CN115134371A - Scheduling method, system, equipment and medium containing edge network computing resources - Google Patents
Scheduling method, system, equipment and medium containing edge network computing resources Download PDFInfo
- Publication number
- CN115134371A CN115134371A CN202210676053.XA CN202210676053A CN115134371A CN 115134371 A CN115134371 A CN 115134371A CN 202210676053 A CN202210676053 A CN 202210676053A CN 115134371 A CN115134371 A CN 115134371A
- Authority
- CN
- China
- Prior art keywords
- scheduling
- network
- computing
- job
- resource
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1023—Server selection for load balancing based on a hash applied to IP addresses or costs
Abstract
The invention discloses a scheduling method, a system, equipment and a medium containing edge network computing resources, which comprises the following steps: sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user; according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node; and carrying out load balancing scheduling according to the probabilities of different nodes. The heterogeneous computational power scheduling method comprises heterogeneous computational power scheduling of the computational power resources of the edge network, and realizes load balancing of different computational nodes and efficient execution of the operation by defining computational power resource allocation processing flows under different scenes and calculating the probability of current operation being allocated to different nodes.
Description
Technical Field
The invention relates to the field of communication network computing, in particular to a scheduling method, a system, equipment and a medium containing edge network computing resources, and relates to a heterogeneous computing power scheduling strategy of the edge network computing resources.
Background
With the continuous development of communication technology, edge computing has become an important business model in the 5G era, and can provide more service capabilities and wider application scenarios. With the continuous development of information technology and the wide application of various intelligent devices, the data scale generated by people is increasingly huge, and increasingly generated mass data needs a stronger computing platform as a support.
In the current platforms for processing multiple data, computational power resources are distributed dispersedly, and particularly, the situation that the computational power of a core network and the computational power of an edge network exist simultaneously exists. The increasing demand for computing power has made centralized management and scheduling of distributed computing power even more urgent. In recent years, with the continuous development of edge computing, the scheduling of heterogeneous resources evolves from the direction of core network computing power to the direction of joint scheduling of core network computing power and edge network computing power.
The first method adopted at present is a resource scheduling method applied to a centralized environment, which is applied to a resource scheduling program, runs on a server for heterogeneous computation, obtains a resource demand amount according to a task amount calculated by a current target computing unit, and generates respective resource allocation information of different computing programs in a current scheduling period according to comparison between the resource demand amount and a pre-allocated resource demand. The second type is a heterogeneous resource-oriented multi-dimensional scheduling system, which comprises a cluster management module, a container management module, a resource scheduling module, a hardware abstraction module, a relational database and a component communication module, and the cluster resources are uniformly scheduled and managed by designing the resource scheduling into a two-stage scheduling architecture.
The resource scheduling method is applied to a centralized environment, the scheduled resources are heterogeneous resources on each server, and the scheduling method is mainly used for scheduling heterogeneous resources such as GPU. Heterogeneous resources between different servers lack uniform scheduling. The scheduling method takes the servers as the center, and each server schedules the resource requirements according to the running calculation tasks on the server, so that the problem of unbalanced use of heterogeneous resources among different servers can occur. The multi-dimensional scheduling for heterogeneous resources mainly aims at the heterogeneous resources in a cluster range and is performed through a two-stage resource scheduling framework. The applicable scene of the scheduling method is limited to the interior of the cluster, and the global scheduling of the heterogeneous resources in a larger range is not considered.
Disclosure of Invention
In view of the above-mentioned shortcomings, an object of the present invention is to provide a scheduling method, system, device and medium including edge network computing resources, which implement load balancing and improve processing efficiency by calculating the probability of current operations being allocated to different nodes in different scenarios.
In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:
a scheduling method for computing power resources of an edge network comprises the following steps:
sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user;
according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and carrying out load balancing scheduling according to the probabilities of different nodes.
As a further improvement of the present invention, before the calculation of the computing resource policy and the probability that the current job is allocated to different nodes in different scenarios by the computing resource scheduling method, the method further comprises:
according to the job type and the job requirement, the network address of the job access node is analyzed to obtain the position of the job access node, and then the calculation resources are distributed according to the position of the job access node by a calculation resource scheduling method.
