233results about How to "To achieve load balancing" patented technology

This invention publishes a locked distribution method of shared resources in a type of computer integration system including: making more than one node of system members as locking management node of all the shared resources in the system, meanwhile each shared resource matches a locking management node; when one apply asks for applicant or release shored resources lock, and send applicant to the corresponding locking management node of this shored resources, the corresponding locking management node will finish the distribution and release of lock's applicant. This invention also publishes the corresponding structure of computer integration system. The invention can be used for realizing high usable locking service in the integration, and it has a high expansibility and realizes a balance of load.

Disclosed is an operation cross-domain control method in a grid computing environment; the method uses a trust mechanism to realize the operation control in the grid environment, makes trust evaluation to the available resources in the grid, analyzes the operation to be processed, and uses a mobile agent to transfer the operation to an appropriate resource for execution, according to the information provided by a control system in the grid. The proposal overcomes the disadvantages of poor reliability, non-guaranteed response time, over-long running time on the resources and imperfect fault management in other operation control proposals, and can achieve self-adaptation to the resource and operation control in the grid, reduce the grid traffic and improve the utilization rate of the network to form an operational parallel solution, thereby achieving the purpose of improving the utilization efficiency of grid resources and the execution efficiency of grid computing, speeding up the execution of tasks and improving the accuracy of the results so as to enhance the processing efficiency of distributed systems.

The invention discloses a memory shared system and a device, as well as a method in a multi-core NUMA system. The system comprises a plurality of nodes; the operating system of each node comprises a memory shared device, the memory shared device comprises: a memory information collection and processing module, a memory application module, as well as a module establishing a distant end memory dynamic management structure and connected with a communication module. The method comprises the following steps: step S1, when the memory of a front node is deficient, the memory deficiency information is sent to other nodes, the node information of vacant memory in other nodes is collected, then the memory application information is sent to the nodes with vacant memory requesting the use of distant end memory; step S2, after receiving the memory application information, the nodes with vacant memory distributes the distant end memory with the current node, and let the current node dynamically manages and exclusively uses the distant end memory. The invention borrows the distant end vacant memory to accomplish the load leveling of the entire system.

The invention provides an AEDF task scheduling method based on an SMP, and designs a dynamic load balancing method which runs on the SMP and is based on self-adapting. The method can track the current load condition of all the processors in the SMP in real time, and reasonably allocates different tasks into the processor with less loads in a dynamic way; in the process of task scheduling, by counting the runtime of the hung task, the method regulates time slot, period and blocking time of the next time of running in a dynamic way. The AEDF task scheduling method based on the SMP can comprehensively analyze the current load condition of the processors, task-fulfilling situation and the running condition of the system, and reasonably allocates resources of the processors, thus effectively improving the utilization rate of the resources of all the processors in the SMP, shortening execution time needed by completing all the tasks, and leading the operating efficiency and the using effectiveness of the SMP to be effectively improved; meanwhile, by adopting the execution sequence of dynamic scheduling task resources, the real-time requirements for executing urgent task can be guaranteed in a preferential way.

The invention discloses a cloud storage method for a cloud computing server, belonging to the technical field of computer networks. The method comprises the following steps: (1) a cloud computing server group server and a client registers on a middleware server respectively; (2) the middleware server searches a local database and looks up an operable file and a usable cloud computing server group for a current user; (3) the user selects an upload file, fragments the file and transmits the file fragments to the middleware server; (4) the middleware server verifies the upload file, requires the user to retransmit if the upload file is wrong, and transmits data to the cloud computing server group usable for the user if the upload file is correct; (5) the cloud computer server group rearranges and stores the received data and sends a success message to the middleware server; and (6) the middleware server establishes mapping between the name of the upload file and the user. According to the invention, the correctness and integrity of a data transmitting process and the privacy of data are ensured, and high reliability of normal work of the server group is ensured.

