659 results about "Timeout" patented technology

Techniques for cache management for LLM processing are described. Example embodiments include a signal hashing model that generates a key for particular context data. An LLM output corresponding to the context data is stored in a cache along with the key. For a user input received by the system, a cache lookup is performed using a key for context data corresponding to the received user input. For a cache hit, the stored output is used to respond to the user input. For a cache miss, a LLM processes the context data and the user input to generate an output within a first timeout. If the LLM is unable to generate an output within the first timeout, then in some cases, the LLM is allowed to continue processing until a second timeout, and a final or partial output from the LLM is stored in the cache.

The invention relates to the technical field of elevator group control, in particular to an elevator group control method and system based on deep learning, and the method comprises the steps: building an elevator taking behavior time sequence data set through collecting historical elevator taking behavior data and combining with an elevator calling request and the number of people waiting for an elevator obtained through image recognition; the elevator taking frequency, the elevator calling interval and the passenger number fluctuation are modeled by adopting a long-short-term memory network, passenger flow periodical characteristics in workdays, holidays and peak periods are extracted, the passenger flow of each floor in the future is predicted, and pre-scheduling deployment of idle elevators is completed in advance. In the scheduling stage, state input containing the elevator running state and predicted flow is constructed, a reinforcement learning model is introduced for strategy optimization, a reward function is constructed based on passenger waiting time, response timeout and the service completion number, a scheduling strategy is driven to continuously converge, elevator response distribution is optimized, and position presetting and path adjustment are preset. And the scheduling intelligence and the service efficiency are improved.

The invention discloses a method for increasing and suppressing the number of alarms through an alarm source, and relates to the technical field of communication network operation and maintenance, and the method specifically comprises the steps: constructing a network real-time topology, building a set membership graph between a router and a downlink device, and when the heartbeat overtime alarm of the downlink device occurs in the network, starting the router according to the set membership graph. And counting the proportion of the lost equipment in a sliding time window taking the parent router as the dimension, when the proportion of the lost equipment exceeds a threshold value, judging that the router is a root cause, sending out an early warning of the router meeting the condition of a root cause determiner, and shielding the overtime alarm of the router. According to the method, the root cause event of'all the downlink devices of the same route are disconnected 'is dynamically identified in the real-time alarm flow, and a high-severity router fault alarm is used for replacing a plurality of device heartbeat overtime alarms, so that the alarm noise is remarkably reduced, and the operation and maintenance efficiency is improved.

Aspects include processing the query from a client device to divide the query into a plurality of subqueries, identifying a data source that corresponds to each of the plurality of subqueries, and identifying an expected response time for each of the data sources. Based on a determination that at least one of the expected response times is greater than a default timeout threshold, aspects include classifying the query as a delayed split query and updating a timeout value of the query to be a delayed timeout threshold, which is greater than the default timeout threshold. Aspects also include performing the delayed split execution of the query.

The invention discloses a starting method for a pre-starting execution environment, computer equipment and a related product, and relates to the technical field of computers, the method comprises the following steps: when a first network driver triggers a target unloading process due to loading of a second network driver and the first network driver has sent a network release message in the target unloading process, starting the target unloading process; monitoring an overtime event and a network reply message event of the monitoring timer; when the timeout event is monitored and the network reply message event is not monitored, triggering the first network driver to resend the network release message; and when the retransmission times exceed a preset retransmission threshold value, forcibly terminating the target event polling operation, and transferring the network control right from the first network driver to the second network driver, so that the second network driver takes over and continues to execute the starting process based on the own network protocol stack. The technical problem of deadlock in the starting process of the pre-starting execution environment in related technologies is solved, and the technical effect of improving the starting reliability of the pre-starting execution environment is achieved.

The invention discloses a dual-channel adaptive network stability testing method and system, and belongs to the technical field of computer network testing. The system adopts a dual-channel asynchronous communication architecture in which a control channel and a data channel are physically separated, instructions and state information are transmitted through an independent TCP control channel, and high-speed test data streams are transmitted through a UDP or TCP data channel. The method comprises the following steps: dynamically and adaptively adjusting a timeout parameter based on real-time network delay detection; nanosecond constant interval packet sending control is realized by using a high-precision monotonic clock source and busy waiting compensation mechanism; a packet loss detection, positioning and ACK confirmation mechanism based on the serial number is realized in an application layer; and real-time statistical display and idle detection automatic termination are realized through multi-thread parallel processing. According to the invention, the precision, the stability and the adaptive capability of network testing are remarkably improved, and the method is suitable for long-time stability testing in various scenes such as a high-speed Ethernet, a satellite link, an industrial Internet of Things gateway and the like.

