDU required for the first VNF ​​in the above embodiment, and will not be repeated here.

[0218] In the above embodiments, the request information received by the VNFM is input by an administrator on the VNFM. In some implementations, the NFVO can receive the request information input by the administrator, and the request information can be transmitted to the VNFM via the NFVO. Accordingly, the deployment software package and VNFD of the VNF to be deployed are also uploaded through the NFVO's deployment interface, and the NFVO can also implement group scheduling verification functions. For example, in step S301 in Figures 3 and 4, the NFVO can receive a first request, which is used to trigger the deployment of a virtual deployment unit (VDU) for a first virtual network function (VNF). The NFVO, using the VNFM as a relay, sends a second request to the infrastructure service manager via the VNFM, requesting the deployment of a VDU for the first VNF.

[0219] In some implementations, the NFVO may interact directly with the infrastructure service manager. In this case, the VNFM in the above embodiment may be replaced by the NFVO, and the related methods executed by the VNFM in the embodiment method may be executed by the NFVO.

[0220] Figure 13 is a schematic diagram of the structure of a virtual network function deployment device provided by one embodiment of the present application. The deployment method implemented by VNFM in Figures 3 to 5 of the embodiments is described below in conjunction with Figure 13. As shown in Figure 13, the device 1300 of this embodiment may include: a receiving module 1301 and a sending module 1302.

[0221] Among them, the receiving module 1301 is used to receive the third request information, and the third request information is used to trigger the deployment of a virtual deployment unit VDU for the first virtual network function VNF.

[0222] The sending module 1302 is used to send a first request message to the infrastructure service manager, where the first request message is used to request deployment of a VDU for the first VNF.

[0223] In some implementations, the third request information and the first request information carry first information, and the first information indicates that the load group object identifier associated with the VDU required by the first VNF ​​is the first load group object identifier, and the load group object identifier is used to indicate the load group object.

[0224] In some implementations, the sending module 1302 is further configured to send a second request message to the infrastructure service manager, where the second request message is used to request a query for a load group object list.

[0225] In some implementations, the receiving module 1301 is further configured to receive fourth request information, where the fourth request information is used to query a load group object list.

[0226] In some implementations, the receiving module 1301 is also used to receive a fifth request message, the fifth request message is used to request the creation or update of a load group object whose load group object identifier is the first load group object identifier, the fifth request message carries second information or third information, the second information includes attribute information that the load group object whose load group object identifier is the first load group object identifier should include, and the third information includes attribute information that the load group object whose load group object identifier is the first load group object identifier should update.

[0227] Figure 14 is a schematic diagram of the structure of a virtual network function deployment apparatus provided in another embodiment of the present application. The deployment method implemented by the infrastructure server in Figures 3 to 5 of the embodiments is described below in conjunction with Figure 14. As shown in Figure 14, apparatus 1400 of this embodiment may include: a receiving module 1401, a processing module 1402, and a verification module 1403.

[0228] The receiving module 1401 is used to receive first request information from the VNFM, where the first request information is used to request deployment of a VDU for the first VNF.

[0229] The processing module 1402 is used to create M VDUs for the first VNF ​​and deploy the M VDUs using a group scheduling method. The group scheduling method is that when there are multiple VDUs to be deployed and resource allocation fails for the first VDU among the multiple VDUs, resources are reallocated to the second VDU among the multiple VDUs. The second VDU includes the VDUs among the multiple VDUs to which resources have been successfully allocated before the first VDU.

[0230] In some implementations, the verification module 1403 is used to create a load group object whose load group object identifier is the first load group object identifier based on the second information, or to update the load group object whose load group object identifier is the first load group object identifier based on the third information.

[0231] In some implementations, the receiving module 1401 is configured to receive second request information from the VNFM, where the second request information is used to request a load group object list.

[0232] In some implementations, the receiving module 1401 is used to receive a fifth request message from the VNFM, and the verification module 1403 is used to create a load group object whose load group object identifier is the first load group object identifier based on the second information carried in the fifth request message, or to update the load group object whose load group object identifier is the first load group object identifier based on the third information carried in the fifth request message.

[0233] It should be understood that apparatus 1300 and apparatus 1400 are embodied in the form of functional modules. The term "module" may refer to an application specific integrated circuit (ASIC), an electronic circuit, a processor (e.g., a shared processor, a dedicated processor, or a group processor, etc.) and memory for executing one or more software or firmware programs, combined logic circuits, and / or other suitable components that support the described functionality.

[0234] The aforementioned apparatus 1300 and apparatus 1400 have the function of implementing the corresponding processes and / or steps in the aforementioned method embodiments; the aforementioned functions may be implemented by hardware, or by hardware executing corresponding software. The hardware or software may include one or more modules corresponding to the aforementioned functions.

[0235] Figure 15 is a schematic diagram of the structure of a virtual network function deployment device provided in yet another embodiment of the present application. Device 1500 shown in Figure 15 can be used to execute any of the aforementioned methods performed by the virtual network function deployment device. The virtual network function deployment device of Figures 13 and 14 of the embodiments is described in detail below with reference to Figure 15.

