71results about How to "Solve the scalability problem" patented technology

The invention relates to a method for implementing the high availability of the system in a multiprocessor environment. In the multiprocessor environment, one ore M standby computers correspond to N active computers and deliver the standby service for N active computers; when one standby computer corresponds to N active computers, the standby computer monitors the heartbeat with the active computer in the mode of the multicast and the standby computer monitors the working state of the active computers in the mode of the heart beating line; when the standby computer monitors that the active computer malfunctions, the standby computer takes over the work of the active computer and provides the service provided originally by the active computer to the outside so as to implement the continuous application under the circumstance that the user can not feel the service interception. The invention has the beneficial effects that the scalability of the high-availability system is solved and the high availability of the system can be realized with lower cost in a complicated multiprocessor environment having a plurality of applications, and the utilization ratio of the system resources can be enhanced.

The invention proposes a visual modeling method for a distributed simulation platform and aims to provide a visual modeling method capable of lowering the development complexity of all nodes of distributed simulation. The method is realized by the following technical scheme: creating a simulation platform framework and visual modeling related components; performing distributed simulation deployment visually; dividing the simulation platform framework into an application layer, a management layer, a service layer, an interaction layer, a resource layer and a simulation model resource pool from top to bottom, wherein the management layer comprises a model interaction component and a node deployment component; generating an icon arranged in the simulation model resource pool by the model interaction component according to an interface attribute of a simulation model; seamlessly connecting logic deployment and physical deployment of simulation system nodes; automatically generating corresponding software interfaces and corresponding service components by simulation sub-nodes participating in distributed simulation through a node deployment graph; and performing visual modeling in a model interaction view to construct a visual model convenient for a user to design the simulation model interaction view.

The invention discloses area control equipment, domain control equipment and a control system for SDN (Software Defined Networking). The area control equipment is used for collecting equipment state and network topology information of a data plane, and comprises an area host management module, a link discovery module, an area topology management module and an area routing computation module. The domain control equipment is used for collecting a network state and information mastered by a lower-layer controller and acquiring network information mastered by other domain controllers in a distributed way, and comprises a domain host management module, a domain topology management module and a domain routing computation module. Super-linear growth of SDN controller plane computation complexity is lowered to linear growth through construction of an abstract hierarchical view, thereby solving the expandability problem of a control plane in a control domain when the SDN is applied to a large-scale network.

The invention provides a flow sampling method and system for a software-defined network, and relates to the field of communication, in particular to a flow sampling method and system for a software-defined network. The method comprises the following steps: an SDN controller is used for acquiring flow sampling requirement from a north interface or a management interface; the SDN controller is used for matching the flow sampling requirement with current flow table entry information and SDN forwarding device information, generating a set table entry and a guide flow table entry for performing sampling behavior, and sending the set table entry and the guide flow table entry to an SDN forwarding device, wherein the set table entry comprises a sampling ratio, and the guide flow table entry comprises flow information and associated set table entry ID; the SDN forwarding device is used for sampling the flow according to the guide flow table entry and the set table entry after receiving the set table entry and the guide flow table entry. The technical scheme provided by the invention is applicable to the software-defined network; according to the method, the SDN flow sampling with service continuity is achieved.

The invention belongs to the technical field of network communication, particularly relates to a subnet access method based on an identity-position separate mapping mechanism. The method is used for accessing a terminal into a virtual tree network structure comprising a switch router and a subnet router to realize a micromotion protocol. The method comprises the steps of: establishing the virtual tree network structure in a default route mode; setting the route mode of the subnet router into a host route mode and setting the route mode for accessing the switch router into a policy route mode; and accessing the terminal into the subnet router. By the identity-position separate mapping mechanism, the invention realizes the micromotion protocol, solves the extensible problem of a network route, reduces the switching delay of the terminal among networks, overcomes obstacles existing when a heterogeneous terminal is accessed into a network, and avoids the generation of triangular route.

The invention relates to a control system of Ethernet and data report transmission method. Wherein, it is characterized in that: the control system according to the control protocol of two layers of Ethernet obtains the chain information of Ethernet; according to the chain information, building and maintain the network topology of Ethernet, the control system according to the connection built by said network topology requests to calculate the route between the terminal bridges, and transfer the path information according to the route setting; each network bridge data system according to the transfer path information to transfer the data report. The invention can realize the management of transmitted data in the Ethernet, to control the data flux, without changing the packing structure of Ethernet data plane, and without lower cost.