The invention relates to an operation scheduling system which is applicable to the grid environment and based on the reliability cost; as indicated in graph 1, the whole system includes three layers: the first layer is an operation submission interface module 1; the second layer is an operation scheduling module 2 and the grid resource platform 7 as the substrate layer. From the perspective of the operational principle, the core of the invention is the operation scheduling module in the second layer, which includes a pre-scheduling module 3, a scheduling strategy module 4, an operation finish time prediction module 5 and a resource information module 6. The operation scheduling system in the invention proposes an operation running time prediction model and a resource usability prediction model; the operation running time prediction model based on the mathematical model and the resource usability prediction model based on the Markov model have high accuracy and high generality. The operation scheduling system adopts the copy fault-tolerance strategy, the primary copy asynchronous operation fault-tolerance strategy and the retry fault-tolerance respectively according to different operation service quality requirements and resource characteristics so that the operation scheduling system has high flexibility and high validity; meanwhile, the operation scheduling system supports the computation-intensive operation and the data-intensive operation to have good generality. Compared with the scheduling system in the prior art, the operation scheduling system has the advantages of supporting more concurrent users, improving the resource utilization rate, good generality, good extensibility and high system throughput.

The invention provides a hadoop-platform-based method for improving video transcoding efficiency and relates to hadoop-platform-based technologies for improving video adaptive technology efficiency. The hadoop-platform-based method for improving the video transcoding efficiency comprises subjecting a video file to storage segmentation processing through a hadoop platform; subjecting video clips generated by segmentation to FFmpeg (Fast forward moving pictures experts group) transcoding; and consigning transcoding results to an HDFS (Hadoop Distributed File System) for video clip combination, and accordingly transcoding tasks of a large amount of video data can be completed efficiently. According to the efficient video transcoding of the hadoop-platform-based video adaptive technology, the video transcoding speed is fast, the network bandwidth traffic consumption is low, the load balancing processing is appropriate and the transcoding time consumption is reduced significantly.

The present invention discloses a transmission method in a wireless relay system, which jointly transmits data between a source node and a destination node through a plurality of relay stations. These relay stations transmit same data on the same time frequency resources and transmit different data on varied time frequency resources. The destination node demodulates and decodes received superposed signals of the same data transmitted by the RS on the same time frequency resource. In addition, the destination node demodulates and decodes different data transmitted by the RS on the varied time frequency resource and matches demodulated and decoded data on demand and finally acquires the transmitted data. Moreover, the present invention discloses a wireless relay transmission system. The transmission method and the transmission system provided by the present invention can improve link transmission reliability, realize load balance amongst a plurality of relay stations, avoid overload of single relay station and enhance service quality for the destination node.

A predictive system and method for aircraft load balancing and handoff management leverages the aircraft flight plan as well as channel occupancy and loading information. Several novel techniques are applied to the load balancing and handoff management problem: Use of aircraft position and flight plan information to geographically and temporally predict the appropriate ground stations that the aircraft should connect to for handoff, and monitoring the load of ground stations and using the ground-requested, aircraft initiated handoff procedure to influence the aircraft to connect to lightly loaded ground stations.

A load balancer is provided wherein packets are transmitted to a burst distributor and a hash splitter. The burst distributor consults a flow table to make a determination as to which forwarding engine will receive the packet, and if the flow table is full, returns an invalid forwarding engine. A selector sends the packet to the forwarding engine returned by the burst distributor, unless the burst distributor returns an invalid forwarding engine, in which case the selector sends the packet to the forwarding engine selected by the hash splitter. The system is scalable by adding additional burst distributors and using a hash splitter to determine which burst distributor receives a packet.

The invention provides a wireless access point accessing method which comprises the steps of receiving a detection requesting frame sent by a terminal to judge whether self-capacity is larger than a preset threshold, when the self-capacity is larger than the preset threshold and the detection requesting frame meets preset conditions, calculating sending parameters, and sending detection response frame to the terminal according to the sending parameters, so that the terminal can access to wireless access points with light load according to the sending parameters. The wireless access point accessing method can actively inform current capacity situation of the terminal, and the terminal can access into access points (AP) with light load according to sending parameters of different APs, thereby avoiding the fact that the terminal sends detection requesting frame to APs with heavy load for multiple times and is rejected for multiple times, further balancing load among APs, and avoiding the fact that users cannot get access to the internet for a long time.

