521 results about "Scheduling (computing)" patented technology

A distributed task scheduling method and system that separates and performs task management and task execution on separate computing devices and distributes task execution over multiple computing devices. The distributed task scheduler includes two-tier architecture having at least one execution host and at least one data broker. The execution hosts handle the tasks and the data broker manages the task schedule. The data broker determines any scheduled tasks that need to be executed. Once an available task is found, the data broker dispenses the scheduled task to an execution host. A timeout period is selected for each assigned task. If the assigned execution host does not report back to the data broker within the timeout period the completion of the assigned task, the data broker is free to assign the task to another execution host to ensure reliable execution of the task.

An architecture and design called Resource control programming (RCP), for automating the development of multithreaded applications for computing machines equipped with multiple symmetrical processors and shared memory. The Rcp runtime (0102) provides a special class of configurable software device called Rcp Gate (0600), for managing the inputs and outputs and user functions with a predefined signature called node functions (0500). Each Rcp gate manages one node function, and each node function can have one or more invocations. The inputs and outputs of the node functions are virtualized by means of virtual queues and the real queues are bound to the node function invocations, during execution. Each Rcp gate computes its efficiency during execution, which determines the efficiency at which the node function invocations are running. The Rcp Gate will schedule more node function invocations or throttle the scheduling of the node functions depending on the efficiency of the Rcp gate. Thus automatic load balancing of the node functions is provided without any prior knowledge of the load of the node functions and without computing the time taken by each of the node functions.

The invention discloses a management scheduling technology based on a hyper-converged framework. The method comprises a hyper-converged system architecture design, resource integrated management basedon a hyper-converged architecture, unified computing virtualization oriented to a domestic heterogeneous platform, storage virtualization based on distributed storage, network virtualization based onsoftware definition and a container dynamic scheduling management technology oriented to a high-mobility environment. According to the management scheduling technology based on the hyper-converged framework provided by the invention, the virtualization capability and the management capability of the tactical cloud platform are improved; a key technical support is provided for constructing army maneuvering tactical cloud full-link ecology; an on-demand flexible virtualized computing storage resource pool is provided, heterogeneous fusion computing virtualization is achieved, meanwhile, a distributed storage technology is used for constructing a storage resource pool, a software definition technology is used for constructing a virtual network, a super-fusion resource pool is formed, localization data and network access of application services are achieved, the I/O bottleneck problem of a traditional virtualization deployment mode is solved, and the service response performance is improved.

The invention discloses a heuristic type coarse grain parallel grid task scheduling method which is characterized by comprising the following steps: firstly, a task submitter inputs a to-be-schemed task set, a usable computing resource set, an execution time set of tasks on the computing resource, a maximum number of iteration times, a threshold value delta and an entropy value epsilon; secondly, a task scheduler represents a scheduling problem of allocating resources to execute the tasks to be a standard minimum solving problem under task optimization and constraint conditions; thirdly, a grid task scheduling problem is solved by an iterative process of the heuristic type coarse grain parallel algorithm; and fourthly, after the algorithm is finished, the task scheduling result is output. In the way, the invention provides a new multipoint crossing method and the increasing property of a population optimal solution is maintained by an elitist strategy; during a variation stage, a directed variation method based on task immigration is adopted to prevent degeneration of the population; and the method is high in performance and better than the traditional random algorithm in computing power and convergence rate.

Techniques for network state-aware and network scheduling policy-aware task scheduling are provided. In one aspect, a method for scheduling tasks in a distributed computing network includes the steps of: collecting a pool of candidate resources in the distributed computing network for performing a given one of the tasks; predicting a performance of each of the candidate resources in performing the given task based on both i) a state and ii) a scheduling policy of the distributed computing network; and selecting a best candidate resource for the given task based on the performance. A system for scheduling tasks in a distributed computing network is also provided which includes a task scheduler; and a network scheduler, wherein the task scheduler is configured to schedule the tasks in the distributed computing network based on both i) the state and ii) the scheduling policy of the distributed computing network.

