29 results about "Weighted round robin" patented technology

The invention relates to the technical field of solid-state storage, in particular to a solid-state disk data consistency control method, and aims to solve the problems that a solid-state disk is poor in data consistency, high in power failure loss risk and difficult to balance writing efficiency and consistency under the scenes of power failure, parallel writing and garbage collection. Full-process consistency control is realized by constructing a'cache-persistent cache 'layered architecture and combining a multi-channel write synchronization, power failure emergency migration and periodic check repair mechanism: receiving data and storing the data in a cache firstly, distributing a multi-channel write task according to weighted polling and synchronizing the state; the standby capacitor is triggered to supply power during power failure, and the cache data is migrated to the persistent cache according to the FIFO; and according to a configurable period of 1-24 hours, flash memory block data are verified through SHA-256 hash, and when the flash memory block data are inconsistent, the flash memory block data are repaired from a persistent cache / backup block. According to the method, the data consistency compliance rate of the solid-state disk is larger than or equal to 99.9%, the power-down data loss rate is smaller than or equal to 0.01%, the write-in delay is reduced by 15%-20%, and the method is adaptive to consumption-level, enterprise-level and industrial-level solid-state disks.

The invention discloses a multi-AGV cooperative scheduling system for realizing load balancing. The multi-AGV cooperative scheduling system comprises a task management module, a state monitoring module, a load balancing scheduling module, an AGV cooperative scheduling module, an edge server optimization module, a task compression module, a task parallel processing module and a performance evaluation module. Wherein the load balancing scheduling module adopts a mode of combining first-stage weighted polling initial allocation and second-stage optimal allocation based on multi-agent deep reinforcement learning, the AGV collaborative scheduling module adopts an AxTD3 method to divide an Actor network into two parallel sub-networks, and joint training is carried out by sharing a Critic network. The edge server optimization module takes execution time deviation as a state, constructs a lightweight state vector in combination with a server isomerism index, and introduces federated learning and multi-Critic network architecture. According to the invention, adaptive collaborative scheduling of multiple AGVs in an edge computing environment is realized, and the system load balancing capability, the resource utilization efficiency and the task processing response speed are effectively improved.

The invention discloses a question and answer system knowledge base shunting method based on a weighted polling algorithm, and belongs to the technical field of artificial intelligence retrieval enhancement generation. In order to overcome the defects of a single knowledge base architecture of a traditional RAG system in the aspects of multi-source knowledge management, gray updating and load balancing, a cluster architecture comprising a main knowledge base, an auxiliary knowledge base and a test knowledge base is constructed, and intelligent flow distribution is achieved through a flow distribution controller. The system adopts a multi-dimensional weight calculation model, fuses a user role coefficient, a knowledge base state coefficient and a performance coefficient, and dynamically calculates the weight of each knowledge base according to a weight calculation formula; using a weighted polling algorithm to accurately distribute user requests; and automatically adjusting the weight of the knowledge base in the gray scale test period. The vector retrieval module converts questions into vectors to retrieve in a target library, and a result fusion engine resolves multi-library answer conflicts through weighted voting. After the method is applied, the update failure rate of the knowledge base is reduced from 25% to 5%, the question and answer accuracy is improved by 22%, and the response speed is improved by 40%.

This invention discloses a dynamic optimization method for network switch port WRR scheduling resources based on convex optimization modeling. Applied to the field of switching equipment scheduling and control, it addresses the problem that existing scheduling weight configurations cannot adapt to environmental changes, resulting in unstable forwarding performance of switching equipment. This invention establishes constraint functions based on port bandwidth, token resources, and queue number, collects the arrival probability distribution of data packets in each queue within the switch port to establish a distribution model, and combines the constraint functions and the collected model to construct a convex optimization problem for switch port queue resource allocation. The objective of the optimization problem is defined as minimizing the sum of queuing probabilities of the switch queues. The problem is solved using an existing convex optimization solver, and the results are configured as the weights of each port queue based on the Weighted Round Robin (WRR) scheduling algorithm. This invention solves the problem of adaptive resource allocation during dynamic network traffic exchange, enabling network switching equipment to provide stable network performance under different service loads.

