88results about How to "Avoid load imbalance" patented technology

A method and apparatus for robustly enhanced Class of Service (COS) at the application layer permits highly flexible privilege based access and enables implementation of complex policies and rules for classification and differentiation of services. Differentiation facilitates categorization of traffic to permit flexible design and implementation of multiple Class of Service levels.

A distributed data storage method and system based on multilayer consistent hashing are provided. The system includes a plurality of storage nodes for providing data storage and redundancy protection;a plurality of management nodes for maintaining storage node attributes and virtual groups-storage node mapping information; a plurality of monitoring nodes for maintaining storage node state and processing state changes such as storage node addition, deletion, and failure; and one or more clients that provide storage system access points for applications or users. Storage nodes are organized ina tree structure, with each storage node in each layer of the tree being assigned with multiple identities and a hash space maintaining consisteny hash space. Instead of sharing a hash space between all storage nodes, multiple hash spaces in the storage node architecture tree that are consistent across each layer. The system guarantees the reliability and availability of the stored procedure and avoids the problem of load imbalance when accessing the system.

The invention discloses a method for balancing a load of a network GIS (Geographic Information System) heterogeneous cluster server. In the method, built-in attributes which accord with the Zipf distribution rule and a server heterogeneous processing capacity are accessed on the basis of GIS data; the method adapts to dense access of a user in the aspect of the cluster caching distribution; when the cache hit rate is improved, the access load of hot spot data is balanced; the minimum processing cost of a cluster system, which is required by a data request service, is solved from the integral performance of a heterogeneous cluster service system, and user access response time is optimized when the load of the heterogeneous cluster server is balanced; and the distribution processing is carried out on the basis of data request contents and the access load of the hot spot data is prevented from being excessively centralized. The method disclosed by the invention highly accords with the large-scale user highly-clustered access characteristic in a network GIS, well coordinates and balances the relation between the load distribution and the access local control, ensures the service efficiency and the optimization of the load, effectively promotes the service performance of the actual network GIS and the utilization efficiency of the heterogeneous cluster service system.

The invention relates to an implementation method and a system for NGINX server load balancing, and belongs to the field of computer network communications. The implementation method provided by the invention comprises the following steps: adopting a triggering mechanism to dynamically update resource information of a backend server cluster via an NGINX server, then calculating a load value of a backend server under each service type according to the resource information of the backend server cluster and different service types, and processing an HTTP request by selecting a backend server of which the load value is the smallest in the service type to which the HTTP request belongs. The invention further provides a corresponding system for NGINX server load balancing. According to the implementation method and the system, requests of a user can be distributed reasonably to ensure that the load of the back-end server cluster is in a relatively balanced state. Therefore, the purpose of reasonable use of the server resource is achieved.

The invention discloses an intelligent optimized scheduling system for logistics. The system is composed of a basic data module, an intelligent scheduling module and a board management module. The basic data module consists of a line management unit, a vehicle type management unit, an address management unit, a car pooling rule unit, a scheme design unit, a transport capacity management unit, a data dictionary unit, and an area management unit. The intelligent scheduling module is used for realizing the complete intelligent scheduling function and is mainly responsible for order extraction and scheduling calculation, carrying out examination on a calculation result, carrying out proper arrangement after successful examination, and writing back to a TMS. The board management module mainly includes an arrangement checking unit and a TMS uploading instruction checking unit; the arrangement checking unit is used for realizing a function of real-time arrangement situation checking of a warehouse in the system; and the TMS uploading instruction checking unit is used for realizing a function of uploading instruction checking. According to the system, the inventory, manpower and transport costs can be lowered; the punctual delivery ratio can be improved; the profitable operation can be realized; and the competitive advantage can be obtained.

The invention discloses a passive optical network system, an optical line terminal, optical network units and a method for realizing load balance. The system comprises the optical line terminal OLT and the optical network units ONU, wherein the OLT is used for monitoring the uplink flow rate of at least one ONU, counting the uplink flow rates of uplink wavelengths according to the monitored uplink flow rates of the ONUs, judging whether uplink flow rate unbalance occurs among the uplink wavelengths according to preset conditions, determining the ONUs needing wavelength adjustment if the uplink flow rate unbalance occurs, generating wavelength adjustment messages and transmitting the messages to the ONUs needing wavelength adjustment; and the ONUs receive the wavelength adjustment messages from the OLT and perform corresponding wavelength adjustment according to the wavelength adjustment messages.

