84 results about "Cloud scheduling" patented technology

The invention relates to cloud platform technology, solves the problems of low task dispatching speed and low resource utilization rate of the conventional cloud platform during scheduling, and provides a cloud scheduling system, a cloud scheduling method and a multistage cloud scheduling system. The invention adopts the technical scheme that: the cloud scheduling system comprises a task transmitting end, a scheduler, a manager, n virtual machines and an external network server, wherein the scheduler is connected with the manager; the n virtual machines are connected with the scheduler; the task transmitting end is connected with the manager; the n virtual machines are connected with the manager; the external network server is connected with the n virtual machines; and n is a positive integer. The cloud scheduling system has the advantages that the system can schedule large-scale tasks efficiently at a high speed and is applicable to the cloud platform.

The invention belongs to the technical field of cloud computing and network multimedia, and particularly provides an auto-expanding/shrinking cost-optimized content distribution service method based on a hybrid cloud scheduling model. The method comprises: a future number of user visits is predicted on the basis of historical data and provides basis for auto-expanding/shrinking of resources; according to the predicted value and a long-term scheduling algorithm, a rough plan of resource booking strategy is obtained through calculation; wherein a short-term scheduling model is introduced to reduce the prediction error and to improve the precision of resource supply and the quality of service. In the long-term scheduling algorithm, a locality-aware booking model is set up to derive the locality-aware resource booking algorithm. A resource prediction algorithm employs the ARIMA model. In a short-term adjustment algorithm, virtual machine status parameters are designed and a content missing algorithm is provided, so that the user experience of the entire system is further improved. The method enables hybrid cloud technology to support streaming media content distribution applications efficiently with auto-expanding/shrinking functions and optimized costs.

The invention relates to the technical field of computers. The invention discloses a GPU service redirection system and method based on cloud and terminal GPU fusion. A redirection module is deployedin a virtual machine on a cloud computing server; a scheduling module is deployed on the cloud scheduling server; The method comprises the following steps: deploying a service module on a cloud computing server configured with a GPU; and deploying a service module and a window management module on the terminal computer, and redirecting a graphics library API called by the virtual machine to a cloud computing server or the terminal computer configured with a GPU according to a fusion scheduling rule to provide GPU acceleration support for graphic intensive applications running in the virtual machine. According to the invention, the unification and sharing concept of the cloud and the terminal GPU is utilized; According to the GPU redirection technology, the GPU resource scheduling strategyand the application window and cloud desktop fusion method, effective and low-cost graph acceleration support is provided for a cloud desktop solution, the computing load of a cloud server is reduced,and the resource utilization rate of a display card of a terminal computer is increased.

The invention provides a cloud scheduling system based on mass video structured processing. The cloud scheduling system comprises a cloud scheduling management module and a plurality of analysis modules, wherein the cloud scheduling management module conducts logic cutting for massive videos upon reception of tasks, and allocates the massive video subjected to logic cut to the analyzing node of each analysis module according to the load capability of each analysis module. Certain cloud scheduling mechanism is combined with data intelligent cutting and analyzing technology to effectively improve the time-effectiveness and stability of massive video processing.

The invention provides a cloud scheduling method and a system thereof. The method comprises receiving a cloud scheduling mission, acquiring a user level and a priority attribute setting information of the cloud scheduling mission; dividing the cloud scheduling mission into a plurality of primary missions; respectively scheduling corresponding hardware resources and software resources in a cloud network system to conduct parallel processing to the plurality of primary missions according to the user level and the priority attribute setting information and obtaining processing results; and combining the processing results of the plurality of primary missions and generating a processing result of the cloud scheduling mission. According to the cloud scheduling method and the system thereof, the cloud scheduling mission can be divided into the plurality of primary missions in parallel processing, the different hardware resources and the software resources are scheduled according to requirements for different cloud scheduling missions, so that processing efficiency is improved and the requirements for individualized services for the different cloud scheduling missions are satisfied.

The invention discloses a cloud scheduling method for supercomputing resources. The method is suitable for being executed in a cloud scheduling center. The cloud scheduling center is connected with the client. The supercomputing center is connected with other cloud scheduling centers. The supercomputing center comprises at least one computing node and one resource computing node. The method comprises the following steps: receiving a job submission instruction from a client, and the instruction comprises attribute information of a submitted job and a performance index for executing the job; retrieving at least one supercomputing center according to the performance indexes; obtaining resource use information of the supercomputing center from the retrieved resource computing nodes of the supercomputing centers; at least one supercomputing center used for executing the job is matched from the retrieved supercomputing centers in combination with the attribute information of the job and theacquired resource use information; and submitting the job to the matched supercomputing center. The invention also discloses a corresponding cloud scheduling center and system.

