Real-time cloud rendering multi-tenant sharing quantitative computing power scheduling method

Abstract

The invention discloses a real-time cloud rendering multi-tenant sharing quantitative computing power scheduling method. According to the invention, in a public cloud platform environment, aiming at a multi-tenant, multi-application and multi-specification complex amount-based payment scene, queuing logic in a machine resource limited scene is realized. According to the method, a consumption queue is provided, the task processing efficiency of the system is improved through logic decoupling on the basis of a message queue achieved through the technology, and the polling processing complexity of the tenant N * application N * specification N is reduced to the single-dimension polling scheduling of the specification N.

Description

Technical Field

[0001] The present invention relates to the field of cloud rendering technology, and in particular to a scheduling method for real-time cloud rendering multi-tenant sharing of volume computing power. Background Art

[0002] In real-time cloud rendering public cloud platform services, when multiple tenants use pay-as-you-go billing at the same time, the same tenant may need to use machines of different specifications to run different cloud applications, and different cloud applications between different tenants may need to use machines of the same specifications. Therefore, in multi-tenant scenarios with limited pay-as-you-go computing resources, the existing technical solutions lack effective message queue strategies to meet the pay-as-you-go resource scheduling needs of the above complex scenarios, and can simultaneously meet the high efficiency of resource scheduling and scheduling fairness among tenants.

[0003] The demand for on-demand computing power scheduling in general scenarios only involves the demand for a single machine specification, and different tenants are allocated on-demand computing power based on the principle of "first come, first served". However, in the on-demand scenario of real-time cloud rendering, multiple services involving multiple tenants require multiple machine specifications. In this scenario, business needs cannot be met simply based on a single dimension of first come, first served.

[0004] The cloud applications of all tenants of the platform cannot use one queue. When the machine specifications required by the cloud applications at the front of the queue have no resources, the machine specifications required by the applications at the back of the queue have resources but cannot be allocated to them, causing the queue to be stuck.

[0005] All tenants on the platform cannot use the same machine resource specification queue. Different tenants purchase different quantities of pay-as-you-go services. When the pay-as-you-go services of the tenants at the front of the queue reach the upper limit of the paid quantity, but there are still resources for the required machine specifications, the tenants at the back of the queue cannot be allocated, causing the queue to get stuck.

[0006] The platform uses one queue for each tenant and each machine specification. If different tenants need to use the same machine specification, the system needs to traverse all message queues, which reduces the efficiency of system queue processing.

[0007] Therefore, the prior art has defects and needs to be improved. Summary of the invention

[0008] The technical problem to be solved by the present invention is: to provide a scheduling method for real-time cloud rendering multi-tenant shared metered computing power, to meet the resource scheduling and allocation in complex public cloud metered resource scenarios, while meeting the requirements of efficiency and fairness; and to solve the possible message queue jam problem.

[0009] The technical solution of the present invention is as follows: a scheduling method for real-time cloud rendering multi-tenant shared volume computing power is provided, comprising the following steps.

[0010] S1: Platform tenants complete the deployment and service configuration of cloud applications.

[0011] S2: Cloud application access requests that require the creation of machine resources are assigned to the message queue.

[0012] S3: The scheduling system determines the availability of real-time cloud services, including the upper limit of resource usage and whether resources are idle; the scheduling system initializes a consumption queue for each machine resource specification in the resource pool.

[0013] Determine whether the cloud application in the current message queue has reached the upper limit of machine resource usage; if the upper limit has been reached, the request will remain in the message queue, and the queue position information will be displayed on the client according to the position of the message queue where the current cloud application access request is located, and then re-polling every A seconds; if the upper limit has not been reached, determine whether the machine resources required for the current cloud application access request are idle; if they are idle, the cloud application access request will be started and run directly; if they are not idle, synchronize the current cloud application access request to the consumption queue, and display the queue position information on the client according to the position of the message queue where the current cloud application access request is located, and then re-polling every B seconds.

[0014] S4: The consumer queue in the scheduling system receives the cloud application access request synchronized to the consumer queue in the message queue; and allocates the request to the designated consumer queue according to the machine resource specification information carried in the cloud application access request synchronized to the message queue.

[0015] S5: The scheduling system monitors the resource pool in real time. When resources are idle, it allocates the required resources to the specified consumption queue according to the consumption queue information.

[0016] The specific steps of step S1 are: S11: complete the deployment of cloud applications in the cloud; S12: configure the real-time cloud rendering machine resource specifications for use when running on a volume basis according to the computing power required for the cloud application; S13: configure the upper limit of machine resource usage for the cloud application according to actual business needs, and each machine resource can support access by one terminal user.

[0017] The specific steps of step S2 are: S21: determine the tenant to which the cloud application access request belongs; S22: determine the machine resource specifications required for the cloud application access request; S23: according to the tenant and machine resource specifications, request tasks with the same tenant and the same machine resource specifications enter the same message queue.

[0018] In step S3, the range of A is 1-10; the range of B is 1-10;

[0019] Step S4 also includes: the consumption queue records the synchronization time for each newly synchronized cloud application access request and marks it as "consumption time".

