41 results about "Burstiness" patented technology

The invention discloses an industrial data transmission optimization management system based on service link quality, and relates to the technical field of data transmission. Weak delay oscillation and a periodic packet loss rule are reserved through multi-scale processing, so that the system can discover precursor anomalies which are difficult to capture by a traditional monitoring mechanism; minute-level even second-level early warning of link disturbance is realized, and data congestion caused by sudden quality reduction is avoided; a dynamic discrimination mechanism driven by a precursor score is utilized, so that a scheduling decision is converted from passive response to active adjustment, and the probability of cascade failure caused by mass mutation is greatly reduced; an interference spectrum adaptive strategy enables the system to adaptively match the optimal coding mode and the redundant structure when facing narrow-band noise, periodic interference or transient impact, so that the anti-interference capability and the bandwidth utilization efficiency are improved; priority adjustment is carried out in combination with a link risk trend and a state evolution rule, and the protection timeliness and resource allocation efficiency of the key link are remarkably improved.

The invention discloses a flow behavior perception-based photoelectric double-plane cross-layer interconnection scheduling method, which comprises the following steps of: acquiring communication service data flow in a distributed training task, and generating flow behavior time sequence characteristics; performing flow behavior time sequence feature modeling based on the flow behavior time sequence features; identifying flow behavior time sequence characteristics of different types of communication service data streams, and dividing the communication service data streams into lightweight control streams, periodic synchronization streams and sudden transport streams; the method comprises the following steps: acquiring a network state of a link in a data transmission plane, establishing a transmission path reachability constraint, allocating capacity for the link, detecting a link capacity occupation value and a link capacity upper limit, and generating a scheduling strategy; the scheduling strategy is converted into a mapping instruction, and the data transmission plane executes distribution of a transmission path of the communication service data flow according to the scheduling strategy; according to the invention, data stream differential transmission and cross-layer cooperative control based on photoelectric biplanes are realized, the training iteration time is remarkably shortened, and the utilization rate of network resources is improved.

The present application relates to the technical field of semantic processing, and more particularly to a word element aggregation and distribution method and system based on multi-level cache and semantic preheating, which obtains historical word element request interaction logs, analyzes the access characteristics of different business scenarios and context fragments; in combination with semantic dependency relationships, identifies regular repetition and sudden hotspot characteristics of word element access, determines a preloading target set; real-time acquisition of context semantic data of the current inference request, analysis of word element prediction demand parameters to be generated for output; fusion of word element prediction demand parameters and preloading target set, dynamic generation of word element aggregation parameters and distribution scheduling strategy. The present application realizes forward-looking preheating and accurate scheduling of word element resources by deeply mining historical rules and real-time semantic intent, effectively solves the problems of weak semantic perception, low hit rate and response lag of traditional cache strategies, significantly reduces the inference delay of large models, and improves the throughput and stability of the system in high-concurrency scenarios.

The invention relates to the field of electric power business scheduling, and particularly discloses an electric power system business collaborative scheduling method based on a 5G technology, which constructs a high-low layer collaborative double-layer intelligent scheduling architecture. In a high level, situation prediction is carried out by fusing power grid operation data and historical traffic, and the macroscopic resource ratio of network slices is adjusted prospectively to cope with the sudden demand of power business, so that the problem of insufficient resources of static resource allocation in sudden scenes such as faults is solved. And at the lower layer, service semantic analysis is carried out deeply in the data packet, so that the real-time emergency level of key services such as differential protection and the like is accurately identified, and resources are further refined into a micro resource pool special for the services in combination with slice configuration issued by the higher layer, so that real-time scheduling and preemption based on semantics are realized. Through the combination of macroscopic predictive reservation and microscopic semantic scheduling, it is ensured that high-priority power services can obtain communication guarantee under any network load.