The invention discloses a multi-network-based data transmission method and a multi-network-based data transmission device. The multi-network-based data transmission method comprises the following steps: when a terminal receives a data transmission request of a user, querying the load conditions of all networks in sequence according to the decreasing sequence of the transmission performances of all the networks by the terminal; when the load condition of a current network is queried to be smaller than a first load threshold, regarding the current network as an access network by the terminal; and establishing a connection between the access network and the terminal and carrying out data transmission through the access network. According to the invention, the problem that multi-mode terminals in relevant technologies cannot use a network with a superior transmission performance is solved, and the terminal can select an appropriate network according to the load conditions of the networks, so that the network access quality of the terminal is increased and the load balance among heterogeneous networks is realized.

The invention discloses a sparse matrix data storage method based on a ground power unit (GPU). The method comprises the following steps of: 1), sequencing the line length array length [] according to ascending order; 2), classifying the array length [] into four sections of [0, 8), [8, 16), [16, 32), [32, +infinity) according to the number of every line of non-zero element; respectively combining the 32nd, 16th, 8th, 4th lines in every section; 3], zeroizing the line in every data section and performing the line filling operation on every data section, wherein the element of the filled line is zero completely; 4], generating three one-dimensional arrays of cval [], ccol_ind [], crow_ptr [] of the SC-CSR format. In the method of the invention, the line length change amplitude of every line is reduced via segment treatment, thereby reducing the load unbalance between the thread bunch and the thread block; the adjacent lines are staggered and combined to avoid the resource waste of the thread bunch calculation when the non-zero element is less than 32, and to improve the efficiency of joint access of the CUDA display memory and decrease the step of calculating kernel and reducing lines, and therefore obviously improving the calculating performance of the vector multiplication of the sparse matrix.

The invention discloses a virtual machine cluster resource allocation and scheduling method. The method comprises the following steps: S1, seeking a load alert time point of a virtual machine from a historical load database, wherein the load alert time point is a moment when a virtual machine resource utilization Dvm of more than one virtual machine reaches an alert value ALR; S2, presetting a first time length mu, starting a physical machine at a moment mu before the load alert time point, wherein the added physical resource SN of the newly started physical machine is required to be greater than N*Mvm*A2, and migrating the virtual machine to the newly started physical machine. By using the virtual machine cluster resource allocation and scheduling method, the load alert time point is determined according to historical data, and the physical machine is started in advance to perform shunt migration on the virtual machine to guarantee that the resource utilization rate of each virtual machine is reduced in advance; the physical machine is started in advance and the virtual machine is smoothly migrated, so that the purpose of load balance is achieved.

The invention discloses a parallel method for simulating and optimizing the computer architecture scheme based on a cluster system, and aims to provide a parallel method for simulating and optimizing the design scheme of the computer architecture. The technical scheme is that a parallel computer system which consists of a main control node and simulation nodes and is provided with a remote command execution environment is firstly built, and a global configuration program, a simulated configuration file generating program, a task dispatching program and a result analyzing program are arranged on the main control node, wherein, the global configuration program is used for arranging global configuration; the simulated configuration file generating program is used for generating all simulated configuration files; the task dispatching program distributes simulation evaluation tasks to each node, controls each simulation node and performs simulation evaluation; and the result analyzing program searches simulation result files sent from the simulation nodes for statistics, screens out optimal configuration parameter values, and outputs a report. By adopting the invention, the time for evaluation and optimization can be reduced, and the selection accuracy is improved.

The invention provides a method and a device for realizing multi-core parallel at a TEE side, a computer system and the like. The method comprises the following steps: a TEE creates a plurality of sub-threads, wherein the sub-threads are used for realizing sub-functions of a TA deployed at a TEE side; for each sub-thread, the TEE triggers a rich execution environment (REE) to generate a shadow thread corresponding to the sub-thread, and the operation of the shadow thread promotes a core running the shadow thread to enter the TEE; and the TEE schedules the created sub-thread to the core where the corresponding shadow thread is located for execution. By utilizing the method, a plurality of service logics in the service with high performance requirements can be operated in parallel in the TEE, and the TEE triggers the REE to generate the thread and automatically enters the TEE side, so that active kernel adding at the TEE side is realized, and the parallel flexibility of the TEE side is improved.