As a further improvement of the present invention, before the calculation of the computing resource policy and the probability that the current job is allocated to different nodes in different scenarios by the computing resource scheduling method, the method further comprises:
and analyzing the job requirements according to the job types and the job requirements, and then distributing the computing power resources according to the analyzed computing power requirements by a computing power resource scheduling method.
As a further improvement of the present invention, before analyzing the job requirement, the method further includes:
and according to the job type and the job requirement, analyzing the network address of the job access node, then analyzing the job requirement, and then distributing the computing resources according to the analyzed computing requirement by a computing resource scheduling method.
As a further improvement of the present invention, the computing resource scheduling method includes:
when the computing tasks are all in the cluster of the core network and new jobs are submitted, the core network has N nodes, N respectively 1 ，N 2 ，…，N n Selecting k nodes from the resource requirements, M respectively 1 ，M 2 ，…，M k The loads distributed thereto are respectively L 1 ，L 2 ，…，L k (ii) a The specific operation allocation strategy is as follows: taking the current task with probability P j Is allocated to node M j ，P j Obtained from the following equation (1):
x in the formula (1) j Representing total load of a job to be de-allocatedThe proportion of the load on the current node j to the total load of the job is obtained by the following formula (2):
l in the formula (2) full For total load of work, L j Load distributed to the current node j; l is full Obtained by the formula (3):
l in the formula (3) i Load distributed to the current node i;
the proportion X of the load of the current node to the total load is calculated and obtained by the formula (2) j Formula (1) by X j And obtaining the probability of scheduling the tasks to different nodes.
As a further improvement of the present invention, the load balancing according to the probabilities of different nodes is implemented by that when the load ratio of the rest nodes of the cluster where the core network is located is larger, the probability that the job is scheduled to the node is larger, so as to implement the load balancing of different nodes.
As a further improvement of the invention, when the computing task part is completed in the edge network and the edge network completion part is completed in the core network, the current total load of the edge network is L E The total current load of the core network is L C Selecting m nodes from the edge network according to the resource requirement, wherein the nodes are respectively E 1 ，E 2 ，…，E m The loads distributed thereto are respectively L 1 ，L 2 ，…，L m Selecting n nodes, C respectively, from the core network 1 ，C 2 ，…， C n The loads distributed thereto are respectively L 1 ，L 2 ，…，L n Then L is E Calculated from equation (4):
L C calculated from equation (5):
for the node l in the edge network, the proportion of the load except the current node to the total load is X l Calculated by the formula (6),
for node k in the core network, the proportion of the load of the current node to the total load is X k Calculated by the formula (7),
the edge network assigns tasks to it with a probability P l To node E l ，P l Obtained from equation (8) as follows:
the core network assigns tasks thereto with a probability P k To node C k ，P k Obtained from equation (9) as follows:
and obtaining the probability of the task to be scheduled to different nodes.
As a further improvement of the present invention, the load balancing according to the probabilities of different nodes is implemented by that when the load ratio of the rest nodes of the cluster where the edge network or the core network is located is larger, the probability that the job is scheduled to the node is larger, so as to implement the load balancing of different nodes inside the core network or the edge network.
A scheduling system including edge network computing resources, comprising:
the computing resource sensing module is used for acquiring the operation type and the operation requirement of a user;
the computing resource management module is used for obtaining a computing resource strategy and the probability that the current operation is distributed to different nodes in different scenes by computing through a computing resource scheduling method according to the operation type and the operation requirement, and further performing network scheduling by combining a network scheduling method; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and the balanced scheduling module is used for carrying out load balanced scheduling according to the probability of different nodes.
Optionally, the computing resource management module includes:
the computing resource scheduling module is used for scheduling and managing computing resources of the core network and the edge network, allocating operating resources to users according to the operation types and the operation requirements of the users, and dynamically adjusting the resource allocation of the user operation according to the requirements;
the network scheduling module is used for sinking the service gateway to a node where a user is located and routing the job related data to a processing node by combining the computing resource scheduling module after determining the position of the user request;
and the job deployment module is used for deploying the jobs to the specified nodes.
An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the scheduling method including edge network computing resources when executing the computer program.
A computer-readable storage medium, having stored thereon a computer program for implementing the steps of the scheduling method involving edge network computing power resources when executed by a processor.
