78 results about "Cloud workflow" patented technology

The invention discloses a cloud workflow scheduling method based on collected discrete particle swarm optimization. Aiming at cloud workflow scheduling with different QoS requirements, the method adopts a novel collected discrete particle swarm optimization algorithm, through redefinition for the speed, the position and relative update operation of particle swarm optimization in collected space, application requirements for cloud workflow scheduling optimization in discrete space can be met, and efficiency of cloud workflow scheduling is improved.

The invention discloses a multi-target cloud workflow scheduling method based on an improved non-dominated genetic algorithm. By introducing a scoring mechanism idea and considering the influence of the current population and historical population information on individual dominant information, the accuracy of population individual evaluation is improved, and the efficiency of iterative search isimproved. The method comprises the following steps of constructing population hierarchy so as to directly depict diversity and optimality of an optimal solution traversed by algorithm search, and dynamically updating a population hierarchical structure according to the degree of approaching Pareto optimality of an offspring individual in an iteration process by improving a parent individual selection mode, so that the possibility that the found solution approaches Pareto optimality is improved. Meanwhile, a search direction self-adaptive adjustment strategy based on optimal level individual monitoring is provided.By setting local optimum and divergence detection parameters, relevant parameters can be adjusted in time when the search is trapped in the local optimum or tends to be divergent,and the optimization direction is changed to jump out of the local optimum or regression convergence.

The present invention discloses a scheduling method for cloud workflow job. The method comprises: decomposing each cloud workflow job according to a DAG(Directed Acyclic Graph) flow chart of each cloud workflow job so as to obtain a plurality of mutually independent sub cloud workflow jobs; estimating the execution time of each sub cloud workflow job, wherein the sub cloud workflow jobs enters a waiting queue; according to SLA(Service Level Agreement), restraining and assigning priorities of the sub cloud workflow jobs; and scheduling the sub cloud workflow jobs by using reinforcement learning. The scheduling method for cloud workflow jobs provided by the present invention improves the resource utilization rate and the service quality in a cloud computing system.

The invention discloses a method for optimizing a dynamic cloud workflow by using a genetic algorithm. The algorithm aims at minimizing periodic average expense of the cloud workflow in iteration execution, and limiting the total execution time that the workflow is executed for on period in a dynamic environment each time within a maximum execution deadline defined by a user. As the execution mode of the workflow is dynamic and variable in the cloud realization environment, the method carries out overall modeling on all flow topology results which may occur, then the characteristics of dynamic time-variation of the cloud workflow are comprehensively considered, the average performance of the execution of the cloud workflow in the dynamic environment is optimized by using the genetic algorithm, and the execution efficiency of the cloud workflow is improved.

The invention discloses an energy consumption-oriented cloud workflow scheduling optimization method. The method comprises the following steps: (1) establishing energy consumption-oriented cloud workflow process model and resource model; (2) calculating a task priority; (3) taking out a task t with highest priority from a task set T, finding out a virtual machine set VMt capable of executing the task t, and calculating energy consumption for distributing the task t to each virtual machine in the VMt and completing all distributed tasks; (4) finding out a vm with minimal energy consumption, if only one vm has the minimal energy consumption, distributing the task t to the vm, and if a plurality of vms have the minimal energy consumption, distributing the task t to the vm with characteristics that the vm has the minimal energy consumption and a host in which the vm is located has highest performance per watt; deleting the task t from the task set T, and if the task set T is not null, going to the step (3), or otherwise, going to the step (5); and (5) outputting a workflow scheduling scheme. According to the scheduling optimization method provided by the invention, an energy consumption factor is considered, so that the energy consumption for task processing by the host is effectively reduced while workflow execution time efficiency is kept.

