299results about How to "Reduce bandwidth pressure" patented technology

The present invention discloses a co-processor IP core of a programmable convolutional neural network. The invention aims to realize the arithmetic acceleration of the convolutional neural network on a digital chip (FPGA or ASIC). The co-processor IP core specifically comprises a global controller, an I / O controller, a multi-level cache system, a convolution unit, a pooling unit, a filling unit, a full-connection unit, an internal interconnection logical unit, and an instruction set designed for the co-processor IP. The proposed hardware structure supports the complete flows of convolutional neural networks diversified in scale. The hardware-level parallelism is fully utilized and the multi-level cache system is designed. As a result, the characteristics of high performance, low power consumption and the like are realized. The operation flow is controlled through instructions, so that the programmability and the configurability are realized. The co-processor IP core can be easily applied to different application scenes.

A method for playing streaming media is disclosed, which includes: determining at least two storages storing a media file according to information of the media file (400); obtaining media streaming data corresponding to the media file from at least one of the determined storages and playing the obtained media streaming data (401), thereby solving the problem in the prior art that the bandwidth pressure on the streaming media server is excessively heavy if the same streaming media server is accessed when the same media file is obtained when a large number of users perform the data transmission service at the same time in the present streaming media system. A system and a device for playing streaming media are also disclosed.

An embodiment of the invention discloses a difference service pack generating method, a difference service pack downloading method, a file difference comparison server and a file difference comparison client-side. The difference service pack generating method comprises the following steps of acquiring binary files of a historical edition of software if the file difference comparison server receives a binary file, which is uploaded by an entrance server, of a latest edition of the software; comparing the binary file of the latest edition to the binary files of the historical edition by using a preset binary difference algorithm; generating difference service packs of the binary files of the historical edition; storing a corresponding relation between an edition number of the historical edition and numbers of the difference service pack into an upgrading configuration database; transmitting the generated difference service packs to the entrance server; and uploading the difference service packs to a content distribution network (CDN) downloading source server by using the entrance server. The consumed flow and the consumed time of the client-side on downloading of the difference service packs can be effectively reduced, and the bandwidth pressure of the server can be effectively reduced.

The invention belongs to the technical field of the Internet and provides a content delivery method and system of a content delivery network. The content delivery network (CDN) comprises a CDN data source server and multiple CDN nodes, wherein the P2P protocol is executed between the multiple CDN nodes, and the CDN nodes receive a content delivery command and obtain data files relevant to the content delivery command from the CDN data source server or other CDN nodes so that the data files relevant to the content delivery command can be downloaded by a client side. Through the content delivery method and system of the content delivery network, the load of the CDN data source server can be effectively reduced, the bandwidth pressure of the CDN data source server is reduced, and content delivery efficiency and quality are improved.

The invention discloses an Internet of things networking system and method based on edge computing. The method comprises the following steps that: terminal equipment transmits scene data to intelligent gateway equipment, wherein the scene data include numerical data collected by a sensor as well as streaming media data collected by a pickup and the camera; the intelligent gateway equipment performs service processing on the received scene data, and is responsible for uplink and downlink data transmission; and a data aggregation platform receives periodically reported data from the intelligentgateway equipment and stores the received data in a database for an intelligent algorithm platform to train and optimize an algorithm file, wherein the algorithm file is used for updating an intelligent decision module of the intelligent gateway equipment to optimize service scenes. The scheme realizes the edge computing of the service scenes through the intelligent gateway equipment, thereby effectively improving the processing efficiency and real-time performance. In conjunction with the updating of a service algorithm module of an intelligent gateway, flexible adaptation to the service demands and environmental changes of different scenes can be realized, and the flexibility of the intelligent gateway is improved. When a center or network fails, normal operation of terminal scene services can be ensured, thereby achieving relatively high practical and popularization value.

The invention discloses a security and protection video monitoring system based on cloud edge cooperative computing and an implementation method thereof. The security and protection video monitoring system comprises a video acquisition layer, an edge computing layer, an intelligent judgment module, a cloud platform processing module, an intelligent service module and an application terminal module. The objective of the invention is to solve the problem in a centralized processing mode of an existing monitoring system. In order to solve the problems of huge pressure in transmission, analysis and storage of massive videos, poor real-time performance of information processing feedback and the like, the invention provides an intelligent monitoring system implementation scheme based on cloud-side cooperative computing, and the problems of transmission and processing of massive data can be effectively solved.