[0020] The specific steps of step S5 are: the scheduling system monitors in real time whether there are idle resources in the platform's volume resource pool; when idle resources appear, the requests of the corresponding consumption queue are allocated according to the "consumption time" on a first-come, first-served basis according to the machine resource specifications;

[0021] For cloud application access requests that have obtained resource allocation, they are dequeued from the message queue and the cloud application is started based on the tenant and machine resource specification request information; if the cloud application access request also exists in the consumption queue, it is also removed from the consumption queue.

[0022] By adopting the above scheme, the present invention provides a scheduling method for real-time cloud rendering multi-tenant shared pay-as-you-go computing power. On the basis of using one queue for each tenant and each machine specification, the concept of "consumption queue" is innovatively proposed. After each tenant's access task to each machine specification enters the corresponding queue, the scheduling system first performs a round of system service logic judgment, and puts the tasks that meet the system conditions into the consumption queue in the order of "queue entry time", and re-marks the "consumption time" for each task. The scheduling system only needs to schedule the consumption queue in the order of the consumption time of the task, which can meet the resource scheduling requirements in the above complex many-to-many pay-as-you-go scenario. In the public cloud platform environment, the present invention realizes the queuing logic in the scenario of limited machine resources for complex pay-as-you-go scenarios with multiple tenants, multiple applications, and multiple specifications. The present invention proposes a "consumption queue". On the basis of the message queue realized by the technology, the system task processing efficiency is improved through logical decoupling, and the polling processing complexity of [tenant N*application N*specification N] is reduced to a single-dimensional polling scheduling of only [specification N]. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0024] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0025] See also Figure 1 The present invention provides a scheduling method for real-time cloud rendering multi-tenant shared computing power, including the following steps.

[0026] S1: Platform tenants complete the deployment and service configuration of cloud applications.

[0027] S2: Cloud application access requests that require the creation of machine resources are assigned to the message queue.

[0028] S3: The scheduling system determines the availability of real-time cloud services, including the upper limit of resource usage and whether resources are idle; the scheduling system initializes a consumption queue for each machine resource specification in the resource pool.

[0029] Determine whether the cloud application in the current message queue has reached the upper limit of machine resource usage; if it has reached the upper limit, the request will stay in the message queue, and according to the position of the message queue where the current cloud application access request is located, the queue position information will be displayed on the client, and then re-polling will be performed every A seconds; if it has not reached the upper limit, determine whether the machine resources required for the current cloud application access request are idle; if they are idle, the cloud application access request will be directly started and run; if they are not idle, synchronize the current cloud application access request to the consumption queue, and according to the position of the message queue where the current cloud application access request is located, the queue position information will be displayed on the client, and then re-polling will be performed every B seconds. In this step, the range of A is 1-10; the range of B is 1-10;

[0030] S4: The consumer queue in the scheduling system receives the cloud application access request synchronized to the consumer queue in the message queue; and allocates the request to the designated consumer queue according to the machine resource specification information carried in the cloud application access request synchronized to the message queue.

[0031] S5: The scheduling system monitors the resource pool in real time. When resources are idle, it allocates the required resources to the specified consumption queue according to the consumption queue information.

[0032] In the present invention, a message queue is allocated to a cloud application access request, and then the cloud service availability is judged. If the cloud service has reached the upper limit of machine resource usage, it means that there is no idle machine, and then the round-robin waiting is performed directly. If the upper limit of machine resource usage has not been reached, it means that there are idle resources, and then it is judged whether the machine resources required by the current cloud application access request are idle; if there are idle resources, the cloud application access request is directly started and run; if there are no idle resources, it means that there are no machine resources in the cloud service that meet the requirements of the cloud application access request; these cloud application access requests are synchronized to the consumption queue, and the consumption queue is created according to each machine resource specification, that is, it starts to queue under each different resource specification; the scheduling system monitors the resource pool in real time, and when the resources are idle, the amount of resources will be allocated to the designated consumption queue according to the consumption queue information, and machine resources are allocated to these cloud application requests. That is, the present invention proposes a "consumption queue", which, based on the message queue realized by technology, realizes the improvement of system task processing efficiency through logical decoupling, and reduces the complexity of polling processing from [tenant N*application N*specification N] to a single-dimensional polling scheduling of only [specification N] (i.e., each machine resource specification).

[0033] The specific steps of step S1 are: S11: complete the deployment of cloud applications in the cloud; S12: configure the real-time cloud rendering machine resource specifications for use when running on a volume basis according to the computing power required for the cloud application; S13: configure the upper limit of machine resource usage for the cloud application according to actual business needs, and each machine resource can support access by one terminal user.

[0034] The specific steps of step S2 are: S21: determine the tenant to which the cloud application access request belongs; S22: determine the machine resource specifications required for the cloud application access request; S23: according to the tenant and machine resource specifications, request tasks with the same tenant and the same machine resource specifications enter the same message queue.