The invention discloses a two-stage media edge cloud scheduling method and a two-stage media edge cloud scheduling device, belonging to the field of multimedia network communication. According to the method and device, an online service migration method or a dynamic content deployment method is flexibly adopted to perform resource scheduling according to different variation amplitudes of a user access mode; meanwhile, service flow influenced by the adjustment of a film deployment can be processed by utilizing the online service migration method in the dynamic content deployment process. Therefore, by virtue of the cooperation of the online service migration method and the dynamic content deployment method, firstly, the smooth transition of the scheduling process from the migration stage to the deployment stage can be realized in the process that the user access mode is varied along with the time, and the high request acceptance rate of media edge cloud can be guaranteed by virtue of reasonable cost; secondly, complete transparency of the scheduling process to the user service can be realized; thirdly, the influence of the jitter of the user access mode on the request acceptance rate can be effectively avoided.

The invention provides an intelligent traffic self-networking system with a cloudlet function. The system comprises an intelligent traffic information acquisition and analysis node and a cloudlet scheduling center; the intelligent traffic information node utilizes an RFID (Radio Frequency Identification Device) as a data acquisition technology, the data acquisition technology is compared and fused with a GPS (Global Positioning System) data acquisition technology and a data processing analysis platform (such as a GIS (Geographic Information System)); the automatic self-networking function is realized by utilizing a cloudlet self-networking circuit; and information processing, analysis and forwarding of a single node and information comparison processing of multiple nodes can be realized through processing and calculating of a cloudlet topological structure so as to carry out functions of flow control, path planning, fault analysis, time prediction and the like. The nodes are communicated with the cloudlet scheduling center through a network and a radio so that section control is realized and the scheduling center processes priority levels of various kinds of information on the intelligent traffic information node.

The invention relates to a publishing method based on a distributed storage container cloud application. The publishing method comprises the following steps: 1) constructing a container cloud platform, wherein the container cloud platform comprises a distributed storage server, a container cloud management sub-platform and a container cloud scheduling system; 2) creating a directory: creating a shared storage specification directory in a distributed storage server for storing a program package; 3) uploading the program package: uploading the program package required by the application to a path of a shared storage specification directory; and 4) application release: mounting the distributed storage in a corresponding container instance in a volume form, copying a program package in the volume into the container by executing an application starting script when the application release is started, performing related program package decompression processing, and finally executing an application starting command to complete the application release. Compared with the prior art, the method has the advantages that distributed storage and centralized management are achieved, storage resources are saved, program package decompression processes are not exclusive to each other, and the release efficiency is improved.

The invention discloses a shared cloud intelligent garbage can operation management system capable of walking autonomously. The system comprises: a user terminal which comprises an owner terminal anda sorter terminal; client software that is used for realizing cloud scheduling sharing of the cloud intelligent garbage can and user management at the user terminal, and receiving the operation of theowner terminal and the operation of the sorter terminal; server software that is used for receiving a command request of the client software and feeding back an execution state to the client software; cloud intelligent garbage can control software that is used for receiving a scheduling instruction sent by the server software and feeding back an execution condition to the server software; the shared cloud intelligent garbage can that can autonomously enter and exit the elevator according to an instruction and collect garbage in a door-to-door classified manner; a cloud elevator intelligent garbage can elevator riding system that comprises an elevator and a cloud background. According to the invention, garbage classification can be effectively implemented, and all building owners can stillconveniently throw garbage in a classified manner in various regions and under various weather conditions throughout the country, and the whole system can be popularized.

The invention discloses a cloud scheduling sharing method of social idle vehicles. The cloud scheduling sharing method comprises the following steps: 1, building a remote automatic monitoring system of shared vehicles; 2, generating a to-be-rent vehicle management list; 3, selecting a vehicle by a demand side and paying money for the vehicle, and then automatically freezing a state of the selectedvehicle in the to-be-rent vehicle management list; 4, starting the vehicle, and automatically performing whole-process real-time monitoring on the vehicles; 5, sending an instruction of car-turn success by a remote monitoring center module only when the demand side parks in a recommended stopping point; 6, performing multi-attribute labeling on various module data information through a data processing system and carrying out selection and arrangement and statistic analysis; 7, automatically releasing the frozen state of the selected vehicle in the to-be-rent vehicle management list after thevehicle is returned successfully. The invention aims to provide a cloud scheduling method of temporarily transferring vehicles and sharing social vehicle resources through a social idle vehicle operating company.