Compared with the prior art, the invention has the following beneficial effects:
the invention relates to a heterogeneous computing power scheduling strategy containing computing power resources of an edge network, which realizes load balance of different computing nodes, realizes efficient execution of operation and realizes the computing power resource allocation processing strategy under different scenes by defining computing power resource allocation processing flows under different scenes and calculating the probability of current operation allocated to different nodes; particularly, the method solves the problem that the computing task is completed in the cluster where the core network is located, the computing task is completed in the edge network, and the computing task is partially completed in the core network at the edge network completion part, and realizes the scheduling strategy of computing resources, thereby realizing load balance and improving the processing efficiency.
Drawings
FIG. 1 is a flow chart of a scheduling method including edge network computing resources according to the present invention;
FIG. 2 is a process flow of computing power resource allocation in scenario one of the embodiments of the present invention
FIG. 3 is a process flow of computing power resource allocation in scenario two of the present invention
FIG. 4 is a process flow of computing power resource allocation in scenario three of the present invention;
FIG. 5 is a scenario diagram illustrating a process flow for resource allocation for four calculations according to the present invention;
fig. 6 is a schematic diagram of an electronic device according to the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood by those skilled in the art, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The invention aims to provide a scheduling strategy of heterogeneous computational power resources of an edge network, which realizes load balancing and improves processing efficiency by calculating the probability of current operation distributed to different nodes under different scenes.
As shown in fig. 1, the present invention provides a scheduling method for computing resources including an edge network, including:
sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user;
according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and carrying out load balancing scheduling according to the probabilities of different nodes.
The method comprises heterogeneous computing power scheduling of computing power resources of the edge network, and realizes load balance of different computing nodes and efficient execution of the operation by defining computing power resource allocation processing flows under different scenes and calculating the probability of current operation allocated to different nodes.
The calculation resource management module deploys the operation according to the type of the operation, the operation requirement and the like, and the method is specifically divided into the following scenes:
(1) in a first scenario, under the condition that the calculation power demand and the network resource demand of the job are clear, the calculation power resource management module automatically schedules calculation power resources and networks according to the job type and the job demand, and deploys the job to the nodes. The scheduling method for this scenario is the same as above.
(2) And in a second scenario, the computing power requirement of the job is clear, the network requirement is not clear, the network address of the job access node is analyzed, then computing power resources are distributed nearby according to the position of the job access node by the computing power resource scheduling module, and the job is deployed to the node by the joint network scheduling module. The scheduling method of the scene is based on the steps, and the following steps are added:
before the calculation of the computing resource strategy and the probability of distributing the current operation to different nodes under different scenes by the computing resource scheduling method, the method further comprises the following steps:
according to the operation type and the operation requirement, the network address of the operation access node is analyzed to obtain the position of the operation access node, and then the calculation resource is distributed by the calculation resource scheduling method according to the position of the operation access node.
(3) And in a third scenario, the network requirement of the operation is clear, the computing power requirement is not clear, the operation requirement is analyzed firstly, then the computing power resource is distributed by the computing power resource scheduling module according to the analyzed computing power requirement, and the operation is deployed to the node by the combined network scheduling module. The scheduling method of the scene is added with the following steps on the basis of the steps:
before the computing power resource strategy and the probability of distributing the current operation to different nodes under different scenes are obtained by the computing power resource scheduling method, the method also comprises the following steps:
and analyzing the job requirements according to the job types and the job requirements, and then distributing the computing power resources according to the analyzed computing power requirements by a computing power resource scheduling method.
(4) And in a fourth scenario, the network requirement of the job is not clear, the computing power requirement is not clear, firstly, the network address of the job access node is analyzed, then, the job requirement is analyzed, next, the computing power resource is distributed by the computing power resource scheduling module according to the analyzed computing power requirement, and the job is deployed to the node by the combined network scheduling module. The scheduling method of the scene is based on the steps, and the following steps are added:
before the analysis of the operation requirement, the method further comprises the following steps:
and according to the job type and the job requirement, analyzing the network address of the job access node, then analyzing the job requirement, and then distributing the computing resources according to the analyzed computing requirement by a computing resource scheduling method.
The invention is described in further detail below with reference to the figures and examples. The invention is not limited to the examples given.
The resource scheduling framework comprises modules of computing power node resources, computing power resource perception, computing power resource management, application interface abstraction and the like. The computing resources are computing resources such as a CPU (X86, ARM and the like), a GPU, an NPU, a TPU and the like, and the computing resources are distributed on different computing nodes.
