188 results about "Virtual machine scheduling" patented technology

A system, computer program product and method for dispatching work items in a virtual machine operating system. The virtual machine operating system defines first and second virtual machines. First and second work queues are created in a memory. The first virtual machine assigns a first work item to the first work queue and a second work item to the second work queue. The first virtual machine schedules work items from the first work queue for execution by the first virtual machine. The first and second work items originate from the first virtual machine. The first and second work queues are directly accessible by the first virtual machine. The second virtual machine assigns a third work item to the first work queue and a fourth work item to the second work queue. The second virtual machine schedules work items from the second work queue for execution by the second virtual machine. The third and fourth work items originate from the second virtual machine. The first and second work queues are directly accessible by the second virtual machine. The first virtual machine is not able to schedule work items from the second work queue, and the second virtual machine is not able to schedule work items from the first work queue. A shared work queue in shared memory is also disclosed.

The invention relates to a multi-agent reinforcement learning scheduling method, a multi-agent reinforcement learning scheduling system and electronic equipment. The method comprises the following steps: step a, collecting server parameters of a network data center and virtual machine load information running on each server; b, establishing a virtual simulation environment by using the parametersof the server and the load information of the virtual machine, and establishing a deep reinforcement learning model of multiple agents; c, performing offline training and learning by utilizing the deep reinforcement learning model of the multiple agents and a simulation environment, and training an agent model for each server; and d, deploying the agent model to a real service node, and schedulingaccording to the load condition of each service node. The service running on the server is virtualized through the virtualization technology, load balancing is carried out in a virtual machine scheduling mode, resource distribution is more macroscopic, and a strategy that multiple agents generate cooperation in a complex dynamic environment can be achieved.

The invention belongs to the technical field of computing system virtualization, and particularly relates to a virtual machine scheduling method for a public cloud platform. The technical scheme is that: a plurality of scheduling servers are set to process user requests. By the multi-scheduler scheduling method, the optimization of platform resources and the setting of virtual machine migration, the problem that congestion is easy to cause by a single-scheduler scheduling method is effectively solved; a quick response to the user request can be made to improve service efficiency; and the method can be widely applied to the virtual machine scheduling of the public cloud platform.

The invention provides a dispatching method of a virtual machine under a multi-core environment. The dispatching method divides zones for CUP resources according to dispatching strategy types when a system is started, monitors the CPU loading condition of all divided zones in real time when the system operates, and dynamically adjusts the size of the CPU resources in the divided zones. The dispatching method carries out dispatching in the same divided zone by using the same dispatching strategy, thus improving dispatching efficiency, achieving the purpose of resource loading balance by dynamically adjusting the size of the CPU resources of the divided zone, realizing the full utilization of resources and reducing waste of resources.

The invention discloses a two-level management method for a hybrid desktop cloud service platform. The two-level management method for the hybrid desktop cloud service platform comprises the following steps: building a hybrid desktop cloud data center; sending a login request to a central management node (Vmm-Server) by a server node; receiving and processing the login request of the server node by the central management node; uniformly receiving user requests by the central management node through a service interface layer; responding to the user requests by the central management mode according to a two-tier scheduling method; maintaining system load balance by the central management node according to the current system state and the running state of a virtual machine. According to the two-level management method disclosed by the invention is designed and realized upon the open system architecture of a hybrid desktop cloud server, and the scheduling and migration efficiency of the virtual machine can be improved.

The invention relates to the system-level virtual technology and the virtual machine scheduling technology on a multi-core processor in the field of computer system structures and discloses a method for ensuring fault isolation of virtual machines of a cache-aware multi-core processor. The method comprises the steps of: carrying out state abnormity monitoring and heartbeat information detection on all the running virtual machines by a fault detector; when detecting an abnormal virtual machine, firstly recording the distribution condition of the fault virtual machine on a core of the processorby an isolation scheduler; and when finding that the loads of the other virtual machines share one L2 cache with the fault virtual machine, transferring the fault virtual machine to a core corresponding to an independent cache by the isolation scheduler and carrying out VM-core scheduling on the other normal virtual machines according to cache sensitivity characteristics. The invention discloses a method for the dynamic isolation scheduling of the virtual machines on the multi-core processor, thereby preventing the other normal virtual machines from being polluted by the cache, reducing the running influence of the fault virtual machine on the normal virtual machines and enhancing the whole isolation of the system.

