296results about How to "Improve load balancing" patented technology

A method and apparatus for dynamically balancing the loading of data storage facilities is described. A listing is acquired of locations and loading of all segments of a requested data object including all copies of the segments of the requested data object. Those storage devices containing copies of each segment of the data object having a least loading is selected. If the loading of the storage devices is greater than the maximum loading for the storage devices, the segment is designated to be copied. The presence of all segments of the requested data object is determined. If there are missing segments of the requested data object, each of those missing segments is assigned a file identification and file location, such that those missing segments are assigned to data storage devices having the least loading. The missing segments are retrieved from a back-up storage device.

Disclosed is a system for routing print jobs to one of a plurality of printers or print queues, such as a logical print queue. After providing a print job, a set of printers capable of printing the print job is determined. The set of printers is queried to determine status information thereon. A criterion is then applied to the status information for the set of printers to determine a subset of printers. The determined subset is dependent upon the determined status information. A printer is selected from the subset and the print job is routed to the selected printer.

The embodiment of the invention discloses a method and a device for controlling the carrier frequency of a multi-carrier/subdistrict system. The method comprises the following steps: receiving the signal path quality indicating CQI information of the carrier frequency, which is reported by a terminal; carrying out deactivation or activation judgment for the carrier frequency according to the signal path quality indicating CQI information; and ordering the terminal to deactivate or activate the carrier frequency according to the judgment result. The method flexibly activates and deactivates a certain carrier frequency according to carrier frequency quality reporting and threshold comparison in the multi-carrier/subdistrict system, thereby enhancing the system capacity and having better load balancing effect.

The invention relates to a resource scheduling method based on kubernetes. The method comprises steps of calculating all Node scores in the cluster according to a fixed rule to generate a first node priority queue, and obtaining a Pod priority queue by using a dynamic priority algorithm, wherein the two queues filter Node which cannot be scheduled to generate a second node priority queue; selecting the node with the highest priority from the Pod priority queues to bind with the Pod popped by the Pod priority queues; and if the binding succeeds, entering the next Pod scheduling cycle, if the binding fails, adopting a built-in priority algorithm to optimize Node binding from the second node priority queue, and if the binding fails again, enabling the Pod to run without an appropriate Node, and entering the next Pod scheduling cycle. The method comprises the steps of static scheduling and dynamic resource load balancing, scheduling efficiency is improved, task deployment efficiency is improved, task operation integrity and load balance of the whole cluster are improved, load balance of the cluster is actively adjusted, and resource utilization efficiency of the cluster is improved.

A technique dynamically splits Traffic Engineering (TE) Label Switched Paths (LSPs) in a computer network. According to the novel technique, a head-end node may determine that a TE-LSP to a destination needs to be sized to a larger bandwidth (a “larger” TE-LSP) than currently available over a single path to the destination (e.g., a path that may also be required to meet other constraints, such as cost, delay, etc.). In response, the head-end node may dynamically “split” the larger TE-LSP, and create a first split TE-LSP over a best (e.g., shortest) available path (e.g., that meets other constraints). The first split TE-LSP may reserve a maximum available bandwidth for that best available path. The head-end node may then continue recursively creating subsequent split TE-LSPs for any remaining bandwidth of the larger TE-LSP over available paths until the larger TE-LSP may no longer be split (e.g., all bandwidth has been placed, configurable maximum number of splits reached, etc.). Notably, the split TE-LSPs may be “re-groomed” at a later time to recombine and reorganize the split TE-LSPs.

Embodiments of the present invention provide a system, method, and computer program product that enables applications transferring data packets over a network to a multi-processing system to choose how the data packets are going to be processed by, e.g., allowing the applications to pre-assign connections to a particular network thread and migrate a connection from one network thread to another network thread without putting the connection into an inconsistent state.