The present invention relates to method and arrangements to measure connectivity between Two Way Active Measurement Protocol TWAMP entities comprising a session-server (15) and session-reflectors (16-23) in a communication network. The method comprises steps like: dividing in the session- server (15), the session-reflectors (16-23) into at least one defined group (11-14)of session-reflectors (16-17,18- 9,20-21,22-23); transmitting a test session (1;1-4) from the session-server (15) to the session-reflectors(16- 7;16-23) by using Source Specific Multicasting, said test session comprising at least one test packet (1a-1c;1a-4c) 1 being destined to at least one session-reflector (16-17;16- 23) in at least one defined group (11;11-14); replying to the at least one test packet (1a-1c;1a-4c) by the at least one session-reflector (16-17;16-23).

The present invention discloses a distributed duplicated data deleting system and method based on a Hadoop platform. The system comprises a client, a master node and worker nodes. Distributed parallel duplicated data deleting processing is implemented by a MapReduce parallel programming frame of the Hadoop platform. The implementation method comprises: the client sends a file to the master node; the master node completes file fragmentation and data distribution and construction of a file metadata table; and each worker node, according to fine granularity, performs blocking on data fragmentation, calculates fingerprint values of fine-grained data blocks, carries out query and comparison in an index of a database Hbase, stores new data blocks in a distributed file system HDFS and feeds back index information to the master node. The system and method can have high throughput and excellent expansibility while ensuring a high deduplication rate.

The invention discloses a data center heat management method based on temperature prediction. According to the local characteristics of heat transfer in the data center, a joint prediction framework is established to implement the data center temperature prediction algorithm based on neural network. The real-time running data of server node, wireless sensor data and CFD simulation data are used totrain the artificial neural network prediction model to predict the inlet temperature distribution of the data center server. Then feedback control strategy is used to manage the temperature of eachnode in the data center, and the heat management is carried out by adjusting the workload of the node. The invention is devoted to the research of the heat management method of the data center, and under the premise of ensuring the thermal safety requirement of the data center, the energy consumption of the refrigeration equipment is minimized.

The invention discloses a large-scale repeated video retrieval method based on random multi-view hash. Video key frames are extracted through a time interval sampling method, and color histogram HSV and local binary pattern LBP characteristics of the key frames are extracted; the HSV and LBP characteristics are mapped to quasi-hash space through linear mapping and the sigmoid function, and hash codes representing videos finally are obtained through quasi-hash codes generated through thresholding processing; the Hamming distance between a test video hash code sequence and other video hash code sequences is calculated, and approximate repeated videos of the test video are retrieved according to the Hamming distance. Multi-view characteristic information of the videos is used, retrieval efficiency and accuracy are improved through hash mapping, and large-scale quick retrieval of the approximate repeated videos is achieved.

The invention discloses an intelligent QoS routing optimization method and system based on deep reinforcement learning in an SDN environment, and the method comprises the following steps: expressing all streaming media services in a network as a service request set, and then for each request, searching a path meeting the network service quality from a streaming media server to a heterogeneous client; sequentially determining the route of each flow request, and finally constructing a multicast tree by adopting a QoS route optimization algorithm. For a network congestion link or a malicious node, the most suitable next node at present can be found for routing through a deep reinforcement learning method. By adopting the method of combining deep learning and reinforcement learning, the transmission delay of the video stream can be effectively reduced, and the accuracy of routing decision can be improved. Meanwhile, the design of a distributed control plane is adopted and can be realized in various network topologies, so that the network congestion can be avoided, the expandability of the network is improved, the interaction with a single controller is reduced, and the overall utilityof the network is improved.

The invention provides an M2M (Machine-to-Machine) security method for a cotton spinning production CPS (Cyber Physical System) based on a block chain technology. A block chain facing the M2M communication security problem is designed according to the characteristic and principle of the block chain technology, an M2M communication system is divided into a public network area, an equipment area anda private area, a machine equipment block chain is designed in the public network area, and a communication block chain is designed in the private area. The block chain technology solves the problemsof equipment change, dynamic maintenance of cotton spinning production materials and security of M2M communication of the cotton spinning production process under the cotton spinning production CPS architecture. The method overcomes the defects of the prior art, and the block chain technology can effectively solve the extensibility of the M2M system in the cotton spinning production CPS and the timeliness and security of M2M communication.