The invention discloses an addressing system and method aimed at layering host marking and based on overlaying DHT; structure of each addressing server in the system is identical and is provided with a node accessing and exiting processing module, an information maintenance module, a route searching module, a mapping record information memory module, and an overlaying DHT structure establishing and maintenance module. The addressing method is to set up the overlaying DHT first: non-semantic host identifiers used in the network are hierarchically transformed; area marks indicating the locating area position of each node are added; then each node is divided into different areas, and each area DHT is set up respectively; finally, all the area DHT are overlaid together to constitute a global DHT; when routes are looked for and transmitted, a next hop is chosen in priority in the area DHT based on the overlaying DHT; trans-regional routing times should be reduced to the fully extent to improve routing efficiency; each operation of registration/updating is carried out in two DHT simultaneously to realize equity of load on balance and economy.

The invention discloses a big data environment oriented metadata organization method, which comprises the following steps: a client-side receives a file creation request from a user, and sends the file creation request to an index server; and the index server, according to an absolute path of a to-be-created file in the file creation request, obtains a global ID (Identification) and a distribution code list of a parent directory of the corresponding file, obtains a key value right key according to the global ID of the parent directory and a filename of the to-be-created file, saves file index information of the to-be-created file according to the key, and obtains a distribution code which is used for storing metadata, wherein if the distribution code is newly added or more than one distribution code exists in the distribution code list of the parent directory, a bloom filter and the global ID of the file are used for updating a bitmap of the distribution codes. According to the method and the system disclosed by the invention, the technical problem that the existing method, in order to guarantee the locality of reference of the metadata, has unbalanced load in large directory metadata centralized storage can be solved.

The invention discloses a Storm-based stream computing bipartite graph task scheduling method. The method is characterized in that Storm DAJG (Directed Acyclic Job Graph) nodes and cluster physical machine UNG (Undirected Node Graph) nodes are seen as two types of vertexes of a bipartite graph; a bipartite graph model is built; and in combination with computing capabilities of nodes of physical machines in a cluster and network transmission delay in the cluster, and according to schedulable relationships between tasks and node resources, task scheduling is performed by adopting an algorithm for solving maximum weight matching of the bipartite graph. According to the method, under the condition of ensuring resource load balance of the physical machines in the cluster, the network delay in adata stream migration process during task execution is shortened, so that the overall performance of a system is improved.

The invention relates to a task scheduling method and system, in particular to a multi-timing-sequence task scheduling method and system. The method comprises the steps that node information is collected, and according to the node information, the operational capability of each node is set; task information is collected, and according to the task information, the grid number of each task is set; according to operational capability, all nodes are sorted; according to the grid number of each task, all tasks are sorted; and according to the load state of each node, a greedy algorithm is used, and according to the mode that nodes with high operational capability process tasks with the large grid numbers, the tasks are distributed to the nodes. Resource distribution on a current platform and the loading capacity and states of the nodes in the platform are considered, during task scheduling, tasks with high calculated quantity are distributed to node with high computing capacity, the problems of starvation states and the low system resource using rate can be effectively avoided, overall operational efficiency is improved, and operation time is optimized.

The invention discloses a method and a device for selecting an access point (AP). The method for selecting the AP during a basic service set (BSS) switching process comprises the following steps that: a Station selects the AP with the minimum search delay from a Cache and performs authentication and association when performing BBS switch, wherein the search delay is a time interval value starting from the time when a search request is sent by the Station to the time when a search response frame is received from the AP. Through the method and the device, the condition of a wireless link is taken into comprehensive consideration and the load imbalance of the AP is effectively avoided; moreover, in the method, a 802.11 protocol does not need to be modified and the method has high deployability.

The embodiment of the invention discloses a CTDB virtual IP equalization allocation method, device and equipment, and a computer readable storage medium. The CTDB virtual IP equalization allocation method comprises the following steps: obtaining a loaded virtual IP and an idle virtual IP to be allocated in a current distributed storage cluster when adding a new storage node; and retaining the loaded virtual IP on a source storage node without drifting, and allocating each virtual IP to be allocated to each storage node of the distributed storage cluster evenly. The technical scheme provided bythe application can ensure that the loaded virtual IP do not drift, the client service is not interrupted, and the allocation of the virtual IP can be balanced; therefore, the case that the virtual IP is assigned to the same node and the corresponding load is assigned to the same node is avoided, the data processing performance of the distributed storage cluster is improved, and the stability andhigh availability of the distributed storage cluster are improved.