The application relates to a multi-dimension dynamic weighted polling scheduling method and system for bank-enterprise direct connection query, which comprises the following steps: acquiring account information of all accounts of an enterprise and initializing a sorting value of each account; screening out accounts to be queried in a current batch according to the initialized sorting value of each account; calling a bank-enterprise direct connection interface to query the balance or transaction details of the accounts in the current batch; calculating an average transaction number interval based on the transaction number of each account and adopting a zero-elimination average method; respectively calculating a new basic weight value, a historical transaction weight value and a real-time transaction weight value by adopting a specific calculation method; generating an account weight value; generating a new sorting value for each account in the current batch according to the account weight value; screening out accounts to be queried in the next batch of the current batch according to the new sorting value; and stopping until a preset iteration number is reached, completing the polling scheduling of the enterprise account query and being suitable for efficient query scheduling scenes in a large-scale bank account environment.

A device includes a plurality of digital signal processors (DSPs) and analog crossbars in communication with the DSPs. The DSPs and crossbars dynamically adjust power consumption based on traffic load and / or a signal quality, leveraging traffic prediction models to optimize resource allocation. The system supports features such as powering down unused lanes, asymmetrically deactivating transmission or receiving lanes, and preemptively scaling resources during predicted traffic increases. Dynamic voltage and frequency scaling (DVFS) optimizes power usage based on link utilization and traffic priority, minimizing latency impacts during transitions. The switch controller integrates traffic prioritization algorithms, including weighted round-robin (WRR), to allocate power and bandwidth efficiently to high-priority traffic flows. Redundant crossbars handle overflow traffic or failover scenarios, transitioning between standby and active states dynamically.

The invention relates to a QOS (Quality of Service) strategy selection and queue scheduling method of a base switching network, which is characterized by comprising the following steps of: firstly, after network equipment receives a data frame, executing a QOS strategy selection operation by the network equipment, selecting a priority mapping table corresponding to the data frame from five different priority mapping tables, querying in a priority mapping table corresponding to the data frame to obtain a priority ID of the data frame; then, the network equipment queries a priority ID queue mapping table based on the priority ID of the data frame to obtain an output queue ID, and inserts the data frame into an output queue corresponding to the output queue ID; and finally, the network equipment transmits the output queue with the data frames to receiving equipment at an output port through a weighted polling scheduling (WRR) algorithm. According to the method and the device, a plurality of priority mapping tables are preset, so that the QOS model can process the data frames in various formats, and the types of compatible and supported data traffic are expanded.

The invention discloses a method for releasing a server program by one key, which realizes full-process automatic release through cooperation of six core modules: an interactive trigger module supports multi-end triggering of a Web interface, an API interface and a command line, provides lightweight operation only needing eight core configurations, and supports historical configuration multiplexing and network interrupt caching; the configuration analysis and verification module is used for presetting multiple types of application analysis rules, executing multi-level verification of a basic format, service logic and resource availability, and providing an automatic correction suggestion when verification fails; the environment adaptation and resource scheduling module remotely detects server hardware, software and network states, automatically adapts to scripts corresponding to an operating system and an application type, allocates resources through a weighted polling algorithm, and supports cross-regional network optimization; the real-time monitoring and feedback module collects logs and state indexes in a second level. The whole release process is free of manual intervention, the deployment time consumption is remarkably shortened, and the error rate is reduced to a low level.

The invention provides an alarm computing load balancing method, which belongs to the technical field of cloud computing and distributed monitoring operation and maintenance, adopts a distributed architecture, and comprises a plurality of Prometheus examples, a scheduling management module, a task distributor, an equal division algorithm executor and a shared storage component. The scheduling management module dynamically fragments the alarm rules according to rule types, data source features or time dimensions, and adjusts the number and size of fragments according to real-time loads. And the equal division algorithm executor combines the load, the performance index and the resource utilization rate of each instance, adopts strategies such as consistent Hash, polling or weighted polling to calculate an assignment relationship, and supports newly-added rule allocation and fragment migration. Dynamic management and real-time synchronization of rules are achieved through the shared storage component, stateless deployment of Prometheus instances is achieved, requests are distributed through the load balancer, and therefore the load balancing capacity, the rule management efficiency and the alarm timeliness are improved.