This invention discloses a VLIW with single-instruction multiple data streams converging computing cluster structure, which includes the data buffer and microcontroller connected with main controller, and their connective computing cluster. The main controller are responsible for the instruction and data mobile, instruction loads in the microcontroller, data loads in the data buffer and receives their output. The data buffer receives main controller data, supplies operation data for computing cluster and receives its result, and outputs to the main controller. Microcontroller receives the VLIW sequence form main controller, decodes and broadcast to each processing unit for parallel execution. The computing group includes a number of same processing units with a number of dealing components respectively and the same instruction sequence, and the data is from different modules of data buffer.

The invention provides a synchronous method of an enterprise organization structure address book, which comprises the following steps that: a client terminal sends a synchronous request, and the synchronous request comprises a time stamp; a server terminal screens the address book data in incremental synchronization according to the time stamp of the synchronous request submitted by the client terminal; the server terminal issues the address book data in incremental synchronization; and the client terminal receives the address book data in incremental synchronization and updates a local address book copy. The invention also provides a synchronous method system of the enterprise organization structure address book. The synchronous method of the enterprise organization structure address book and the system thereof can mitigate the server and network bandwidth pressure caused by data synchronization in the instant communication application of large enterprises, ensure the normal use of the server, and meet the needs of the address book data synchronization of the organization structures of the large enterprises.

The invention discloses a three-dimensional model display system, which is divided into a client side, a server side and a data service side, wherein the client side is used for achieving displaying of operation data in a three-dimensional scene; the server side is used for achieving integration of operation data; and the data service side is used for providing a source of the operation data. The invention further correspondingly provides a method for achieving three-dimensional model scheduling. By the display system and the scheduling method disclosed by the invention, high-precision original three-dimensional model files with different formats, which are submitted by different design units can be split and optimized into a model format which can be supported by a three-dimensional model scheduling engine; and three-dimensional models of all production equipment on the display system can be displayed on the same scene.

A service system of stream media based on controllable P2P technique comprises bureau center, sub center, client end and network connected with three said units on sense of space or service providing layer, service support layer and network transmission layer on sense of logical function. Its realizing method utilizes dispatch server to control P2 P transmission between client ends according to control policy.

The invention relates to a task scheduling and resource allocation method and system for a real-time cloud platform. According to the task scheduling and resource allocation method and system for the real-time cloud platform, a global state storage module obtains the running state of the cloud platform and reports the running state to a global state monitoring module; the global state monitoring module develops a corresponding scheduling strategy through a task allocation matrix, a task adjacency matrix and a mask matrix on the basis of the running state; node-driven and / or task-driven task scheduling and resource allocation are performed on the real-time cloud platform on the basis of the scheduling strategy. The task scheduling and resource allocation method and system for the real-time cloud platform can take the relationship among tasks into full consideration during task allocating and reduce traffic among nodes and bandwidth pressure so as to improve the performance of the platform, can perfectly adapt to all situations of dynamic scheduling of the cloud platform to ensure that the cloud platform can maintain high computing performance and resource utilization rate at all times during operation, and are low in time complexity and applicable to cloud environments with large-scale nodes and large amount of tasks.

The invention relates to the technical field of monitoring devices, in particular to intelligent video image processing equipment based on edge calculation. The equipment comprises an image channel processing module, an image coding processing module, a video stream processing server, an edge intelligent processing module and the like. After the data acquisition module acquires original image data, the original data is shunted by the image channel processing module and is transmitted to the video stream processing server and the edge intelligent processing module, and the edge intelligent processing module processes the encoded data according to a corresponding intelligent algorithm. According to the invention, the problem that the existing traditional video image processing equipment haspoor image processing capability is solved; the problem that massive big data cannot be processed in time due to development of a high-definition camera is solved, the effects of reducing uploading ofvideo data, reducing video storage space and efficiently processing and extracting useful video information are achieved, and meanwhile the technical effects of meeting real-time transmission of thevideo data, relieving bandwidth pressure and saving the storage space of a monitoring terminal are achieved.

In an embodiment, a compression unit is provided which may perform compression of images with low latency and relatively little hardware. Similarly, a decompression unit may be provided which may decompress the images with low latency and hardware. In an embodiment, the transmission of compressed coefficients may be performed using less than two passes through the list of coefficients. During the first pass, the most significant coefficients may be transmitted and other significance groups may be identified as linked lists. The linked lists may then be traverse to send the other significance groups. In an embodiment, a color space conversion may be made to permit filtering of fewer color components than might be possible in the source color space.