Aiming at the task scheduling problem of an edge computing server, the invention provides a mobile edge computing system task scheduling method based on migration and reinforcement learning. The method comprises the following steps: constructing an Actor-Critic network for each server to train a scheduling strategy of the server, wherein the Actor network determines the action according to the state of the Actor network, and the Critic network evaluates the quality of the action according to the actions and the states of all the servers. And all the servers share one Critic network. When a plurality of edge server scheduling strategies are trained by using multi-agent reinforcement learning, a strategy network with the same structure is constructed for the scheduling strategy of each server. These policy networks not only have the same network layer, but also have the same number of nodes in each layer. The strategies are trained by using a centralized training decentralized executionmechanism, so that the problem of dimensionality disasters caused by too many servers is avoided.

The invention discloses a cloud computing associated task scheduling method and device based on time constraint. The method comprises the steps that firstly, the associated relation between tasks in an associated task set is expressed by building a DAG; the thought of layering the DAG is adopted to enable unassociated tasks in the same layer to be fallen into the same task set; hardware resources are virtualized, and a virtual machine cluster is built so as to provide task execution places; finally, the layering scheduling method based on the time constraint is adopted to schedule the tasks in the task set in each layer to the most appropriate virtual machine to be executed so as to ensure that the tasks are completed on schedule. According to the technical scheme, the problem that execution of the cloud computing associated tasks is delayed in the scheduling process in the prior art is effectively solved, the delay phenomena of the tasks can be reduced to the maximum, it is ensured that the tasks are completed within the time expected by a user, and meanwhile virtual machine resources are effectively utilized.

The invention discloses a method and system for scheduling routine work and scheduling nodes. The method for scheduling the routine work comprises the steps of obtaining resource information of all computational nodes, collecting work information of the routine work required to be scheduled, dividing the routine work into groups according to the dependency relationship of the routine work, determining a key path of the routine work of each set according to a key path method, arranging each set to be in the routine work on the key path, judging whether all the routine work is arranged or not, if all the routine work is arranged, informing the corresponding computational nodes of obtaining the arranged routine work, and therefore enabling the corresponding computational nodes to conduct work processing according to the arranged routine work. Due to the fact that the routine work is arranged through the key path method, the processing blank of the routine work in a cloud computing environment is filled up, and the defects that in the past, due to manual arrangement, resources are busy and idle unevenly, and the utilization rate is low are overcome.

The invention discloses a distributed real-time task scheduling method and device, equipment and a storage medium. The distributed real-time task scheduling method comprises the following steps: receiving each task to be distributed through a Master node, and obtaining a task queue; distributing the task queue to a Redis node; caching a task queue through a Redis node; distributing the tasks to bedistributed in the task queue; according to a preset mode, submitting to a main load computing center; receiving a task submitted by the Redis node through a main load calculation center as a basic task; acquiring state information of each node server in the cluster from a work cluster statistics center as the current state; according to the basic task and the current state, performing task scheduling to obtain a target task; and dynamically distributing the target task to each node server according to the current state of each node server. The invention further relates to a blockchain technology, wherein the basic task, the current state and the target task can be stored in a blockchain node. The distributed real-time task scheduling method and device can improve the task scheduling efficiency.

The embodiment of the invention provides a task scheduling method and device and an electronic terminal, and relates to the technical field of computing. The method comprises the steps: a scheduling task configuration request is received, and a current scheduling task configuration table is output; Then, operation information for the current scheduling task configuration table is obtained, and theoperation information comprises task scheduling parameters; According to task scheduling parameters in the operation information, a target scheduling task configuration table is generated, and the target scheduling task configuration table at least comprises n timing tasks, whether each timing task is effective or not and execution starting time; Therefore, m specified timing tasks with effectiveidentifiers are screened out from the target scheduling task configuration table, the execution starting time of the m specified timing tasks is at the current time or after the current time, and them specified timing tasks serve as m to-be-processed tasks; And then m to-be-processed tasks are executed. Therefore, according to the technical scheme provided by the embodiment of the invention, theswitch workload and problems in the development process can be reduced, and the requirements of flexibly increasing and modifying the quartz scheduling are met.