The invention discloses a file storing system based on a dynamic transcript strategy. The file storing system comprises a file popular degree calculating module, a plurality of storage modules, a master control module and an encoding and decoding module. The file popular degree calculating module is used for calculating the popular degree of files stored within the periodic time and the average popular degree of all the files stored within the periodic time. The storage modules are used for storing and having access to data. The file popular degree calculating module and the storage modules are respectively connected with the master control module, and the master control module is used for dynamically adjusting the number of transcripts of the storage files according to the popular degree of the storage files. The encoding and decoding module is connected with the master control module and used for carrying out erasure code encoding on the storage files conforming to the encoding conditions according to the number of the transcripts of the storage files. The invention further discloses a file storing method. According to the file storing system and the file storing method, the number of the complete transcripts of the high-popular-degree files is increased, the access speed and load balancing of the high-speed files are improved, the number of the complete transcripts of the low-popular-degree files is reduced, and due to the fact that storing is carried out after the RS erasure code encoding, the storing tolerance capability and reliability are improved.

The invention relates to a load balancing software architecture based on a multi-core platform and a method therefor, comprising the following content: 1) the multi-core platform and load balancing equipment are taken as framework elements, and a kernel thread is initialized for every network card on the load balancing equipment to be taken as the soft interruption routine of the network card; the kernel threads respectively process input and output of the corresponding network cards, and the kernel threads are independent; 2) an independent channel is set up for the information exchange between every two network card threads, and a group of bidirectional ring-shaped arrays can be set up in each channel and taken as a coupler between the multi-core platform and load balancing equipment; the ring-shaped arrays can maintain a writing point and a reading point and are respectively operated by two threads, so that the information exchange between the threads is not needed to be locked; 3) routing policy of balanced load of a server can be realized in the soft interruption routine of the input of the network card, and all work of an application layer is executed in a link layer. The multithreading lockless software architecture is adopted to realize the balanced load of the server; furthermore, the practice proves that the invention can remarkably improve the performance of the load balancing equipment.

The invention relates to the technical field of cloud computing, in particular to a virtual machine load balancing method under a cloud environment. The method comprises the following steps that a random forest regression model is built according to a historical task data set of a cloud platform; a current task set and a current virtual machine set of the cloud platform are obtained; by using thebuilt random forest regression model, execution time of each task in the current task set of the cloud platform is forecast; according to the CPU application amount, memory application amount and taskdata processing scale which are requested to the cloud platform by each task in the current task set of the cloud platform and forecast task execution time, a particle swarm optimization algorithm isused, the load balance degree of a virtual machine in the cloud platform serves as an optimization objective, and a globally optimal particle code is obtained; the globally optimal particle code is decoded into a task-to-virtual machine allocation scheme and output. Accordingly, the total task execution time can be shortened while the resource utilization rate of the virtual machine is balanced.

The invention provides a cloud-based user visit request scheduling method, and has the following key points: in the process of carrying out scheduling on the user requests, judging the load situation of service nodes by a request scheduling module through detecting average request execution time and request queue length to be executed of each service node so as to carry out corresponding adjustment on the weight and further calculate the load interval of each service node, simultaneously generating a random number within 0-1 range to the request to be processed, searching the load interval corresponding to the random number, searching the service node corresponding to the load interval, then sending the user request to the service node for execution, monitoring the execution result and renewing the average request execution time and the request queue length to be executed of the service node, if new increased user visit request exists during the time, adding the user visit request to the user request queue to be processed, and repeating the process until all the requests are executed. Therefore, the load balancing capacity and responding speed of a cloud computing system are obviously improved.

An embodiment of the invention provides a server of a distributed storage system and a data synchronizing method between servers. The method includes: the first server receives a data updating request and searches the data fragment corresponding to the keyword information in the data updating request; the data fragment is updated according to the updating data in the data updating request; if the fact that the second server with other data fragments identical with the data fragment in name is offline is detected, the data updating operation is recorded into the first log of the first server, wherein the data fragments with the same name are disposed in the first server and the second server; a log synchronization request transmitted by the second server is responded to, and the first log is returned to the second server. The method has the advantages that the second server only needs to perform data synchronization according to the first log, data synchronization efficiency is increased greatly, the servers can provide data services for users as soon as possible, and user waiting time caused by server offline can be reduced.