A resource management system in a data center one or more data storage resource providers and a transaction server. The transaction server is configured to receive, from a client, a request for read and / or write access for a data storage resource, the request comprising one or more specifications, to provide, to the one or more data storage resource providers, at least a portion of the request, and to receive, from the one or more data storage resource providers, respective responses to the request, the responses respectively comprising one or more allocation options. The transaction server is further configured to select one of the one or more allocation options for registration, and register the selected allocation option with a data manager. At least one of the one or more data storage providers is configured to provide the data storage resource in accordance with the registered allocation option.

The invention provides an NAT (network address translation) traversal method, system and relay server by a relay mode. The NAT traversal method by a relay mode comprises the following steps: a first host transmits a relay request message to service address of the relay server by utilizing an NAT address through NAT equipment; and the relay server returns a relay response message to the first host, wherein the relay response message comprises an address list, and the address list comprises a relay address and the NAT address.

The invention provides a network intrusion detection method for a parallel AP cluster based on the MapReduce. The detection method includes the following steps that firstly, intrusion detection sample data is preprocessed, and numeralization and normalization of feature data are finished; secondly, the number of intrusion detection samples is compressed through the parallel AP cluster based on the MapReduce; thirdly, efficient detection is achieved through the compressed data samples by means of a KNN or SVM classifier. According to the network intrusion detection method for the parallel AP cluster based on the MapReduce, good expansibility is achieved in the data sample processing process, mass data samples can be effectively compressed, the detection speed is increased, and the detection accuracy is improved.

The invention discloses a mechanism for increasing hierarchically-distributed SDN (Software Defined Network) control plane routing efficiency, and aims to solve the problem of extensibility in SDN controller deployment under a large-scale OpenFlow network environment. A routing request in a network is partitioned into three levels according to a source IP address and a destination IP address of a data packet; a CSR algorithm is adopted for a routing request in a local area network; and a global routing request is processed as completely-distributed control routing. According to the mechanism, a BI (Blocking Island) graph is built according to a BI theory, and a global network is subjected to topology aggregation in order to reduce search space of a routing algorithm.

The invention belongs to the technical field of an FPGA (field-programmable gate array) and in particular relates to an FPGA interconnection structure supporting the time division switching. On the basis of the structure of an assembly line, time division multiplexing and source-synchronous interconnection, the time slot switching principle which is commonly used in a communication system is combined for providing a novel FPGA interconnection resource structure, the time division switching is used for replacing an original traditional space switching interconnection structure, and the structure adds part of time slot switching on a switching matrix in time division multiplexing interconnection resources, thereby being capable of greatly simplifying the switching matrix and the design of an input and output selector. The novel FPGA interconnection structure can reduce the area of the interconnection resources and improve the logic density of the FPGA, and has the characteristics of being high in performances, reducing the complexity in design, improving the reliability and being low in power consumption.

The invention discloses a web push method based on an event bus. The method comprises the following steps of S1, service container stack address binding; S2, asynchronous processing of messages; S3, releasing of application service messages; S4, web side message subscription; and S5, event bus pushing. The difficulty of upgrading and capacity expanding can be reduced by the technical scheme, and the workload of upgrading can be reduced so that message sending and receiving are enabled to be more timely; the services can be quickly deployed with increasing of the number of the registered users; the concurrency of the system can be enhanced through asynchronous processing; the TPS (transaction per second) of the system can be obviously enhanced; and service calling is safer.

The invention relates to a manageable P2P (Peer-to-Peer) streaming media live broadcasting system based on an IMS (IP Multimedia Subsystem) in an NGN (Next Generation Network) and a realization method thereof. The system provided by the invention expands an overlay network management function entity and a network resource inquiry function entity on a streaming media service layer. The method provided by the invention comprises the following steps: using user terminal equipment capable of carrying out P2P as a calling request to establish a streaming media live broadcasting session; utilizing a streaming media service control function entity as a back-to-back user agent to select a content transmitter for calling user terminal equipment and establish a media transmission session with the content transmitter; utilizing the user terminal equipment capable of carrying out P2P to carry out a channel switching request updating session; and utilizing the streaming media service control function entity to select the content transmitter of a new channel for the user terminal equipment capable of carrying out P2P and establish the media transmission session with the content transmitter of the new channel. According to the invention, the centralized management of a P2P overlay network is realized; the problem of inconformity in a P2P distribution network and a physical network is solved; and simultaneously, the QoS (Quality of Service) of P2P media transmission is ensured by utilizing the SIP / SDP (session initiation protocol / session description protocol) session consultation mechanism in the IMS and the network control and resource reservation provided by the NGN.