The embodiment of the invention relates to an unmanned driving based garbage classification cleaning method which comprises the following steps of receiving reservation information sent by a client bya cloud dispatching platform; generating a dispatching instruction according to state information and reservation information fed back by a vehicle, and sending the dispatching instruction to a target vehicle through a vehicle ID; enabling the target vehicle to drive according to the dispatching instruction, and judging whether the target vehicle arrives a designated sweeping position or not is according to self position information and sweeping position information; when the target vehicle arrives the designated sweeping position, generating arrival information, and sending the arrival information to the cloud dispatching platform; and verifying a user identity, and when verification passes, opening a garbage cabin door corresponding to the garbage classification information. The unmanned driving based garbage classification cleaning method is suitable for the condition without a garbage can in an area; when a user wants to throw garbage, the user can make the reservation through mobile equipment, the vehicle can arrive the designated place for garbage collection in a reserved time period, a work place is more flexible, and an application range is wider.

The invention provides a quick start method which is applied in a terminal. The method includes the steps of uploading a first list of starting sequence of starting items of the terminal to cloud, wherein starting time of the first list based on the starting sequence is a first time, receiving cloud feedback, acquiring a second list, wherein starting time based on the second list is a second time, and the first time is longer than the second time, and starting according to the second list. According to the method, starting speed of the terminal can be increased only by a user uploading starting sequence list of the terminal to the terminal, when in starting, loading of new starting items is achieved only by cloud scheduling, without downloading new software to mend, and thus the operation is simple and practical. The invention further provides the terminal.

The invention relates to the technical field of network cloud safety protection and specifically relates to an anti-large-scale DDoS attack defense system based on cloud + end device continuous linkage mode, which comprises a cloud platform defense system and a local end defense system. The cloud platform defense system comprises a cloud scheduling center, a cloud cleaning node, a DNS near-sourceresolution server. Compared with the single local hardware end device defense, the system in the invention can use the advantages of large-capacity cleaning capability of the cloud distributed cleaning nodes and break through the anti-D defense bottle neck of local bandwidth limit; compared with the single continuous access cloud defense, cloud node distributed cleaning and secondary fine filtering of the local end defense system can be used to achieve more comprehensive DDoS attack defense and better cleaning effect; compared with the single continuous access cloud defense, the system and themethod in the invention convert the data between the cloud cleaning node device and the local end defense system based on four-layer protocol port processing and based on a seven-layer protocol proxyprocessing mode, which is more rapid.

A method and system for schedule Internet of things resource computer-readable storage media and terminals, The method comprises: a cloud scheduling center uses historical Internet of Things resourcescheduling information training to generate cloud resource prediction model, and optimizes the cloud resource prediction model by using the corresponding computing performance and storage performanceof edge nodes, and generates edge node resource prediction model corresponding to each edge node, and sends it to the corresponding edge node step by step; a plurality of edge nodes respectively receive their own corresponding edge node resource prediction models; the information of the user's resource usage in the current cycle is obtained, and the resource demand in the next cycle is forecaste by using its own edge node resource forecasting model; the resource of the next cycle is allocated to the user based on the information of the predicted resource demand of the next cycle. The scheme can improve the speed and efficiency of resource scheduling in the Internet of Things.

The invention discloses a method for removing tight coupling of a use case and an environment. The method comprises the steps of: scanning a stored test task periodically, triggering the test task to run if a trigger condition is met, and placing the use case of the test task in a waiting queue; and matching resources required by the use case in the waiting queue with physical resources of the test environment when a use case request sent by the test environment is received, and sending the use case in the waiting queue to the test environment if the physical resources of the test environment meet requirement of the use case. The invention further discloses a device for implementing the method and a cloud scheduling system at the same time.

The embodiment of the invention provides a multi-level data center master-slave switching control method and a cloud scheduling director, which are used for solving the technical problem of relativelylow switching efficiency during master-slave equipment switching in the prior art. The cloud scheduling director obtains scene state information of a first node in the first data center, wherein thefirst data center is any level of data center in the multiple levels of data centers, the first node is any one of a plurality of nodes included in the first data center, and the scene state information is used for indicating the operation health state of a main device of the first node in a cloud scheduling switching scene; based on the scene state information, judging whether the main equipmentof the first node has a fault or not; and if yes, switching the master device of the first node to a first slave device in the plurality of slave devices of the first node, the first slave device being any one of the plurality of slave devices of the first node.