The invention discloses a CXL protocol extension method, device and system based on the Ethernet. The method comprises the following steps of: carrying out long-distance expansion interconnection on a CXL (Compute Express Link) protocol by using Ethernet as a physical transmission medium; by bearing the CXL protocol on the Ethernet, the range of memory pooling is expanded from the rack level to the whole data center (kilometer level), and the potential of the CXL technology is thoroughly released; through protocol translation, label management and flow control realized by pure hardware, an operating system kernel and a software protocol stack are bypassed, so that the end-to-end memory access delay is far lower than that of any software-based network storage scheme, and a reliable request-response tracking and timeout retransmission mechanism is realized through a hardware label manager. By distinguishing the control flow and the data flow, differentiated QoS guarantee is provided for different types of memory flows, and the stability of the system under high load is ensured.

The invention discloses a management method and device for a process instance in a semiconductor computer integrated manufacturing system, and the method comprises the steps: taking a process timeout time as a global configuration parameter, and configuring the global configuration parameter in a process template in a process template design stage; generating corresponding process instances through the process template, and creating an independent runtime context for each process instance; storing the process instance in a process instance container module; before each process node instance is executed, the timeout state of the process instance to which the process node instance belongs is judged through a unified preposition method; periodically scanning a process instance container module through an independent timing compensation task, and carrying out timeout judgment on the process instance in a running state; receiving an external instruction through a manual intervention interface provided by the management module; all modification operations on the process instance state keep the isolation of the instance level. According to the invention, flexible manual intervention and timeout termination mechanisms are supported.

The invention relates to the technical field of communication equipment operation and maintenance, in particular to a serial port equipment firmware upgrading method and system. The serial port equipment firmware upgrading method comprises the steps that an upgrading task instruction issued by a remote management platform is received, and after equipment identity verification and upgrading file integrity verification are executed, system protection measures are started; creating a security isolation sandbox environment, forbidding a high-risk function module in a script execution environment, and defining addressing parameters of target serial port equipment; standard scripts are loaded, and serial port read-write, file operation and state reporting type security interaction interfaces are injected; calling a mode switching function of the standardized script, so that the serial port equipment is switched to a firmware upgrading mode from an operation mode; cyclically calling a data transmission function of the standardized script to transmit firmware data in blocks, and synchronously implementing timeout monitoring; and after upgrading is completed or abnormal interruption occurs, the security isolation sandbox environment is destroyed, the original business is recovered, and an upgrading result is reported to the remote management platform.

The invention relates to a function operation duration detection and stack automatic collection method and device based on an eBPF technology, and aims to solve the problems that a traditional tool cannot accurately monitor function operation timeout, cross-state stack collection lags and the like. The method comprises the steps that an eBPF program is injected into an objective function (a kernel mode passes through kprobe, and a user mode passes through uprobe) entry, and a thread PID / TID and an entry timestamp are recorded into an eBPF hash table; registering a periodic perf event to trigger timed detection, calculating a function operation time length and comparing the function operation time length with a threshold value; when overtime triggering is carried out, a kernel mode stack and a user mode stack are synchronously collected through bpfgetstackid, and after an event is packaged, the kernel mode stack and the user mode stack are reported to the user mode through bpfperfeedback; and analyzing and displaying the stack in the user mode. The scheme has the advantages of non-return dependency detection (covering a blocking scene), cross-state tracking (kernel / user state function), low-overhead event driving (adapting to a production environment), cross-state stack association (accurately positioning a jamming root), low intrusion, high real-time performance and the like, and the system performance analysis efficiency is effectively improved.