The invention discloses a distributed traffic simulation system and a simulation method based on variable region division. According to the invention, a long time segment is divided into a plurality of continuous short time segments; change of distribution of people and vehicle on road network in the short time segments is much smaller than that in the total time segment; and in each short time segment, a region division scheme is adopted according to the distribution rule of people and vehicles, and the change of the number of people and vehicle in the time segments in the region is decreased, therefore, the purpose of balancing the load of each simulation server is achieved.

The invention discloses a high-efficiency text data mining method and belongs to the technical field of information. The method comprises the following steps of: 1) in a file pre-processing stage, combining original files of which contents are subjected to word segmentation into a plurality of new files; 2) in a data mapping stage, computing the total frequency number of each word in the new files, the frequency number of each word in each original file, relative frequency pr and the like, and sending a result to a re-orientation module; 3) in a re-orientation stage, computing the payload of each Reduce task, and arranging a payload indicator payi for each Reduce task; 4) judging whether the current word is allocated to the Reduce task, if the current word is not allocated to the Reduce task, allocating the current word to a Reducej task, wherein payj plus pr*100 is less than or equal to the payload, then updating a payload indicator payj of the Reducej task, and otherwise, allocatingthe current word to a corresponding Reducei task; 5) in a data protocol stage, computing parameters such as the final frequency number and the like of the allocated word; and 6) according to a data protocol result, extracting the word of which the frequency number is greater than a set threshold value in a set range. By the method, frequency number computing efficiency and data mining efficiency are greatly improved.

The invention discloses an index construction method for distributed picture search and a device. The index construction method comprises the following steps of: extracting more than one feature of all pictures in a picture database of the server, and converting the features into a sight word set to obtain an original sight word dictionary and a total index table corresponding to the original sight word dictionary; acquiring an effective sight word dictionary representing the original sight word dictionary, wherein each sight word in the effective sight word dictionary is built with a respective corresponding inverted record table; and uniformly distributing each sight word in the effective sight word dictionary to each index server by adopting a first preset rule so that each index server obtains a sub sight word dictionary consisting of the sight words therein and the inverted record tables of the sight words in the sub sight word dictionary form an index table of the sub sight word dictionary. The method can improve the index efficiency of the server and meet the real-time response requirement on a user in searching.

The invention discloses a method for achieving UE (user equipment) random access of user equipment and a system. The method is applicable to the network scene of a macro base station and a micro base station which share IDs of cells. The method comprises the following steps: the UE obtains time-frequency synchronous downlink according to monitored information of a cell broadcast system and determines the strongest frequency position; when the strongest frequency position is determined to be in a micro cell, the random access process is accomplished through information interaction with the micro base station for providing service to the micro cell; and when the strongest frequency position is determined to be in a macro cell, the random access process can be accomplished through information interaction with the macro base station for providing service to the macro cell. According to the technical scheme provided by the invention, the base station for providing service to the strongest frequency position is preferentially in charge of the UE random access in a coverage area, so that the problem of long time delay during treatment since both the macro base station and the micro base station participate at the same time due to the adoption the conventional random access technique can be solved.

The embodiment of the invention provides a port resource allocation method and device. The port resource allocation method comprises the following steps: detecting the port resource use condition of each network equipment in a dual-master network; determining first network equipment with the minimum use of port resources in the dual-master network according to the port resource use condition of each network equipment; and distributing port resources which are not used on the first network equipment to each network equipment in the dual-master network. According to the scheme of the invention,the stability of continuously providing network services by the dual-master network can be improved.

The invention discloses a use method of an intelligent lock low-power-consumption remote control system and system and the system. The method includes the following steps that firstly, a user terminal is matched with an intelligent lock and subjected to authority checking through Bluetooth short-range communication, and the user terminal obtains an authority credential of the intelligent lock after authority checking is successful; secondly, the user terminal conducts WiFi configuration on the intelligent lock after Bluetooth matching; thirdly, the intelligent lock sends a connection building request to a central control system and is connected with an information transfer server; fourthly, the user terminal is connected into the information transfer server through WiFi or a 3G/4G network and sends the authority credential to the information transfer server, meanwhile a central server cluster of the central control system informs the intelligent lock of sending an authority credential on the intelligent lock, when the central server cluster judges that the two authority credentials are consistent, remote connection is successfully built, and the user terminal can remotely control the intelligent lock. The method has the advantages of being good in stability, quick in responding and low in power consumption.