The embodiment of the invention provides a streaming media processing method, a server and a browser. The method comprises the following steps that when the situation that the live broadcast connection is established between a browser of a live broadcast terminal and a master server according to a live broadcast logo is detected, the master server acquires streaming media for live broadcast recorded by the browser of the live broadcast terminal based on a webpage real-time communication technology; the master server receives a streaming media acquisition request sent by a browser of at least one receiving terminal, and the streaming media acquire the live broadcast logo carried in the request; the master server forwards the streaming media recorded by the browser of the live broadcast terminal to the browser of the at least one receiving terminal. According to the streaming media processing method, the server and the browser provided by the invention, the load of the live broadcast terminal is reduced, and the bandwidth pressure of the live broadcast terminal is also reduced.

The invention relates to a direct-broadcast flow transmission method and apparatus. The method comprises: a current state of a client is obtained; on the basis of the obtained current state of the client, statistics of a ratio of the number of clients in a freezing state to the total number of clients is carried out; when the ratio exceeds a first threshold value, a current code rate of a direct-broadcast flow is obtained; and when the current code rate is higher than a predetermined code rate threshold value, the client is notified to adjust a transmission strategy to improving the transmission capability and enhance a proportion of direct-broadcast flow data obtained by the client from a content distribution network (CDN) server to direct-broadcast flow data obtained by a point-to-point (P2P) data transmission mode. According to embodiments of the invention, the transmission strategy of the client is adjusted based on the direct-broadcast flow code rate and the client state, so that the service and bandwidth resources are saved and the direct broadcasting smoothness is guaranteed.

The invention discloses a method, device and equipment for avoiding repeated alarms in a human body inspection and quarantine system. The method for avoiding repeated alarms in the human body inspection and quarantine system comprises the following steps that infrared video image data and visible light video image data of a human body are acquired, and heating personnel in the infrared video imagedata are measured by using an infrared thermal imaging temperature measurement method; when the heating personnel are monitored, a face image area of the heating personnel in a visible light video image is detected by using the matching relationship between an infrared image and the visible light image, and whether the face image area is tracked or not is judged; if the face image area is not tracked, screenshots of the face image area are saved, an alarm is given, and the face image area is tracked at the same time; and if the face image area is tracked, no alarm is given. The human body inspection and quarantine system is prevented from repeatedly alarming the same alarm personnel.

The embodiments of the invention disclose a data compensation method of a streaming media of a mobile terminal and a data compensation system thereof, and a content distribution network (CDN). The method comprises the following steps that: a personal computer (PC) sends a first node obtaining request message to a node scheduling server, receives a first obtaining response message returned by the node scheduling server, and determines whether there is a peer-to-peer (P2P) node storing low-bit rate content data of a streaming media program; if so, the low-bit rate content data of the streaming media program are downloaded from the P2P node and are stored; if not, the low-bit rate content data of the streaming media program are downloaded from a content server and are stored; the node scheduling server is required to add a first shared entry to a program shared state table, so that a follow-up mobile terminal can download the low-bit rate content data of the streaming media program from the PC. According to the embodiments of the invention, P2P sharing of content data of a streaming media program can be realized between a mobile terminal and a PC and thus sharing of the P2P can be effectively utilized to save network resources, thereby reducing the bandwidth stress and lowering system deployment and cost.

The invention discloses a video monitoring system. The video monitoring system disclosed by the invention can realize any promotion of video streams, and has low cost and good expansibility. The videomonitoring system disclosed by the invention solves the control problem of the video streams of a remote camera device by using a video stream pushing terminal, stores the video streams by using a streaming media server, controls the flow direction of the video streams by using a streaming media playing platform, and can push the specified video stream to a specified address by implementing an instruction through interaction control middleware, thereby pushing the video streams to any network node or providing video services externally; and the traditional tree structure feature is modified into a reticular structure by using the cloud service of the streaming server, thereby greatly reducing the pressures of the server and a network bandwidth and supporting rapid capacity expansion on demand.

An intelligent network frame device consists of remote server and client end including interactive layer for receiving operation request of display operation unit then sending it to core processing unit and for receiving data operation request of core processing unit then sending data operation result to display operation unit according to data operation request , and core processing unit for carrying out relevant data operation according to operation request sent by interactive layer and for ensuring synchronous refreshment of local data and remote server .