The invention discloses a short message sending method and device, equipment, a storage medium and a program product, and the method comprises the steps: segmenting short message service flow according to preset time slice configuration, and obtaining short message data under a plurality of time slices; performing rate segmentation on the plurality of channel resources according to a preset rate arrangement strategy and a thread concurrency number to obtain rate configuration of a plurality of threads of each channel resource; wherein the rate arrangement strategy is used for indicating rate configuration of each channel resource; binding the short message data with a plurality of threads so as to enable the threads to execute sending on the bound short message data in each time slice according to own rate configuration; according to time arrangement and careful division of the channel rate, the sending efficiency of a multi-short-message service large-batch sending scene can be improved, sudden surge short message data can be effectively dealt with, multi-short-message service channel blockage caused by short message backlog is avoided, and the success rate of short message service sending loss is improved.

The application provides a mobile edge computing dynamic service deployment method based on deep reinforcement learning. In the learning process, a migration conflict resolution mechanism is introduced to maintain the service performance of the user. Specifically, not only the network state is predicted, but also the user connected to each edge node is calculated according to the user trajectory, and it is determined which node is the crowded node according to the user request and the service on the edge node, the resource reservation is carried out for the crowded node, and whether the conflict will be produced is detected. In order to achieve this goal, first, the network architecture of edge computing is determined, then the system model and the quality of service (QoS) model are established, and the problem to be solved is formulated. The application studies the joint optimization problem of delay and cost under the constraints of physical resources and operation cost, taking the minimum total cost of users as the target, and considering guaranteeing that multiple service requests can be responded in time under the sudden aggregation.

The invention discloses a non-cellular intelligent beam management-oriented generative learning-assisted twinning enhancement pre-training method, which comprises the following steps of: firstly, constructing a long short-term memory network fusion diffusion model generative model as a channel prediction digital twinborn body, and predicting a future channel state in a probability mode; secondly, constructing an autoregressive comprehensive moving average model and a variational auto-encoder mixed model to carry out generative flow prediction, and accurately capturing the periodicity and the burstiness of the flow; and finally, jointly applying channel and flow prediction results to a monotonic Q value function decomposition beam selection algorithm after state enhancement. According to the invention, through prior knowledge provided by generative learning, the problems of difficult data sample acquisition, high training cost, slow convergence and poor dynamic traffic adaptability faced by training a beam selection algorithm in a real environment are solved. Compared with a traditional method, the training efficiency, the convergence speed and the final performance in a complex dynamic scene of a beam selection algorithm can be remarkably improved.

The application discloses a time slot allocation method and device, equipment and medium, and relates to the field of time slot allocation. The method comprises the following steps: an unmanned system generates a backbone network based on network topology information of a plurality of communication nodes in a previous time frame through a center node, so as to dynamically determine backbone nodes of a current time frame by using the backbone network; the center node dynamically predicts time slot segments allocated to the center node and the backbone nodes in the current time frame based on traffic demands of the plurality of communication nodes in the previous time frame; and each node in the center node and the backbone nodes dynamically predicts time slots allocated to each node and management sub-nodes of each node in a time slot segment corresponding to each node based on traffic demands of each node and the management sub-nodes of each node in the previous time frame, so as to meet the networking communication demand of the unmanned system with high reliability and low delay, better adapt to the non-stationary environment with strong business burstiness and frequent topology changes of the unmanned system, and realize efficient and reliable time slot allocation for large-scale unmanned system networking.

The application discloses a kind of dynamic protocol identification method and system based on multi-modal data, belong to the technical field of data protocol identification, identification method includes: S1: calculate data packet arrival interval time standard deviation, judge communication mode is periodic or burst communication mode;S2: data packet is divided according to length, calculate and count the probability distribution of each interval, and calculate the Shannon entropy value of data packet;S3: according to data packet arrival interval time standard deviation, the Shannon entropy value of data packet and the probability distribution of each interval, generate protocol fingerprint;S4: using edge AI classification engine classifies data packet.The application can improve the accuracy of protocol identification, improve network security protection capability.

The invention provides a PCIe (Peripheral Component Interconnect Express) dynamic bandwidth allocation and low-delay communication method and system based on an FPGA (Field Programmable Gate Array), and the method comprises the steps: carrying out the real-time statistics of the data transmission quantity of a plurality of virtual channels, and obtaining the flow data of the plurality of virtual channels; dynamically allocating bandwidth weights of the plurality of virtual channels according to the traffic data of the plurality of virtual channels to obtain credit values of the plurality of virtual channels; according to the credit values of the plurality of virtual channels, controlling the data transmission burst amount of the plurality of virtual channels, and performing hardware unloading processing on a transaction layer data packet received by the PCIe interface through a transaction layer packet analysis operation to obtain target address information and a data load; according to the target address information, the data load is directly written into an internal storage area of the FPGA through memory mapping operation, and transmission completion data is obtained. The method and the device can adapt to sudden data transmission requirements, and the bandwidth utilization rate is improved.