The invention discloses a cell reselection method and device applied to a multi-carrier condition. According to the cell reselection method and device, UE adjusts cell reselection strategies according to acquired cell load information, priority information and other information, so that the problem of cell overload can be avoided; cell priority configuration related to the capabilities of the UE is introduced, so that UE supporting specific functions can preferentially reside in cells supporting the functions, and therefore, the improvement of user experience can be benefitted. The present invention discloses a method and device for reducing the access load of cells. According to the method and device for reducing the access load of the cells, as for cells where overload has already occurred, the current load of the cells can be effectively alleviated through limiting access requests or paging for services with high time delay tolerance. According to the cell reselection method, the cell reselection device, the method and device for reducing the access load of the cells provided by the technical schemes of the invention, the advantages of existing cell selection strategies are retained, and at the same time, the problem of load unbalance of the cells can be avoided; an existing system is just slightly modified, and the compatibility of the system will not be affected. The methods and devices have the advantages of simplicity in implementation and high efficiency.

The invention discloses a parallel symbol execution system based on multi-Agent distributed scheduling. The system comprises a distributed scheduling and monitoring unit, an execution track generating unit, a symbol execution unit and a constraint solving unit. The distributed scheduling and monitoring unit is used for achieving pipeline parallel scheduling on the execution track generating unit, the symbol execution unit and the constraint solving unit, the parallel operating degree of the system is increased, and loads are balanced. The execution track generating unit is used for actually executing specific input drive measured programs, detecting whether abnormalities are generated or not, recording a program execution instruction sequence and generating a program execution track in an intermediary language expression form through analysis and transformation. The symbol execution unit is used for conducting symbolic simulation execution on the program execution track, collecting constraint conditions of route branches and generating a corresponding constraint expression. The constraint solving unit is used for conducting depth-first negation solving on the constraint expression, generating a new input sample, marking the negated expression and avoiding repeated route analysis. A realized task scheduling policy can realize pipeline parallelism, and the problem of load imbalance caused by the fact that only program execution sub-trees serve as scheduling loads is avoided.

The invention relates to a sparse tensor canonical decomposition method based on data division and task allocation. The sparse tensor canonical decomposition method comprises the following steps: initially, performing multi-stage division and task allocation on a plurality of processing cores on a core group according to the many-core characteristics of an SW processor; initially, performing multi-stage segmentation processing on sparse tensor data; designing a communication strategy aiming at sparse tensor canonical decomposition by utilizing the register communication characteristics of the SW processor SW26010; aiming at the common performance bottleneck of different sparse tensor canonical decomposition methods, namely different requirements (whether tensor elements need to be randomly extracted for calculation) of matrix tensor multiplied by Khatri-Rao product (MTTKRP for short) during specific operation, different calculation schemes of the MTTKRP process are designed by utilizing the characteristics of a SW processor. According to the method, the characteristics of the SW system structure are fully excavated, the calculation requirements of sparse tensor decomposition are fully considered, multiple sparse tensor canonical decomposition calculation methods can be completed on the SW system structure in parallel and efficiently, and dynamic load balance is guaranteed to the maximum extent.

The embodiment of the invention provides a thread scheduling system, computer equipment and a storage medium. The thread scheduling system comprises a cluster scheduling module, a data distribution module, a resource management module and a core module, the cluster scheduling module is used for synchronizing dynamic data of any node thread pool model of a service system to middleware in real time,and then the middleware sends changed data to other nodes; the data distribution module is used for acquiring node dynamic data from middleware and adjusting a thread pool model of the current node in real time; the resource management module is used for pooling the whole system resources by the resource management layer and uniformly dividing all the resources of the thread pool into various types of groups; the core module is used for the core layer to be responsible for creating and destroying threads, and when the number of the core threads meets the requirement, the task is executed by the idle threads; when the number of the core threads is not enough, a new thread is created to execute the task; dependence on hardware resources is reduced, and hardware cost is saved for enterprises.