The invention discloses a method for dynamically setting dog feeding time of a watchdog circuit, which comprises the following steps of: collecting and monitoring power consumption of a CPU (Central Processing Unit) and a corresponding load, and fitting the load and power of the CPU according to monitoring data so as to reflect the running state condition of a system; power consumption of a monitoring CPU is collected in real time, the real-time CPU load is calculated according to the fitting relation between the CPU load and power, and the watchdog feeding time of a watchdog circuit is determined according to the CPU load; a square wave period is set, square waves are collected to set timeout time, and the timeout time is kept consistent with dog feeding time, determined according to CPU loads, of a watchdog circuit. According to the method, the CPU power consumption and the load are fitted, the CPU load is monitored in real time, the square wave period is adjusted, the dog feeding time is dynamically adjusted according to the CPU load, resource waste is avoided, and the method is suitable for the fields of embedded systems, industrial control and the like.

The invention provides an end-to-end reliable retransmission method, a computing node and a system for an AI intelligent computing network. The method comprises the following steps: a source end and a destination end establish synchronization through a handshake mechanism; the source end sends a data packet to the destination end, and locally caches the sent data packet; the destination end receives the data packet, compares the serial number of the data packet with the expected serial number of local maintenance, and returns an acknowledgement message or a retransmission request to the source end according to a comparison result, and the acknowledgement message carries the current expected serial number; in response to the confirmation message, the source end releases all the data packets of which the serial numbers are smaller than the expected serial number carried by the confirmation message in the local cache; or, in response to the retransmission request or transmission timeout, acquiring the corresponding data packet from the local cache for retransmission; and after the source end sends the last data packet, if the confirmation message is not received within the preset time, actively sending a state query request to the destination end to obtain the receiving state of the last data packet.

The invention discloses an HTTP (Hyper Text Transport Protocol) text restoration method based on DPDK (Data Platform Development Kit), which relates to the technical field of computer networks, and adopts a multi-process decoupling architecture to respectively run processes such as flow collection and protocol analysis, fragmentation recombination and text restoration, formatted message splicing, text restoration timeout waiting and the like in independent CPU (Central Processing Unit) cores; and efficient and lock-free data transmission is carried out by using a ring mechanism of the DPDK. Different from a traditional memory splicing scheme, the data offset is calculated based on the TCP serial number, the load content is directly written into the corresponding position of the file, the problems of message disorder, packet loss, retransmission and the like are effectively solved, an atomic renaming mechanism from the file ID1 (intermediate state) to the file ID2 (complete state) is adopted, and accurate migration of the file state is ensured. Besides, the invention innovatively provides a bidirectional asynchronous scheduling mechanism based on data export and text restoration state monitoring, so that the data export process can continuously keep high throughput, and the overall processing capacity of the system is ensured.

The invention provides a multi-ring state sensing communication network optimization method, relates to the field of industrial network communication, and solves the technical problems of communication uncertainty and insufficient system reliability caused by message conflicts in the prior art. The method comprises the following steps of: when a communication system is initialized, logically dividing a plurality of nodes into a plurality of data transmission rings by a centralized management node on a single physical bus, and initializing an annular frame for controlling a sending permission for each transmission ring; each node transmits a token based on the ring frame, and after the nodes in the ring obtain the token of the ring, the nodes obtain the bus sending authority, construct a data frame containing a target node address and send the data frame to the bus to perform data communication between the nodes; each node continuously detects communication states of adjacent nodes and a transmission ring through a heartbeat mechanism and a ring frame timeout mechanism; and automatically triggering a data ring reconstruction process and updating a token transmission path when detecting that the node fails, the node is added or the centralized management node is disconnected.

The invention discloses a Raft consensus algorithm-based dynamic management method and a related device, and relates to the technical field of communication, when a slave node server does not receive a heartbeat packet of a master node server overtime, the slave node server is automatically converted into a candidate node server and broadcasts an election request; and if the number of the received feedback information for confirming that the candidate node server is changed into the master node server is greater than half of the total number of all the slave node servers, determining that the candidate node server is changed into the master node server. The delay risk that a master-slave centralized architecture depends on manual monitoring and manual switching is thoroughly eliminated, fault discovery and response are solidified on an algorithm level, and automatic disaster recovery without human intervention is achieved.