The present application relates to a kind of value burst flow detection method based on P4 and Sketch.Break through the detection scheme of traditional based on successive burst, propose value burst flow definition, observe the burst of traffic on the flow level, process the input data packet under the limited memory overhead by the data structure design based on Sketch, detect burst data packet according to the time interval between data packet, and then detect value burst flow. When data packet arrives, each data packet is mapped to a bucket set, but at most to one bucket is operated, under appropriate conditions, can vote according to the voting mechanism based on burst; when each measurement period ends, traverse all buckets, obtain the value burst flow detected by query operation and feedback to control plane. The present application reduces bandwidth overhead and user burden in burst detection process, while only limited memory occupation and low time complexity are generated.

The application discloses a multi-channel satellite frequency point file processing method and system based on threshold optimization, relates to the technical field of satellite communication data processing, and comprises the following steps: calculating the burst position and burst length of a main channel file; counting the burst quantity in a historical main channel file, and calculating a threshold according to the burst quantity in the historical main channel file; if the burst quantity of the main channel file is smaller than the threshold, a low-burst processing mode is selected, otherwise, a high-burst processing mode is selected; and according to the selected mode, the data corresponding to the burst is read from a main auxiliary channel file and processed. The dynamic threshold is estimated, and the reading mode is selected according to the dynamic threshold, which is suitable for efficiently processing the burst data in the multi-channel satellite frequency point file, and solves the problems of burstiness, low-delay requirement of data access and accurate reading according to position in the satellite positioning scene.

The invention discloses a burst degree control and delay guarantee parameter configuration method for a time-sensitive network, which comprises the following steps of: directly constraining an accumulated upper limit of a credit value in a CBS mechanism by introducing a credit upper limit parameter CLP which is independent of a credit growth rate and can be dynamically adjusted, and further limiting excessive occupation of burst flow on channel resources from the source; meanwhile, a worst-case delay analysis model of an integrated credit upper bound CLP is constructed based on a network calculation theory, end-to-end delay constraints of traffic can still be strictly met while traffic burst control is implemented, and burst control and delay guarantee under a dynamic load network are realized. The method has the advantages that through cooperative regulation and control of a single additional parameter CLP, unification of burst control and deterministic time delay guarantee is achieved on the premise that existing CBS standardized parameter configuration is not changed, and the performance and deployment flexibility of the TSN in a dynamic load environment are remarkably improved.

The invention discloses a thread pool computing power collaborative redistribution method and device, equipment and a medium, and relates to the technical field of computers, and the method comprises the steps: generating a time sequence vector covering an object distribution rate, an Eden region growth slope, an old memory occupancy rate and a Survivor region object promotion rate through a monitoring layer, the probability of occurrence of a garbage collection event in a future set time period is pre-judged through a neural network model of a prediction layer, when the probability reaches a set threshold value, a decision layer generates an adjustment instruction containing a thread pool vertical shrinkage proportion and a computing power resource horizontal migration proportion, and then an execution layer calls a thread pool regulator and a k8s regulator to implement adjustment. Therefore, the problem that the computing power is suddenly caused by the garbage collection event is fundamentally solved, and the prospective redistribution of the computing power resources is realized; violent fluctuation of calculation power can be effectively restrained in the garbage recycling pause period, and abrupt reduction and recovery lag of the calculation power are avoided; meanwhile, system delay is reduced and resource utilization rate is improved.

