CN114490000A - Task processing method, device, equipment and storage medium - Google Patents
Task processing method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN114490000A CN114490000A CN202210145211.9A CN202210145211A CN114490000A CN 114490000 A CN114490000 A CN 114490000A CN 202210145211 A CN202210145211 A CN 202210145211A CN 114490000 A CN114490000 A CN 114490000A
- Authority
- CN
- China
- Prior art keywords
- task
- target
- controller
- task processing
- physical cluster
- 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
- 238000003672 processing method Methods 0.000 title claims abstract description 46
- 238000012545 processing Methods 0.000 claims abstract description 215
- 238000000034 method Methods 0.000 claims description 56
- 230000008569 process Effects 0.000 claims description 27
- 238000012544 monitoring process Methods 0.000 claims description 22
- 230000008859 change Effects 0.000 claims description 16
- 238000012508 change request Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 9
- 238000013473 artificial intelligence Methods 0.000 abstract description 72
- 238000005516 engineering process Methods 0.000 abstract description 16
- 238000007726 management method Methods 0.000 description 24
- 238000004364 calculation method Methods 0.000 description 12
- 238000002955 isolation Methods 0.000 description 12
- 238000012217 deletion Methods 0.000 description 11
- 230000037430 deletion Effects 0.000 description 11
- 238000012795 verification Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 8
- 230000003993 interaction Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000003924 mental process Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3006—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3051—Monitoring arrangements for monitoring the configuration of the computing system or of the computing system component, e.g. monitoring the presence of processing resources, peripherals, I/O links, software programs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The disclosure provides a task processing method, a task processing device and a task processing storage medium, and relates to the technical field of computers, in particular to the technical field of artificial intelligence such as distributed computing and big data. The specific implementation scheme is as follows: acquiring a task processing request sent by a user side; selecting a target task controller of the user side from candidate task controllers according to the control identifier in the task processing request; and sending the task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call resources in a target physical cluster to execute a target task in the task processing request. According to the technology disclosed by the invention, the task in the artificial intelligence platform can be reasonably executed, and a new solution is provided for processing the task in the artificial intelligence platform.
Description
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to the field of distributed computing and artificial intelligence technologies such as big data, and relates to a task processing method, apparatus, device, and storage medium.
Background
With the technical development of artificial intelligence technology, an artificial intelligence platform gradually becomes a runtime base for applications in the fields of images, voice, text and the like, and the artificial intelligence platform needs to bear various tasks such as data processing, model training, model prediction and the like from different users. Under the condition of more and more tasks, how to reasonably execute the tasks in the artificial intelligence platform is very important.
Disclosure of Invention
The disclosure provides a task processing method, a task processing device, a task processing equipment and a storage medium.
According to a first aspect of the present disclosure, there is provided a task processing method, including:
acquiring a task processing request sent by a user side;
selecting a target task controller of the user side from candidate task controllers according to the control identifier in the task processing request;
and sending the task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call resources in a target physical cluster to execute a target task in the task processing request.
According to a second aspect of the present disclosure, there is provided a task processing method, the method including:
receiving a task processing request sent by a unified controller;
and calling resources in the target physical cluster to execute the target task in the task processing request.
According to a third aspect of the present disclosure, there is provided a task processing system, the system comprising: a unified controller and a candidate task controller, wherein,
the unified controller is configured to execute the task processing method according to the first aspect;
the candidate task controller is configured to execute the task processing method according to the second aspect.
According to a fourth aspect of the present disclosure, there is provided an electronic apparatus including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of task processing according to any one of the embodiments of the present disclosure.
According to a fifth aspect of the present disclosure, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a task processing method according to any one of the embodiments of the present disclosure.
According to the technology disclosed by the invention, the task in the artificial intelligence platform can be reasonably executed, and a new solution is provided for processing the task in the artificial intelligence platform.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.
Drawings
The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:
FIG. 1 is a flow chart of a task processing method provided according to an embodiment of the present disclosure;
FIG. 2 is a flow chart of another task processing method provided according to an embodiment of the present disclosure;
FIG. 3 is a flow chart of yet another task processing method provided in accordance with an embodiment of the present disclosure;
FIG. 4 is a flowchart of another task processing method provided in accordance with an embodiment of the present disclosure;
FIG. 5 is a flow chart of yet another task processing method provided in accordance with an embodiment of the present disclosure;
FIG. 6 is a flowchart of yet another task processing method provided in accordance with an embodiment of the present disclosure;
FIG. 7 is a diagram of a task processing system architecture provided in accordance with an embodiment of the present disclosure;
FIG. 8 is a schematic structural diagram of a task processing device provided in accordance with an embodiment of the present disclosure;
FIG. 9 is a schematic structural diagram of another task processing device provided in accordance with an embodiment of the present disclosure;
fig. 10 is a block diagram of an electronic device for implementing a task processing method of an embodiment of the present disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 is a flowchart of a task processing method, which is applicable to a case how a task is processed according to an embodiment of the present disclosure. Optionally, the whole set of task processing method may be executed by a unified controller, a plurality of task controllers, a physical cluster, and the like in the artificial intelligence platform in a matching manner. Wherein, the unified controller can be used for interfacing with a user, such as receiving a task processing request of the user; the unified controller may also be used to interface task controllers, such as may forward a task processing request of a user to a corresponding task controller, and the like. The task controller can be upwards butted with the unified controller and downwards butted with the physical cluster; optionally, in this embodiment, a user (also referred to as a tenant) renting a resource in the artificial intelligence platform has its own task controller, that is, task controllers of different users are different; further, each task controller is dedicated to process task related matters of the user corresponding to the task controller, such as task processing requests and the like. The physical cluster is computing equipment at the bottom layer of the artificial intelligence platform; optionally, in this embodiment, the computing devices in one region are divided into one physical cluster, for example, one computer room may be taken as one physical cluster.
The task processing method provided by the embodiment can be applied to a unified controller in an artificial intelligence platform, and the method can be executed by a task processing device, which can be implemented in a software and/or hardware manner and can be integrated in an electronic device, such as a unified controller. As shown in fig. 1, the task processing method of the present embodiment may include:
s101, acquiring a task processing request sent by a user side.
In this embodiment, the user side may be an intelligent device owned by a user who rents resources in the artificial intelligence platform. Optionally, the user side may be configured with a task interaction tool of the artificial intelligence platform, where the task interaction tool is a front end of the platform facing the user, and is a bridge for interaction between the user side and the artificial intelligence platform. Further, the task interaction tool can be presented in the form of a standalone APP, or can also be presented in the form of a browser interface, or can also be presented in the form of an applet, and the like.