The embodiment of the invention relates to an electricity meter data complementary reading method, which is applied to a concentrator terminal and comprises the following steps: sending a data reading instruction to all electricity meters according to a preset physical address sequence at the beginning of each preset meter reading period; if the effective collection data fed back by the electricity meter is not received within the preset overtime threshold value, establishing a collection missing queue according to the electricity meter information, the data information and the time information; and in the current meter reading period, after the current round of data acquisition is completed, calculating the residual idle time of the current meter reading period and the estimated single-time additional reading time consumption, and performing data additional reading according to the residual idle time of the current meter reading period and the estimated single-time additional reading time consumption. According to the technical scheme provided by the embodiment of the invention, the data which is not successfully acquired is subjected to supplementary reading by utilizing the remaining time of the current meter reading period and the subsequent meter reading period, a layered priority supplementary reading mechanism is provided, the timeliness and fairness of data supplementary reading are considered, and the loss risk of long-term missing acquisition of data can be avoided.

The invention relates to an RS-485 bus anomaly detection and self-recovery method and device, electronic equipment and a storage medium. The method comprises the following steps: setting a transceiving state machine comprising an overtime timer in an MCU (Microprogrammed Control Unit) in the uninterruptible power supply, and controlling the transceiving state of the transceiving state machine and / or the level of a DE pin of RS-485 based on the monitoring time of the overtime timer; a timing protection circuit is arranged between the MCU and the DE pin, and the level state of the DE pin is adjusted based on the first duration for maintaining the high level of the DE pin and the timing protection circuit; setting a periodic monitoring mechanism in the MCU, and monitoring communication between the uninterruptible power supply and the RS-485 based on the periodic monitoring mechanism to obtain a monitoring result; based on the monitoring result, determining a corresponding grading recovery strategy; the hierarchical recovery strategy is used for recovering communication between the uninterruptible power supply and the RS-485. Therefore, the reliability and the stability of the communication link between the UPS and the external monitoring equipment are improved, and the maintenance cost of the equipment is reduced.

The invention relates to the technical field of PCIe (Peripheral Component Interconnect Express), in particular to a PCIe dynamic degradation type link resetting method and system. Comprising the following steps: S1, continuously reading the LTSSM state of a PCIe link, recording the duration of an abnormal state, and when the duration exceeds a preset threshold value, judging that timeout abnormity exists; s2, after the abnormal early warning is received, a dynamic degradation reset instruction for the NVMe controller is initiated; s3, the state machine control unit takes over the reset link initialization process and follows the physical layer priority gt; the link layer is preferentially gt; a priority strategy is simplified by a transaction layer, and a centralized resource guide link rapidly completes training and enters an L0 state. Compared with the prior art, the method has strict protocol compliance, the reset efficiency is improved, and the overall stability and performance of the system are optimized.

The invention discloses a decentralized high-performance multicast method and system based on Gossip and RDMA, and relates to the field of computer network communication. The method has better reliability, realizes zero-loss transmission of data packets through message de-duplication, timeout retransmission, ACK / NACK convergence and a dynamic topology adaptation mechanism, ensures data consistency, and solves the problem of insufficient reliability of traditional UDP multicast; the method has the advantages that the real-time performance is high, the delay is low, the throughput is high, the zero copy and kernel bypass characteristics of the RDMA and the batch processing mechanism of the MPMC are combined, the average delay is lower than 10 microseconds, the throughput in a large data packet scene reaches 374GB / s, and the low-delay requirements of high-frequency transaction, high-performance calculation and the like are met; the method has high expansibility, is based on a decentralized architecture of an improved Gossip protocol, does not need core node maintenance topology, supports dynamic joining or quitting of nodes, can adapt to large-scale cluster expansion, and solves the expansibility bottleneck of centralized multicast.

The invention relates to the technical field of fusion terminals, and particularly provides a fusion terminal data reporting missing rereporting method and system and a medium, and the method comprises the steps: through a reporting task initialization mechanism, when a fusion terminal is started or a task scheme and a table file change, emptying historical data and carrying out initialization again; adopting an acquisition scheme matching strategy based on the matching degree and the interrogation interval to accurately select a target acquisition scheme; in combination with a two-stage compensation mechanism of self-compensation and failure linked list compensation, compensation is scheduled according to protocol timeout in the self-compensation stage, and periodic compensation, forced compensation and redundant data cleaning are executed in the failure linked list stage. Through the above steps, the integrity and reliability of data reporting of the fusion terminal are effectively improved, and the accuracy of collection success rate statistics of the master station is guaranteed.