The invention provides a load balancing method of a concentrated-type wireless local area network and an access controller. The method includes the steps of obtaining a wireless access point with a load larger than a preset first overload threshold value, controlling the wireless access point to stop responding to an access request of a new user terminal, obtaining a candidate wireless access point in an accessed user terminal list which records information of correlated candidate wireless access points of user terminals, obtaining a new wireless access point with a load meeting a preset threshold value of the new user terminal from the candidate wireless access point, removing the user terminal from the wireless access point accessed currently and controlling the new wireless access point to have access to the user terminal. When a network topological structure is changed, loads of APs are adjusted in time, and loads in the whole network are balanced.

The present invention discloses a multiple customer terminals interactive load balacing method and its system. It is implemented by using several customer terminals, a process front-end and several logic operation servers and their connection, in which the system loads can be distributed on several logic operation servers, according to the load condition the quantity of servers can be increased or reduce,d and the state information of every logic operation server is stored in the process front-end, and the curtomer can directly create connection with correspondent logic operation server according to the indication of process front-end, and the process front-end can be connected with every logic operation server by means of local area network of interconnection network. It can implement load balance, and its logic operation servers can be separately distributed so as to can fully utilize bandwidth of network.

The invention discloses a parallel task dynamical allocation method which is suitable for a parallel system including a multilevel user process. The multilevel user process comprises at least two levels of main processes and common processes. The method includes that the main processes arrange parallel tasks to be allocated to form task queues to be allocated at the local level, task request information and the number of requested tasks sent by a next level of user processes are received, beginning from heads of the task queues to be allocated at the local level, the tasks are allocated to a next level of user processes according to the number of the requested tasks of the next level of user processes, and the task queues to be allocated at the local level are updated. The parallel task dynamical allocation method has the advantages that the efficiency of dynamical task allocation can be improved, and the load balance among mass computing resources is achieved.

Load-balancing a distributed system by replacing overloaded servers, including the steps of retrieving, by an assembling device using a fragment pull protocol, erasure-coded fragments associated with segments, from a set of fractional-storage servers. Occasionally, while retrieving the fragments, identifying at least one server from the set that is loaded to a degree requiring replacement, and replacing, using the fragment pull protocol, the identified server with a substitute server that is not loaded to the degree requiring replacement. Wherein the substitute server and the remaining servers of the set are capable of delivering enough erasure-coded fragments in the course of reconstructing the segments.

A wireless network system and a wireless access point (AP) device and a wireless terminal device thereof are provided. The wireless network system includes at least a wireless AP device and at least a wireless terminal device. Each wireless AP device broadcasts a beacon including a load state content of the wireless AP device. Each wireless terminal device receives beacons of all wireless AP devices, and ranks load states of all wireless AP devices in a load list according to at least CPU utilization rates in the load state contents of all wireless AP devices respectively. When a wireless terminal device intends to establish a connection with one of the wireless AP devices, the wireless terminal devices searches through the load list to select a wireless AP device being in a low load state, and transmits a connection request message to the selected wireless AP device.

The invention discloses a partitioned management method for a multi-tenant shared data table, a server and a system. The method includes the steps that the inquiry expenditure for data of each tenant in the shared data table is recorded, wherein the data of each tenant are stored in one partition in the shared data table; statistics is conducted on the inquiry expenditure of each partition in the shared data table within a preset time period; the average inquiry expenditure of each partition is calculated according to the corresponding inquiry expenditure and the preset time period; whether the average inquiry expenditures exceed a preset inquiry expenditure threshold value or not is identified; a new partition is built for the shared data table to migrate the data of one or more tenants to the newly built partition by responding to the condition that the average inquiry expenditure exceeds the preset inquiry expenditure threshold value. According to the technical scheme, load balance of the partition data is achieved according to the data scale and inquiry frequency of tenant differentiation, and the data inquiry performance is promoted.