A communication system which includes plural radio systems, and which connects to a terminal apparatus corresponding to the plural radio systems, comprising: access points for the each radio system; gateways; an authentication apparatus; and a home agent. The authentication apparatus being shared by the plural radio systems: assigns an IP address to the terminal apparatus in a radio system which is initially connected among the plural radio systems according to an IP address assignment request from the terminal; judges whether an IP address assignment request for a radio system to be subsequently connected is from the same terminal apparatus by using an unique identifier of the terminal; and assigns the IP address assigned in the initially connected radio system to the terminal apparatus in the radio system to be subsequently connected in a case where the assignment request is determined to be from the same terminal apparatus.

The present invention provides bidirectional load equalization system implementation method in multi computer server. It contains using one or more load equalization unit node composing load equalization system and connected with outer network; load equalization unit node to make load equalization to client-side request; through modifying data package target MAC address to make data package dispensing; when fed back data package in through load equalization system, dispensing to each load equalization unit node to make whole server system having bidirectional load equalization function; when load equalization unit node gone wrong gone wrong, through migrating request data package and feeding back data package to other load equalization unit node to realize high usability.

An adaptive print driver receives printing instructions relating to a document and outputs print data pages to an imager. The adaptive printer driver includes a printing instruction parser for parsing printing instructions, a format controller for assigning a data format to the parsed print instructions and transferring the parsed print instructions according to the data format to a plurality of format generators. The plurality of format generators generates print data which is passed to a coalescer that coalesces the received print data into at least one print data page. The adaptive print driver uses a data format determination analysis process to balance rasterization on the adaptive print driver to achieve an optimal load balance between rasterization to take place on the adaptive print driver and rasterization to take place on the imager. The data format determination analysis considers the complexity of the document as well as information pertaining to the performance capabilities of the host, the performance capabilities of the imager, the load capacity of the host, and the load capacity of the imager, if such information is available. The present invention is also directed to a method for preparing a document for imaging on an imager.

In a trunking-port network environment, a ticket assigner and a ticket distributor are included in a packet switch load balance device together to achieve an output load balance during the transmission process of packet switching. The ticket assigner assigns one of the tickets to a packet received by a logical port, and the ticket distributor determines which one of the physical output ports to transmit the packet according to the assigned ticket. In the transmission process, a load monitor dynamically adjusts the ticket distribution scheme used for the ticket distributor, and thus the load balance is optimized.

The present invention provides a multi-microgrid flexible interconnection system and a control method thereof. The system includes a plurality of microgrids and multi-port flexible interconnectors, wherein the multi-port flexible interconnectors are arranged between the feeder lines of the plurality of microgrids; and the multi-port flexible interconnectors are connected into the tail ends of the adjacent feeder lines of the plurality of microgrids, so that the flexible interconnection of the plurality of microgrids is realized. According to the multi-microgrid flexible interconnection system and the control method thereof of the invention, the multi-port flexible interconnectors, replacing traditional contact switches, are connected into the tail ends of the adjacent feeder lines between the different microgrids, so that the flexible interconnection of the plurality of microgrids is realized; due to the controllability of power electronic devices, power flow interaction among the plurality of microgrids can be realized. With the multi-microgrid flexible interconnection system and the control method thereof of the invention, multi-microgrid loop network operation can be realized with no short-circuit current increased; the local absorbing capacity of new energy resources can be improved; load balance between the feeder lines can be improved; mutual backup and mutual power supply of the adjacent microgrids can be realized; line loss can be decreased; the reliability of power supply can be improved; and voltage quality at the tail ends of the feeder lines can be maintained.

The invention discloses a cluster compression method based on data similarity. By analyzing file data, structural characteristic vector of characteristic fingerprint is extracted from files to calculate the data similarity; files are input in cluster by utilizing a graph partitioning method with a restricted condition, so that a plurality of categories in even sizes are formed; and finally, compression is respectively performed on each category by utilizing compression methods, such as BMCOM, so as to remove the redundant data in category interior. The invention adopts a clustering method based on data sampling; and key data with a high condensability serves as sample data. Firstly, clustering is performed on the sample data; then, the remaining data is classified through a marriage stabilizing method, thereby improving clustering efficiency under a condition that the compressing effect is not reduced. As a compressing and filing method, the invention can be applied to a distributed storage system, so that the problem of uneven data dependence and load in the prior method can be solved.