The application belongs to the technical field of network operation and maintenance, and specifically discloses a switch flow adaptive control method and system, which samples data frames of a switch, strictly distinguishes the priority of data transmitted by the switch, and performs flow rate determination and rate overrun analysis based on the priority distinguishing result, suspends the reception or discards low-priority flow when the rate is overrun, avoids excessive consumption of the transmission bandwidth of the switch by the low-priority flow, and ensures the continuity of high-priority flow, so as to solve the problem that the performance of key services cannot be guaranteed and network resources are easily exhausted by abnormal flow when the flow of the switch is congested in the prior art; the rate limitation and smooth output control of high-priority flow can eliminate the burst problem, completely block the impact of abnormal flow on the switch, stabilize the output flow, and reduce the pressure on downstream equipment; and all determination processes are based on simple rate conversion, are suitable for hardware high-speed implementation, do not require complex queue management states, and are convenient to deploy and apply.

The invention relates to the field of big data services, in particular to a dynamic resource scheduling and optimizing method for big data services, which comprises the following steps: constructing a time sequence feature set, outputting a resource demand predicted value in a future preset advance time window, calculating a reservation proportion of an elastic buffer pool, and executing container preheating operation before a load peak value arrives; wherein the reservation proportion is dynamically adjusted according to the confidence coefficient of the prediction result, and when the prediction confidence coefficient is higher than a preset threshold value, the scale of the preheating container is automatically expanded to cover the predicted load increment. According to the method, the time sliding window is constructed, the active prediction model is utilized to analyze the periodicity and the burstiness of the load, the container preheating is executed before the peak value arrives based on the pre-reaction principle, and the advanced time window is introduced, so that the cold start time of the traditional passive expansion is eliminated, the resource is used as required, and the resource utilization rate is improved. And the response speed of the system to the tidal flow is obviously improved.

The application discloses an online operation monitoring system and method based on a cable branch box, and relates to the technical field of power monitoring.The system performs online operation monitoring on multiple parameters of the cable branch box, accurately identifies whether the cable branch box has dominant operation abnormalities, effectively avoids sudden failures, and after eliminating the dominant abnormalities, relies on a digital twin body constructed by a cable branch box simulation monitoring model to perform long-term trend simulation on progressive and implicit defects such as busbar deformation and conductor creep, performs dynamic similarity analysis on historical data and simulation data by tracing the monitoring unified value system, amplifies early weak abnormal signals and quantifies them into identifiable risk indexes, relies on the digital twin model to simulate the long-term trend of the progressive and implicit defects, combines the dynamic similarity analysis to quantify the risk indexes, realizes the leap from "passive maintenance" to "active prevention", and improves the reliability and operation and maintenance efficiency of the distribution network.

The invention discloses a large model reasoning method and system based on active load awareness, and belongs to the technical field of large model reasoning services. Aiming at the problem that reasoning service is difficult to consider resource efficiency and service quality due to burstiness and heterogeneity of mixed loads in the prior art, the invention provides a closed-loop method of cycle planning, online scheduling and self-adaptive adjustment and optimization. In the period planning stage, prospective instance pre-configuration is carried out by utilizing long-term load prediction; in the online scheduling stage, the request is routed and scheduled through response length sensing and a heterogeneous scheduling strategy; and in the self-adaptive tuning stage, fine-grained correction is carried out on the instance scale based on peak working set pre-judgment. According to the method, resource supply and request scheduling are coordinated, the SLO is guaranteed, the delay peak is effectively inhibited, the GPU utilization rate is improved, and the method is suitable for an enterprise-level LMaaS platform.

The application provides a method and system for supporting S-UPF policy control of satellite-ground computing integration, relates to the technical field of satellite-ground communication, and comprises the following steps: constructing a link interruption feature vector sample and training an interrupted link prediction model; analyzing current satellite-ground link multi-dimensional operation real-time parameters; identifying a data control strategy and performing analysis when the interruption risk probability is greater than the expected risk probability; jointly reconstructing a transmission path structure, a transmission cache state and a redundancy protection strategy based on a link interruption data recovery cost model, obtaining an optimized data control strategy, and updating the S-UPF policy. The application solves the technical problem that, in the prior art, satellite-ground link interruption is sudden, data transmission is interrupted passively and needs to be rescheduled, and satellite-ground computing collaboration efficiency is affected. Through the construction of an interruption prediction mechanism and the pre-position optimization of a strategy, the application realizes the rapid recovery of data transmission with minimum scheduling resources, and improves the satellite-ground computing collaboration efficiency.