The computing resource management module is responsible for scheduling computing resources, network scheduling and job deployment to nodes. And (4) scheduling the computing resources, namely performing network and computing combined scheduling according to the operation requirement comprehensive network performance measurement and the computing resource available condition.
The calculation resource scheduling module: and the system is responsible for the calculation resource scheduling management of the core network and the edge network. And allocating operation resources for the user according to the operation type and the operation requirement of the user, and dynamically adjusting the resource allocation of the user operation according to the requirement.
A network scheduling module: after the position of the application request is determined, the joint calculation resource scheduling module sinks the service gateway to the node where the user is located, and the data related to the operation is routed to the processing node.
The operation deployment module: is responsible for deploying jobs on the designated nodes.
The calculation resource management module deploys the operation according to the type of the operation, the operation requirement and the like, and the operation is specifically divided into the following scenes:
(1) in a first scenario, under the definite conditions of the computing power requirement, the network resource requirement and the like of the job, the computing power resource management module automatically schedules the computing power resource and the network according to the job type and the job requirement, and deploys the job to the nodes. The specific process is shown in FIG. 1.
(2) And in a second scenario, the calculation power requirement of the job is clear, the network requirement is not clear, the network address of the job access node is firstly analyzed, then calculation power resources are distributed nearby by the calculation power resource scheduling module according to the position of the job access node, and the job is deployed to the node by combining the network scheduling module. The specific flow is shown in fig. 2.
(3) And in a third scenario, the network requirement of the job is clear, the calculation requirement is not clear, the job requirement is firstly analyzed, then the calculation resource scheduling module distributes the calculation resource according to the analyzed calculation requirement, and the joint network scheduling module deploys the job to the nodes. The specific flow is shown in fig. 3.
(4) And in a fourth scenario, the network requirement of the job is not clear, the computing power requirement is not clear, firstly, the network address of the job access node is analyzed, then, the job requirement is analyzed, next, the computing power resource is distributed by the computing power resource scheduling module according to the analyzed computing power requirement, and the job is deployed to the node by the combined network scheduling module. The specific flow is shown in fig. 4.
In the aspect of computing resource scheduling, under the condition that computing tasks are all completed in a cluster where a core network is located, when new jobs are submitted, the core network is assumed to have N nodes, wherein the N nodes are respectively N 1 ，N 2 ，…，N n Selecting k nodes from the resource requirements, wherein the k nodes are M respectively 1 ，M 2 ，…，M k The loads distributed thereto are each L 1 ，L 2 ，…，L k The specific operation allocation strategy is as follows: taking the current task with probability P j Is allocated to node M j Here, P is j This is obtained by the following equation (1):
x in the formula (1) j Representing total workload removal distribution to current nodeThe proportion of the load on j to the total load of the job is obtained by equation (2):
l in the formula (2) full For total load of work, L j Is the load allocated to the current node j. L is full Obtained from formula (3):
l in the formula (3) i Is the load allocated to the current node i.
The proportion X of the load of the current node to the total load is calculated by the formula (2) j Formula (1) by X j And obtaining the probability of scheduling the tasks to different nodes. When the load proportion of other nodes of the cluster where the core network is located is larger, the probability that the job is scheduled to the node is larger, and the load balance of different nodes is realized through the mode.
Under the condition that the computing tasks are all completed in the edge network, the scheduling strategy is similar to the condition that the computing tasks are all completed in the cluster where the core network is located, and load balancing of different nodes of the edge network is achieved.
In the case that the computation task is partially completed in the edge network and partially completed in the core network, assume that the current total load of the edge network is L E The current total load of the core network is L C Selecting m nodes from the edge network according to the resource requirement, wherein the nodes are respectively E 1 ，E 2 ，…，E m The loads distributed thereto are respectively L 1 ， L 2 ，…，L m Selecting n nodes, C respectively, from the core network 1 ，C 2 ，…，C n The loads distributed thereto are respectively L 1 ，L 2 ，…，L n Then L is E Can be calculated from equation (4):
L C can be calculated from equation (5):
for the node l in the edge network, the proportion of the load except the current node to the total load is X l Calculated by the formula (6),
for node k in the core network, the proportion of the load of the current node to the total load is X k Calculated by the formula (7),
the edge network assigns tasks to it with a probability P l To node E l Here P l Obtained from equation (8) as follows:
the core network assigns tasks thereto with a probability P k To node C k Here P k Obtained from equation (9) as follows:
in the scheduling strategy, when the load proportion of the rest nodes of the cluster where the edge network or the core network is located is higher, the probability that the job is scheduled to the node is higher, and the load balance of different nodes in the core network or the edge network is realized through the mode.