The present invention discloses a method and a route device which can realize the transmission of flow media data. The method is applicable to the route device between the flow media resource and the customer terminal to provide the data transmission. The aforesaid method comprises the steps as follows: terminating a request order of the terminal customer to the flow media resource, recording the information of the flow media resource and the customer terminal and respectively maintaining the connection between the customer terminal and the flow media resource; copying the data received from the flow media resource and send the data to the relative customer terminal according to the customer terminal information of the flow media resource. The present invention can be used for transmitting the data between the flow media resource and the customer terminal. The present invention comprises a recording unit, a customer terminal maintaining unit, a flow media resource maintaining unit and a copy and transmission unit. The present invention reduces the requirement to the capability of the flow media resource and the bandwidth pressure of the medium network, and the present invention ensures the quality of the service.

The invention discloses a coordinated transmission method for a streaming media file, a system and a device. The method, the system and the device are used for solving the problem that a present network bandwidth is insufficient to support a streaming media service with a higher demand on a transmission bandwidth. The coordinated transmission method comprises the following steps: obtaining downloading identifiers from a streaming media server by a master terminal, wherein the quantity of the downloading identifiers is as same as the quantity of terminals participating in the coordinated transmission; allocating the obtained downloading identifiers to each of the terminals participating in the coordinated transmission according to a principle of occupying different downloading identifiers by each of the terminals participating in the coordinated transmission; establishing a media downloading control linkage and a media downloading linkage by the master terminal and the streaming media server according to the allocated downloading identifiers; receiving a packet sent by the streaming media server based on the established linkage; and receiving the packet participating in the coordinated transmission sent from the terminals. According to the coordinated transmission method, a bandwidth pressure of the media linkage between the streaming media server and the master terminal is relieved and the requirement for the streaming media service with the higher demand on the transmission bandwidth is met.

A method and an apparatus for buffering and obtaining resources as well as a resource buffering system are disclosed. The method includes: obtaining an access popularity value of a resource from a database at intervals of a first preset time, and obtaining load values of cache servers from the database if determining that the access popularity value reaches a preset first threshold; and determining a duplication cache server for duplicating the resource according to the access popularity value and the load values, and notifying the duplication cache server to duplicate the resource from a cache server that stores the resource and buffer the resource. The method overcome the problems in the prior, namely, the same resource is buffered in only one cache server and the relatively idle cache servers are unable to share loads with busy cache servers. The method improves extensibility and working efficiency of the buffering system.

The present invention provides an information dissemination system based on the integration of the broadcast network and the Internet, including: a terminal storage module, a regional storage module, a statistics-and-analysis module, a recommendation-and-analysis module, a media content management module and a network scheduling module. By two-level storing and playing of media contents, along with the statistics and analysis for on-demand behaviors of users, the intervention of the operator, and the recommendation and analysis performed by the recommendation-and-analysis module, the media contents are classified. Further, according to the classification, a media content is transmitted to the terminal storage module via the Internet; or, is transmitted to the terminal storage module via the Internet and pushed to the terminal storage module via the broadcast network; or, is pushed to the terminal storage module or the regional storage module at a predetermined time via the broadcast network. Thus, the respective advantages of the broadcast network and the Internet can be integrated and also the bandwidth pressure and the pressure of concurrent accesses on the Internet can be alleviated.

The invention discloses a distributed type software upgrade method and device based on fingerprint identification. The method includes the following steps that S101, broadcasting is conducted in the local area network, response information is obtained, according to the response information, for a terminal, whether software is upgraded through a server nor the local area network is determined; when it is determined that the terminal upgrades the software through the server, the server is connected, and files needing upgrading are obtained from the server one by one; S105, when it is determined that the terminal upgrades the software through the local area network, the upgraded terminal in the local area network is connected, files needing upgrading are obtained from the upgraded terminal one by one; when the files needing upgrading are obtained, fingerprints of the files needing upgrading are checked and version information of a file corresponding to the local terminal is upgraded. a variety of optional download sources can be configured, the single service data size and the server bandwidth pressure are reduced, and the safety of file upgrading is improved.