[0236] As shown in Figure 15 , the apparatus 1500 of this embodiment includes a memory 1501, a processor 1502, a communication interface 1503, and a bus 1504. The memory 1501, the processor 1502, and the communication interface 1503 are connected to each other via the bus 1504.

[0237] The memory 1501 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1501 may store a program. When the program stored in the memory 1501 is executed by the processor 1502, the processor 1502 is configured to execute any of the aforementioned methods.

[0238] The processor 1502 may be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit, or one or more integrated circuits to execute related programs.

[0239] The processor 1502 may also be an integrated circuit chip with signal processing capabilities. In the implementation process, the various related steps in the embodiment of the present application may be completed by hardware integrated logic circuits in the processor 1502 or software instructions.

[0240] The processor 1502 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The processor 1502 may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor.

[0241] The steps of the method disclosed in conjunction with the embodiments of the present application can be directly implemented as being executed by a hardware decoding processor, or can be implemented by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in memory 1501, and processor 1502 reads the information in memory 1501 and, in combination with its hardware, completes the functions required to be performed by the units included in the device of the present application.

[0242] The communication interface 1503 may use, but is not limited to, a transceiver or other transceiver device to implement communication between the apparatus 1500 and other devices or apparatuses.

[0243] The bus 1504 may include a path for transmitting information between various components of the device 1500 (eg, the memory 1501 , the processor 1502 , and the communication interface 1503 ).

[0244] An embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored. When a processor executes the computer instructions, each step of the method in the above embodiment is implemented.

[0245] An embodiment of the present application further provides a computer program product, including computer instructions, which, when executed by a processor, implement the various steps of the method in the above embodiment.

[0246] It should be noted that the modules or components shown in the above embodiments may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits, or one or more microprocessors, or one or more field programmable gate arrays. For another example, when a module is implemented by a processing element calling program code, the processing element may be a general-purpose processor, such as a central processing unit or other processor that can call program code, such as a controller. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0247] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, software modules or any combination thereof. When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrations. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0248] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0249] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for deploying a virtual network function, applied to an infrastructure service manager, characterized in that: The method comprises: receiving a first request message sent from a virtual network function manager VNFM, wherein the first request message is used to request deployment of a virtual deployment unit VDU for a first virtual network function VNF; Create M VDUs for the first VNF, where the M VDUs are associated as VDUs in the same group, where M is a positive integer; The VDUs in the same group are deployed using a group scheduling method, wherein the group scheduling method allocates resources to each VDU in the same group of VDUs within a calculation cycle until each VDU in the same group of VDUs obtains corresponding resources or the resource allocation of the entire group of VDUs in the same group fails, and the calculation cycle is used to calculate the corresponding resources for each VDU in the same group of VDUs.

2. The method according to claim 1, characterized in that The first request information carries first information, and the first information indicates that the load group object identifier associated with the VDU required by the first VNF ​​is a first load group object identifier, and the load group object identifier is used to indicate a load group object, and the load group object includes the following attribute information: the maximum number of VDUs that can be associated with the load group object identifier, the scheduling policy when the VDU associated with the load group object identifier is deployed using the group scheduling method, the scheduling priority when the VDU associated with the load group object identifier is deployed using the group scheduling method, and the scheduling timeout time when the VDU associated with the load group object identifier is deployed using the group scheduling method; The step of creating M VDUs for the first VNF ​​includes: Creating the M VDUs, and associating each of the M VDUs with the first load group object identifier; The M VDUs are deployed using a group scheduling method, including: Using the group scheduling method, the VDU associated with the first load group object identifier is deployed according to the attribute information of the load group object indicated by the first load group object identifier.

3. The method according to claim 2, characterized in that The method of using the group scheduling method to deploy the VDU associated with the first load group object identifier according to the attribute information of the load group object indicated by the first load group object identifier includes: Get the undeployed VDU; When the load group object identifier associated with the undeployed VDU is the first load group object identifier, adding the undeployed VDU to the scheduling queue associated with the obtained first load group object identifier; Using the group scheduling method, the VDUs in the scheduling queue are deployed according to the attribute information of the load group object indicated by the first load group object identifier.

4. The method according to claim 3, characterized in that The using the group scheduling method to deploy the VDUs in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier includes: When the number of VDUs in the scheduling queue is equal to the maximum number in the load group object indicated by the first load group object identifier, the group scheduling method is used to deploy the VDUs in the scheduling queue according to the scheduling policy and scheduling priority in the load group object indicated by the first load group object identifier until each VDU in the scheduling queue is successfully deployed or the deployment duration exceeds the scheduling timeout duration in the load group object indicated by the first load group object identifier.

5. The method according to any one of claims 2 to 4, characterized in that The method further comprises: Receive second information from the VNFM, where the second information is used to instruct creation of a load group object identified as the first load group object identifier.

6. The method according to claim 5, characterized in that The method further comprises: Sending a load group object list to the VNFM; The receiving the second information from the VNFM includes: When the load group object list does not include the load group object identified as the first load group object identifier, receiving the second information from the VNFM.