The invention belongs of the field of Iaas (infrastructure as a service) cloud computation, and relates to a hybrid genetic algorithm based dynamic cloud-computing virtual machine scheduling method. The method includes the steps: monitoring load information of physical machines and virtual machines in the cloud computing environment, analyzing a load cycle and changes of each host, and determining a cycle of load changing; computing a virtual machine placement combination appearing in cloud computation through a hybrid genetic algorithm with multiple fitness degrees. According to three optimization objectives, an optimized virtual machine placement strategy is computed to serve as a final result; the algorithm is executed periodically, the virtual machines are reasonably placed through dynamic migration of the virtual machines, resource utilization rate is increased, and resources are saved. By the method, the problem that resource utilization rate in a current cloud computing center can be solved, and the requirement on automatic management of the current cloud computing center is met.

The invention discloses a system, a platform and a method for virtual machine scheduling decision-making. The above-mentioned virtual machine scheduling decision-making system includes: a template management module for selecting a policy template customized according to user needs; a policy management module for creating and activating a strategy by setting parameters in the policy template; a scheduling decision calculation module for Use the strategy in the active state to perform exhaustive calculations on the current resource information and cluster operation data to obtain the optimal decision. In the present invention, the virtual machine scheduling policy is developed and deployed using the policy template of the script, the development speed is fast and the deployment is flexible, and rich alternative policies can be provided to meet the needs of users at different levels, thereby improving the system performance of the virtual machine.

The invention provides a power supply and performance management system for a virtualization server cluster. The power supply and performance management system comprises an acquiring module, a performance optimizing module and an energy consumption optimizing module, wherein the acquiring module is used for acquiring operation data of the virtualization server cluster; the performance optimizing module is used for carrying out energy efficiency modeling on performance and resource utilizing conditions of a virtual machine according to an original energy source and resource utilization relationship and outputting resource configuration parameter of the virtual machine; and the energy consumption optimizing module is used for receiving the resource configuration parameters of the virtual machine, which are from the performance optimizing module, establishing an energy consumption model according to the resource configuration parameters of the virtual machine and carrying out scheduling and resource re-allocation on the virtual machine according to limiting conditions of a resource threshold and the scheduling limiting condition of the virtual machine. According to the power supply and performance management system disclosed by the invention, due to the combination of a control theory with a linear programming technology, the considerable energy-saving effect can be achieved while application performances are ensured. Particularly, the effect of saving energy by 43 percent can be achieved by integrating a performance controller with an energy optimizer.

The invention provides a resource scheduling method, which relates to the computer application field, and solves the problem of the present virtual machine that the resource is unreasonable to use during the scheduling process. The method comprises the following steps that: resource utilization rate of each calculation node on which a virtual machine runs is detected in a resource pool; the detected resource utilization rate of each calculation node is compared with a preset integrated standard value, and a calculation node with the utilization rate being lower than the integrated standard value is determined as a low utilization rate calculation node; and the virtual machine on the low utilization rate calculation node is integrated and centralized onto partial calculation nodes in the resource pool. The technical scheme is applicable to a scheduling system, and the resource scheduling with high utilization rate can be realized.

The invention provides an Xen virtual machine scheduling control method in multi-core environment. The method comprises the following steps: (I) carrying out scheduling initialization; (II) carrying out initialization of scheduling information statistics, namely monitoring the virtual CPU (central processing unit) in a management domain 0, and sending the monitoring result to a VMM (virtual machine monitor) in real time; (III) creating a virtual machine, allocating the virtual CPU, storing the virtual CPU to a red black tree of a corresponding CPU group according to the priority level and a credit value, and starting all timers set in the initialization stage; and (IV) scheduling the virtual machine. According to the invention, the time slice executed by a logic CPU in the group can be dynamically adjusted, the time consumed by the scheduling method in the execution process is effectively reduced, relatively short response time for the IO (input/output) task is realized, and the method adapts to any type with relatively high requirement on time.

The invention aims to provide an implementing method for network element management system. The system comprises virtualization of input equipment, output equipment, storage equipment and communication resource, each target management object in the network element management system is processed by an independent virtual machine and the virtualized resource is subject to exclusive access. The whole system is composed of human-computer interface management, equipment communication management, virtual machine scheduling management, resource management, system management and one or a plurality of virtual network element supervisors. The processing method for the system comprises: packing network element management executing code for managing specific equipment and data resource to manufacture the virtual network element supervisors; loading the network element supervisors into the network element management system to configure input and output equipments, communication interface, file system and database resource; starting the network element supervisors to operate, communicating with the equipments to be managed by the virtual network element supervisors, and managing the equipments by network element management interface. In the invention, implementation based on virtualization technology is adopted to concentrate network element management function for specific equipments in a single virtual machine, and virtual machines are completely isolated, thus ensuring safety of system, greatly enhancing expandability of system and fully utilizing processing capacity of computer.