The invention provides a backup mechanism for repeating data deleting of a cloud service, and belongs to the field of cloud storage service data backup. According to the backup mechanism, a data repeating deleting mechanism is carried out on a cloud service end, data are stored to a cloud storage pool in a unification mode, and only a file pointer is set in a personal folder of a client. The bandwidth pressure occupied by data uploading is greatly relieved, the load, for carrying out data repeating deleting, of a cloud storage server is reduced, and the storage cost is saved.

Embodiments of the present invention disclose a content replacement method, system and node in a CDN (Content Delivery Network), wherein the method comprises: a current node in the CDN receives a content access request sent from a user terminal; the current node queries whether a content fragment requested by the user is stored in the current node; if the content fragment is stored in the current node, the content fragment requested by the user is sent to the user terminal, and a popularity score for the content fragment requested by the user is updated in a popularity score table; otherwise, the content fragment requested by the user is pulled from an upper node and sent to the user terminal, an index and a popularity score for the content fragment requested by the user are established in the popularity score table, and the popularity score for the content fragment requested by the user is updated in the popularity score table. Through the adoption of the embodiments of the present invention, the content replacement in the CDN can be triggered according to the user access behavior so that current existing contents meet user requirements as far as possible, thereby saving the memory and bandwidth resources.

The invention discloses an image data generating system for stereo display. The image dagta generating system for stereo display comprises a two-dimensional image acquisition machine for acquiring the two-dimensional images of a target scene, a light ray shielding information acquisition machine for acquiring the continuous two images of the target scene and an image signal processor for acquiring the mutual shielding information of a scene object according the two-dimensional images of the target scene and getting the shadow information of the scene object according to the continuous two images of the target scene, wherein the continuous two images comprise target scene images with shadows and target scene images without shadows; and the image signal processor combines the shielding information and the shadow information to generate the depth image of the target scene. The image data generating system for the stereo display can simplify the data processing and reduces the transmission bandwidth.

The invention provides a large-amplitude regular movement background intelligent analysis and control system, which comprises a characteristic database, a video collection module, a quality enhancing module, a foreground extracting module, a characteristics analysis module, a characteristics recognition module, a control trigger rule judging module and a real-time control platform, wherein the characteristic database comprises a plurality of target and / or behavior model characteristics; the video collection module is used for processing the collected video signal and for obtaining the digital video sequence; the quality enhancing module is used for enhancing the quality of the video digital sequence; the foreground extracting module is used for eliminating the interference of the movement background and for extracting the foreground; the characteristics analysis module is used for the video digital sequence to perform the characteristic matching and classification; the characteristics recognition module is used for comparing the video object with a plurality of target and / or behavior model bank according to the parameters set by the user; the control trigger rule judging module is used for judging whether the rule is broken according to the rules and depth of field which are previously set by the user together with the video characteristics, and if the rule is broken, abnormal information is issued; and the real-time control platform is used for receiving the video analysis result and for issuing a control order according to the analysis result.

The invention relates to a distributed network measured data statistical analysis method. In the method, a two-stage distributed processing way consisting of probe equipment pre-processing and data analysis server post-processing is adopted. The method comprises the following steps that: probe equipment performs matching, pre-processing and fusing operations on acquired original data according to a data analysis rule issued by a data analysis server; and after the operations are finished, a front-end processing result and a part of fused data are transmitted to the data analysis server in the form of a report message through service Ethernet, and the data analysis server receives the fused data, performs post-processing on data pre-processing results summarized by probes and the fused data, and finally submits post-processed results and data to a user. Through adoption of the distributed network measured data statistical analysis method, the processing capability of the probe equipment can be brought into full play on the premise of keeping existing equipment and an existing configuration method, and the aims of lowering transmission overload of the service Ethernet, lowering computing overhead of the data analysis server and improving the overall performance of an analysis system are fulfilled.

The invention discloses a data processing method and device in a multistage server and a data processing system. The data processing method in the multistage server includes: setting a logic rule for dividing data into different data types for various stages of servers, and selecting data of the second type as data to be reported by the various stages of servers; dividing the data in each server into two kinds according to the logic rule; reporting second type of data in the next stage of server to the current stage of server, and storing the data in the current server to serve as data of the current stage of server; judging whether required data can be obtained from the data of the current stage of server when the current stage of server receives data access requests of a request side, returning the data obtained from the current stage of server to the request side if the required data can be obtained, otherwise obtaining the required data from a next stage of server of the current stage of server, and returning the data to the request side.