The invention discloses a water conservancy model cloud computing method and device based on a model combination, and relates to the technical field of model calculation.The water conservancy model cloud computing method and device based on the model combination are based on an OpenMI standard and workflow, establishment, editing, viewing and verification to the model combination can be provided, and model combination facing business requirements and facing different scales, different spaces and different dimensions is conducted on a model component subjected to standard processing; meanwhile, the parallel processing efficiency of the water conservancy model combination is improved through a cloud scheduling method, more complex application scenes are met, and complex calculation requirements on high availability and high expansibility at present and in the future can be better adapted.

The present invention discloses a cloud scheduling management method which comprises a step of receiving the scheduling request sent by a main service node and obtaining the virtual machine node information in the scheduling request, and a step of sending the virtual machine node information to a data center and requesting the data center to carry out the operation of creating or closing a virtual machine node. The invention also discloses a cloud scheduling management center and a cloud scheduling management system. According to the cloud scheduling management method, the management center and the management system, the dynamic allocation of resources is realized, the resource waste caused by excessive deployment is avoided, and the problem of a low resource utilization rate is solved.

The invention discloses a cloud scheduling method and device for a delivery robot, and a server. The method comprises the following steps: aggregating waybills in a waybill pool according to a similarity, and forming a waybill package set including at least one optimized waybill package; scheduling and sorting all optimized waybill packages in the waybill package set; and sequentially distributingthe optimized waybill packages to the delivery robot according to a sorting result of the optimized waybill packages. According to the invention, order pressing, order merging and waybill package sorting are carried out on multiple delivery orders from the perspective of waybill scheduling, screening, order superaddition and order redistribution are carried out on multiple robots from the perspective of transport capacity scheduling, order superaddition and order redistribution are carried out in sequence so that a waybill aggregation result is maximized, a waybill carrying amount of the robots is increased, and the operation cost of robot delivery in practice is reduced; and meanwhile, the delivery efficiency and the service experience are improved.

The invention discloses an identity recognition system for a bus driver, and the system comprises a cloud scheduling module which is used for determining a vehicle identification according to the identification information transmitted by an identity authentication controller, and transmitting the vehicle identification and verification information to the identity authentication controller; matching the vehicle identifier with the identity information sent by the intelligent terminal, and if matching succeeds, sending verification information to the intelligent terminal; an intelligent terminal which is used for collecting identity information of a vehicle driver and sending the received verification information to the identity authentication controller; and an identity authentication controller which is used for comparing the verification information sent by the cloud scheduling module with the verification information sent by the intelligent terminal, wherein if the verification information sent by the cloud scheduling module and the verification information sent by the intelligent terminal are the same, identity authentication of the driver succeeds, and a whole vehicle activation instruction is sent to the whole vehicle control unit. According to the system, the driving management and control of the bus can be effectively improved, and meanwhile, the mobile phone of the driver is used for biological recognition, so that the system cost is reduced.

The invention provides a cloud optimization scheduling method supporting priority and antiaffinity. The method is used for solving the problem that the anti-affinity constraint and priority constraintof an existing cloud scheduling system on tasks cannot be effectively met, converting the anti-affinity and priority representation of task cluster information into a graph model language in a unified format, solving a maximum flow graph, and carrying out task deployment according to an obtained optimal solution. The method has the characteristics of multi-dimensional resource representation andgood expansibility, can improve the capability of meeting constraints between tasks and physical resources in current cloud optimization scheduling management, and provides more task scheduling opportunities. The invention further provides a cloud optimization scheduling system supporting priority and antiaffinity. The cloud optimization scheduling system comprises a cloud cluster manager, an antiaffinity rule parser, a priority rule parser, an optimization scheduling model generator and an optimization scheduling model solver.

The invention relates to the field of intelligent automobile wash equipment, in particular to an automobile wash service system based on cloud control and a control method thereof. The automobile washservice system based on cloud control is characterized in that intelligent wash equipment and a control system are arranged on a wash service platform; the control system is connected with the intelligent wash equipment and feeds working state of the wash service platform back to a cloud scheduling platform; a client platform is docked to the cloud scheduling platform to acquire working state information of the wash service platform and to start and/or stop the wash service platform; the cloud scheduling platform provides unified monitoring for the working state of the wash service platform and feeds monitoring information back to the client platform. The automobile wash service system based on cloud control and the control method thereof have the advantages that a user gains real-time understanding of the information of the wash service platform through a cellphone, settlements can be made automatically, a car wash process has no need for manual intervention, and car wash services can be better enjoyed.