A communication prediction-based energy saving method and an apparatus, the method including receiving, by a first radio access network device, from a second network device a communication prediction result that includes a communication period, traffic, a periodic communication indication, or a burst communication indication, where the traffic represents a predicted uplink or downlink data traffic volume, where the periodic communication indication represents whether communication predicted to occur is periodic, and when the communication predicted to occur is periodic, a communication interval, or the uplink or downlink data traffic volume of the periodic communication, where the burst communication indication represents whether the communication predicted to occur is bursty, and, when the communication predicted to occur is bursty, a time point, the uplink or downlink data traffic volume of the burst communication, and performing, by the first radio access network device, an energy saving operation based on the communication prediction result.

Devices, systems, and methods to support bursty traffic via low-power wake-up signals for a wireless transmit / receive unit (WTRU). A first LP-WUS is received during a first type LP-WUS MO. The first type LP-WUS includes a first indication and a second indication. A transmission is monitored for, based on the first indication and the second indication. The monitoring includes skipping monitoring for a PDCCH transmission and receiving a second type LP-WUS, based on the first indication having a first value, or monitoring for a PDCCH transmission based on the first indication having a second value. In some implementations, the monitoring comprises monitoring second type LP-WUS MOs sparsely, based on the first indication having the first value and the second indication having the first value, or the monitoring comprises monitoring all second type LP-WUS MOs, based on the first indication having the first value and the second indication having the second value.

The invention relates to the technical field of routers, in particular to a packet transmission optimization method based on intelligent traffic prediction, which comprises the following steps: S1, extracting traffic characteristics; s2, generating a flow prediction result; s3, dynamically generating a packet transmission optimization strategy for the preset time window in the future; and S4, performing active optimization on packet transmission. The invention relates to a packet transmission optimization method based on intelligent traffic prediction, which realizes the fundamental change of network traffic management, that is, a traditional passive and reactive management mode is upgraded to a prediction-based active and preventive optimization mode, and the network traffic is optimized through a closed loop of'acquisition-prediction-strategy generation-execution '. According to the invention, the router can pre-configure resources and adjust the strategy before traffic congestion or decline of service quality, so that network jitter is significantly reduced, transmission delay is reduced, sudden packet loss is avoided, and smoother, more reliable and predictable network experience is provided for users.

The invention discloses a traffic scheduling method and system, relates to the technical field of network communication, and solves the problem that high-dynamic and sudden traffic scheduling requirements are difficult to meet in the prior art. According to the embodiment of the invention, through the multi-objective optimization function, the variance of the link load can be effectively inhibited, the link time delay and packet loss rate are reduced, and the network quality is improved; by introducing an optimization algorithm combining a Lagrangian multiplier method and a real-time gradient descent method, a flow distribution scheme can be quickly adjusted based on network change conditions, and high-dynamic and sudden flow scheduling requirements can be better met.

The invention discloses a multi-channel satellite frequency point file processing method and system based on threshold optimization, and relates to the technical field of satellite communication data processing, and the method comprises the steps: calculating the burst position and burst length of a main channel file; counting the number of emergencies in the historical main channel file, and calculating a threshold value according to the number of emergencies in the historical main channel file; if the burst number of the main channel file is smaller than a threshold value, a low burst processing mode is selected, and otherwise, a high burst processing mode is selected; and according to the selected mode, reading corresponding burst data from the main and auxiliary channel files and processing the data. The method is suitable for efficiently processing burst data in a multi-channel satellite frequency point file, and solves the problems of data access burstiness, low delay requirements and accurate reading according to positions in a satellite positioning scene.

The invention relates to the technical field of content distribution network acceleration, and discloses an intelligent acceleration method and system for a short play content distribution network. The method comprises the steps that historical request records are collected from edge nodes, a request interval set is generated according to request intervals, a periodic request mode and a sudden request mode are recognized, and then a mixed request mode graph is constructed in a combined mode. And calculating the connection strength and the path complexity of each mode node in the map, and generating an activeness score according to the connection strength and the path complexity. And pruning the atlas according to the score, and reserving high-activity mode nodes and connection edges to form a core request mode network. And predicting request traffic distribution in a future time window according to the core network, generating a prefetch instruction list, and finally scheduling cache resources to perform distributed preloading on the movie content. According to the method, the key user behavior mode can be finely identified and preferentially processed, and the content prefetching accuracy and the distribution network efficiency are improved.