Specifically, when a user has a task processing requirement, the user can fill in task related information on a task creating interface provided by a task interaction tool on the user side, click and submit the task related information to trigger generation of a task processing request, and send the task processing request to a unified controller of the artificial intelligence platform; and then the unified controller can obtain the task processing request sent by the user side.
Optionally, the task processing request may include a target task to be processed; the task type of the target task may be, for example, an image recognition task, a data calculation task, a project planning task, a model training task, or the like, and the task type of the target task is not limited in the embodiments of the present disclosure.
Further, the task processing request may further include a user identifier and task resource requirement information. The user identification can be used for representing the user identity, can be a user ID, or can also be an access token issued by an artificial intelligence platform to the user, and the like; the task resource demand information is resource information required for executing the target task.
And S102, selecting a target task controller of the user side from the candidate task controllers according to the control identifier in the task processing request.
The control identifier is used for identifying the task controller; in this embodiment, each task controller has a control identifier, and the control identifiers of different task controllers are different. Optionally, each user renting resources in the human intelligence platform may be assigned a task controller dedicated to performing the user's associated tasks. Further, after the task controller is allocated to the user, the user may be informed of the control identifier of the task controller to instruct the user to carry the control identifier when sending the relevant task request to the artificial intelligence platform.
Optionally, the task processing request may further include a control identifier.
Each task controller in the artificial intelligence platform in this embodiment can be used as a candidate task controller. The target task controller is the task controller allocated to the user associated with the user side.
Specifically, after the task processing request sent by the user side is obtained, the control identifier may be extracted from the task processing request, and the candidate task controller whose control identifier is equal to the extracted control identifier in the candidate task controllers may be used as the target task controller of the user side.
S103, sending a task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call the resources in the target physical cluster to execute the target task in the task processing request.
In this embodiment, the target physical cluster may be one or more of the physical clusters that the artificial intelligence platform can provide.
Optionally, after the target task controller is determined, the task processing request may be directly sent to the target task controller; or, the data format conversion may be performed on the task processing request, for example, unstructured data in the task processing request may be converted into structured data, and then the converted task processing request is sent to the target task controller; alternatively, the control identifier and the like in the task processing request may be removed, and the task processing request from which the control identifier is removed may be sent to the target task controller and the like.
For example, after receiving the task processing request, the target task controller may determine a target physical cluster and call a resource in the target physical cluster to execute a target task in the task processing request; further, the target task controller may invoke resources in the target physical cluster according to the task resource requirement information in the task processing request to execute the target task.
According to the technical scheme provided by the embodiment of the disclosure, the target task controller of the user side is selected from the candidate task controllers according to the control identification in the task processing request sent by the user side, the task processing request is sent to the target task controller, and the target task controller calls the resources in the target physical cluster to execute the target task. According to the technical scheme, by introducing the task controllers and distributing different task controllers for different users to schedule the tasks of the resource execution users in the physical cluster, compared with the situation that the tasks of all the users are executed by uniformly scheduling the resources in the physical cluster by only one controller, the problem that one user monopolizes the resources of a control layer can be avoided under the condition that the number of the users is large and the number of the tasks of all the users is large, isolation of the users on the control layer is achieved, namely different task controllers are adopted to process task processing requests of different users on the control layer, and balanced and efficient processing of the tasks of all the users in the artificial intelligence platform is guaranteed.
Fig. 2 is a flowchart of another task processing method provided according to an embodiment of the present disclosure, and this embodiment adds a permission verification process to the above embodiment. As shown in fig. 2, the task processing method of this embodiment may include:
s201, acquiring a task processing request sent by a user side.
S202, determining whether the user side has the resource access authority or not according to the control identification in the task processing request. If yes, go to S203; if not, go to S205.
In order to ensure the security of resources in the artificial intelligence platform, the security of tasks and the like, after the task processing request sent by the user side is obtained, whether the user associated with the user side has the right to access the resources in the artificial intelligence platform or not can be determined.
In one implementation, the control identifier may be extracted from the task processing request, and it is determined whether the extracted control identifier is a control identifier of any task controller in the artificial intelligence platform; if not, it is indicated that the user associated with the user side does not rent resources in the artificial intelligence platform, that is, it is determined that the user associated with the user side does not have resource access authority. Otherwise, determining that the user associated with the user side has the resource access right.
In another possible implementation manner, whether the user side has the resource access right may also be determined according to the user identifier and the control identifier in the task processing request. For example, the association relationship between the user identifier and the control identifier stored in advance may be searched to determine whether an association relationship exists between the user identifier and the control identifier in the task processing request; and if the association relationship does not exist, determining that the user associated with the user side does not have the resource access authority. Otherwise, determining that the user associated with the user side has the resource access right. It can be understood that, in the embodiment, whether the user side has the resource access right is determined based on the incidence relation between the user identifier and the control identifier, and the security of resources and tasks in the artificial intelligence platform is greatly improved.
For example, in order to reduce the burden of the unified controller, the unified controller may perform permission verification by means of a tenant management module in the artificial intelligence platform. The tenant management module can be used for managing users of resources in the tenant artificial intelligent platform, for example, the tenant management module can be used for checking resource access authority of the users. Optionally, the unified controller and the tenant management module may be configured in the same server, or may be configured in different servers, and the like, which is not limited in this embodiment.
For example, the control identifier and the user identifier may be extracted from the task processing request, and the extracted control identifier and the user identifier may be sent to the tenant management module; the tenant management module carries out authority verification according to the control identification and the user identification based on the authority verification function, and feeds back the verification result to the unified controller; and then the unified controller can determine whether the user side has the resource access authority according to the verification result fed back by the tenant management module.
Further, in the case that it is determined that the user associated with the user side has the resource access right, S203 may be executed; in the case that it is determined that the user associated with the user side does not have the resource access right, the notification that the task cannot be executed may be directly fed back to the user side, that is, S205.
And S203, selecting a target task controller of the user side from the candidate task controllers according to the control identifier in the task processing request.
And S204, sending a task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call the resources in the target physical cluster to execute the target task in the task processing request.
And S205, feeding back the notice that the task cannot be executed to the user side.
Optionally, the task non-execution notification may include a reason why the task cannot be executed, so that the user can intuitively know the reason why the task cannot be executed.
According to the technical scheme provided by the embodiment of the disclosure, whether the user side has the resource access authority is determined according to the control identification in the task processing request sent by the user side, and under the condition that the user side is determined to have the resource access authority, the target task controller of the user side is selected from the candidate task controllers, the task processing request is sent to the target task controller, and the target task controller calls the resource in the target physical cluster to execute the target task. According to the scheme, the safety of resources, tasks and the like in the artificial intelligence platform is further ensured by adding the permission verification process to the user side.
