72results about How to "Achieve horizontal expansion" patented technology

The invention discloses an intelligent load balancing method based on a c/s (Client/Server) architecture. The method comprises the following steps that: (1) a communication connection is established between all task nodes and a load balancer by employing a c/s star structure, wherein the task nodes and the load balancer are initialized, and a long connection is maintained; (2) the task nodes periodically transmit node information to the load balancer; (3) the load balancer calculates a performance value of each task node according to the node information and determines distribution of the tasks according to the variation tendency of the performance value of each task node, if the performance of a certain task node is improved, tasks in a total task queue are distributed to the task nodes, and if the performance of a certain task node is reduced, the tasks in the total task queue are paused to be distributed to the task nodes; (4) the task nodes receive the distributed tasks and feed back a task execution log to the load balancer. According to the method, dynamic regulation of the tasks can be realized, and fine intelligent load balancing is achieved.

The invention provides a user request processing method and device. The method comprises the steps that a first container corresponding to a user request is determined after the user request is received; a logical container corresponding to the first container is determined by using the preset relationship between the first container and the logical container; a container cluster corresponding to the logical container is acquired by using the logical address corresponding to the logical container, wherein the container cluster includes at least two second containers; and the user request is processed through the second containers. With application of the technical scheme, the user request can be simultaneously processed by the first container and at least two second containers, all the containers do not independently work, and the first container can call at least two second containers in the container cluster so that data sharing between different containers can be realized and the same application can be completed together; and transverse capacity expanding and scaling of the second containers in the container cluster can be realized and thus the processing capacity of the user request can be enhanced.

The invention provides a metadata group design method based on a real-time application group. The particular design process comprises the step of initial deployment, wherein initial configuration is carried out on a metadata group, the metadata group is composed of a plurality of metadata nodes, a configuration file is generated and synchronized to each metadata node, and the initialization of metadata storage is finished on each metadata node; the step of sequentially starting each metadata node when the group is started, wherein the metadata node with the smallest number in the configuration file automatically obtains the hegemony and guides the whole metadata group to build an overall consistency buffering and sharing lock; the step of fault recovery; and the step of adding new metadata nodes. Compared with the prior art, the metadata group design method based on the real-time application group achieves transverse expansion of the metadata group in a group file system, effectively resolves the performance bottleneck of metadata in a large-scale group file system, is high in practicability and is easy to popularize.

The invention belongs to the technical field of communication, and specifically discloses a large-scale SDN controller cluster and a network system. The cluster comprises an SDN controller group; a buffering server group which is used for storing the control data of the SDN controller group, wherein the control data at least comprises configuration information, a control rule, routing informationand topological data. According to the invention, the control logic and data storage of the SDN controller are separated, i.e., enabling the data storage to be set in one distributed buffering system.Each SDN controller just has the control logic, and does not has the data storage, so the redundancy of data does not increase all the time when the large-scale SDN controller cluster is needed. Also, different types of SDN control data can be placed in different buffering systems, thereby achieving the lateral expansion of data storage.

The invention discloses a service gateway system of internet of things and an application method. The gateway system comprises a terminal access-in adaptation layer subsystem, a core layer subsystem, a capability layer subsystem and an application layer subsystem, wherein the terminal access-in adaptation layer subsystem further comprises a terminal access-in gateway module, the core layer subsystem further comprises a routing control module and a data storing module, the capability layer subsystem further comprises a capability gateway module and a real-time data distributing module, and the application layer subsystem further comprises an application and development tool module. The technical scheme is adopted, an interface between a terminal and an application is enabled to be standardized, difference among the terminals can be avoided, and developers are enabled to be capable of rapidly constructing applications of the internet of the things with high quality.

The invention relates to a service processing method and system, a server and a storage medium, and belongs to the technical field of networks. The method is applied to a service processing system, and comprises the following steps that a front-end server selects a target scheduling server from a plurality of scheduling servers by adopting a first load balancing scheduling algorithm, and sends a service request to the target scheduling server through a switching unit; the target scheduling server selects a target service server from the plurality of service servers by adopting a second load balancing scheduling algorithm, and sends a service request to the target service server; the target service server receives the service request; business requests are processed, processing results areobtained, the plurality of scheduling servers are arranged, even if the number of the service requests is large, the service requests can be evenly distributed to the plurality of scheduling servers,and then the plurality of scheduling servers evenly distribute the service requests to the plurality of service servers, so that the problem that the load capacity of a single scheduling server is toolarge is prevented, and the normal work of the scheduling servers is ensured.