As shown in fig. 2 to 4, the present invention further provides a scheduling system including an edge network computing resource, including:
the computing resource sensing module is used for acquiring the operation type and the operation requirement of a user;
the computing resource management module is used for obtaining a computing resource strategy and the probability that the current operation is distributed to different nodes in different scenes by computing through a computing resource scheduling method according to the operation type and the operation requirement, and further performing network scheduling by combining a network scheduling method; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and the balanced scheduling module is used for carrying out load balanced scheduling according to the probability of different nodes.
In an embodiment of the present invention, the computing resource management module includes:
the computing resource scheduling module is used for being responsible for computing resource scheduling management of a core network and an edge network, allocating operating resources for users according to the operation types and the operation requirements of the users, and dynamically adjusting the resource allocation of the user operation according to the requirements;
the network scheduling module is used for sinking the service gateway to a node where a user is located and routing the job related data to a processing node by combining the computing resource scheduling module after determining the position of the user request;
and the job deployment module is used for deploying the job to the specified node.
In another embodiment, the computing resource management module further includes:
the operation access network point analysis module is arranged in front of the computing resource scheduling module and used for analyzing the network address of the operation access node according to the operation type and the operation requirement to obtain the position of the operation access node, and then computing resources are distributed according to the position of the operation access node by the computing resource scheduling method.
Wherein, the invention provides a third embodiment, and the computing power resource management module further includes:
and the calculation force demand analysis module is arranged in front of the calculation force resource scheduling module and is used for analyzing the job demand according to the job type and the job demand and then distributing the calculation force resources according to the analyzed calculation force demand by the calculation force resource scheduling method.
In a fourth embodiment, the computing resource management module further includes:
and the operation access network point analysis module is arranged in front of the computing power demand analysis module and is used for analyzing the network address of the operation access node according to the operation type and the operation demand, then analyzing the operation demand and distributing computing power resources according to the analyzed computing power demand by the computing power resource scheduling method.
As shown in fig. 6, the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the scheduling method including the edge network computing resource when executing the computer program.
The scheduling method of the computing power resource containing the edge network comprises the following steps:
sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user;
according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and carrying out load balancing scheduling according to the probabilities of different nodes.
The present invention also provides a computer-readable storage medium, which stores a computer program, which when executed by a processor implements the steps of the scheduling method including edge network computing resources.
The scheduling method of the computing resources comprising the edge network comprises the following steps:
sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user;
according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and carrying out load balancing scheduling according to the probabilities of different nodes.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention has been described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, it should be understood by those of ordinary skill in the art that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (12)
1. A scheduling method for computing resources of an edge network comprises the following steps:
sensing the resources of the whole computing power resource pool in real time, and acquiring the operation type and the operation requirement of a user;
according to the operation type and the operation requirement, a calculation resource scheduling method calculates to obtain a calculation resource strategy and the probability of distributing the current operation to different nodes under different scenes, and then a network scheduling method is combined to perform network scheduling; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and carrying out load balancing scheduling according to the probabilities of different nodes.
2. The method of claim 1, wherein the edge network computing resources are distributed in a distributed manner,
before the computing power resource strategy and the probability of distributing the current operation to different nodes under different scenes are obtained by the computing power resource scheduling method, the method also comprises the following steps:
according to the operation type and the operation requirement, the network address of the operation access node is analyzed to obtain the position of the operation access node, and then the calculation resource is distributed by the calculation resource scheduling method according to the position of the operation access node.
3. The method of claim 1, wherein the edge network computing resources are distributed in a distributed manner,
before the calculation of the computing resource strategy and the probability of distributing the current operation to different nodes under different scenes by the computing resource scheduling method, the method further comprises the following steps:
and analyzing the job requirements according to the job types and the job requirements, and then distributing the computing power resources according to the analyzed computing power requirements by a computing power resource scheduling method.
4. The method according to claim 3, wherein the scheduling method comprises the edge network computing resources,
before the analysis of the job requirement, the method further comprises the following steps:
and according to the job type and the job requirement, analyzing the network address of the job access node, then analyzing the job requirement, and then distributing the computing power resource according to the analyzed computing power requirement by a computing power resource scheduling method.