The invention discloses a load balancing method based on vector mapping. The method comprises the following steps of: obtaining the load information of physical nodes and virtual machines in a virtual cluster system; obtaining the physical nodes, for which the load condition needs to be regulated, according to the resource information, and calculating a balance degree to obtain the virtual machines to be migrated; converting and normalizing the three-dimensional resource of the virtual machines to be migrated and all the physical nodes, representing the converted and normalized resource in a space coordinate system in the form of three-dimensional vector; projecting the space coordinate system and the resource vector, grouping the projection vectors, and obtaining the group match of the physical nodes and the virtual machines; and for each group match, scheduling with an optimal descending order bin packing virtual machine to get the decision result. The technical scheme of the invention is used in a distributed cluster system, is supportive to differentiated, global and expandable scheduling of multi-dimensional resource, and is featured by high speed of decision making and high accuracy.

The invention provides an energy saving method, an energy saving management server and a system under a desktop virtual environment, wherein the method comprises the following steps of: acquiring configuration information of a plurality of physical servers and resource occupation information of virtual machines running on the physical servers by the energy saving management server; determining the resource occupation type of a corresponding virtual machine according to the resource occupation information; forming a virtual machine group with the plurality of virtual machines which run on different physical servers and have complementary resource occupation types; determining a target physical server for the virtual machine group according to the configuration information; indicating a virtual machine scheduling server to transfer the plurality of virtual machines in the virtual machine group to the target physical server; and when the physical server has no running virtual machine, indicating the physical server to execute the shutdown operation by the energy saving management server. With the method, the server and the system, power consumption is reduced, and an energy saving effect is improved.

The invention discloses a share cache perception-based virtual machine scheduling method and device, and aims to solve the problem that the performance between virtual machines is mutually interfered caused by resource competition of share caches, and thus the operating efficiency of the system is degraded. In some feasible modes of execution of the invention, the method comprises the steps: determining multiple virtual machines in need of scheduling; obtaining the use condition of all the share caches by each virtual machine of the multiple virtual machines in need of scheduling; clustering the multiple virtual machines in need of scheduling into multiple categories according to the use condition of all the share caches by each virtual machine; assigning virtual machines in each category uniformly to multiple nodes in sequence so that the share cache use condition of the one or more than one virtual machine assigned to each node is matched; the multiple virtual machines in need of scheduling are scheduled to respective assigned modes.

The present invention provides a novel cloud environment virtual machine scheduling algorithm, which is capable of improving a utilization rate of cloud platform host resources and maintaining load balance of a system and reducing energy consumption. According to characteristics of virtual machine migration, the present invention separately studies and explains three problems, i.e. selection of a to-be-migrated node, selection of a virtual machine in a node, and selection of a target node, takes system power energy consumption, CPU resources, memory resources and bandwidth resources into comprehensive consideration, and eventually makes a final choice by means of a probability roulette wheel. The method provided by the present invention is capable of improving stability of a cloud platform, and is beneficial for improving operating efficiency and reducing hardware loss.

The invention discloses a multi-field supportable virtual machine dispatching device and a multi-field supportable virtual machine dispatching method. The multi-field supportable virtual machine dispatching device comprises a virtual machine cluster information central access module, a virtual machine dispatching module and a virtual machine dispatching information central output module, wherein the virtual machine cluster information central access module is used for receiving input cluster information, the virtual machine dispatching module is used for analyzing cluster information and generating dispatching policy required by fields, and the virtual machine dispatching information central output module is used for centrally outputting policy information. The virtual machine cluster information central access module receives processing requests of a virtualization cluster system and preprocesses the same into a uniform input information format. The virtual dispatching module classifies states of the virtualization cluster system according to input information, generates corresponding dispatching policy for processing the requests according to classifications and outputs dispatching commands through the information central output module. Dispatching virtualization clusters under multiple fields can be realized well, and the purpose of dispatching under different fields is achieved. By modular design, uniform interfaces are used among modules, transparency in design of dispatching algorithm is improved, and application range is wide.

The invention discloses a group virtual machine scheduling policy for a cloud computing environment. The policy comprises the following steps of S1, establishing a feasible decision space of virtual machine scheduling; S2, minimizing total flow of a network where group virtual machines are located, and establishing an objective function for optimizing the total flow of the network; S3, minimizinga maximum link utilization rate in the network, and establishing an objective function for optimizing the maximum link utilization rate; and S4, establishing an overall objective function, and solvingthe overall objective function in combination with an ant colony algorithm and a simulated annealing algorithm to obtain an optimal solution of the function and mapping relationships between the virtual machines and a physical machine. Under the condition of fully considering resource constraints, the control of the total flow of the network and the balance of flow distribution on a network linkare defined as a combination optimization problem, and solving is performed in combination with the ant colony algorithm and the simulated annealing algorithm. According to the scheduling policy provided by the invention, the performance of the network where the group virtual machines are located can be better optimized; the congestion is reduced; and the service quality of users is effectively improved.