The invention belongs to the technical field of network attack and defense simulation, and discloses a network attack and defense simulation engine system based on discrete event driving, which comprises an event scheduling module, a simulation network construction module, an attack and defense simulation module, a simulation kernel module and a panoramic situation awareness module, the event scheduling module is in the form of MITRE ATTamp; the CK is constructed as a theoretical base and is used for converting all network activities into a dispatchable discrete event sequence; the network activities comprise data packet sending, protocol timeout and attack triggering; and the simulation network construction module can construct a simulation network containing complete service logic. According to the application, the MITRE ATTamp is scheduled by the event scheduling module; the CK tactical intention is converted into an ordered event stream, a hierarchical task network planner in the attack and defense simulation module dynamically decomposes and re-plans a high-level target, and ATTamp can be utilized; and the CK technology ID reconstructs an attack chain in real time.

The invention relates to a token-based task scheduling method and system and a medium, and the method comprises the steps: initializing the token capacity of a token bucket and the number of available tokens, and entering a slow start process; when the current round of task is issued, the tasks in the cache queue are issued to an algorithm module for processing based on the number of the available tokens, and the number of the available tokens is updated; before the next round of task is issued, the token capacity is updated, the number of result tokens is determined based on the task processing result returned by the algorithm module, and the number of time tokens is determined based on the dynamic timeout time; and updating the number of the available tokens in the token bucket based on the number of the result tokens, the number of the time tokens and the updated token capacity, so that the task sending rate can be dynamically adjusted, the computing power resource utilization rate is improved, and the fault risk is reduced.

The invention provides a low-power-consumption scheduling method and system for an edge heterogeneous computing unit, and the method comprises the steps: constructing a resource sensing model of heterogeneous hardware, and collecting the performance parameters of a processing unit in real time; establishing a task priority evaluation model, and calculating a task priority by adopting an analytic hierarchy process based on the resource perception model and the task priority evaluation model; designing a collaborative optimization model; based on the collaborative optimization model and by adopting an improved particle swarm optimization algorithm, solving an optimal allocation scheme of tasks in the heterogeneous hardware, and dynamically adjusting the working frequency of the heterogeneous hardware; monitoring states of task execution and heterogeneous hardware in real time, judging whether resource overload or task timeout risks occur in the states or not, and if yes, triggering scheduling scheme updating. According to the method, local optimum is effectively avoided, the solution efficiency and the solution quality are remarkably improved, and the frequency-to-voltage linkage dynamic power consumption regulation and control closed loop is realized by triggering scheduling scheme updating.

The invention discloses a cascade energy storage system-oriented collaborative test control method, and relates to the technical field of cascade energy storage system collaborative test control. Detecting the state of the energy storage module, verifying communication delay, marking an overtime module, triggering self-inspection, and recording initial parameters of a normal module; selecting a test mode, and configuring cooperative control and data acquisition parameters; the main controller generates a total power instruction and dynamically distributes the total power instruction, sends a synchronous clock signal to trigger the module to perform synchronous execution, gradually reduces power for charging and discharging and performs soft cut-off, and sends an isolation instruction for fault simulation; receiving real-time data, calculating indexes such as voltage balance and SOC deviation, and analyzing efficiency and communication stability; balancing current is dynamically adjusted according to a result, and a shutdown instruction is sent abnormally; and if a termination condition is met, stopping testing, and generating a report containing performance indexes and fault records. According to the invention, the test collaboration and precision are improved, the fault early warning and data utilization are optimized, the safety is enhanced, and the operation and maintenance cost is reduced.

A computer-implemented method, system and computer program product for handling eviction of subprocess pods of a node. An indication to evict pods of a node in a container environment is received. Upon receiving such an indication, process information pertaining to the evicted pods is retrieved. In scenarios in which some of the evicted pods include subprocess pods, based on such processing information, those subprocess pods with a timeout that is less than a threshold value may be tagged for timeout handling. Such timeout handling involves selecting a number of subprocess pods to ensure that their parent process will successfully complete its assigned job. A schedule plan may then be generated for such selected subprocess pods, where the schedule plan updates the timeout for such subprocess pods to enable them to complete execution of their assigned tasks prior to their expiration time (timeout) even though they were evicted.