The invention discloses a power distribution communication network resource allocation method and device, and belongs to the technical field of network resource allocation, and the method comprises the steps: obtaining the business content of a new access business and a target resource requirement; judging the service type of the new access service based on the service content; when the service type of the new access service is an instruction control type, dividing the new access service to a core guarantee layer of the power distribution communication network, and distributing network resources to the new access service according to a target resource requirement by adopting a preemptive priority scheduling method; and when the service type of the new access service is a data processing service, dividing the new access service to a data support layer of the power distribution communication network, and distributing network resources to the new access service according to the target resource requirement by adopting a weighted polling method. Therefore, by implementing the method and the device, reasonable resource scheduling can be realized, so that the resource utilization rate is effectively improved.

The invention relates to the technical field of enterprise management systems, in particular to an enterprise management system optimization method based on background data, which comprises the following steps of: performing hierarchical encryption on sensitive data, dividing a core original layer, an aggregation middle layer and an abstract display layer, and matching a differentiation algorithm; then calculating a permission score through a hybrid model, and dynamically controlling data presentation; an encryption queue is optimized based on real-time load activation weighted polling or a minimum connection number algorithm, an access range is shrunk by combining access map pruning, and threats are isolated; switching a national cipher SM9 protocol and dual-channel redundancy transmission according to channel quality, and realizing full-life-cycle anomaly detection through a data blood relationship chain and a graph neural network; finally, the dynamic presentation rule is optimized through reinforcement learning iteration. According to the method, accurate protection of sensitive data, dynamic adaptation of system loads, controllability of authority fine grit and traceability of a data circulation full link are achieved, the service adaptability and user experience are improved, compliance requirements are met, and stable and efficient support is provided for digital transformation of enterprises.

The invention discloses a multi-protocol intelligent cockpit interconnection integrated system, which relates to the field of communication compatibility, and comprises the following steps: S1, a hardware interface module of a universal interconnection interface layer receives an access request of a mobile phone, and establishes communication connection between the mobile phone and a vehicle machine; original data transmitted by the mobile phone is subjected to standardized primary processing through a unified data transmission protocol of the software interface module; s2, a dynamic identification module of the protocol adapter layer adopts a feature code scanning technology to identify the type of an interconnection protocol accessed to a mobile phone within 100ms; the type of the interconnection protocol comprises a CarPlay, an Android Auto, a HiCar and a CarLink; calling a protocol conversion engine to carry out adaptation aiming at different protocols, and converting data in a mobile phone protocol format into data in a system general data format; and S3, a resource scheduler of the central control layer adopts a DWRR dynamic weighted polling algorithm to dynamically allocate CPU and memory resources of the vehicle machine according to the protocol type identified in the step S2 and the current service scene.

The application provides an implementation method and device for application atomization fusion arrangement based on OPENAPI, relates to the technical field of computer application, and covers service disintegration and interface definition, service registration and discovery, arrangement and adaptive execution, data processing and mapping, security control, integrated deployment and operation and maintenance steps; the application splits complex business into atomized services, defines interfaces according to the OpenAPI specification, facilitates service management, reduces integration cost, and improves system flexibility and expansibility; load balancing is realized by constructing a service registration center and combining a weighted round robin algorithm to guarantee efficient and stable service calling; API gateways, containerization technology and automatic operation and maintenance systems are used to realize rapid service deployment, elastic scaling and intelligent operation and maintenance; meanwhile, data is processed through semantic analysis and format conversion, and security is guaranteed through multi-level identity authentication and encryption.

The application discloses a time slot scheduling system and method, wherein the system comprises a time slot management module and a weighted polling scheduler; the time slot management module is used for determining the data interaction path delay between a master device and a slave device based on the time sequence delay of a hardware unit in the data transmission path between the master device and the slave device; the data interaction path delay between the master device and each slave device is used for performing the arbitration of the weighted polling scheduling authority, generating a weighted polling request scheduling signal based on the arbitration result, and sending the weighted polling request scheduling signal to the weighted polling scheduler; the weighted polling scheduler is used for polling and selecting the slave device command queue participating in the weighted polling scheduling, and determining the arbitrated slave device command queue; the application can dynamically control the scheduling time slot of the slave device based on the write command length and the data interaction path delay between the master device and the slave device, and inhibit the local congestion caused by the continuous data flow injection into the on-chip interconnection network.