The invention discloses a virtual machine scheduling method based on combination optimization. The method can be used for real-time virtual machine scheduling decisions in a virtual cluster, the starting positions of virtual machines can be decided in real time according to the current resource utilization conditions of the virtual cluster so that the number of occupied physical hosts can be minimum; in this way, the resource utilization rate can be increased, and meanwhile energy consumption can be lowered. According to the virtual machine scheduling method, simple mathematical calculation is adopted, a scheduling result can be obtained within limited comparison frequency, and the comparison frequency is related to the type of the virtual machines and the type of the physical hosts. Compared with an artificial intelligence algorithm adopted in the process of traditionally solving a combination optimization problem, the virtual machine scheduling method has high real-time performance and can decide the positions where the virtual machines are placed in real time and meet the requirement for high real-time performance of a virtual machine scheduling system.

The invention discloses a virtual machine coexisting scheduling method based on processor performance monitoring. The virtual machine coexisting scheduling method comprises the following steps: based on performance events provided by a hardware processor performance monitoring unit and based on pressure values and tolerance degrees of VCPUs (Virtual Central Processing Units) running on a multi-core physical server, calculating to obtain the pressure coefficients and tolerance coefficients of the VCPUs, determining a VPCU to be moved out from high-load physical CPUs (Central Processing Units) according to the pressure coefficients and the tolerance coefficients, and through coexisting coefficient prediction, moving the VCPU to be moved out to a physical CPU with a minimum coexisting coefficient. When the virtual machine coexisting scheduling method disclosed by the invention is used for predicting the performance of the VCPU coexisting combination, VCPU coexisting coefficients are introduced to represent the performance of different VCPU coexisting combinations, a coexisting coefficient prediction method based on the VCPU coexisting combinations is provided, favorable decision information is provided for virtual machine scheduling, a virtual machine in match with the use situation of a shared cache is scheduled to the same CPU of the shared cache for execution, and the purpose of improving the operation efficiency of the virtual machine is achieved.

The invention discloses a virtual machine scheduling method applied to a cloud computing environment. The method comprises: 1) modeling of the virtual machine: pre-creating multiple virtual machines; 2) new task scheduling: searching for virtual machines with processing capabilities matched with tasks based on sizes of tasks among pre-created virtual machines; 3) processing of new tasks: processing tasks by searched virtual machines. The virtual machine scheduling method applied to the cloud computing environment has following beneficial effects: virtual machines can be matched with tasks so that reasonable utilization of cloud resources is achieved; and service quality of cloud computing is increased.

A virtualization apparatus is provided. The virtualization apparatus includes a plurality of virtual machines (VMs), a process scheduler configured to schedule processes to be executed on the respective virtual machines, a virtual machine monitor (VMM) configured to provide each of the virtual machine with a virtualized execution environment, a virtual machine scheduler configured to schedule the virtual machines to run in the virtual machine monitor, and a synchronization unit configured to synchronize a process schedule time which is scheduled by the process scheduler and a virtual machine schedule time which is scheduled by the virtual machine scheduler, or to change the virtual machine schedule time in consideration of the process schedule time.

The embodiment of the invention provides a method and system for dispatching a virtual machine. The method comprises the steps as follows: analyzing a running log of a virtual machine cluster in an offline state to obtain running characteristics of virtual machine operations in the virtual machine cluster; fitting according to the running characteristics of the virtual machine operations to obtain a characteristic curve of the load related parameters of the virtual machine; dividing a running cycle of the virtual machine into a plurality of initial stages according to the characteristic curve and carrying out resource demand calculation on the divided initial stages in the characteristic curve to obtain a first histogram; carrying out optimal adjustment of local resource distribution on the first histogram by calling a dynamic programming algorithm to obtain a second histogram and carrying out final stage division according to the second histogram; and solving a stage division graph model according to the second histogram and determining a virtual machine migration path by solving the stage division graph model. According to the invention, the energy-saving pertinence is achieved and the energy-saving effect can be greatly improved.

The invention discloses a system for quickly constructing a virtual machine in a large-scale scene of a network target range. The system mainly comprises a network target range platform and a virtualmachine scheduling module which are deployed at a network target range control node, and a virtual machine management module and a quick construction module which are deployed at a network target range computing node. According to the method, when a user starts a network target range scene topology, virtual machines with the same dependence software are dispatched to the same computing node preferentially, and CDROM equipment is added to the virtual machines when the virtual machines are started; repeated software packages are filtered according to the rapid construction information of the virtual machine, dependent software packages are downloaded, and an iso file is generated; and a CDROM path mounted on the virtual machine is modified after the virtual machine is started, and virtual machines with the same information is quickly constructed on the same computing node to share one iso file. According to the method, agent-free construction of the virtual machine in a large-scale sceneof a network target range can be realized, various initialization files can be supported, and network congestion is avoided.