Fig. 3 is a flowchart of another task processing method provided according to an embodiment of the present disclosure, and this embodiment further optimizes based on the foregoing embodiments, and introduces a corresponding relationship between a task controller and a physical cluster. As shown in fig. 3, the task processing method of the present embodiment may include:
s301, acquiring a task processing request sent by a user side.
S302, according to the control identification in the task processing request, selecting a target task controller of the user side from the candidate task controllers.
And S303, sending a task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to select a target physical cluster from the selectable physical clusters according to the resource condition of the selectable physical cluster corresponding to the target task controller and the task resource requirement information in the task processing request, and controlling the target physical cluster to process the target task.
The task resource requirement information is resource information required for executing the target task, and may include, but is not limited to, resources such as calculation and storage required for executing the target task.
In this embodiment, for each user, after the task controller is allocated to the user, the allocated task controller may be initialized. For example, a virtual node may be created in the task controller based on the usage resource requirement information for the user. The resource utilization requirement information is the resource information of the artificial intelligence platform which needs to be rented by the user, and may include, but is not limited to, the number of the computing nodes, and information of a CPU, a memory, a single core, a double core, and the like of each computing node.
Each virtual node corresponds to a compute node that uses the resource demand information. Furthermore, the virtual nodes themselves do not have actual computing and storage resources, and the resources of the virtual nodes can be bound to the physical cluster at the bottom of the artificial intelligence platform. In this embodiment, for each virtual node, a physical cluster may be bound to the virtual node according to the relevant information of the virtual node and the resource status of each physical cluster in the artificial intelligence platform.
Optionally, different virtual nodes under the same task controller may be bound to the same physical cluster; further, virtual nodes under different task controllers may also be bound to the same physical cluster. In this embodiment, the physical clusters in the artificial intelligence platform may be divided into two types, i.e., a common physical cluster and an exclusive physical cluster. The common physical cluster is a physical cluster allowing different users to share cluster resources; an exclusive physical cluster is a physical cluster in which the cluster resource is used by only one user. Further, for a common physical cluster, virtual nodes of different task controllers can be bound; only virtual nodes of the same task controller can be bound for an exclusive physical cluster.
Further, for each task controller, the physical cluster bound by the virtual node under the task controller is the physical cluster bound by the task controller. In this embodiment, the selectable physical cluster corresponding to the target task controller is a physical cluster bound to the virtual node under the target task controller. For each optional physical cluster, the resource status of the optional physical cluster may include, but is not limited to, information about currently remaining available resources of the optional physical cluster, and the like.
Specifically, after the target task controller is determined, a task processing request may be sent to the target task controller; and the target task controller selects a target physical cluster from the selectable physical clusters according to the resource status of the selectable physical clusters corresponding to the target task controller and the task resource demand information in the task processing request, and controls the target physical cluster to process the target task.
For example, initializing the task controller may further include creating an AI task controller module (or may be referred to as an AI load object controller) in the task controller, for interfacing with the unified controller, and parsing the task processing request. Further, initializing the task controller may further include creating a tenant scheduler module in the task controller, which is used to interface the physical cluster, for example, the resource state of the physical cluster bound to the virtual node under the task controller may be synchronized in real time, or the resource states of all the physical clusters under the artificial intelligence platform may be synchronized in real time.
Further, the unified controller sends a task processing request to the target task controller; an AI task controller module in the target task controller receives a task processing request sent by the unified controller, analyzes task resource demand information in the task processing request to determine resources such as calculation, storage and the like required by executing the target task, and informs a tenant scheduler module in the target task controller of the determined resources such as calculation, storage and the like required; and a tenant scheduler module in the target task controller selects a target physical cluster from the selectable physical clusters according to the required resources such as calculation, storage and the like and the resource condition of the selectable physical cluster corresponding to the target task controller, and controls the target physical cluster to process the target task. For example, the target physical cluster may be notified of resources such as computation and storage required to execute the target task, the target physical cluster may create a resource object, and the target task may be executed based on the created resource object.
According to the technical scheme provided by the embodiment of the disclosure, a target task controller of a user side is selected from candidate task controllers according to a control identifier in a task processing request sent by the user side, and the task processing request is sent to the target task controller; and the target task controller selects the target physical cluster according to the resource condition of the optional physical cluster corresponding to the target task controller and the task resource demand information in the task processing request, and calls the resources in the target physical cluster to execute the target task. According to the scheme, the task controller is introduced, so that the isolation of the user at the control layer is realized; meanwhile, the corresponding relation between the task controller and the physical cluster is introduced, so that the isolation of the users on the bottom resource level is realized to a certain extent, and the mutual influence of tasks among different users is greatly reduced.
For example, on the basis of the above embodiments, as an implementable manner of the embodiments of the present disclosure, the method may further include: and selecting an optional physical cluster corresponding to the target task controller from the public physical cluster and the exclusive physical cluster according to the use resource demand information of the user side. In this embodiment, the number of the common physical clusters and the number of the exclusive physical clusters may be multiple.
Specifically, after the user associated with the user side successfully registers on the artificial intelligence platform, a target task controller can be allocated to the user associated with the user side according to the resource use requirement information provided by the user side; in the process of initializing the target task controller, the resource use property can be determined according to the resource use requirement information; selecting a candidate physical cluster from the public physical cluster and the exclusive physical cluster according to the resource use property; and binding a physical cluster for each virtual node in the target task controller according to the used resource demand information and the resource condition of the candidate physical cluster, namely determining the optional physical cluster corresponding to the target task controller.
Wherein, the resource usage property can be unique or sharable; optionally, if the resource usage property is unique, each exclusive physical cluster in the artificial intelligence platform may be used as a candidate physical cluster; if the resource usage property is shareable, each common physical cluster in the artificial intelligence platform can be used as a candidate physical cluster.
Optionally, in order to reduce the burden of the unified controller, the unified controller may further allocate an optional physical cluster to the target task controller by using a tenant management module in the artificial intelligence platform. For example, the unified controller may send the usage resource demand information of the user side to the tenant management module, and the tenant management module selects, according to the usage resource demand information of the user side, an optional physical cluster corresponding to the target task controller from the common physical cluster and the exclusive physical cluster.
It should be noted that, in this embodiment, for a common physical cluster, different task controllers can be scheduled at the same time, so that resources can be fully utilized; furthermore, each task in a single common physical cluster can be isolated in a resource space based on the existing related technology (such as a name space), that is, the common physical cluster is introduced in the embodiment, so that not only can resources be fully utilized, but also the isolation of a user on a bottom resource level can be realized to a certain extent.