The invention discloses a customized recommendation method based on user interest. The method comprises that dishes are classified; users are divided into different expert user classes in each restaurant; an interest model is established for each user; a candidate expert user with authority in a present restaurant in the aspect of user interest is selected, a preliminary recommendation dish of the candidate expert user is obtained; the similarity between the user and a candidate expert is calculated; and a final dish is recommended. According to the method, user interest, suggestions of expert users of the restaurant and dining time of the user are combined, dishes that satisfy customized demands are recommended for the user, the dish recommending efficiency and accuracy are improved, the dishes are recommended for the user in real time, and a time factor is introduce to improve an interest sensing algorithm and a dish preference prediction effect and further to improve the recommendation accuracy.

The invention discloses a big data platform constructing method based on S-PLUS. A big data platform comprises a physical layer, a virtualization layer, a service layer and an application layer, wherein heterogeneous hardware resources are deployed on the physical layer; Eucalyptus is used to construct a virtual machine cluster on the virtualization layer, and a Hadoop environment is deployed on the virtual machine cluster; the service layer integrates S-PLUS language, realizes a data mining function and provides services; a clear operation interface is provided to a user on the application layer which has the functions of a data storage and processing system, an array operation tool, a complete and coherent statistical analysis tool and excellent statistical cartography. S-PLUS is a simple and powerful programming language and can control data input and output and realize functions of branching, circulating and user customizing. The big data platform can effectively process targets such as information system services, information monitoring and public safety which need to be paid attention to by an intelligent city system, the targets are displayed in a big data manner and results are analyzed, and high processing efficiency is achieved.

The invention discloses a distributed system based on a graph database. The system comprises a client, a database management layer, a computing layer, a storage layer, a resource management module anda data import layer, and is characterized in that the computing layer comprises a plurality of distributed computing clusters, each distributed computing cluster consists of a query control node anda graph query node, and subsequent processing such as screening, aggregation and sorting is performed on the data extracted by the storage layer, the storage layer comprises a storage control node anda graph storage node, and the graph storage node undertakes storage, query and update operations of graph data. According to the system, the performance is improved by better utilizing the proximityof the graph data through an operator push-down scheme, meanwhile, the performance overhead caused by data communication between the storage layer and the computing layer is also reduced, the problemof expandability of the distributed graph data is solved through a thought of separating the storage layer from the computing layer, and the storage layer solves the problems of data storage and fragmentation through the thought of combining key value pairs with hash partitioning.

The embodiment of the invention provides a multi-party unsupervised learning joint modeling method based on an X86 computing chip, and the method comprises providing a trusted measurement manager based on an X86 architecture chip, a distributed unsupervised machine learning framework and a privacy computing interconnection system, wherein the trusted measurement manager based on the X86 architecture chip comprises a plurality of servers based on the X86 architecture chip, one server serves as a service party, and the other servers serve as calculation parties; realizing multi-party local unsupervised model training based on local data, performing interaction with a parameter server at the same time, realizing multi-party local model safe aggregation, and building a global model. All parties involved in the method carry out data transmission and model parameter aggregation through a privacy computing interconnection system constructed based on MPC, a trusted execution environment or an encrypted connection mode, and the privacy of related data is ensured. And data availability and invisibility, data algorithm credibility measurement and multi-party joint unsupervised learning modeling are ensured.

The embodiment of the invention discloses a system, a method and servers for implementing a cloud classroom. The system comprises a first server, at least one second server and more than two third serves, wherein the first server is used for sending a first request message to the second server according to a cloud classroom establishment request sent by a user, and the first request message is used for requesting to allocate data resources to a cloud classroom; the first server is further used for receiving a first response message sent by the second server, establishing connection with the at least one corresponding third server according to the first response message, and establishing the cloud classroom according to the at least one third server; the at least one second server is used for receiving the first request message sent by the first server and sending the first response message to the first server, and the first response message comprises information of the at least one third server allocated to the cloud classroom; and the more than two third serves are used for providing data resources for the first server to establish the cloud classroom.