The embodiment of the invention provides a heartbeat interval adjusting method and device and computing equipment, the heartbeat interval adjusting method is applied to a sending end of a network transmission system, the network transmission system further comprises a receiving end, and the method comprises the following steps: responding to establishment of network connection between the sending end and the receiving end; obtaining a first heartbeat interval and a timeout duration threshold value of network connection; sending a heartbeat packet to the receiving end through the network connection based on the first heartbeat interval; under the condition that a heartbeat packet acknowledgement message fed back by the receiving end is received, determining a response duration; determining an adjustment direction based on the response duration and a timeout duration threshold, and adjusting the first heartbeat interval according to the adjustment direction; and returning to execute the step of sending the heartbeat packet to the receiving end through the network connection based on the first heartbeat interval until the network connection reaches a stable state, and obtaining a target heartbeat interval. The effectiveness of network connection is ensured, unnecessary resource consumption is avoided, and self-adaptive optimization of network connection is realized.

The invention discloses a remote secure connection method and system for an industrial internet of things, and belongs to the technical field of remote secure connection, and the method comprises the steps: obtaining the polling interval data of target industrial equipment, analyzing a polling mode characterization factor, judging a polling interval execution scheme of the remote connection of the target industrial equipment, and carrying out the remote connection of the target industrial equipment; synchronously analyzing a polling mode characterization factor deviation value, executing dynamic polling interval adjustment of remote connection of the target industrial equipment, analyzing dynamic polling interval adjustment effect information, analyzing a network state characterization factor of remote connection, and executing dynamic protocol timeout adjustment of remote connection of the target industrial equipment. And analyzing the dynamic protocol overtime adjustment effect information to obtain a protocol overtime execution scheme of remote connection of the target industrial equipment. According to the invention, the dynamic polling interval is dynamically changed, the inherent rule is broken, the protocol timeout time is dynamically adjusted, the security and real-time performance of remote secure connection are improved, and the compatibility and standardization of the system protocol are enhanced.

A method, computer program product, and computer system for handling timeouts of client requests to a server. The method includes: accessing timeout metadata of a synchronous client request at a server queue; determining that the client request has a timeout risk; identifying a response storage location for a request processor to use to store a response to the client request when the response is provided after the timeout of the client request; responding to the client request indicating a user of a response storage location for asynchronous response retrieval; and providing the identifier of the response storage location as metadata in the client request when sending the client request to the request processor for processing.

The invention discloses a PLC and MES simulation interaction protocol conversion gateway realization method, and relates to the field of data interaction, and the method comprises the following steps: 1, collecting an original data flow of a PLC side heterogeneous industrial bus protocol and an interaction instruction of an MES side manufacturing execution layer protocol in real time, dynamically scanning the data flow through an abnormality diagnosis engine, and carrying out the dynamic scanning of the data flow through an abnormality diagnosis engine; identifying missing fields, transmission timeout or numerical value border crossing abnormal states in the PLC data, and generating an abnormal feature vector set carrying timestamps and associated point position labels; 2, constructing and pre-training an anomaly prediction model through a deep learning algorithm, and inputting an anomaly feature vector to the model; trend prediction and confidence evaluation of missing data are realized through a pre-training deep learning model, quantitative credibility support is provided while data continuity is ensured, it is ensured that an MES system can still operate continuously based on credible simulation data when equipment is abnormal, and the risk of production stagnation caused by forced interruption of a traditional gateway is avoided.

The invention provides an adaptive high-performance consistent control method for a multi-agent system under denial of service (DoS) attack, and a sensing unit comprises a state sensing module carried on each agent, obtains body state information in real time, and shares information with other agents through an adjacent communication network; the information is used for error calculation and local control decision, and meanwhile, communication quality and whether interruption occurs are judged, so that the observability and anti-disturbance capability of the system are enhanced; a distributed control unit controller is deployed in each intelligent body, error mapping and performance constraint are carried out, a time-varying performance boundary is constructed, a control law and a self-adaptive estimation law are designed, a communication execution unit is responsible for transmitting a trigger signal, state estimation or control input generated by the controller among the intelligent bodies, intermittent DoS attacks can be tolerated, and the control law and the self-adaptive estimation law are designed. And an automatic retransmission and timeout judgment mechanism is provided, so that reliable transmission of key data is guaranteed. State estimation, performance recovery and efficient communication can be realized in a DoS attack environment.