The application discloses a kind of intelligent method and system based on flow, belong to distributed computing and resource scheduling technical field.The method includes: in gateway level, the business flow data of each center system is collected in real time;Maximum QPS performance benchmark based on pressure test is established for each center application;Real-time load rate of each center is dynamically calculated based on sliding time window algorithm;When load rate meets preset double threshold trigger condition, combined with performance benchmark and real-time load rate, the target allocation proportion of high-load center to low-load center is calculated by the algorithm of proportion distribution according to system processing capacity;Finally, cross-center system flow allocation operation is executed through dynamic weighted round robin mechanism.The application realizes the second-level monitoring of flow, intelligent decision and smooth scheduling, effectively solves the problems of response lag, dependence on manual intervention and unreasonable resource allocation in the prior art, significantly improves resource utilization and system stability.

The invention discloses a task processing method and device and electronic equipment, and relates to the technical field of artificial intelligence, and the method comprises the steps: determining the real-time performance information of a plurality of server nodes, and calculating the dynamic load factors of the server nodes according to the real-time performance information, the dynamic load factors being used for reflecting the current task processing capability of the server nodes. A weight value is distributed to each server node, the initial weight value of each server node is determined based on the dynamic load factor, and the weight value of each server node is dynamically updated based on a smooth weighted polling algorithm. And a task processing request is sent, the task processing request is used for requesting a target server node to execute the to-be-processed task, and the target server node is the server node with the highest current weight value. By applying the technical scheme of the invention, the large-scale multi-modal task processing efficiency can be improved, resource redundancy is reduced, and the accuracy of a processing result is enhanced.

The application discloses a large file cross-unit transmission method based on a business cooperation scene, relates to the technical field of cross-organization data transmission and resource scheduling, and takes a business context as a core to map an event type, an emergency level and an SLA deadline time into a calculable priority, and is linked with real-time network measurement results, so that scheduling no longer stays in static file attributes, but can instantly improve the service share of a key request according to events such as out-of-stock early warning and promotion start, thereby reducing the probability that a high-emergency task is blocked by regular transmission. By introducing unit-level quotas, admission and interruptible preemption on the side of a bottleneck link, long-term occupation caused by first-come-first-served is avoided, basic fairness between different units is ensured, and bandwidth can be quickly transferred to more urgent flows at a business peak moment. Weighted round-robin weights are dynamically updated according to the emergency degree, the SLA urgency degree and the quota surplus degree, hysteresis and minimum interval control are superimposed, and the agility of scheduling is improved.

The invention discloses a large file cross-unit transmission method based on a service collaboration scene, and relates to the technical field of cross-organization data transmission and resource scheduling, in the invention, a service context is taken as a core, an event type, an emergency level and an SLA deadline are mapped into computable priorities, and linkage with a real-time network measurement result is carried out; scheduling does not stay at a static file attribute any more, the service share of a key request can be instantly improved along with out-of-stock early warning, promotion starting and other events, and therefore the probability that a high-emergency-degree task is blocked by conventional transmission is reduced; by introducing unit-level quota, admission and interruptible preemption at a bottleneck link side, long-term occupation caused by first arrival and first arrival is avoided, basic fairness among different units is ensured, and bandwidth can be quickly transferred to a more urgent flow at a service peak moment; the weighted polling weight is dynamically updated along with the urgency degree, the SLA urgency degree and the quota redundancy, hysteresis and minimum interval control are overlaid, and scheduling agility is improved.

The application provides a short video full-process customer acquisition system based on agent cooperation, and relates to the technical field of artificial intelligence. The input layer of the system is used for receiving task data input by a user. The scheduling layer adopts an agent scheduling engine. The agent scheduling engine is used for determining at least one subtask and a subtask execution sequence according to the task data, and distributing the subtask based on a publish / subscribe mode and allocating computing power resources based on a weighted round robin algorithm. The function layer includes an account positioning module, a benchmark account disassembly module, an account foundation building module, a topic library building module, a video creation module and a clipping and publishing module which are independently encapsulated in the form of agents. Based on the subtask execution sequence and the computing power resources, the account positioning module, the benchmark account disassembly module, the account foundation building module, the topic library building module, the video creation module and the clipping and publishing module execute corresponding subtasks. The application can automatically create short videos to acquire customers and improve efficiency.