The invention relates to a task scheduling method and system, in particular to a multi-timing-sequence task scheduling method and system. The method comprises the steps that node information is collected, and according to the node information, the operational capability of each node is set; task information is collected, and according to the task information, the grid number of each task is set; according to operational capability, all nodes are sorted; according to the grid number of each task, all tasks are sorted; and according to the load state of each node, a greedy algorithm is used, and according to the mode that nodes with high operational capability process tasks with the large grid numbers, the tasks are distributed to the nodes. Resource distribution on a current platform and the loading capacity and states of the nodes in the platform are considered, during task scheduling, tasks with high calculated quantity are distributed to node with high computing capacity, the problems of starvation states and the low system resource using rate can be effectively avoided, overall operational efficiency is improved, and operation time is optimized.

An embodiment of the present application discloses a task scheduling method and device based on a hadoop cluster. A specific embodiment of the method includes: deploying a plurality of virtual nodes according to a service type of each task to be scheduled, wherein each service type corresponds to at least one virtual node; receiving a task to be scheduled sent by a user, and determining a virtualnode for allocating the task to be scheduled according to the service type of the task to be scheduled; using the determined virtual node, the task to be scheduled is assigned to a proxy node corresponding to the virtual node so that the proxy node submits the task to be scheduled to the hadoop cluster, wherein each virtual node corresponds to at least one proxy node. The embodiment determines a virtual node and a corresponding proxy node for submitting a task to be scheduled to a hadoop cluster according to the service type of the task to be scheduled, and can realize the parallel computing requirements of a large number of tasks to be scheduled of different service types.

The invention discloses an intelligent planning space-time cloud GIS platform based on a Docker container and a micro-service architecture, and the GIS platform employs a hierarchical architecture, and sequentially comprises a basic bottom layer which is used for providing basic services suitable for providing basic supporting facilities and task scheduling and elastic computing resources for theGIS platform; a Linux bottom layer which comprises a control group, a namespace, application protection, equipment and network connection; a data layer which is a data storage engine of the GIS platform; a Docker layer which comprises a Docker client side, a Docker process and a Docker registration information side; and an application layer which adopts a micro-service architecture, and is used for packaging each application program, wherein each application program communicates with the Docker client of the Docker layer, and a request is sent to the Docker server through the Docker client, sothat the Docker container corresponding to each Docker mirror image is managed. According to the invention, the problems of poor flexibility, expansibility and coupling of the existing GIS platform are solved.

The invention discloses a management warehouse for scheduling an intelligent analysis algorithm in a complex scene, and a scheduling method. The management warehouse comprises a global scheduling layer, a computing resource node layer and a shared file storage layer. The global scheduling layer receives the service request; wherein the service request comprises a to-be-executed service; determining a computing resource node for executing the service according to the service request; the service request is sent to the computing resource node layer; one or more intelligent analysis algorithms are called from the shared file storage layer to execute the to-be-executed service through the computing resource node executing the service in the computing resource node layer, so that the problem that unified management of computing resources cannot be achieved in the prior art is solved, multiple algorithms are fused, and the resource utilization rate is increased.

The invention provides a scheduling method and device, electronic equipment and a readable storage medium, and the method comprises the steps: obtaining a to-be-processed task and to-be-processed data, and obtaining a target learning model needed by the to-be-processed task from a plurality of pre-stored deep learning models according to the correlation between different pre-stored deep learning models and different processing tasks; scheduling the target learning model to the corresponding computing unit based on a pre-trained scheduling model according to the resource utilization rate of each computing unit so as to process the to-be-processed data to obtain an output result. In this way, when the processing task is executed in the end-side equipment, the scheduling model is establishedin advance, and the optimized scheduling strategy is obtained based on the real-time resource utilization rate, so that the overall execution efficiency of the equipment is improved.

The invention provides a method, a device and a system for realizing task scheduling. The method comprises the following steps: acquiring a target data processing task submitted by a client and running parameters of the target data processing task; obtaining resource use information of the available computing cluster; selecting a target computing cluster from available computing clusters accordingto the running parameters of the target processing task and the resource use information of the available computing clusters; and scheduling the target data processing task to a target computing cluster for running.