The invention provides a self-adaptive multi-object evolution method adopting constraint cloud workflow scheduling. The overall detection and local mining capability of the multi-object evolution method can be improved. The multi-object evolution method comprises the steps that S1, the evolution states of populations in the evolution process are detected according to the number of Pareto solutions and Pareto entropies, and corresponding individual evaluation strategy processing constraint conditions are self-adaptively utilized to sort individuals in the populations according to the detected evolution states of the populations in the evolution process, wherein a constraint violation processing method is adopted to process the constraint conditions in individual evaluation strategies; S2, according to individual sorting results, individuals are selected from the populations to perform genetic manipulation, and sub-populations are obtained, wherein genetic manipulation parameters are self-adaptively adjusted according to the evolution states of the populations in the evolution process during genetic manipulation. The self-adaptive multi-object evolution method is suitable for solving the multi-object evolution problem having constraints and can be applied to the technical field of workflow scheduling in a cloud computing environment.

The invention relates to a service level agreement SLA-based cloud workflow engine resource scheduling and controlling method. The method comprises the following steps of: dividing a resource distribution process into two flows of carrying out regular check and requesting to carry out scheduling; in the flow of carrying out regular check, loading all the lessees in a system and SLAs of the lessees by the system, obtaining a real-time speed of each engine, combining the real-time speeds into real-time speeds of the lessees, calculating the differences between SLA speeds of the lessees and the real-time speeds of the lessees, for the lessees, the differences of which exceed the upper and lower limits of a threshold value, executing a double-threshold value-based engine controlling method to increase and decrease engines in a cluster, executing a genetic algorithm-based scheduling method according to the distribution of a physical machine and a virtual machine of the current engine so as to calculate a new distribution, and completing engine migration; and in the flow of requesting to carry out scheduling, loading a mutual exclusion rule set and the current user table after a case request is received, traversing the mutual exclusion rule set and the current user table, selecting the engine which does not trigger mutual exclusion rules or has minimum performance influence, recording performance snapshots of a user and the engine, updating a mutual exclusion rule list, and processing the request and responding the user.

The invention provides a scheduling method for a cloud workflow. A scheduling model that sets resource demand minimization as a goal and uses a task deadline as a constraint condition is established, an optimal solution to the scheduling model is searched, and according to a scheduling scheme corresponding to the optimal solution, a cloud workflow is scheduled; and the optimal solution to the scheduling model is searched by using a hybrid adaptive iterative local searching method. According to the method, compared with the prior art, the provided scheduling method for searching a multi-mode cloud service workflow by using iterative local searching has higher efficiency in terms of a searching speed and solution efficiency; and a resource distribution method can be found rapidly within the task deadline, so that the user's task can be completed efficiently with the lowest costs. Therefore, the time cost and the economic cost are saved for the user.

In order to solve the problem of cloud scientific workflow scheduling time with budget cost constraint, the invention provides a time optimization scheduling algorithm R-ACO based on an ant colony algorithm and in combination with a traditional task probability upward weight, so as to reduce the whole cloud workflow task scheduling time. According to the algorithm, the characteristic of mutual constraint among tasks in scientific workflows is considered, the execution sequence of the tasks is ranked by utilizing traditional probability upward weights, and then the ant colony algorithm is usedfor carrying out scheduling between tasks and resources with the optimized scheduling time as the target under the budget cost constraint. The method provided by the invention can effectively shortenthe scheduling time.

The invention belongs to the technical field of cloud workflows, and particularly relates to a cloud workflow implementation method capable of supporting various engines. The cloud workflow implementation method is characterized in that appropriate engines are found in an engine group to complete various types of given tasks through mode conversion. The beneficial effects are that the method realizes decoupling of engines and data in a workflow system, so that a task submitted by a user is not limited to be executed by a specific engine any more, a purpose that any engines in the workflow engine group can provide services for any users is achieved, tasks, which are originally executed by a specific engine, are distributed to more engine to execute through analysis, data extraction, engine dispatching and adaptation between the tasks and the engines for the tasks, and thus various types of engines are supported.