The invention relates to a telemedicine-oriented 5G multi-link network dynamic adaptation method and device, and the method comprises the steps: building a mathematical model based on a WRR weighted polling scheduler, putting a to-be-sent message into a to-be-sent cache stack, enabling a pointer of the stack reading the message to correspond to a first message, creating two WRR queue weights to correspond to two scheduling message queues, and enabling the WRR queue weights to correspond to two scheduling message queues; a message in a cache stack is subjected to polling scheduling and split into two message queues, the two message queues correspond to two transmission links, a message reading pointer of each transmission link points to the head of the message queue, and by integrating a 5G multi-link aggregation algorithm, the communication bandwidth of a single service is increased by two times. According to the technical scheme, the bandwidth requirements of the upper and lower networks of the medical equipment in a mobile environment are met, SIM cards of different operators can be inserted, the problems that the bandwidth is limited and the time delay is too large in real-time data communication interaction faced by the current medical terminal equipment are solved, and a new technical scheme is provided for the development of remote medical services.

The invention provides a channel selection method and a related device, and relates to the technical field of communication. The method comprises the following steps: under the condition that first data to be transmitted is detected, selecting a first network channel from a plurality of network channels based on a smooth weighted polling algorithm to carry out data transmission; in the process of transmitting the first to-be-transmitted data through the first network channel, if it is detected that the first network channel is abnormal, reducing the effective weight of the first network channel; and when the first to-be-transmitted data fails to be transmitted or second to-be-transmitted data is detected, network channel selection is carried out based on the smooth weighted polling algorithm and the adjusted effective weights of the plurality of network channels. According to the technical scheme provided by the invention, the problem of how to better realize load balancing for a plurality of network channels in the prior art can be solved.

This invention discloses a joint deployment optimization method for hybrid expert model inference in heterogeneous GPU environments. The method first uses offline trajectory statistics to determine expert activation frequencies and collects resource profiles of heterogeneous devices. Then, it constructs a joint optimization model with the objective of minimizing the weighted sum of inference latency surrogate terms and quantization error penalty terms. Under memory constraints, a pruned greedy search algorithm is used to solve for the placement location, number, and quantization accuracy of each expert's replicas. During the online inference phase, a routing mechanism combining local priority and load-aware weighted round-robin is used for dynamic token distribution. This invention can simultaneously address inference latency, memory usage, and accuracy preservation in heterogeneous GPU environments, improving the processing efficiency of popular experts, reducing cross-device communication overhead and slow device trailing effects, and is suitable for distributed large-scale model inference deployment scenarios.

This invention discloses a QoS scheduling method, apparatus, device, and storage medium for a switch, relating to the field of network communication technology. The method includes: performing deep detection on input data packets based on a preset service feature library to obtain classified data packets with priority tags; writing the classified data packets into corresponding priority queues based on the priority tags, and reserving exclusive buffer resources for target queues in the priority queues; dynamically calculating the weights of real-time bandwidth utilization and queue waiting time to obtain a target weight sequence; performing weighted round-robin transmission on non-target queues in the priority queues according to the target weight sequence; and interrupting weighted round-robin transmission when a data packet is detected in the target queue in the priority queue and the queue waiting time exceeds a preset delay threshold, prioritizing forwarding data packets from the target queue, thus ensuring the upper limit of real-time service latency and eliminating low-priority traffic starvation under high-load burst traffic conditions.

The invention discloses an RNDIS virtual network card communication method based on a double-SoC heterogeneous architecture, and relates to the technical field of Internet of Things terminal equipment communication, a virtual Ethernet communication tunnel is established between a video SoC and a master control SoC, two ends of the virtual Ethernet communication tunnel are logically a master local area network node and a slave local area network node, and Socket-based communication connection is established; the service layer classifies original service data packets to be transmitted, executes an active subpackage strategy, splits the original service data packets into scheduling segments, and encapsulates a unified private message header for each scheduling segment; constructing multi-level priority queues, and respectively storing the packaged scheduling fragments or non-subpackage messages into the corresponding multi-level priority queues according to service types; and performing serial scheduling on the multi-level priority queue by utilizing a single sending task, writing a scheduling fragment into a Socket interface by adopting a strategy based on combination of a fragment-level preemptive rule and weighted polling, and sending the scheduling fragment to a receiving end.