It can be understood that, in the embodiment, two physical clusters with different properties, namely a common physical cluster and an exclusive physical cluster, are introduced into the artificial intelligence platform, so that the requirements of different users can be met; furthermore, a public physical cluster is introduced, so that not only can the isolation of users on the bottom resource layer be realized to a certain extent, but also the bottom resource can be fully utilized, and the resource utilization rate of the whole platform can be further improved.
For example, for each task controller, the physical cluster to which the task controller is bound may also change according to the dynamic change of the user's usage resource demand information. Furthermore, in an implementation manner, the task processing method provided by the present disclosure may further include: and responding to a used resource change request sent by the user side, and updating the optional physical cluster corresponding to the target task controller according to the resource change information in the used resource change request.
The resource use change request is a request sent to the unified controller when a user has a resource requirement in the artificial intelligence platform rented by the change. Optionally, the request for changing the used resource may be a request for deleting the used resource, specifically, a request sent by the user when the user does not want to rent the resource in the artificial intelligence platform; the request for changing the use resource can also be a request for newly adding the use resource, and specifically can be a request sent when the resource in the artificial intelligence platform rented by the user cannot meet the self requirement; the request for changing the used resources can also be a request for reducing the used resources, and specifically can be a request sent when the user cannot fully utilize the resources in the rented artificial intelligence platform; the request for changing the use resource may also be a request for replacing the use resource, specifically, a request sent when a user wants to replace a certain virtual node, and the like.
Optionally, the request for changing the resource may include a user identifier and resource change information. Further, the resource change request includes different resource change information. For example, if the request for changing the used resource is a request for adding a new used resource, the resource change information may include the number of computing nodes to be added, information related to each computing node, and the like; for another example, if the request for changing the used resource is a request for reducing the used resource, the resource change information may include a computing node that needs to be deleted, and the like; for another example, if the request for changing the used resource is a request for deleting the used resource, the resource change information may include a control identifier and the like; for another example, if the request for changing the used resource is a request for replacing the used resource, the resource change information may include identification information of the original computing node, information related to the new computing node, and the like.
Further, for each task controller, if the demand of the user for using the resource changes, the number of the computing nodes changes, and further the number of the virtual nodes changes accordingly, and the physical cluster bound by the virtual nodes also changes inevitably.
Specifically, after receiving a resource usage change request sent by a user, resource change information may be extracted from the resource usage change request, and a virtual node in the target task controller is updated according to the resource change information, and an optional physical cluster corresponding to the target task controller is updated according to the updated virtual node.
For example, if the request for changing the used resource is a request for newly adding the used resource, a corresponding virtual node may be newly created in the target task controller according to the number of newly added computing nodes in the resource change information, the related information of each computing node, and the like, and a physical cluster may be bound to each newly created virtual node according to the related information of each newly created virtual node and the resource status of all physical clusters, that is, an optional physical cluster corresponding to the target task controller is updated.
Optionally, in order to reduce the burden of the unified controller, the unified controller may further update the optional physical cluster corresponding to the target task controller by using a tenant management module in the artificial intelligence platform. For example, the unified controller may send a request for changing the resource used by the user side to the tenant management module, and the tenant management module updates the optional physical cluster corresponding to the target task controller according to the resource change information in the request for changing the resource used.
It can be understood that, in this embodiment, the physical cluster bound to the task controller can be dynamically updated, so that the flexibility of the scheme is increased, and the user experience is further improved.
Fig. 4 is a flowchart of another task processing method provided according to an embodiment of the present disclosure, and this embodiment adds a task state determination process to the above embodiment. As shown in fig. 4, the task processing method of this embodiment may include:
s401, a task processing request sent by a user side is obtained.
S402, selecting a target task controller of the user side from the candidate task controllers according to the control identification in the task processing request.
And S403, sending a task processing request to the target task controller, wherein the task processing request is used for instructing the target task controller to call the resource in the target physical cluster to execute the target task in the task processing request.
S404, responding to a task state query request sent by a user side, and determining the current task state of the target task according to the monitoring condition of the target task controller on the target physical cluster.
The task state query request is a request sent by a user when the user wants to know the execution condition of the target task. Optionally, the task state query request may include a control identifier, a user identifier, a task identifier of the target task, and the like. And the task identifier is used for uniquely positioning the target task.
Optionally, in this embodiment, in the process that the target physical cluster executes the target task, the target task controller may monitor the target physical cluster in real time. Further, the target virtual node in the target task controller monitors the target physical cluster, and specifically, may monitor resources required for executing the target task in the target physical cluster. The target virtual node is any virtual node in the target task controller corresponding to the target physical cluster.
Specifically, after receiving a task state query request sent by a user side, a task identifier, a control identifier and the like can be extracted from the task state query request; positioning a target task controller according to the control identifier; acquiring the monitoring condition of the target physical cluster from the target task controller according to the task identifier; and then, analyzing the acquired monitoring condition to determine the current task state of the target task. The current task state may be in execution or completed, etc.
Furthermore, in order to reduce the burden of the unified controller, the unified controller can also determine the current task state of the target task by means of a tenant management module in the artificial intelligence platform. For example, the unified controller may send a task state query request of a user side to the tenant management module, and the tenant management module determines and feeds back the current task state of the target task according to the monitoring condition of the target task controller on the target physical cluster; and the unified controller can then obtain the current task state from the tenant management module.
S405, feeding back the current task state to the user side.
Specifically, after the current task state of the target task is determined, the current task state can be directly fed back to the user side; or the current task state may also be processed, for example, the current task state may be processed according to a set output format, and the processed current task state is fed back to the user side.
According to the technical scheme provided by the embodiment of the disclosure, a target task controller of a user side is selected from candidate task controllers according to a control identifier in a task processing request sent by the user side, the task processing request is sent to the target task controller, and the target task controller calls resources in a target physical cluster to execute a target task; and then, if a task state query request sent by the user side is received, determining the current task state of the target task according to the monitoring condition of the target task controller on the target physical cluster, and feeding back the current task state to the user side. According to the scheme, the monitoring function of the task controller on the physical cluster is introduced, so that a user can accurately know the task state in real time.
Further, in the process of executing the target task by the target physical cluster, if the user wants to abandon the execution of the target task, a task deletion request can be sent to the unified controller through the user side. For example, the task processing method provided by the present disclosure may further include: and forwarding the task deleting request sent by the user side to the target task controller so that the target task controller releases the resources in the target physical cluster.
The task deleting request may also include a control identifier, a user identifier, a task identifier of the target task, and the like.
Specifically, after receiving a task deletion request sent by a user side, a target task controller can be determined according to a control identifier in the task deletion request; forwarding the task deletion request to a target task controller; and releasing the resources in the target physical cluster occupied by the target task controller.