The invention belongs to the technical field of cloud workflow scheduling in container environments, and discloses a method and a system of cloud workflow scheduling in a container environment. The system of cloud workflow scheduling in the container environment includes a resource acquisition module, a container performance monitoring module, a main control module, a resource allocation module, adata mirror image optimization module, an allocation report module, a data storage module and a display module. According to the system, container performance can be monitored through the container performance monitoring module while performance of a server can be simply and conveniently monitored, error warnings can be generated, and a method of visually presenting parameters through forms of graphs and tables can be implemented; at the same time, a mirror image is enabled to still ensure good re-editability after packaging and publishing through the data mirror image optimization module; and all file drives of existing Docker are supported, and through firstly containerizing a local mirror image and then carrying out exporting after abridging processing, no specific file drive is depended on.

The invention discloses an SLA-based stateless cloud workflow load balancing scheduling method. According to the requirements of different tenants on the business scene and the flow model, different SLA levels are selected, different SLA levels are adopted; the cloud workflow system provides different request throughput services for the tenants; grading service is carried out on different processrequests; engine load real-time monitoring and process model distribution condition on an engine are realized by combining a shared memory; the overall memory overhead of the engine cluster is reducedwhile the request wave crest of the engine service is reduced; Therefore, the load balancing capability of the cloud workflow under the multi-tenant architecture is improved, so that the flow serviceprovider can provide services for more tenants on the basis of meeting the analysis execution performance requirements of different tenants on request throughput and different flow definitions.

The invention provides a computing intensive cloud workflow scheduling method based on an owl search algorithm, and belongs to the technical field of cloud computing. By modifying a population iteration updating formula in an owl search algorithm, each scheduling scheme is updated according to the influence of an optimal scheduling scheme on the scheduling scheme, so that optimization is more targeted; in a population iteration updating mechanism, randomness is introduced by utilizing a genetic variation thought, a search process is prevented from falling into local optimum, an optimal scheduling scheme can be obtained in a shorter time, reasonable distribution of virtual machines is realized, and tasks are efficiently scheduled. The method can effectively overcome the defects that an optimal solution in an existing method is high in searching randomness, prone to falling into local optimum and low in convergence speed, the searching efficiency is improved, the searching time is shortened, a better scheduling scheme can be found in a shorter time, and the time cost of workflow scheduling is reduced.

The invention discloses a multi-target cloud workflow scheduling method based on a joint reinforcement learning strategy, and the method comprises the steps: building a reinforcement learning agent joint strategy model through the extension of the attributes and methods of a workflow request and a cloud resource, and enabling the scheduling model to be more suitable for an actual workflow application scene. According to the method, the influence of the scheduling process, each decision sub-network and historical decision information is comprehensively considered during behavior selection, so that the finally selected behavior is more reasonable, the dominance and diversity of generating a non-dominated solution set by the algorithm are further improved, and the practicability of the method is effectively improved.

The invention relates to a cloud workflow task execution time prediction method based on limit gradient improvement, and belongs to the technical field of cloud computing. According to the method, influence factors of task execution time are classified from three levels of workflow task composition, resources on which task operation depends and a physical execution environment of the resources, and comprehensive modeling of the influence factors of the task execution time is achieved. Secondly, aiming at the condition that the sample data set has a data missing value, the data set with the missing value is complemented by adopting a machine learning method; and finally, by means of the multi-type data processing capability of the extreme gradient lifting algorithm, the parameter design isrelatively simple, the calculated amount is small, the advantages of a serial learner and a parallel learner are combined, and a cloud workflow task execution time prediction model is trained by adopting the extreme gradient lifting algorithm. Compared with an existing prediction model, limitation on the sample data type is reduced, prediction errors are reduced, and the generalization ability ofthe model is further improved.

The embodiment of the invention provides a cloud workflow scheduling method and device based on multi-objective optimization, which relate to the technical field of high-performance computing, can carry out multi-objective optimization based on local search and weight vector adjustment, and obtain faster convergence speed and better diversity of individuals at the same time. The method includes establishing a simulation cloud computing resource model and a cloud workflow task model; then the population is initialized and the fitness of each individual is calculated. Further evolutionary processing for each individual; then, the local search is carried out to obtain two optimal individuals in the neighbors of each sub-problem, and the new individuals are calculated to update the individualsagain. And when the updated individual satisfies the convergence condition, the evolutionary population is adjusted by weight vector, which includes deleting subproblems of congested regions and adding new subproblems to sparse regions; finally outputting optimal individual. The technical proposal provided by the embodiment of the invention is applicable to a cloud workflow scheduling process.