5. The method for scheduling computational power resources of an edge network according to any one of claims 1 to 4, wherein the method for scheduling computational power resources comprises:
when the computing tasks are all in the cluster of the core network and new jobs are submitted, the core network has N nodes, N is respectively 1 ，N 2 ，…，N n Selecting k nodes from the resource requirements, M respectively 1 ，M 2 ，…，M k The loads distributed thereto are respectively L 1 ，L 2 ，…，L k (ii) a The strategy of the specific operation allocation is: the current task is divided into a probability P j Is allocated to node M j ，P j Obtained from the following equation (1):
x in the formula (1) j The total load of the job is divided by the load distributed to the current node j, and the total load of the job is obtained by equation (2):
l in the formula (2) full For total load of operation, L j Load distributed to the current node j; l is a radical of an alcohol full Obtained by the formula (3):
l in the formula (3) i Load distributed to the current node i;
the proportion X of the load of the current node to the total load is calculated and obtained by the formula (2) j Formula (1) by X j And obtaining the probability of scheduling the tasks to different nodes.
6. The scheduling method including the edge network computing resource according to claim 5, wherein the load balancing according to the probability of the different nodes is realized by that when the load ratio of the rest nodes of the cluster where the core network is located is larger, the probability that the job is scheduled to the node is larger.
7. The method according to any of claims 1 to 4, wherein the edge network computing resources are distributed in a distributed manner,
the calculation task part is completed in the edge network and the part is completed in the core networkWhen the network is finished, the current total load of the edge network is L E The current total load of the core network is L C Selecting m nodes from the edge network according to the resource requirement, wherein the nodes are respectively E 1 ，E 2 ，…，E m The loads distributed thereto are respectively L 1 ，L 2 ，…，L m Selecting n nodes, C respectively, from the core network 1 ，C 2 ，…，C n The loads distributed thereto are respectively L 1 ，L 2 ，…，L n Then L is E Calculated from equation (4):
L C calculated from equation (5):
for the node l in the edge network, the proportion of the load except the current node to the total load is X l Calculated by the formula (6),
for node k in the core network, the proportion of the load of the current node to the total load is X k Calculated by the formula (7),
the edge network assigns tasks to it with a probability P l To node E l ，P l Obtained from equation (8) as follows:
the core network assigns tasks thereto with a probability P k To node C k ，P k Obtained from equation (9) as follows:
and obtaining the probability of scheduling the tasks to different nodes.
8. The scheduling method including the edge network computing resource according to claim 7, wherein the load balancing according to the probability of the different nodes is realized by that when the load ratio of the rest nodes of the cluster where the edge network or the core network is located is larger, the probability that the job is scheduled to the node is larger, so that the load balancing of the different nodes inside the core network or the edge network is realized.
9. A scheduling system including edge network computing resources, comprising:
the computing resource perception module is used for acquiring the operation type and the operation requirement of a user;
the computing resource management module is used for obtaining a computing resource strategy and the probability of distributing the current operation to different nodes under different scenes by a computing resource scheduling method according to the operation type and the operation requirement, and further performing network scheduling by combining a network scheduling method; obtaining job deployment information based on the computing resources and the network resources and outputting the job deployment information to an execution node;
and the balanced scheduling module is used for carrying out load balanced scheduling according to the probability of different nodes.
10. The scheduling system of claim 9 wherein the computing resource management module comprises:
the computing resource scheduling module is used for scheduling and managing computing resources of the core network and the edge network, allocating operating resources to users according to the operation types and the operation requirements of the users, and dynamically adjusting the resource allocation of the user operation according to the requirements;
the network scheduling module is used for sinking the service gateway to a node where the user is located and routing the data related to the operation to the processing node by combining the computing resource scheduling module after the position of the user request is determined;
and the job deployment module is used for deploying the job to the specified node.
11. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the scheduling method including edge network computing power resource of any one of claims 1-8 when executing the computer program.