It can be understood that, by releasing the resources occupied by the task in time when the user wants to abandon the executed task, on one hand, the resources can be prevented from being wasted; on the other hand, the system resources are enriched in time, so that other tasks can be performed in sequence.
Fig. 5 is a flowchart of another task processing method provided according to an embodiment of the present disclosure, which is applicable to a case how a task is processed. Optionally, the whole set of task processing method may be executed by a unified controller, a plurality of task controllers, a physical cluster, and the like in the artificial intelligence platform in a matching manner. The task processing method provided by the embodiment can be applied to a target task controller in an artificial intelligence platform; the target task controller is one of the task controllers, and specifically is a task controller allocated to a user associated with a user side that sends a task processing request.
The method may be performed by a task processing device, which may be implemented in software and/or hardware, and may be integrated in an electronic device, such as a target task controller. As shown in fig. 5, the task processing method of this embodiment may include:
s501, receiving a task processing request sent by the unified controller.
S502, calling the resources in the target physical cluster to execute the target task in the task processing request.
Specifically, after acquiring a task processing request sent by a user side, the unified controller selects a target task controller of the user side from the candidate task controllers according to a control identifier in the task processing request, and sends the task processing request to the target task controller.
Furthermore, the target task controller receives a task processing request sent by the unified controller, determines a target physical cluster, and calls resources in the target physical cluster to execute a target task in the task processing request; further, the target task controller may call resources in the target physical cluster according to the task resource requirement information in the task processing request to execute the target task.
According to the technical scheme provided by the embodiment of the disclosure, the target task controller receives the task processing request sent by the unified controller, and calls the resources in the target physical cluster to execute the target task. According to the technical scheme, by introducing the task controller, the phenomenon that one user monopolizes resources of the control layer can be avoided under the condition that the number of users is large and the number of tasks of each user is large, isolation of the users on the control layer is achieved, namely different task controllers are adopted to process task processing requests of different users on the control layer, and balance and efficient processing of the tasks of each user in the artificial intelligence platform are guaranteed.
Fig. 6 is a flowchart of another task processing method provided according to an embodiment of the present disclosure, and this embodiment further explains in detail how to "call a resource in a target physical cluster to execute a target task in a task processing request" on the basis of the above embodiment. As shown in fig. 6, the task processing method of the present embodiment may include:
s601, receiving a task processing request sent by the unified controller.
S602, selecting a target physical cluster from the selectable physical clusters according to the resource status of the selectable physical clusters corresponding to the target task controller and the task resource demand information in the task processing request.
Optionally, the target task controller may include an AI task controller module, a tenant scheduler module, and one or more virtual nodes; each virtual node is bound with a physical cluster; in this embodiment, the selectable physical cluster corresponding to the target task controller is a physical cluster bound to the virtual node under the target task controller.
Specifically, the unified controller sends a task processing request to the target task controller; an AI task controller module in the target task controller receives a task processing request sent by the unified controller, analyzes task resource demand information in the task processing request to obtain resources such as calculation and storage required by execution of the target task, and converts the resources such as calculation and storage required by execution of the target task obtained by analysis according to a resource object format supported by the optional physical cluster to obtain a resource object set. The resource object set is a set of processes and configurations, such as computing power, CPU, memory, etc., required for executing the target task. For example, where the optional physical cluster is the K8S cluster, the set of resource objects may include, but is not limited to, containers Pod, storage Volumes Volumes, and configuration ConfigMaps, among others.
And then, initiating an object creation requirement by an AI task controller module in the target task controller to trigger a tenant scheduler module in the target task controller to select a target physical cluster from the selectable physical clusters according to the resource condition of the selectable physical cluster corresponding to the target task controller.
For example, in an implementation manner, after receiving a task processing request sent by a unified controller, if it is identified that resources of an optional physical cluster cannot meet the execution of a target task according to the resource status of the optional physical cluster corresponding to the target task controller, the target task controller rebunds the physical cluster for a virtual node of the target task controller according to the resource status of all physical clusters under an artificial intelligence platform; and then the target task can be scheduled according to the physical cluster after the rebinding. Namely, the target physical cluster is selected according to the task resource demand information in the task processing request and the resource condition of the re-bound physical cluster.
S603, controlling the target physical cluster to execute the target task in the task processing request.
Specifically, a tenant scheduler module in the target task controller controls the target physical cluster to process the target task. For example, the target physical cluster may be notified of resources such as computation and storage required to execute the target task, the target physical cluster may create a resource object, and the target task may be executed based on the created resource object.
Optionally, as an optional implementation manner of the embodiment of the present disclosure, controlling the target physical cluster to execute the target task in the task processing request may be: sending an object creation request including a set of resource objects to the target physical cluster to cause the target physical cluster to create the set of resource objects to perform the target task; wherein the resource object set is determined according to the task resource demand information.
Specifically, after determining the target physical cluster, the tenant scheduler module in the target task controller may generate an object creation request including a resource object set and a task identifier of the target task based on an object creation requirement initiated by an AI task controller module in the target task controller; and sending an object creation request to the target physical cluster, creating the resource object in the resource object set by the target physical cluster, and executing the target task based on the created resource object set.
According to the technical scheme provided by the embodiment of the disclosure, the target task controller receives the task processing request sent by the unified controller, and the target physical cluster is selected and the resources in the target physical cluster are called to execute the target task according to the resource condition of the optional physical cluster corresponding to the target task controller and the task resource demand information in the task processing request. According to the scheme, the task controller is introduced, so that the isolation of the user at the control layer is realized; meanwhile, the corresponding relation between the task controller and the physical cluster is introduced, so that the isolation of the users on the bottom resource level is realized to a certain extent, and the mutual influence of tasks among different users is greatly reduced.
Optionally, after determining that the target physical cluster has completed the creation of the resource object set, the tenant scheduler module in the target task controller may determine a target virtual node from the virtual nodes owned by the target task controller according to the target physical cluster based on the binding relationship between the virtual node and the physical cluster, and control the target virtual node to record the corresponding relationship between the target physical cluster, the resource object set, and the target task. For example, a correspondence between an identification of the target physical cluster, an identification of each resource object in the set of resource objects, and a task identification of the target task may be recorded.
Further, in the process that the target physical cluster executes the target task, the target task controller may monitor the target physical cluster in real time. Further, the target virtual node in the target task controller monitors the target physical cluster, and specifically, may monitor a resource object executing the target task in the target physical cluster.
For example, on the basis of the above embodiments, as an implementable manner of the embodiments of the present disclosure, the method may further include: and sending the monitoring condition of the target physical cluster to the unified controller so that the unified controller determines the current task state of the target task according to the monitoring condition.