According to the cloud workflow implementation method with intrusion tolerance, the executive programs corresponding to the tasks in the workflow are deployed in the containers respectively, and thenthe containers are transplanted to the virtual machines with the different operating systems, so that the cloud system intrusion tolerance capability can be effectively improved. Furthermore, the subtask results executed by the plurality of virtual machines are judged; calculating the confidence coefficient of the sub-task result generated within the preset time; therefore, the reliability and credibility of the execution of the current subtask are determined, whether the selection result is submitted to the next subtask or the execution subtask of the virtual machine is reset is determined, the reliability and credibility of the output result are ensured through the mutual judgment and verification of the generation results of the plurality of virtual machines, and the invasion toleranceof the cloud system can be truly improved. In addition, the invention provides a system applying the cloud workflow implementation method with intrusion tolerance, and the purpose and the effect of the implementation method are achieved.

The invention discloses a parallel cloud workflow scheduling method based on a reinforcement learning strategy, and the method comprises the steps: introducing a pointer network in a task selection process, taking softmax probability distribution as a pointer for processing variable-length input so that a workflow scheduling model is capable of sensing the dynamic change of a to-be-selected task at different workflow scheduling stages, thus improving the task selection efficiency. More task execution sequence knowledge is learned, and the optimization performance of a scheduling solution is improved.

The invention discloses a cloud workflow management system (hereinafter referred to as a cloud workflow) supporting large-scale living example intensive applications, and belongs to the technical sub-field of workflow management and the technical field of cloud computation in the field of computer software engineering. The system mainly comprises the technical sub-fields of workflow process management, execution service, monitoring, load balance, early warning, automatic telescoping, execution service cloning, charging and the like under cloud computation environment; by utilizing cloud computation, the problem that the existing workflow management technology cannot elastically and economically support large-scale living example intensive workflow applications is solved; and the system is developed by using computer software technologies such as Java, HTML, CSS, Javascript and the like, integrates technologies such as Web services, an Eclipse Rich Client Platform plugin and the like, and is mainly used for business process management, business system integration and the like of government and enterprise units.

The invention discloses a heuristic and intelligent computing fused cloud workflow segmented online scheduling optimization method. The heuristic and intelligent computing fused cloud workflow segmented online scheduling optimization method comprises the following steps: acquiring information required by scheduling optimization; calculating a sorting value and a hierarchical value of the task; generating a first-stage task scheduling optimization scheme based on a heuristic method of dynamic key task priority scheduling; obtaining a task scheduling optimization scheme of a second stage based on a genetic algorithm; and outputting the scheduling optimization scheme. According to the heuristic and intelligent computing fused cloud workflow segmented online scheduling optimization method, thesegmented scheduling optimization method integrating the heuristic method and the intelligent calculation method is adopted, and the solving time is equal to the solving time of the heuristic method,and the quality of the solution is similar to the quality of the solution solved by the intelligent calculation method, so that the understanding quality is effectively improved on the premise of adapting to real-time online scheduling.

System and method for implementing a workflow of a first domain, wherein the workflow is implemented as a series of steps to accomplish a workload and wherein at least one of the steps utilizes a process, are described. In one embodiment, the method comprises establishing a mutual trust relationship between the first domain and a second domain; wherein one of the steps is authored by the second domain, the method further comprising associating with the step authored by the second domain a digital attestation for enabling the first domain to verify authorship and non-modification thereof.

The invention provides a method for realizing load balancing of a flow instance of a cloud workflow system, including the following steps: (1) estimating the computing resource requirement of the currently obtained flow instance in each monitoring period during a running period by using a log analysis method; (2) selecting a suitable workflow engine in a scheduling domain for the currently startedflow instance by using a first adaptive descending strategy with a buffer queue. The method for realizing load balancing of a flow instance of a cloud workflow system can solve the problem that the traditional flow instance load balancing algorithm needs to move the flow instance and affect the user experience.