The application provides a distributed space data processing system based on messages. The system comprises a task management module, a task execution module and a message middleware module, wherein the task management module is used to define a task of processing space data and send task messages to the message middleware module; the task execution module is used to acquire the task messages from the message middleware module, conduct corresponding processing on the space data according to the task messages and feed back task state messages to the message middleware module; and the message middleware module is used to transmit the task messages sent by the task management module to the corresponding task execution module according to preset message transmission rules and is also used to conduct corresponding processing on the task messages according to the task state messages fed back by the task execution module.

The invention provides a resource processing method and device, and belongs to the technical field of computers. The method comprises that a service resource is obtained; the service resource is divided according to different service types to obtain sub-resources, each sub-resource is associated with at least one sub-resource, except itself, in the sub-resources, and break strategy configuration information is preset between the associated sub-resources; each sub resource is arranged in one server; and on the basis of the break strategy configuration information corresponding to the sub-resources in the operation process, and the sub-resource are broken from the associated sub-resource in the operation process. The service resource is divided into the sub-resources, the sub-recourses are arranged in the different servers, the response pressure can be expanded rapidly and horizontally, when one sub-resource is abnormal, services between the sub-resource and the other associated sub-resources are broken via the break strategy, and it is ensured that services of the other sub-resources associated with the sub-resource can be used normally.

The present invention discloses an electronic device, a multi-block integration control method and a computer readable storage medium. When each block generates new blocks, new full-amount blocks areformed according to the newly inserted blocks at all the block chains, the generated new full-amount blocks are inserted to a full-amount chain, and the full-amount chain achieves horizontal expansionof the block chains so as to facilitate storage, backup and management.

The invention provides a model online deployment method and device, and the method comprises the steps: receiving training data, carrying out the training of a model based on TensorFlow according to the training data, and obtaining a trained model; storing the trained model based on a preset path, and packaging the model, the model having an alias and a version number corresponding to the model; generating a corresponding container based on the packaged model, wherein the container calls the packaged model based on a version number and an alias; and respectively deploying the trained model andthe container to an infrastructure of a Paas online platform based on the Kubernetes. Compared with the traditional mainstream mode, algorithm engineers do not need to pay attention to model serviceconstruction and can focus on researching the research itself; model updating and multi-model deployment can be rapid and efficient, short-time online operation and verification of more models becomepossible, scheduling management of the model service is more reliable through combination with Kubernetes, dynamic adaptation to page view change can be realized, and the service is more reliable.

The invention discloses a network data packet analysis method and system. The analysis method includes: acquiring a network data packet; dividing analysis content of the network data packet into multiple tasks according to an analysis object, and determining an execution sequence; creating multiple message queues and multiple computational node groups in one-to-one correspondence to the tasks; caching the network data packet into the message queue corresponding to the starting task; sequentially executing following actions until the last task is executed by each computational node group: acquiring all data in the corresponding message queue corresponding to each computational node group, executing the corresponding task, and sending an execution result to the next message queue. Compared with the prior art, the network data packet analysis method and system has the advantages that the analysis content of the network data packet is divided into the tasks which are distributed into different computational node groups, and the message queues are taken as a core for distributed processing, so that processing capacity is improved greatly, coupling among different computational nodes is reduced, and horizontal extension is realized.

Embodiments of the invention provide a fund investment supervision index obtaining method and device. The method comprises the steps of obtaining a corresponding index generation task based on a received index data processing instruction; splitting the index generation task into a plurality of sub-tasks by applying a preset dimension; according to the fund single-day data corresponding to each sub-task and pre-stored self-defined index configuration information, concurrently executing each sub-task to obtain an index calculation result corresponding to each sub-task; aggregating the calculation results corresponding to the sub-tasks to obtain an index calculation result corresponding to the index generation task, and taking the index calculation result as a fund investment supervision index corresponding to the index data processing instruction and used for fund investment supervision. The method can effectively improve the obtaining accuracy, flexibility and efficiency of the fund investment supervision index, can effectively reduce the development period of the fund investment supervision index, and can effectively improve the reliability and accuracy of fund investment supervision.

The invention discloses a BIM design software security storage and circulation tracing monitoring technology, system and device. A DES module is embedded into BIM software and used for monitoring an instant storage request of a user. After a storage request is found, data encryption and storage capabilities are found. Data volumes with storage format of cim are created. A principle of isolating model data is created from parameter data in a cim format to ensure safety of user data and tracking of a user subsequent turnover process. A load balancing module is suitable for processing concurrentrequests and improving processing speed, and the concept of infinite storage capacity is transversely expanded by the multi-stage elastic storage service based on a relational database and a non-relational database.