The invention discloses a container cloud-oriented task preemption scheduling method and system. The task preemption scheduling method comprises the steps: storing tasks submitted by a user through atask queue; taking out to-be-scheduled tasks from the task queue; judging whether the resources in the container cloud platform can meet the execution of the to-be-scheduled task or not; if the execution of the to-be-scheduled task can be satisfied, scheduling the to-be-scheduled task to the resource node, otherwise, destroying the low-priority task in the container cloud platform to release the computing resource, and scheduling the to-be-scheduled task to the resource node when the resource in the container cloud platform can satisfy the execution of the to-be-scheduled task. According to the task scheduling method and device, the tasks with low priorities can be destroyed and computing resources can be released in a resource full-load operation scene, so the tasks with high priorities are executed in time, meanwhile, due to smooth execution of the tasks, the number of the tasks waiting for being executed in the task queue is greatly reduced, and the task scheduling efficiency is effectively improved.

The invention discloses a scientific workflow scheduling method and device under a network environment. The method comprises the following steps that heterogeneous computing resources used as targets are inquired, and computing power PCj of each heterogeneous computing resource is recorded; all schedulable tasks in workflows used as targets are inquired; the task ratio p (ti, rj) of each task ti in all the schedulable tasks on the j heterogeneous computing resource rj is computed, and ct (ti, rj) = ext (ti, rj) + rt (ti, rj); in all the obtained task ratios, the largest task ratio p (tm, rn) is obtained, and the m task tm is correspondingly scheduled to the n heterogeneous computing resource rn to be executed. By means of the scientific workflow scheduling method and device, good resource load balance is achieved, the method and device can be adaptive to static and moving environments, scheduling time is short, total execution time is short, and combination property is excellent in the existing scheduling algorithms.

The embodiment of the invention discloses a computing power resource allocation method and device, equipment and a storage medium, containers corresponding to different business configurations, a computing power resource quota corresponding to each container configuration converts a first instance into an instance task when the submitted first instance is obtained for a target business, and according to a corresponding relationship between the business and the container, the target container corresponding to the target computing is determined. Iif it is determined that no available computing power resource exists in the target container according to the computing power resource quota corresponding to the target container, the instance task corresponding to the first instance is controlledto enter a task queue of the target container. When it is monitored that the available computing power resources appear in the target container, the target instance task is scheduled from the task queue to be put into operation. Since different services are isolated through the container, examples of the same service share the computing power resources of the same container, even if priority preemption occurs during scheduling, the priority preemption is only constrained in the container, the problem of computing power resource preemption among different services is avoided, and the user experience is improved.

The invention provides a parallel computing method and system based on an AMP framework DSP operating system, and a medium. The parallel computing method comprises the following steps: an operating system use step: a hybrid memory mechanism providing step: completing task distribution in a task pool through a shared memory and distributed memory hybrid mechanism according to a used embedded DSP operating system, outputting a distribution task, and synchronizing a task result after task execution is completed; and a task publishing and executing step: publishing and executing the task based onthe multi-core navigation according to the obtained distribution task. According to the invention, an inter-core synchronization, task distribution and recovery mechanism is introduced into the operating system core, so that the scheduling level and frequency of the system are reduced, and the parallel computing efficiency is improved, and the parallel computing method has very important significance for an embedded DSP operating system with very high real-time requirements.

The embodiment of the invention discloses a stream computing platform and method based on visualization, equipment and a storage medium. The stream computing platform comprises a data development component, a stream computing component and an intelligent scheduling component, wherein the data development component is used for visually developing a corresponding stream computing application througha dragging operator in stream computing, and determining a stream computing engine corresponding to the stream computing application; the stream computing component is used for analyzing scheduling parameters carried in the stream computing application and calling the stream computing engine to execute the stream computing application; and the intelligent scheduling component is used for scheduling execution of the stream computing application according to the scheduling parameters. The technical scheme provided by the embodiment of the invention does not need to be realized only by developers who deeply understand a stream computing application technology, so that the application range of the stream computing platform is expanded, and the expandability of different stream computing engines is supported, and the flexibility of stream computing development is improved.