Specifically, after receiving a task state query request sent by a user side, the unified controller can extract a task identifier, a control identifier and the like from the task state query request; positioning a target task controller according to the control identifier, and sending a monitoring condition acquisition request comprising a task identifier to the target task controller; the target task controller receives the monitoring condition acquisition request sent by the unified controller, acquires the monitoring condition of the target virtual node in the target task controller to the target physical cluster according to the task identifier, sends the acquired monitoring condition to the unified controller, and analyzes the acquired monitoring condition by the unified controller to determine the current task state of the target task.
It can be understood that, in the embodiment, by introducing the monitoring function of the task controller on the physical cluster, the user can accurately know the task state in real time.
Illustratively, after the target task is issued to the target physical cluster, if the tenant scheduler module in the target task controller recognizes that the target physical cluster is removed according to the resource status of the target physical cluster and/or the monitoring condition of the target physical cluster, the target task is rescheduled, that is, the target physical cluster is reselected, and the resource in the target physical cluster is called to execute the target task in the task processing request. It will be appreciated that the benefit of this embodiment of this arrangement is that it ensures that the user's tasks can be performed.
Further, in the process of executing the target task by the target physical cluster, if the user wants to abandon the execution of the target task, a task deletion request can be sent to the unified controller through the user side. For example, the task processing method provided by the present disclosure may further include: receiving a task deleting request of a user side forwarded by a unified controller; and releasing the resources in the target physical cluster. The task deletion request may include a control identifier, a user identifier, a task identifier of the target task, and the like.
Specifically, after receiving a task deletion request sent by a user side, the unified controller can determine a target task controller according to a control identifier in the task deletion request; forwarding the task deletion request to a target task controller; and the target task controller can receive a task deleting request of the user side from the unified controller and release resources in the target physical cluster occupied by the target task.
It can be understood that, by releasing the resources occupied by the task in time when the user wants to abandon the executed task, on one hand, the resources can be prevented from being wasted; on the other hand, the system resources are enriched in time, so that other tasks can be performed in sequence.
Illustratively, the present embodiment provides a preferred example based on the above-described embodiments. Before the preferred embodiment is described, the whole task processing system will be described with reference to the task processing system architecture diagram shown in fig. 7.
As shown in fig. 7, the task processing system may include one unified controller, which may be used to interface users, and a plurality of task controllers, such as task controller a, task controller b, and task controller c, in the task processing system.
Optionally, each task controller is dedicated to process task related matters of the user corresponding to the task controller, such as a task processing request. For example, each task controller may include an AI task controller module, a tenant scheduler module, and one or more virtual nodes, etc. The AI task controller module can be used for interfacing the unified controller and analyzing the task processing request and the like. The tenant scheduler module can be used for butting the physical clusters, synchronizing the resource conditions of the physical clusters in real time and the like; virtual nodes can also be docked, for example, the control virtual node records the corresponding relationship between the physical cluster, the resource object set and the task, and the like.
Furthermore, each virtual node in each task controller can be bound with one physical cluster in the task processing system; in this embodiment, the task processing system includes two types of physical clusters, i.e., a common physical cluster and an exclusive physical cluster, and both the number of the common physical clusters and the number of the exclusive physical clusters may be multiple. Furthermore, for a common physical cluster, virtual nodes of different task controllers can be bound; only virtual nodes of the same task controller can be bound for an exclusive physical cluster.
It should be noted that, in fig. 7, each task controller is exemplified to have two virtual nodes, but it cannot be said that each task controller has only two virtual nodes, there may be more than two virtual nodes, or a certain task controller may also have one virtual node. Furthermore, two common physical clusters, physical cluster 1 and physical cluster 2, respectively, are shown in fig. 7; and two exclusive physical clusters are shown, physical cluster 3 and physical cluster 4 respectively.
Illustratively, the task processing system may further include a tenant management module and a cluster management module. The tenant management module can be used for managing users of resources in the tenant artificial intelligent platform, for example, the resource access authority of the users can be checked, and task controllers allocated to the users can be initialized; the cluster management module can be used for managing the physical cluster under the artificial intelligence platform, such as controlling initialization, registration and removal of the physical cluster.
Optionally, on the basis of the task processing system shown in fig. 7, the whole task processing process is specifically as follows:
when a user has a task processing requirement, the user can fill in task related information on a task creating interface provided by a task interaction tool on a user side, and click and submit the task related information to trigger generation of a task processing request, and the task processing request is sent to a unified controller of the artificial intelligence platform.
The method comprises the steps that a unified controller obtains a task processing request sent by a user side, extracts a control identifier and a user identifier from the task processing request, and sends the extracted control identifier and the user identifier to a tenant management module; the tenant management module carries out authority verification according to the control identification and the user identification based on the authority verification function, and feeds back the verification result to the unified controller; and the unified controller determines whether the user side has the resource access authority or not according to the verification result fed back by the tenant management module.
Under the condition that the user side is determined to have the resource access authority, the unified controller selects a target task controller of the user side, such as a task controller a, from the candidate task controllers according to the control identification in the task processing request; and converting unstructured data in the task processing request into structured data and the like, and then sending the converted task processing request to the task controller a.
An AI task controller module in a task controller a receives a task processing request sent by a unified controller, analyzes task resource demand information in the task processing request to determine resources such as calculation, storage and the like required by executing a target task, and informs a tenant scheduler module in the task controller a of the determined resources such as calculation, storage and the like; according to the required resources such as calculation and storage and the resource conditions of selectable physical clusters corresponding to virtual nodes in the task controller a (for example, a physical cluster 1 corresponding to a virtual node 1 and a physical cluster 2 corresponding to a virtual node 2), a tenant scheduler module in the task controller a selects a target physical cluster (for example, a physical cluster 1) from the selectable physical clusters and controls the physical cluster 1 to process a target task.
It should be noted that, in this embodiment, by introducing the task controller and allocating different task controllers to different users to schedule the resources in the physical cluster to execute the tasks of the users, compared with the case of uniformly scheduling the resources in the physical cluster to execute the tasks of all the users by using only one controller, a phenomenon that one user monopolizes resources of the control plane can be avoided under the condition that there are many users and there are many tasks of each user, isolation of the users at the control plane is achieved, that is, different task controllers are used to process task processing requests of different users at the control plane, and balanced and efficient processing of tasks of each user in the artificial intelligence platform is ensured; in addition, the corresponding relation between the task controller and the physical cluster is introduced, so that the isolation of the users on the bottom resource level is realized to a certain extent, and the mutual influence of tasks among different users is greatly reduced.