A basic feature of the invention is to rely on an AAA infrastructure to 'bootstrap' the HMIPv6 service for a mobile node ( 130 ) that 'roams' in a visited network or the home network. In accordance with a preferred embodiment of the invention, bootstrapping the HMIPv6 service involves authenticating and authorizing the mobile node ( 130 ) for HMIPv6 service based on an AAA infrastructure. In an important scenario, the mobile node is roaming in a visited network, and the AAA infrastructure ( 110, 120, 122 ) links the visited network with the home network of the mobile node. The invention also supports the possibility of having the MAP ( 125 ) located in the home network or other network than the visited network. The reliance on the AAA infrastructure preferably involves transferring HMIPv6-related information required for authenticating and authorizing the mobile node for HMIPv6 service over the AAA infrastructure.

The invention provides a method for solving the scalability problem of a blockchain internet intermediate chain. The intermediate chain system architecture designed by the invention is a hierarchicalarchitecture, a two-layer structure and multi-chain parallelism. Attached drawings of abstract are shown in the specification. The top layer of the architecture is a main chain layer. The main chain layer only has one chain and is a middle chain main chain; and the intermediate chain main chain is maintained by a plurality of nodes of the intermediate chain. The bottom layer of the architecture isa sub-chain layer and is provided with a plurality of sub-chains. Each sub-chain is maintained by part of nodes of the intermediate chain and the link nodes of the two participating chains together,and serves for intercommunication and value transfer between the two participating chains. An intermediate chain main chain of the block chain Internet is responsible for organization and management work of all nodes of the intermediate chain. Meanwhile, information such as registration and reputation records of all participation chains and the chain nodes of the participation chains is recorded on the main chain of the intermediate chain. The intermediate chain sub-chain of the blockchain Internet is only responsible for recording cross-chain transactions between two participating chains related to the intermediate chain sub-chain. In this way, hundreds of participating chain interconnection and intercommunication and value transfer tasks in the block chain internet are shared by numerousintermediate chain sub-chains, so that the problem of intermediate chain performance bottleneck is solved, and the problem of scalability of the intermediate chain of the block chain internet is solved.

The invention discloses an optimization method and system of an operation and maintenance auditing system. A single sign-on service is adopted in an authentication login module in the operation and maintenance auditing system, and a local login mode is switched when the single sign-on service responds mistakenly, so that a unified authentication process capable of realizing disaster recovery is realized. The object storage structure of a log processing module is optimized, so that the log processing module can be called by different service modules, and the problem of capacity expansion is solved. The problems that the capacity expansion of an operation and maintenance auditing system is difficult and the service is unstable when a springboard machine room accesses are solved by utilizinga load balancing scheme capable of automatically identifying newly-added nodes and deploying the nodes, and meanwhile, the data synchronization process is more optimized by adopting a data synchronization mode of an asynchronous message queue.

The invention discloses a five-layer framework structure-based power control software development model. The model comprises a task scene layer, an operation management layer, a data communication layer, a driver abstract layer and a system driver layer distributed in sequence from top to bottom. General demands in the field of power control are summarized; a framework model-based software productcan arrange the layers and software modules in the layers according to actual demands; and the complex control requirements of a power control system and the expansibility, portability and development cycle of control software are met and realized.

The invention discloses a method for multicast control and the method includes the following steps of: receiving a request for joining a multicast group of a first edge node or a request for leaving the multicast group of the first edge node; when the request for joining is received, searching for a nearest joining replication point corresponding to the first edge node; when the request for leaving is received, searching for a nearest leaving replication point corresponding to the first edge node; and reserving bandwidth resource from the searched nearest joining replication point to the first edge node, or releasing bandwidth resource from the searched nearest leaving replication point to the first edge node. The invention also discloses equipment and a system for multicast control. By using a network resource management node to control the multicast admission and release process, the invention realizes the uniform management of multicast bandwidth resource, reduces the interactive transmission of a great amount of management messages in multicast transmission process, dynamically distributes bandwidth and improves the utilization rate of network bandwidth.