The invention provides a user task management system based on message driving. A message system is arranged between specific behaviors of a user and a task system and used for storing various types of specific behaviors of the user, and a user event message queue which is a distributed message queue and used for distributing messages in the message system to the task system is arranged between the message system and the task system. The message system and the user event message queue are added between the behaviors of the user and the task execution, and the user behaviors and the task execution are separated, the pressure of the behaviors of the user on the task system is greatly reduced, and meanwhile by means of the frequency control, the processing speed of tasks is controlled and the safety and reliability of the system are greatly improved. By means of the distributed user event message queue, the transverse expansion of the system is achieved, and the task system can support a large number of data and the high concurrency and can have certain disaster tolerance.

The invention discloses a data storage method and a data storage system, relates to the technical field of data storage, and is used for providing a technical scheme for performing horizontal expansion on storage capacity so as to realize horizontal expansion on the storage capacity by utilizing an existing stand-alone storage engine. In the data storage method, a stand-alone storage engine is deployed in each execution node, for example, when the target execution node receives the storage request scheduled by the scheduling node, the storage request is executed based on the single-machine storage engine deployed in the node, that is, on the basis that the original advantages of the single-machine storage engine are stored, processing of the storage request scheduled to the storage engine can be achieved. Besides, each execution node in the execution cluster adopts a single-machine storage engine, so that the homogeneous storage engines or the heterogeneous storage engines of a plurality of single machines are combined to provide a unified service for the outside, and the horizontal expansion of the storage capacity is realized.

The invention relates to a fishing tackle, in particular to a vertical curved cutwater board, which comprises a baffle board, an upper scupper board, a lower scupper board, a connection board and a netting gear connection board. The cutwater board is characterized in that the baffle board is arc-shaped, the straight edge of the baffle board is connected with the connection board through a U-shaped connection piece and the arc edge of the baffle board is connected with the arc shaped upper and lower scupper boards. The surface area of the lower baffle board is larger than that of the upper baffle board. The included angle between the baffle board and the upper and the lower scupper boards is larger than 90 DEG. The cutwater board is designed from a viewpoint of hydrodynamics, which changes the surface structure and shape of the cutwater board, increases the water flow impulsive force of the cutwater board and fully achieves the horizontal expansion of the netting gear and the multi-layer fishing. The cutwater board has the advantages of high expansion capacity, great stability without tending to side turning, easy operation, long service life, good application effect and, etc.

The invention belongs to the technical field of the communication, and specifically discloses a SDN controller cluster and a network system. The SDN controller cluster comprises a SDN controller group, and a cache server group for correspondingly storing control data of the SDN controller group according to a search rule configured in the SDN controller group, wherein the control data at least comprises configuration information, a control rule, routing information and topological data. Through the SDN controller cluster and network system provided by the invention, the control logic and the data storage of the SDN controller are separably stored, namely.e., the data storage is placed on a distributed cache system in a unified way, each SDN controller only has the control logic and has nodata storage, the redundancy of the data is not increased when the large-scale SDN controller cluster is required; and different SDN control data can be stored on different cache systems so as to realize the lateral extension of the data storage.

The invention discloses a cloud disk system for public security police service operation. The system comprises a document subsystem, a work group subsystem, a public resource subsystem, a message center subsystem, a statistical analysis subsystem, a system management subsystem, a transmission list subsystem and a personal center subsystem, the system depends on supercomputing storage equipment asa basic platform, rapid deployment can be achieved through a reasonable resource distribution strategy, and the system has high expandability and reliability. The storage resources can be distributedaccording to needs, and transverse expansion of the storage resources is achieved. File transmission and a working group can be effectively controlled, uploaded files can be automatically backed up, and the risk of file loss can be effectively avoided.

The invention provides an extended bit line strobe device, which comprises a plurality of strobe sub-modules, wherein each sub-module comprises eight tube units, two or three strobe tubes in each tube unit are connected to continuous uneven bit lines or continuous even bit lines in continuous bit lines, each bit line is connected to a plurality of strobe tubes in tube units; each strobe sub-module is strobe controlled by an individual twelve control signal applying terminal, each control signal applying terminal can control at least one strobe tube; a plurality of strobe sub-modules reuse four operation signal applying terminal, corresponding tube units of the plurality of strobe sub-modules reuse a same operation signal applying terminal; and two strobe sub-modules are connected through two co-used uneven bit lines and two co-used even bit lines. By adopting the extended bit line strobe device, no more than four strobe control signals can simultaneously strobe any four continuous bit lines from at least 20 continuous bit lines, and thus the randomness and flexibility of the bit line strobe are improved.