Fig. 8 is a schematic structural diagram of a task processing device according to an embodiment of the present disclosure. The embodiment of the disclosure is applicable to the situation of how to process the task. The device can be configured in a unified controller in an artificial intelligence platform, the device can be implemented by software and/or hardware, and the device can implement the task processing method described in the embodiment of the disclosure. As shown in fig. 8, the task processing device includes:
a task processing request obtaining module 801, configured to obtain a task processing request sent by a user side;
a controller selection module 802, configured to select a target task controller of a user side from the candidate task controllers according to the control identifier in the task processing request;
and a task processing request sending module 803, configured to send a task processing request to the target task controller, where the task processing request is used to instruct the target task controller to call a resource in the target physical cluster to execute a target task in the task processing request.
According to the technical scheme provided by the embodiment of the disclosure, the target task controller of the user side is selected from the candidate task controllers according to the control identification in the task processing request sent by the user side, the task processing request is sent to the target task controller, and the target task controller calls the resources in the target physical cluster to execute the target task. According to the technical scheme, by introducing the task controllers and distributing different task controllers for different users to schedule the tasks of the resource execution users in the physical cluster, compared with the situation that the tasks of all the users are executed by uniformly scheduling the resources in the physical cluster by only one controller, the problem that one user monopolizes the resources of a control layer can be avoided under the condition that the number of the users is large and the number of the tasks of all the users is large, isolation of the users on the control layer is achieved, namely different task controllers are adopted to process task processing requests of different users on the control layer, and balanced and efficient processing of the tasks of all the users in the artificial intelligence platform is guaranteed.
Illustratively, the controller selection module 802 is further configured to:
determining whether the user side has a resource access right or not according to the control identifier in the task processing request;
and if so, selecting a target task controller of the user side from the candidate task controllers according to the control identification in the task processing request.
Illustratively, the task processing request is further for:
and indicating the target task controller to select the target physical cluster from the selectable physical clusters according to the resource status of the selectable physical cluster corresponding to the target task controller and the task resource demand information in the task processing request, and controlling the target physical cluster to process the target task.
Illustratively, the apparatus may further include:
and the cluster selection module is used for selecting an optional physical cluster corresponding to the target task controller from the public physical cluster and the exclusive physical cluster according to the using resource demand information of the user side.
Illustratively, the apparatus may further include:
and the cluster updating module is used for responding to the resource using change request sent by the user side and updating the optional physical cluster corresponding to the target task controller according to the resource change information in the resource using change request.
Illustratively, the apparatus may further include:
the task state determining module is used for responding to a task state query request sent by a user side and determining the current task state of a target task according to the monitoring condition of a target task controller on a target physical cluster;
and the task state feedback module is used for feeding back the current task state to the user side.
Illustratively, the apparatus may further include:
and the task deletion request forwarding module is used for forwarding the task deletion request sent by the user side to the target task controller so that the target task controller releases the resources in the target physical cluster.
Fig. 9 is a schematic structural diagram of another task processing device provided according to an embodiment of the present disclosure. The embodiment of the disclosure is applicable to the situation of how to process the task. The device can be configured in a target task controller in an artificial intelligence platform, the device can be implemented by software and/or hardware, and the device can implement the task processing method described in the embodiment of the disclosure. As shown in fig. 9, the task processing device includes:
a task processing request receiving module 901, configured to receive a task processing request sent by a unified controller;
and the task processing module 902 is configured to invoke a resource in the target physical cluster to execute a target task in the task processing request.
According to the technical scheme provided by the embodiment of the disclosure, the target task controller receives the task processing request sent by the unified controller, and calls the resources in the target physical cluster to execute the target task. According to the technical scheme, by introducing the task controller, the phenomenon that one user monopolizes resources of the control layer can be avoided under the condition that the number of users is large and the number of tasks of each user is large, isolation of the users on the control layer is achieved, namely different task controllers are adopted to process task processing requests of different users on the control layer, and balance and efficient processing of the tasks of each user in the artificial intelligence platform are guaranteed.
Illustratively, the task processing module 902 includes:
the cluster selection unit is used for selecting a target physical cluster from the selectable physical clusters according to the resource status of the selectable physical clusters corresponding to the target task controller and the task resource demand information in the task processing request;
and the task processing unit is used for controlling the target physical cluster to execute the target task in the task processing request.
Illustratively, the task processing unit is further configured to:
sending an object creation request including a set of resource objects to the target physical cluster to cause the target physical cluster to create the set of resource objects to perform the target task; wherein the resource object set is determined according to the task resource demand information.
Illustratively, the apparatus may further include:
and the monitoring condition sending module is used for sending the monitoring condition of the target physical cluster to the unified controller so that the unified controller can determine the current task state of the target task according to the monitoring condition.
Illustratively, the apparatus may further include:
the task deleting request receiving module is used for receiving a task deleting request of the user side forwarded by the unified controller;
and the resource releasing module is used for releasing the resources in the target physical cluster.
In the technical scheme of the disclosure, the acquisition, storage, application and the like of the related task processing request, the user identifier, the related information of the task controller, the related information of the physical cluster and the like all accord with the regulations of related laws and regulations, and do not violate the good custom of the public order.
The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.
FIG. 10 illustrates a schematic block diagram of an example electronic device 1000 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 10, the electronic device 1000 includes a computing unit 1001 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)1002 or a computer program loaded from a storage unit 1008 into a Random Access Memory (RAM) 1003. In the RAM 1003, various programs and data necessary for the operation of the electronic apparatus 1000 can also be stored. The calculation unit 1001, the ROM1002, and the RAM 1003 are connected to each other by a bus 1004. An input/output (I/O) interface 1005 is also connected to bus 1004.
A number of components in the electronic device 1000 are connected to the I/O interface 1005, including: an input unit 1006 such as a keyboard, a mouse, and the like; an output unit 1007 such as various types of displays, speakers, and the like; a storage unit 1008 such as a magnetic disk, an optical disk, or the like; and a communication unit 1009 such as a network card, a modem, a wireless communication transceiver, or the like. The communication unit 1009 allows the electronic device 1000 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), Complex Programmable Logic Devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server with a combined blockchain.
Artificial intelligence is the subject of research that makes computers simulate some human mental processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), both at the hardware level and at the software level. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing, and the like; the artificial intelligence software technology mainly comprises a computer vision technology, a voice recognition technology, a natural language processing technology, a machine learning/deep learning technology, a big data processing technology, a knowledge map technology and the like.
Cloud computing (cloud computing) refers to a technology system that accesses a flexibly extensible shared physical or virtual resource pool through a network, where resources may include servers, operating systems, networks, software, applications, storage devices, and the like, and may be deployed and managed in a self-service manner as needed. Through the cloud computing technology, high-efficiency and strong data processing capacity can be provided for technical application and model training of artificial intelligence, block chains and the like.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.