Provided are a video cloud platform load balancing method and a video cloud platform load balancing dispatcher. The video cloud platform load balancing method comprises the steps that instantaneous load conditions of all video server nodes themselves at a current moment are obtained, and instantaneous load values of all the nodes are calculated; historical load reference values are obtained according to stored historical load big data information of all the server nodes; comprehensive load values are calculated according to the instantaneous load values and the historical load reference values; load grade thresholds are set, and grades of all the server nodes are determined according to the comprehensive load values and the load grade thresholds. Multiple equivalence classes are divided according to the comprehensive load values of the server nodes, a single instantaneous load value is substituted, and the algorithm superiority is improved.

The invention discloses an optimization control method of a multi-site conveyor belt feeding production processing site system. A reaction item and a diffusion item of a performance function are respectively constructed according to the site self information and the interactive information of adjacent sites, wherein the reaction item is regarded as the feedback of the site self information, and the diffusion item is regarded as the feedback of downstream sites to upstream sites. Compared with a method only relying on the site self information, by introducing a reaction diffusion ideal and designing a new performance function, the invention favorably improves the load balance of the sites, obviously improves the working piece processing rate of the whole system and can effectively solve the problems of dimension disaster and modeling difficulty of a large-scale multi-agent system. Meanwhile, aiming to the problems of two performance criterions of discount and average, through the function constructed in the invention, a uniform continuous-time multi-agent study formula and an optimization algorithm can be established.

The invention discloses a wireless sensor network unmanned aerial vehicle system communication method based on an optimization particle swarm algorithm. The method is used for a wireless sensor unmanned aerial vehicle system network composed of a mobile node unmanned aerial vehicle, a ground communication station (2) and multiple ground wireless sensor nodes (3), and comprises the following steps:(10) network region partition: the whole network region is partitioned into multiple sub-regions by using the optimization particle swarm algorithm; (20), sub-region cluster head selection: a sensornode in each sub-region computes weight sum, and selects a cluster head node; (30) next-hop access node selection: the mobile node accesses the cluster head node with the maximum residual overflow cache time, and then selects the cluster head node needing to be accessed by the next hop of the mobile node; and (40) data transmission: the unmanned aerial vehicle (1) transmits the data to the groundcommunication station (2). Through the wireless sensor network unmanned aerial vehicle communication method disclosed by the invention, the energy consumption is less, the network load equilibrium isgood, and the network life is long.

A new method and framework for implementing network protocol processing utilizing a combination of application threads and a dedicated thread to process IO completions in a completion queue that automatically detects and adjusts thread priorities to alleviate manual intervention. According to the present invention, as data transfer operations are completed by the network interface, completion information identifying the data transfer operations is posted on the completion queue. The completion information is read and processed by a combination of application and dedicated threads running in the system. A method monitors performance of the system to detect whether poor processor utilization or excessive context switches occurs, in which case a different thread is used to process the completion information. In order to context switch to a different thread, the priority level of the dedicated thread is set to a HIGH level to utilize the dedicated thread or the priority level of the dedicated thread is set a LOW level to utilize an application thread.

The invention discloses a task allocation method for a wireless sensor. The task allocation method for the wireless sensor comprises the following steps: establishing a task allocation model of a wireless sensor network, and adjusting the initial task allocation scheme according to the communication relationship between tasks; establishing a multi-objective optimization function combined with thetask load, the communication relationship, the node processing speed and the fault probability an so on; setting constraint conditions of the dynamic multi-objective optimization algorithm according to the number of the sensor nodes and the deadline of completing the task; establishing a dynamic environment detection model for judging whether a new task arrives or not; updating the task allocationscheme by means of the dynamic multi-objective optimization algorithm, and detecting the dynamic environment in real time to obtain the Pareto optimal solution set after every environment change; andselecting the solution with the most balanced energy consumption in the Pareto optimal solution set as the optimal task allocation scheme. According to the invention, the effectiveness indexes such as the energy consumption, the time and the reliability is optimized by means of the dynamic multi-objective task allocation method, so as to obtain the comprehensive optimal scheme with the shortest task completion time, minimum energy consumption and maximum reliability.