7. The method according to any one of claims 2 to 6, characterized in that The method further comprises: Receive third information from the VNFM, where the third information is used to indicate updating of attribute information in a load group object identified as the first load group object identifier.

8. The method according to claim 7, characterized in that The method further comprises: Sending a load group object list to the VNFM; The receiving the third information from the VNFM includes: When the load group object list includes a load group object identified as the first load group object identifier, receiving all the requests from the VNFM The third information.

9. The method according to claim 6 or 8, characterized in that: Before sending the load group object list to the VNFM, the method further includes: A second request message is received from the VNFM, where the second request message is used to request a query of a load group object list.

10. The method according to claim 3, characterized in that: The using the group scheduling method to deploy the VDUs in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier includes: Associating a group scheduling label with a first node by a first scheduler, the first node comprising a candidate deployment node of the VDU in the scheduling queue, the first scheduler being used to allocate resources to the VDU; The group scheduling method is used by a second scheduler to deploy the VDUs in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier, and the second scheduler is a copy of the first scheduler.

11. The method according to claim 10, characterized in that The using the group scheduling method to deploy the VDUs in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier includes: Associating a group scheduling label with a first node by a first scheduler, the first node comprising a candidate deployment node of the VDU in the scheduling queue, the first scheduler being used to allocate resources to the VDU; The group scheduling method is used by at least one scheduling solver to deploy the VDUs in the scheduling queue according to the attribute information of the load group object indicated by the first load group object identifier, and the scheduling solver is an independent third-party scheduler.

12. A method for deploying a virtual network function, applied to a virtual network function manager VNFM, characterized in that: The method comprises: receiving third request information, where the third request information is used to trigger deployment of a virtual deployment unit VDU for the first virtual network function VNF; A first request message is sent to the infrastructure service manager, where the first request message is used to request deployment of a VDU for the first VNF.

13. The method according to claim 12, characterized in that The third request information and the first request information carry first information, the first information indicating that the load group object identifier associated with the VDU required by the first VNF ​​is a first load group object identifier, the load group object identifier is used to indicate a load group object, and the load group object includes the following attribute information: the maximum number of VDUs that can be associated with the load group object identifier, the scheduling policy when deploying the VDU associated with the load group object identifier using the group scheduling method, the scheduling priority when deploying the VDU associated with the load group object identifier using the group scheduling method, and the scheduling timeout time when deploying the VDU associated with the load group object identifier using the group scheduling method.

14. The method according to claim 13, characterized in that The method further comprises: Sending second information to the infrastructure service manager, wherein the second information is used to instruct creation of a load group object identified as the first load group object identifier.

15. The method according to claim 14, characterized in that The method further comprises: receiving a list of load group objects from the infrastructure service manager; The sending of the second information to the infrastructure service manager includes: When the load group object list does not include the load group object identified as the first load group object identifier, second information is sent to the infrastructure service manager.

16. The method according to any one of claims 13 to 15, characterized in that The method further comprises: Sending third information to the infrastructure service manager, wherein the third information is used to instruct to update the attribute information in the load group object identified as the first load group object identifier.

17. The method according to claim 16, characterized in that The method further comprises: receiving a list of load group objects from the infrastructure service manager; The sending of the third information to the infrastructure service manager includes: When the load group object list includes a load group object identified as the first load group object identifier, third information is sent to the infrastructure service manager.

18. The method according to claim 15 or 17, characterized in that Before receiving the load group object list from the infrastructure service manager, the method further includes: Send a second request message to the infrastructure service manager, where the second request message is used to request a query of a load group object list.

19. The method according to any one of claims 12 to 18, characterized in that The third request information is specifically used to request instantiation of the first VNF, or the third request information is specifically used to request expansion of VDU for the first VNF.

20. The method according to any one of claims 12 to 19, characterized in that The third request information is input by an administrator on the VNFM, or the third request information is received from a network function virtualization orchestrator NFVO.

21. A virtual network function deployment device, characterized in that: The device includes a functional module for implementing the deployment method of the virtual network function as described in any one of claims 1 to 11, or includes a functional module for implementing the deployment method of the virtual network function as described in any one of claims 12 to 19.

22. A virtual network function deployment system, characterized in that: It comprises a first deployment device and a second deployment device, wherein the first deployment device is used to execute the virtual network function deployment method as described in any one of claims 1 to 11, and the second deployment device is used to execute the virtual network function deployment method as described in any one of claims 12 to 19.

23. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement a method for deploying a virtual network function as described in any one of claims 1 to 11, or any one of claims 12 to 19.

24. A computer program product, characterized in that It comprises a computer program, which, when executed by a processor, implements the method for deploying a virtual network function as described in any one of claims 1 to 11, or any one of claims 12 to 19.

Citation Information

Patent Citations

  • Method for deploying virtual machine and related device

    CN112306625A

  • Virtual network function deployment method and device, equipment and storage medium

    CN114143184A

  • Containerized VNF deployment method and device

    CN115883283A

  • Method and device for deploying virtualized network function

    CN115994015A

  • Deployment method and deployment device of virtual network function

    CN116723105A