This application discloses a flow control method, apparatus, device, and medium for input / output requests, relating to the field of computer technology. The method includes: responding to an input / output request corresponding to a disk array, obtaining the request type of the input / output request, wherein the request type is an access request or a maintenance request, where an access request is used to enable an external device to perform data access operations on the disk array, and a maintenance request is used to enable internal maintenance operations on the disk array; if the request type is an access request, determining the request queue corresponding to the input / output request as a token bucket queue; or, if the request type is a maintenance request, determining the request queue corresponding to the input / output request as a leaky bucket queue; and adding the input / output request to the request queue corresponding to the input / output request. This application combines the token bucket and leaky bucket mechanisms, balancing the burstiness of access requests with the stability of maintenance requests.

Provided in the embodiments of the present disclosure are an ODUCn mapping method, an electronic device, a storage medium and a program product. The method comprises: acquiring the bandwidth and the degree of parallelism of a target service, determining the multiple yi of the bandwidth with respect to a unit bandwidth, and determining the multiple m×yi of the degree of parallelism of the target service with respect to a predetermined data size; acquiring service data to be processed of the target service within the current clock cycle; determining yi mapping channels for mapping the target service; dividing said service data into m×yi data blocks of equal size, and allocating the m×yi data blocks to the yi mapping channels; and mapping the data blocks allocated to the mapping channels to containers corresponding to the mapping channels. The solution of the embodiment allows a service data transmission process to have good uniformity, effectively reducing the degree of data burstiness, a waste of logical resources caused by time slot interleaving, and the complexity of implementation.

The embodiment of the invention provides a bandwidth adjustment method, system and device, electronic equipment and a computer storage medium, and the method comprises the steps: receiving bandwidth detection data transmitted by a first terminal, transmitting the bandwidth detection data to a second terminal based on a target transmission rate, and receiving state message information fed back by the second terminal; determining continuous packet loss interval distribution data corresponding to the preset time window information under the condition that the state message information indicates that a preset speed limit feature triggering condition is met; and adjusting the target sending rate under the condition that the continuous packet loss interval distribution data meets a preset packet loss sudden judgment condition.

The invention discloses a system and a method for monitoring direct connection equipment by a switch, which belong to the field of network equipment monitoring, and are characterized in that after the switch receives a link signal from the direct connection equipment, link handshake mechanism parameters are recorded; collecting a feature set containing a handshake parameter and an electrical layer connection noise parameter at a set time interval; constructing a basic link behavior model of the direct connection equipment; comparing the newest feature with the model, and generating an initial abnormal identifier if the newest feature exceeds a threshold value; extracting physical layer disturbance data in a corresponding time period, reconstructing an anomaly enhancement vector, and judging an anomaly type through a classification model; when a physical layer intermittent fault appears and repeatedly appears in a plurality of periods, generating an equipment failure early warning signal; when the sudden high-frequency disturbance is identified, the sampling frequency is increased and key monitoring is carried out; according to the invention, active perception and fault early warning of the terminal equipment without the management protocol can be realized, and the link security and the equipment availability are improved.

Devices, systems, and methods to support bursty traffic via low-power wake-up signals for a wireless transmit / receive unit (WTRU). A first LP-WUS is received during a first type LP-WUS MO. The first type LP-WUS includes a first indication and a second indication. A transmission is monitored for, based on the first indication and the second indication. The monitoring includes skipping monitoring for a PDCCH transmission and receiving a second type LP-WUS, based on the first indication having a first value, or monitoring for a PDCCH transmission based on the first indication having a second value. In some implementations, the monitoring comprises monitoring second type LP-WUS MOs sparsely, based on the first indication having the first value and the second indication having the first value, or the monitoring comprises monitoring all second type LP-WUS MOs, based on the first indication having the first value and the second indication having the second value.