[0035] Step S4 also includes: the consumption queue records the synchronization time for each newly synchronized cloud application access request and marks it as "consumption time".

[0036] The specific steps of step S5 are: the scheduling system monitors in real time whether there are idle resources in the platform's volume resource pool; when idle resources appear, the requests of the corresponding consumption queue are allocated according to the "consumption time" on a first-come, first-served basis according to the machine resource specifications;

[0037] For cloud application access requests that have obtained resource allocation, they are dequeued from the message queue and the cloud application is started based on the tenant and machine resource specification request information; if the cloud application access request also exists in the consumption queue, it is also removed from the consumption queue.

[0038] In summary, the present invention provides a scheduling method for real-time cloud rendering multi-tenant shared pay-as-you-go computing power. On the basis of using one queue for each tenant and each machine specification, the concept of "consumption queue" is innovatively proposed. After each tenant's access task to each machine specification enters the corresponding queue, the scheduling system first performs a round of system service logic judgment, and puts the tasks that meet the system conditions into the consumption queue in the order of "queue entry time", and re-marks the "consumption time" for each task. The scheduling system only needs to schedule the consumption queue in the order of the consumption time of the task, which can meet the resource scheduling requirements in the above-mentioned complex many-to-many pay-as-you-go scenarios. In the public cloud platform environment, the present invention realizes the queuing logic in the scenario of limited machine resources for complex pay-as-you-go scenarios with multiple tenants, multiple applications, and multiple specifications. The present invention proposes a "consumption queue". On the basis of the message queue realized by the technology, the system task processing efficiency is improved through logical decoupling, and the polling processing complexity of [tenant N*application N*specification N] is reduced to a single-dimensional polling scheduling of only [specification N].

[0039] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A scheduling method for real-time cloud rendering multi-tenant shared computing power, characterized in that: The following steps are involved: S1: Platform tenants complete the deployment and service configuration of cloud applications; S2: Cloud application access requests that need to create machine resources are assigned to the message queue; S3: The scheduling system determines the availability of real-time cloud services, including resource usage limits and whether resources are idle. The scheduling system initializes a consumption queue for each machine resource specification in the resource pool. Determine whether the cloud application of the current message queue has reached the upper limit of machine resource usage; If the upper limit has been reached, the request will remain in the message queue, and the queue position information will be displayed on the client according to the position of the message queue where the current cloud application access request is located, and then it will be polled again every A seconds; If the upper limit is not reached, determine whether the machine resources required by the current cloud application access request are idle; If there is idle time, the cloud application access request is directly started and run; If there is no idle time, synchronize the current cloud application access request to the consumption queue, and display the queue position information on the client according to the position of the message queue where the current cloud application access request is located, and then re-pollating every B seconds; S4: The consumer queue in the scheduling system receives the cloud application access request synchronized to the consumer queue in the message queue; according to the machine resource specification information carried in the cloud application access request synchronized by the message queue, the request is allocated to the designated consumer queue; S5: The scheduling system monitors the resource pool in real time. When resources are idle, it allocates the required resources to the specified consumption queue according to the consumption queue information.

2. According to claim 1, a scheduling method for real-time cloud rendering multi-tenant sharing of volume computing power is characterized in that: The specific steps of step S1 are: S11: Complete the deployment of cloud applications in the cloud; S12: Based on the computing power required for the cloud application to run, configure the real-time cloud rendering machine resource specifications for the application to be used when running on a volume basis; S13: Configure the upper limit of machine resource usage for cloud applications based on actual business needs. Each machine resource can support access by one end user.

3. According to claim 1, a scheduling method for real-time cloud rendering multi-tenant sharing of volume computing power is characterized in that: The specific steps of step S2 are: S21: Determine the tenant to which the cloud application access request belongs; S22: Determine the machine resource specifications required for the cloud application access request; S23: According to the tenant and machine resource specifications, request tasks of the same tenant and the same machine resource specifications enter the same message queue.

4. According to claim 1, a scheduling method for real-time cloud rendering multi-tenant sharing of volume computing power is characterized in that: In step S3, the range of A is 1-10; the range of B is 1-10.

5. The method for scheduling real-time cloud rendering multi-tenant shared computing power according to claim 1, characterized in that: Step S4 also includes: the consumption queue records the synchronization time for each newly synchronized cloud application access request and marks it as "consumption time".

6. A scheduling method for real-time cloud rendering multi-tenant shared computing power according to claim 5, characterized in that: The specific steps of step S5 are: The scheduling system monitors the platform's resource pool in real time to see if there are idle resources; When idle resources appear, requests to the corresponding consumption queue are allocated on a first-come, first-served basis according to the "consumption time" based on the machine resource specifications.

7. The method for scheduling real-time cloud rendering multi-tenant shared computing power according to claim 1, characterized in that: For cloud application access requests that have obtained resource allocation, they are dequeued from the message queue and the cloud application is started based on the tenant and machine resource specification request information; if the cloud application access request also exists in the consumption queue, it is also removed from the consumption queue.

Images