The application relates to the field of database reading, and particularly discloses a database efficient reading method based on multi-level cache optimization and dynamic index fragmentation, which comprises the following steps: high-frequency query data is accurately identified through a sliding window algorithm, dynamic cache loading and preheating strategies are implemented, query delay is effectively shortened, and the burden of the database is reduced; a consistent hash algorithm is used to realize intelligent distribution of index fragmentation, distributed query routing and transaction processing optimization are matched, and query efficiency is greatly improved; query load balancing is realized by means of a weighted round robin and a least connection strategy, real-time monitoring and an elastic expansion and contraction mechanism are combined, and it is ensured that the system can stably and efficiently operate under a high-concurrency scene. The application significantly enhances the database access efficiency and dynamic adaptability, and is particularly suitable for large-scale distributed data processing environments.

A method for network traffic arbitration includes, at a network router, receiving two or more network packets over two or more input ports. During an observation window, traffic parameters for the two or more network packets are stored in a traffic history table, the traffic parameters including a Quality-of-Service (QOS) priority value for a network packet of the two or more network packets. Based at least in part on the traffic parameters recorded in the traffic history table, including the QoS priority value, arbitration weights are calculated for each of the two or more input ports for a weighted round robin arbitration process.

The invention discloses an SIP (Session Initiation Protocol) load balancing scheduling method and system based on a load balancer, a storage medium and equipment. The method comprises the following steps: sending an SIP signaling request of an SIP client to the load balancer; after receiving the SIP signaling request, the load balancer analyzes an skb message header to obtain a source IP / source port and protocol information of the SIP client, analyzes an SIP message header and reads an SIP call identifier; carrying out Hash calculation on the SIP call identifier and then carrying out modulo operation to obtain an SIP call identifier Hash factor; performing secondary hash on the source IP / source port of the SIP client, the protocol information and the SIP call identifier hash factor to obtain a secondary hash value; whether the secondary hash value exists or not is judged, and if the secondary hash value exists, the message of the SIP signaling is directly forwarded to a real server corresponding to the secondary hash value; and otherwise, a real server is distributed to the message of the SIP signaling by using a weighted polling scheduling algorithm. According to the invention, the SIP balanced scheduling is realized by introducing the SIP call identifier hash factor.

The invention relates to the technical field of transmission control, in particular to an ROS2 lightweight communication method oriented to resource-constrained nodes, which comprises the following steps of: performing timing difference calculation on four durations of transmission, queuing, processing and routing to realize accurate quantization and dynamic classification of task delay; the scheduling system can carry out real-time screening and task marking on the basis of time delay constraint, a time delay control matrix is constructed through a weighted polling scheduling algorithm, the task priority level and the risk weight serve as a calculation core, multi-dimensional decision making of task execution is achieved, and the task execution efficiency is improved. A scheduling mechanism capable of automatically adjusting the task state according to the remaining transmission time and the packet loss proportion is formed, a multi-level feedback queue scheduling algorithm achieves response balance of different task types in a scheduling period through high, middle and low priority weight setting and a time slice migration strategy, self-adaptive adjustment of the weight is achieved in combination with power consumption and memory monitoring, and the scheduling efficiency is improved. And the continuous communication capability of the node in a low-power-consumption state is ensured.

The present application relates to the field of solid state storage technology, and in particular to a solid state disk data consistency control method, aiming to solve the problems of poor data consistency, high risk of power failure loss, and difficult balance between write efficiency and consistency of the solid state disk in the scenarios of power failure, parallel writing, and garbage collection. By constructing a "cache-persistent cache" hierarchical architecture, combined with multi-channel writing synchronization, power failure emergency migration, and periodic verification and repair mechanisms, the full-process consistency control is realized: the received data is first stored in the cache, the multi-channel writing task is allocated according to the weighted round robin and the state is synchronized; when the power fails, the standby capacitor is triggered to supply power, and the cache data is migrated to the persistent cache according to FIFO; according to the 1-24 hour configurable period, the SHA-256 hash check is performed on the flash block data, and when it is inconsistent, it is repaired from the persistent cache / backup block. The present application makes the solid state disk data consistency compliance rate ≥ 99.9%, the power failure data loss rate ≤ 0.01%, and the write delay is reduced by 15%-20%, which is suitable for consumer, enterprise and industrial solid state disks.