12. A computer-readable storage medium, storing a computer program which, when executed by a processor, implements the steps of the method of scheduling comprising edge network computing power resources of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210676053.XA CN115134371A (en) | 2022-06-15 | 2022-06-15 | Scheduling method, system, equipment and medium containing edge network computing resources |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210676053.XA CN115134371A (en) | 2022-06-15 | 2022-06-15 | Scheduling method, system, equipment and medium containing edge network computing resources |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115134371A true CN115134371A (en) | 2022-09-30 |
Family
ID=83378397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210676053.XA Pending CN115134371A (en) | 2022-06-15 | 2022-06-15 | Scheduling method, system, equipment and medium containing edge network computing resources |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115134371A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115550370A (en) * | 2022-12-01 | 2022-12-30 | 浩鲸云计算科技股份有限公司 | Computing power resource optimal scheduling allocation method based on multi-factor strategy |
| CN115883660A (en) * | 2022-11-21 | 2023-03-31 | 中国联合网络通信集团有限公司 | Industrial production computing power network service method, platform, equipment and medium |
| CN116385857A (en) * | 2023-06-02 | 2023-07-04 | 山东协和学院 | Calculation power distribution method based on AI intelligent scheduling |
| CN117611096A (en) * | 2023-12-06 | 2024-02-27 | 广州市烨兴融集团有限公司 | Office data management method and system based on edge calculation |
-
2022
- 2022-06-15 CN CN202210676053.XA patent/CN115134371A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115883660A (en) * | 2022-11-21 | 2023-03-31 | 中国联合网络通信集团有限公司 | Industrial production computing power network service method, platform, equipment and medium |
| CN115550370A (en) * | 2022-12-01 | 2022-12-30 | 浩鲸云计算科技股份有限公司 | Computing power resource optimal scheduling allocation method based on multi-factor strategy |
| CN116385857A (en) * | 2023-06-02 | 2023-07-04 | 山东协和学院 | Calculation power distribution method based on AI intelligent scheduling |
| CN116385857B (en) * | 2023-06-02 | 2023-08-18 | 山东协和学院 | Calculation power distribution method based on AI intelligent scheduling |
| CN117611096A (en) * | 2023-12-06 | 2024-02-27 | 广州市烨兴融集团有限公司 | Office data management method and system based on edge calculation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115134371A (en) | Scheduling method, system, equipment and medium containing edge network computing resources | |
| Singh et al. | Scheduling real-time security aware tasks in fog networks | |
| US9250886B2 (en) | Optimizing provisioning workflows in cloud computing | |
| CN110661842B (en) | Resource scheduling management method, electronic equipment and storage medium | |
| WO2013107012A1 (en) | Task processing system and task processing method for distributed computation | |
| CN103338228A (en) | Cloud calculating load balancing scheduling algorithm based on double-weighted least-connection algorithm | |
| CN110347515B (en) | Resource optimization allocation method suitable for edge computing environment | |
| Chatterjee et al. | Design and implementation of an improved datacenter broker policy to improve the QoS of a cloud | |
| CN114625500A (en) | Method and application for scheduling micro-service application based on topology perception in cloud environment | |
| CN114706689B (en) | Multi-core processor task scheduling method and system based on subtask characteristics | |
| CN112988344A (en) | Distributed batch task scheduling method, device, equipment and storage medium | |
| Mylavarapu et al. | An optimized capacity planning approach for virtual infrastructure exhibiting stochastic workload | |
| CN110167031B (en) | Resource allocation method, equipment and storage medium for centralized base station | |
| Akintoye et al. | Optimization of virtual resources allocation in cloud computing environment | |
| Görlach et al. | Dynamic service provisioning for the cloud | |
| CN109062683A (en) | The method, apparatus and computer readable storage medium of host resource distribution | |
| Gowri et al. | An agent based resource provision for iot through machine learning in fog computing | |
| Abase et al. | Locality sim: cloud simulator with data locality | |
| CN115658263A (en) | Task scheduling method and system for cloud computing platform | |
| Cao et al. | Online cost-rejection rate scheduling for resource requests in hybrid clouds | |
| Djebbar et al. | Task scheduling strategy based on data replication in scientific cloud workflows | |
| Rodrigues et al. | Time-constrained and network-aware containers scheduling in GPU era | |
| Fernández-Cerero et al. | Quality of cloud services determined by the dynamic management of scheduling models for complex heterogeneous workloads | |
| Girase et al. | Dynamic resource provisioning in Cloud Computing environment using priority based virtual machine's | |
| Merkouche et al. | TERA-Scheduler for a Dependency-based Orchestration of Microservices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |