85results about How to "Improve horizontal scalability" patented technology

The present invention is a power load forecasting method based on big data technology and a research application system based on the method. With the support of big data technology, a variety of mature open source products are integrated to form a data source, data integration, data storage, Data calculation, data analysis, implementation of electrical load characteristic analysis and electrical load forecast analysis. The invention effectively improves the efficiency of mass data processing, and solves the limitations of traditional statistical analysis assumptions and judgments. It can scientifically and accurately predict the electricity demand of the future power, which is conducive to the peak-shaving and valley-filling and stable operation of the power grid, and provides decision-making support for the company's power grid planning, equipment maintenance, and power deployment.

The invention provides a TCP long connection communication system and method. The system comprises a TCP communication module, a message and data management module, an information data storage module, a distributed memory cache module and a load shunt module, wherein the TCP communication module is used for receiving data requests sent by terminal equipment, calling corresponding service logical processing interfaces in a service logical processing module according to request information, receiving operation results returned by the message and data management module and sending TCP communication data and instructions or control grammars to the terminal equipment; the message and data management module is used for calling the service logical processing interfaces to conduct management and operation on messages and data; the information data storage module is used for managing equipment information and communication system service data; the distributed memory cache module is an open source distributed memory cache system; the load shunt module is used for dynamically distributing current requests to a server with the current minimum load when receiving new TCP connection requests. The abstract, event-driven and asynchronous TCP long connection communication system and method are based on transport protocols can achieve high concurrency, high real-time performance and dynamic expansion.

The invention discloses a short link processing method and apparatus, and a server. The method comprises the steps of receiving an application request of a requester for a short link, wherein the application request comprises a first long link network address; calculating a corresponding first database table identifier according to the first long link network address; judging whether a first database corresponding to the first database table identifier is in an own access range or not; and if it is determined that the first database is in the own access range, obtaining a corresponding first short link network address in the first database, and sending the first short link network address to the requester. In a sharding mode, the conversion from the long link network address to the corresponding short link network address can be realized, and the corresponding short link network address is unique and short, so that the horizontal expansion capability of a short link storage space can be enhanced and the short link processing speed can be increased.

A big-data-oriented distributed density clustering method comprises the following steps that firstly, environment virtualization is performed, and a Hadoop platform is established; secondly, data are pre-processed and loaded, wherein an original data table is extracted from a database, a needed field is intercepted through a sqoop-query command, and the pre-processed data are directly extracted to an Hdfs; thirdly, a distance matrix is calculated; fourthly, a cut-off distance and dot density are calculated; fifthly, the minimum distance between a dot and a higher-density dot is calculated; sixthly, the critical distance of a critical density point and a critical density point are determined; seventhly, dot clustering is performed, so a final clustering result is obtained; eighthly, off-group points are removed. The big-data-oriented distributed density clustering method is fast and effective when a big data set is processed, and has the effect that input parameters have good robustness on the clustering result.

The invention provides a GaN-based light emitting diode epitaxial structure. From bottom to top, the structure successively comprises a substrate, a stress control layer, an n-type GaN layer, a stress buffer layer, a multi-quantum well active layer, an interval layer, an electron blocking layer, a p-type GaN layer and an ohmic contact layer. The light emitting diode epitaxial structure also comprises a current extension layer whose forbidden bandwidth is greater than the forbidden bandwidth of the n-type GaN and which is arranged below the multi-quantum well active layer. By using the epitaxial structure, electronic leakage in an active region when a light emitting diode works under a heavy current condition can be effectively reduced, radiation recombination efficiency of a hole-electron pair can be effectively increased and an antistatic characteristic of a LED is improved.

Disclosed is a growth method of a light-emitting diode (LED) epitaxial wafer. When growth of an N type doped gallium nitride (GaN) layer with silicon (Si) is carried out, a method of alternate growth of the N type doped GaN layer with Si and a U type undoped GaN layer with no Si is adopted. According to the growth method of the LED epitaxial wafer, the thickness of an original N type GaN layer is kept, a doped GaN layer with Si with continuous growth in a traditional growth method is modified into an alternate structure layer formed by the N type doped GaN layer with Si and a U type GaN layer with no doped Si, and usage amount of dopant is saved to a certain degree. In an alternate structure of the N type doped GaN layer with Si and the U type undoped GaN layer with no Si, the doped GaN lay with Si is low in resistance value, and the undoped GaN layer with no Si is high in resistance value. Due to the facts that the doped GaN layer with Si is low in resistance value, and the undoped GaN layer with no Si is high in resistance value, horizontal expansion capacity of electrons is strengthened, and thus a driving voltage is reduced, and meanwhile luminance and a lighting effect are improved.

The present invention discloses a migration balancing policy based electricity-consuming information distributed file storage method and apparatus. The method comprises: a data acquisition module analyzing original data frames to obtain original data, storing the original data into a local disk module in the form of a file, and uploading the original data frames to a relational database and a MongoDB database; detecting information of the file in the local disk module, and selecting whether to upload the file to a data storage module; the MongoDB database acquiring a node load where a slice is located in mongos by means of routing and by using a migration balancing policy to implement balancing from two aspects: a data volume and a load; a big data management engine regularly monitoring file information that is not subjected to data saving calculation, extracting the file from the MongoDB database, and using time as a tag; converting file data extracted from the MongoDB database into data character string streams, and carrying out calculation to obtain service data; and finally, performing data persistence. Therefore, efficient storage of electricity-consuming information big data is implemented, and electricity-consuming big data storage capacity and storage rate demands of an electricity-consuming acquisition system are met.

The invention relates to an alarm method for an abnormal load of a distribution transformer, and belongs to the field of distribution transformer alarming. According to the invention, periodic fluctuations of the electric power load of the distribution transformer are comprehensively considered through selecting load information data, such as historical load data of the distribution transformer, data of a region to which the distribution transformer belongs, equipment information, customer information data and current and voltage data, of the distribution transformer in a power utilization information acquisition system, a distribution transformer load early warning model is built, and early warning analysis is performed on phenomena such as overloading, excessive loading, heavy loading, underloading and three-phase imbalance so as to solve problems in the prior art that the efficiency is low when mass, fuzzy and messy data is processed, related applications such as big data analysis and mining of the distribution transformer cannot be well supported and the like.

The invention discloses a user page behavior trajectory-based white list generation system and a user page behavior trajectory-based white list generation method and belongs to the white list generation field in video live streaming. According to the system, a trajectory recording unit is arranged in each user terminal; a caching and preprocessing unit, a distributed real-time computing unit and a white list output unit are arranged in a live streaming platform server; the trajectory recording unit is used for recording page skip trajectory information and submitting the page skip trajectory information to the live streaming platform server; the caching and preprocessing unit is used for caching and preprocessing the page skip trajectory information and sending the page skip trajectory information to the distributed real-time computing unit in the form of data segments; and real-time computing subunits in the distributed real-time computing unit can analyze and verify page behavior information in the data segments, and add user IDs which pass verification to a white list. The user page behavior trajectory-based white list generation system and method of the invention have high real-time performance and high precision, can meet application requirements of various special scenes and ensure the effective proceeding of interactive activities.

The invention discloses an active user set maintenance system and method based on a time wheel and a player heartbeat, and relates to the technical field of maintenance of an active user set in live broadcast. The method comprises the steps of: recording heartbeat information by each user terminal and reporting the heartbeat information to a live broadcast platform server; caching and pre-processing the heartbeat information by the live broadcast platform server, and transmitting the heartbeat information to a distributed real-time calculation module by means of a plurality of pieces of data; analyzing and verifying the heartbeat information by the distributed real-time calculation module, and adding users corresponding to the heartbeat information which passes the verification to an active user set slice corresponding to the current time; and taking an active calculation time period as a cycle by an active user set function module, and updating the active user set slice corresponding to the current time to the created time wheel periodically. According to the active user set maintenance system and the method based on the time wheel and the player heartbeat provided by the invention, the active user set is updated and maintained based on the player heartbeat and the time wheel, which is not only timely and efficiently, but also causes no larger load on the system.

The invention provides a system used for processing a high-concurrency data request of a forum. The system comprises a database server, a data cache server, an application server and a page cache server. According to the system, by changing a conventional website architecture, the page cache server is additionally arranged at the front end of the application server, a cache server is additionally arranged between the application server and a storage server, and HTML page source codes are cached by utilizing the page cache server, so that logic operation work of the application server is reduced and the application server is well protected; and page data is cached by utilizing the data cache server, so that the frequency of visiting the database server is reduced, the application server and the database server are well protected, and the pressure resistance and stability of the whole system are improved.

The invention provides an epitaxial growth method capable of improving the anti-static electricity capacity of an III-V class compound semiconductor LED (light emitting diode) chip. An LED epitaxial structure comprises a GaN non-doped layer, an N-type doped GaN layer, an n-type AlGaN / GaN composite insert layer, a high-temperature P-type GaN layer and a P-type contact layer, wherein the n-type AlGaN / GaN composite insert layer specifically comprises the following steps of growing a layer of the n-type AlGaN layer of which the content of Al components is gradually increased; growing an n-type doped AlGaN layer of which the content of the Al components is constant; growing a layer of n-type AlGaN and GaN composite structural layer of which the content of the Al components is gradually reduced, and after the temperature is increased to 1,000-1,200 DEG C, growing a GaN layer. According to the epitaxial growth method, the n-type AlGaN and GaN composite structural layer with the variable content of the Al components is inserted between an N-type GaN layer and a shallow trap layer, so that breakdown of a device, which is caused by ultrahigh regional current, is effectively avoided, and the anti-static electricity capacity of the device is improved; the piezoelectric field effect of the instrument is reduced; the injection efficiency for a current carrier is improved, and the luminous efficiency of a GaN-based LED is improved.

The invention discloses an epitaxial wafer for a gallium-nitride-based light emitting diode, and a preparation method for an epitaxial wafer, and belongs to the technical field of semiconductors. The epitaxial wafer comprises a sapphire substrate, a GaN buffering layer, a non-doped GaN layer, an N-type GaN layer, a multi-quantum-well layer, a P-type electron blocking layer and a P-type layer, wherein the GaN buffering layer, the non-doped GaN layer, the N-type GaN layer, the multi-quantum-well layer, the P-type electron blocking layer and the P-type layer are sequentially stacked on the sapphire substrate. The P-type layer is formed by the alternate stacking of a P-type GaN layer and a graphene film layer. According to the invention, the P-type GaN layer and the graphene film layer are alternately stacked to form the P-type layer, and the graphene film layer can prompt the hole transmission and also can improve the lateral extension capability of hole. Moreover, the graphene film layer also can improve the energy band matching of the P-type electron blocking layer and the P-type GaN layer, improves the electrical contact of the P-type electron blocking layer with the P-type GaN layer, improves the capability of a hole injection multi-quantum-well layer and finally improves the luminous efficiency of a light-emitting diode.

The invention discloses a GaN-based light emitting diode epitaxial wafer and a manufacturing method therefor and belongs to the field of light emitting diodes. The GaN-based light emitting diode epitaxial wafer comprises a substrate, a buffer layer, a three-dimensional growth layer, a u-GaN layer, an n type layer, an n type current expansion layer, a stress releasing layer, a multiple quantum well layer and a p type layer; wherein the buffer layer, the three-dimensional growth layer, the u-GaN layer, the n type layer, the n type current expansion layer, the stress releasing layer, the multiple quantum well layer and the p type layer orderly arranged on the substrate in a covering manner; the n type current expansion layer comprises a first sub-layer covering the n type layer and a second sub-layer covering the first sub-layer; the first sub-layer is an AlGaN layer, the second sub-layer is of a superlattice structure formed by a non-Si-doped GaN layer and an Si-doped GaN layer, the first sub-layer is higher than the n type layer and the second sub-layer in terms of growth temperature, and the second sub-layer is greater than the first sub-layer in terms of growth pressure. The GaN-based light emitting diode epitaxial wafer is capable of overcoming defects and improving photoelectric performance.

The invention discloses an LED chip electrode and a manufacture method thereof. The electrode includes a first metal adhesion layer, a second metal adhesion layer arranged on the first metal adhesion layer, a metal protection layer arranged on the second metal adhesion layer, and a metal routing layer arranged on the metal protection layer. The material of the second metal adhesion layer is the metal Cr. The material of the metal protection layer is the metal Ti. The second metal adhesion layer and the metal protection layer fuse and form a Cr-Ti contact resistor. In the electrode the metal protection layer with material of Ti is formed, the second metal adhesion layer and the metal protection layer in the electrode fuse and form the Cr-Ti contact resistor. The horizontal expansion capacity of the LED chip is effectively improved, and the luminance of an LED chip is enhanced.

The invention discloses a high concurrency implementation method based on a micro-service framework, and relates to the technical field of distributed systems. According to the technical scheme, front-end request shunting is carried out by using nginx. The transverse expansion capability of the distributed system is improved; a rabbion is used for loading a rear-end instance, and starting a multi-instance load for a certain service monomer application with a high-concurrency service requirement. Therefore, the high concurrent processing capability is improved. Read-write separation and master-slave backup are carried out by configuring a database. The occurrence of meter locking is reduced, the query speed is increased, and the consistency of master and slave library data is ensured. Common configuration data are cached by adopting redis, reading from a disk is avoided, the pressure of a database is reduced, the data reading efficiency is greatly improved, and finally, the parameters of middleware are optimized by adjusting the parameters influencing the high concurrency performance in the distributed system, so that the high availability of the distributed system in a high concurrency production environment is realized.

The invention provides a service framework construction method and system based on a plug-in engine, at least one process is constructed, each process is loaded with at least one plug-in, a local forwarding routing table, a network forwarding routing table, a virtual base class and an entity class are constructed at the same time, wherein the entity class inherits from the virtual base class; then an interface parameter containing process or plug-in sending information is provided in the entity class, and a service parameter serving as an information carrier is provided in the virtual base class, thereby realizing communication connection between the process and the plug-in according to the interface parameter; the local forwarding routing table or the network forwarding routing table is selected according to the service parameters to realize transmission of the information carrier between the plug-in and the process. According to the method and the system, service requirements can be quickly responded, the development efficiency is improved, the scalability of the system is enhanced, a framework mode is beneficial to dynamic capacity expansion, original service functions are not influenced, distributed or cluster deployment can be realized, and the lateral expansion capability of services is enhanced.

The invention provides a method, a device, a system and a medium for generating an identifier. The method for generating the identifier comprises the following steps of obtaining first time information representing generation time of a first identifier, obtaining identification information used for generating an instance of the first identifier, obtaining a serial number of the first identifier, and generating the first identifier based on the first time information, the identification information of the instance and the serial number of the first identifier. The invention further provides a method, a device, a system and a medium for generating the identifier. The method for generating an identifier includes that a request is received to generate an identifier for a specific service, in response to the request, one or more instances for providing an identifier generation service for the particular service are determined, and an identifier is generated by one of the one or more instances according to the identifier generation method provided by the present disclosure.

The invention discloses a gallium nitride-based light emitting diode epitaxial wafer and a manufacturing method thereof, belonging to the field of semiconductor technology. An epitaxial wafer includesa substrate, a buffer layer, an N-type semiconductor layer, an active layer and a P-type semiconductor layer. the buffer layer, the N-type semiconductor layer, the active layer and the P-type semiconductor layer are orderly stacked on the substrate. The P-type semiconductor layer comprises M undoped scandium aluminum nitride layers and (M+1) P-type doped gallium nitride layers. The M scandium aluminum nitride layers and the (M+1) gallium nitride layers are alternately laminated, and M is a positive integer. At least one undoped scandium aluminum nitride layer is inserted into the p-type dopedgallium nitride layer, The interface between scandium aluminum nitride layer and gallium nitride layer has strong two-dimensional hole gas, which can effectively improve the lateral expansion abilityof holes in the P-type semiconductor layer, reduce the series resistance of LED, and further reduce the forward voltage of LED, which is conducive to the application of LED in civil lighting.

The invention discloses an active user set maintenance system and method based on a time wheel and user behaviors, and relates to the technical field of maintenance of active user sets in live video streaming. The method comprises the following steps that: user terminals record page behavior information and submit the page behavior information to a live streaming platform server; the live streaming platform server caches and pre-processes the page behavior information, and transmits the page behavior information to a distributed real-time computing module in the forms of a plurality of data slices; the distributed real-time computing module resolves and verifies the page behavior information, and adds users corresponding to page behavior information which passes verification into active user set fragments corresponding to current time; and an active user set functional module periodically updates the active user set fragments corresponding to the current time into a created time wheel by taking active computing time intervals as periods. Through adoption of the active user set maintenance system and method, timeliness and high efficiency are realized; relatively large load is not caused on the system; and interactive activities are performed effectively.

The invention provides a Redis multi-instance monitoring system and method, and relates to the technical field of cluster operation state monitoring, and the system comprises a state acquisition module which is integrated in Redis instances and is used for acquiring performance indexes; an acquisition mode processing module which is used for adjusting the mode of the state acquisition module according to different monitoring scenes; a monitoring core module which is used for calling a performance index, performing state analysis and alarm analysis according to a configured monitoring rule, and generating a state analysis result and alarm information; a data storage module which is used for data persistence; an alarm management module which is used for acquiring alarm information, generating a state adjustment instruction according to the alarm information, and driving a service discovery module to process an abnormal node in a Redis container; the service discovery module which is used for monitoring a container environment where the Redis instance is located and performing operation and maintenance management on nodes of the Redis container; and a visualization module which is used for displaying a monitoring data report through a visualization interface.

The invention discloses an LED epitaxial structure growth method. The method comprises the following steps: processing a substrate, growing a low-temperature buffer layer GaN, growing an undoped GaN layer, growing a Si-doped N-type GaN layer, growing a multi-quantum well layer, growing an AlGaN electronic barrier layer, and growing a Mg-doped P-type GaN layer, and performing cooling, wherein the growth of the multi-quantum well layer sequentially comprises the steps of growing an N2 gradient atmosphere InxGa(1-x)N-1 layer, growing an H2 gradient atmosphere InxGa(1-x)N-2 layer, growing an H2 and N2 mixed atmosphere InxGa(1-x)N-3 layer, growing a GaN layer and growing a GaN layer. According to the method, the problems of low quantum well growth quality and low quantum well radiation recombination efficiency in the existing LED epitaxial growth method are solved, so that the luminous efficiency of the LED is improved, and the forward driving voltage is reduced.

The embodiment of the invention provides a short message channel connection processing method, device and system, equipment and a storage medium. The method comprises the following steps: receiving target state information of a gateway server, wherein the gateway server belongs to at least one gateway server group, the gateway server group corresponds to a short message channel, and the short message channel supports a first number of connections; scheduling connection of a target short message channel in a target gateway server group according to the target state information, wherein the target gateway server side group is a group to which the gateway server side corresponding to the target state information belongs, and the target short message channel is a short message channel corresponding to the target gateway server side group. According to the embodiment of the invention, the actual sending capability of the packet of the gateway server can be matched with the sending capability of the short message channel as much as possible, the resource waste of the short message channel is reduced, and the short message backlog is reduced.

The invention belongs to the technical field of semiconductors, and particularly relates to a light emitting diode and a manufacturing method thereof. The light emitting diode at least comprises an N-type layer, an active layer and a low-temperature P-type GaN layer and a P-type contact layer located on the active layer. A non P-type nitride layer is inserted between the low-temperature P-type GaN layer and the P-type contact layer to replace a high-temperature P-type GaN layer, damages and light absorption to the lattice quality of a quantum well by high-temperature growth can be reduced, the high-density two-dimensional electron gas in the non P-type nitride layer increases the current lateral expansion capability, partial device over heat and easy breakdown features caused by current congestion are improved, and the antistatic ability of the LED device is enhanced.

A method for synchronization of cross-system data based on an account number is characterized in that a service data synchronization task based on the account numbers is formed; synchronization execution periods are configured according to synchronization demands of different services; and the different account numbers can execute synchronization of the service data at regular time in different service scenes. (1) Correlations between the cross-system user account numbers and the services with the data synchronization demands are established; (2) when a time point requiring execution of a synchronous planning task is reached, namely a service data synchronous planning cycle is passed, a synchronous planning task system will trigger establishment of a data synchronous execution task; (3) the synchronous planning task system releases and distributes the synchronous execution task, wherein when the time reaches an expected time point of the synchronous execution task, the synchronous planning task system will plan a task release subsystem from an execution task base; and (4) the synchronous execution system acquires the synchronous execution task sent by the planning system, and the data synchronization can be completed at a designated moment according to specific contents of the task.

The application discloses an LED epitaxial growth method suitable for a small-spacing display screen. The method comprises the following steps: processing a substrate; growing a low-temperature bufferlayer GaN; growing an undoped GaN layer; growing a Si-doped N type GaN layer; growing a multi-quantum well layer; growing an AlGaN electronic barrier layer; growing a Mg-doped P type GaN layer; and then performing cooling. The step of growing the multi-quantum well layer sequentially comprises Mg-doped pretreatment, InGaN cover layer growth, InGaN well layer growth, H2 atmosphere InGaN:Si layer growth, N2 atmosphere InGaN: Mg layer growth, H2 and N2 mixed atmosphere InGaN: Mg / Si layer growth, InGaN protective layer growth and a GaN barrier layer growth. The method provided by the invention solves the problem that the blue shift amount of the light emitting wavelength of the LED is large in the existing LED epitaxial growth, and the light emitting efficiency of the LED is improved, the working voltage is reduced and the anti-static capability is enhanced.

The invention provides a monitoring system, comprising a monitoring device, a control device, an application programming interface (API) device and a monitoring consumption device, wherein the monitoring device is connected with a data bus and a control bus respectively, and used for monitoring resources and sending the collected monitor information to the data bus, and further used for sending the corresponding system information of the monitoring device to the control bus; the control device is connected with the control bus and used for acquiring the system information in the control bus and storing the system information; the API device is connected with the control bus, and used for providing an access interface; the monitoring consumption device is connected with the data bus, and used for acquiring the monitor information in the data bus and storing the monitor information. According to the monitoring system, the fusion of monitoring on different resources can be achieved, so that the user can acquire the monitor information of the different resources in time, and when the monitoring system needs to be expanded, the monitoring device can be quickly added, thus the horizontal expansion capability of the system is effectively improved.

The invention discloses a distributed fault diagnosis system and method for a power transmission line. The distributed fault diagnosis system comprises a monitoring terminal, a front server and a central station server. The central station server comprises a communication micro-service unit, a core micro-service unit and a fault distance measurement micro-service unit; the front server receives the data message sent by the monitoring terminal and stores the data message in a database; the communication micro-service unit reads the data message from the database and forms a file; the core micro-service unit receives an interaction message sent by the communication micro-service unit, finds out a corresponding fault distance measurement micro-service unit based on ID information in the interaction message, and sends a file obtained from the communication micro-service unit to the fault distance measurement micro-service unit; and the fault distance measurement micro-service unit completes fault diagnosis based on the received file. According to the invention, the system access time of fault distance measurement terminals (namely fault distance measurement micro-service units) of different manufacturers can be greatly shortened, and the operation and maintenance efficiency is improved.

The invention discloses a nitride light-emitting diode. The nitride light-emitting diode comprises a substrate and a nucleation layer, a buffer layer, an n-type layer, a light-emitting layer, a P-typelayer and an N-type contact layer which are sequentially arranged on the substrate, the nitride light-emitting diode is characterized in that the N-type contact layer comprises a u-GaN / n-GaN superlattice structural layer. A u-GaN / n-GaN superlattice structure or a u-GaN / n-GaN superlattice structure and an n<+> GaN layer as the N-type contact layer, the contact resistance of a current expansion layer on the N-type contact layer is reduced, the horizontal expansion capability of an injection current is improved, so that the anti-static performance can be improved, and an operating voltage is reduced.

The invention discloses a multi-model intelligent robot system and a construction method, and the system comprises a digital operation service layer which comprises a strategy center, an operation center and basic services; the strategy center is used for scheduling and deciding to use various robot models to cope with different customer responses; the operation center is used for data analysis and model training of the multi-modal intelligent robot; the basic services are used for providing standard management function services of the digital operation service; the system also comprises an intelligent middle platform layer used for responding to the digital operation service layer and providing a plurality of robot models supporting semantic comprehension and task type multi-round voice dialogue services; a data service layer used for providing data storage and query services; and an interface service layer is used for integrating with a corresponding standard service interface for calling. The invention can improve the personification service capability of the intelligent robot, supports the multi-model service combination, achieves the efficient optimization operation of the robot, and is suitable for more business scenes.