The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.
Claims (18)
1. A method of task processing, comprising:
acquiring a task processing request sent by a user side;
selecting a target task controller of the user side from candidate task controllers according to the control identifier in the task processing request;
and sending the task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call resources in a target physical cluster to execute a target task in the task processing request.
2. The method of claim 1, wherein the selecting a target task controller of the user side from candidate task controllers according to the control identifier in the task processing request comprises:
determining whether the user side has a resource access right or not according to the control identification in the task processing request;
and if so, selecting a target task controller of the user side from candidate task controllers according to the control identification in the task processing request.
3. The method of claim 1, wherein the task processing request is further to:
and instructing the target task controller to select a target physical cluster from the selectable physical clusters according to the resource condition of the selectable physical cluster corresponding to the target task controller and the task resource demand information in the task processing request, and controlling the target physical cluster to process the target task.
4. The method of claim 3, further comprising:
and selecting an optional physical cluster corresponding to the target task controller from a public physical cluster and an exclusive physical cluster according to the use resource demand information of the user side.
5. The method of claim 3, further comprising:
and responding to a used resource change request sent by the user side, and updating the optional physical cluster corresponding to the target task controller according to resource change information in the used resource change request.
6. The method of claim 1, further comprising:
responding to a task state query request sent by the user side, and determining the current task state of the target task according to the monitoring condition of the target task controller on the target physical cluster;
and feeding back the current task state to the user side.
7. The method of claim 1, further comprising:
and forwarding the task deleting request sent by the user side to the target task controller so that the target task controller releases the resources in the target physical cluster.
8. A method of task processing, comprising:
receiving a task processing request sent by a unified controller;
and calling resources in the target physical cluster to execute the target task in the task processing request.
9. The method of claim 8, wherein the invoking a resource in a target physical cluster to perform a target task in the task processing request comprises:
selecting a target physical cluster from the selectable physical clusters according to the resource condition of the selectable physical cluster corresponding to the target task controller and the task resource demand information in the task processing request;
and controlling the target physical cluster to execute the target task in the task processing request.
10. The method of claim 9, wherein the controlling the target physical cluster to execute the target task in the task processing request comprises:
sending an object creation request comprising a set of resource objects to the target physical cluster to cause the target physical cluster to create the set of resource objects to perform the target task; and determining the resource object set according to the task resource demand information.
11. The method of claim 8, further comprising:
and sending the monitoring condition of the target physical cluster to the unified controller so that the unified controller determines the current task state of the target task according to the monitoring condition.
12. The method of claim 8, further comprising:
receiving a task deleting request of a user side forwarded by the unified controller;
and releasing the resources in the target physical cluster.
13. A task processing device comprising:
the task processing request acquisition module is used for acquiring a task processing request sent by a user side;
the controller selection module is used for selecting a target task controller of the user side from candidate task controllers according to the control identification in the task processing request;
and the task processing request sending module is used for sending the task processing request to the target task controller, wherein the task processing request is used for indicating the target task controller to call resources in a target physical cluster to execute a target task in the task processing request.
14. A task processing device comprising:
the task processing request receiving module is used for receiving a task processing request sent by the unified controller;
and the task processing module is used for calling resources in the target physical cluster to execute the target task in the task processing request.
15. A task processing system includes a unified controller and candidate task controllers, wherein,
the unified controller is used for executing the task processing method of any one of claims 1-7;
the candidate task controller for performing the task processing method of any one of claims 8-12.
16. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of task processing according to any one of claims 1 to 7 or the method of task processing according to any one of claims 8 to 12.
17. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the task processing method of any one of claims 1 to 7 or the task processing method of any one of claims 8 to 12.
18. A computer program product comprising a computer program which, when executed by a processor, implements a method of task processing according to any one of claims 1-7, or a method of task processing according to any one of claims 8-12.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210145211.9A CN114490000A (en) | 2022-02-17 | 2022-02-17 | Task processing method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210145211.9A CN114490000A (en) | 2022-02-17 | 2022-02-17 | Task processing method, device, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114490000A true CN114490000A (en) | 2022-05-13 |
Family
ID=81481439
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210145211.9A Pending CN114490000A (en) | 2022-02-17 | 2022-02-17 | Task processing method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114490000A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024045646A1 (en) * | 2022-09-01 | 2024-03-07 | 京东科技信息技术有限公司 | Method, apparatus and system for managing cluster access permission |
-
2022
- 2022-02-17 CN CN202210145211.9A patent/CN114490000A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024045646A1 (en) * | 2022-09-01 | 2024-03-07 | 京东科技信息技术有限公司 | Method, apparatus and system for managing cluster access permission |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110310034B (en) | Service arrangement and business flow processing method and device applied to SaaS | |
EP3407548B1 (en) | Chatbot system | |
CN110083455B (en) | Graph calculation processing method, graph calculation processing device, graph calculation processing medium and electronic equipment | |
CN111897638B (en) | Distributed task scheduling method and system | |
US20200110631A1 (en) | Virtual machine deployment method and omm virtual machine | |
US9246920B2 (en) | Cloud resource cloning based on collaborative content | |
US12014216B2 (en) | Method for platform-based scheduling of job flow | |
CN109634764A (en) | Work-flow control method, apparatus, equipment, storage medium and system | |
US9542161B2 (en) | Method and system for generating a source code for a computer program for execution and simulation of a process | |
CN114911598A (en) | Task scheduling method, device, equipment and storage medium | |
CN114490000A (en) | Task processing method, device, equipment and storage medium | |
CN114201294A (en) | Task processing method, device and system, electronic equipment and storage medium | |
CN113608751A (en) | Operation method, device and equipment of reasoning service platform and storage medium | |
CN115309558A (en) | Resource scheduling management system, method, computer equipment and storage medium | |
CN114298313A (en) | Artificial intelligence computer vision reasoning method | |
CN114036250A (en) | High-precision map task processing method and device, electronic equipment and medium | |
CN114564249A (en) | Recommendation scheduling engine, recommendation scheduling method, and computer-readable storage medium | |
CN110351104A (en) | A kind of VIM selection method and device | |
CN114070889A (en) | Configuration method, traffic forwarding method, device, storage medium, and program product | |
CN114090201A (en) | Resource scheduling method, device, equipment and storage medium | |
CN113472638A (en) | Edge gateway control method, system, device, electronic equipment and storage medium | |
US20210075682A1 (en) | Forming groups of nodes for assignment to a system management server | |
CN112749193A (en) | Workflow processing method and device, storage medium and electronic equipment | |
JP2016042248A (en) | Orchestration system, method, and program | |
CN115834714B (en) | Cross-platform task scheduling method, server and system |
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 |