1870 results about "Goodput" patented technology

The invention provides a data management method, device and equipment and a readable storage medium, and the method comprises the steps: obtaining key value cache data generated by a first type of computing power unit in a pre-filling stage, transmitting the key value cache data to a memory pool, and caching the key value cache data in the memory pool; and in response to a data acquisition demand of the second type of computing power unit in a decoding stage, transmitting the specified key value cache data cached in the memory pool to the second type of computing power unit according to the data acquisition demand. According to the technical scheme of the invention, the memory pool is arranged and the aggregated transmission channel is utilized, so that the transmission efficiency of the key value cache data between the computing power devices is effectively improved, the data transmission delay is reduced, and the network transmission protocol overhead is reduced, thereby improving the data interaction efficiency of the computing power devices in a pre-filling stage and a decoding stage, and improving the user experience. The utilization of computing power resources is optimized, and the throughput and performance of large model reasoning are improved.

The invention relates to the technical field of software defined networks, in particular to an SRv6-based SDN (Software Defined Network) data center load balancing method and system, and the method comprises the steps: obtaining node real-time load and SRv6 routing available resource information through an SDN controller, constructing a load balancing model, and recognizing an overload node and a congestion link; generating a flow scheduling path in combination with segmented routing characteristics; in combination with network state information, monitoring the load condition in real time when the node is unbalanced or the link packet loss rate is gt; when 5%, a load balancing adjustment mechanism is triggered, unbalance types are distinguished, a related flow set is determined, optimization is performed by using a DQN algorithm, and a flow scheduling strategy is generated and dynamically adjusted by taking throughput maximization and unbalance degree minimization as targets; and issuing the scheduling strategy to the data center network equipment, updating the flow table through the Netconf protocol, and redistributing the flow. Therefore, the problems of low load balancing efficiency, insufficient dynamic adaptability, low resource utilization rate and the like in the prior art are solved.

The invention discloses a high-dynamic mobile ad hoc network dynamic time slot allocation method and device, equipment and a storage medium, and belongs to the technical field of wireless communication and networks. The device is deployed between an upper-layer application module and a bottom-layer communication module of a communication node, and comprises a topology sensing module, a time frame construction module, an idle time slot distribution module, a GNSS module and an information interaction module. According to the method, neighbor nodes are determined among nodes through interaction information, and node topology change indexes and link states are calculated; the coordination node performs idle time slot allocation according to a request sent by the node in the control time slot, and dynamically adjusts a sensing period according to the current network topology change; and the node sends data in the information time slot according to the time slot allocation table, and carries out information interaction and topology perception according to the perception period. According to the method, the packet loss rate in the link unstable environment is effectively reduced, the data transmission reliability is higher, the method can adapt to link and topology dynamic changes, and the network throughput and the resource utilization rate are improved.

A method, a network device, and a non-transitory computer-readable storage medium are described in relation to an LAS service. The LAS service may include enablement or disablement of the LA service subsequent to an establishment and active traffic flow associated with an application session. The LAS service may be enabled or disabled based on performance metric criteria and threshold values not being satisfied or being satisfied. The LAS service may be implemented to include an exposure function.

The invention relates to the technical field of network management, in particular to an Internet of Things gateway control method and system based on AI computing power, and the method comprises the following steps: obtaining channel real-time signal intensity, noise level and occupancy proportion, monitoring protocol type and transmission condition, analyzing the relationship between retransmission times and interference, deducing protocol combination interference and throughput, and obtaining a real-time channel signal intensity, noise level and occupancy proportion. And configuring an optimal scheduling scheme, evaluating resource consumption and load states, packaging and requesting permission entrustment, and judging permission information in combination with an interaction record and an access control list. According to the method, different protocol distribution combinations are simulated and deduced, the data throughput of the whole network is further estimated and optimized, potential protocol conflicts are reduced, the use efficiency and the overall performance of network resources are effectively improved, consumption of a local computing unit and network bandwidth is comprehensively considered, packaging processing of request information is optimized, and the network performance is improved. And the real-time response capability and manageability of the network state are obviously improved.

The invention discloses a P4 switch processor core-oriented multi-service traffic packet scheduling method, which mainly comprises the following steps of: proposing a feature extraction method based on a statistical interval and Sketch, integrating data packet time correlation and traffic dynamic change, optimizing a Bloom filter conflict detection mechanism, and improving the adaptability and accuracy of an algorithm in a limited resource environment; a decision tree model is combined with scheduling, a decision tree is reconstructed to adapt to an operation mode and a data processing requirement of a P4 switch, and high-precision service flow classification under linear speed operation is ensured; a packet scheduling method AD-PIFO is designed based on a variable priority, the forwarding priority of a data packet is dynamically adjusted, the QoS requirements of different services are met, and the overall performance and efficiency of the switch are improved. The method is outstanding in high-priority traffic throughput and low-delay performance, is superior to similar algorithms in medium-low priority traffic throughput, is stable and effective under different load conditions, is reasonable in extra time overhead of data packet transmission, and is widely applied to SDN network optimization data packet scheduling scenes.

The present disclosure provides connection management techniques based on graph neural networks (GNN) and deep reinforcement learning (DRL) to optimize user association and load balancing. A graph structure of a communication network is considered for the GNN architecture and DRL is used to learn parameters of the GNN algorithm / model. Connection management is defined as a combinatorial graph optimization problem, and the DRL mechanism uses the underlying graph to learn weights of the GNN for an optimal user connections or associations. The connection management techniques can consider local network features to make better decisions to balance network traffic load while network throughput is also maximized. Implementations are provided based on edge computing frameworks include the Open RAN (O-RAN) architecture. Other embodiments may be described and / or claimed.

Various methods and systems of the present disclosure provide a protocol data unit (PDU) set based quality of service (QoS) handling in a wireline access. To support the PDU Set based QoS handling in the wireline access, one or more PDU Set QoS parameters and / or PDU Set information may be provided to one or more wireline nodes (e.g., a wireline gateway access function (W-AGF) and / or a residential gateway (RG) etc.). When the W-AGF receives one or more N2 requests associated with one or more PDU session resources, the W-AGF maps one or more QoS profiles received from 5G core (5GC) to a wireline access user plane (W-UP) QoS level. The W-AGF sets up one or more W-UP resources based on the W-UP QoS level for a PDU session. Further, various methods and systems of the present disclosure also provide low latency, low loss, and scalable throughput (L4S) enabled wireline access.

Devices, networks, systems, methods, and processes for identifying, classifying, and scheduling Low Latency, Low Loss, and Scalable throughput (LAS) traffic are described herein. A device may receive a data flow comprising one or more data packets. The device may parse the one or more data packets to check for a congestion indicator. Upon detecting the congestion indicator in any data packet, the device can classify or mark the data flow as LAS. The device may schedule the data flow by utilizing an Active Queue Management (AQM) queue for L4S data flows. The device can also mark other uplink or downlink data flows associated with the data flow as L4S. The device can also facilitate LAS capability sharing while associating with one or more wireless devices. The device may utilize a management frame with an extended capabilities field to inform the LAS capability of the device.

The invention discloses a TSN scheduling optimization method and device based on flow sensing autonomous learning, equipment and a medium. The method comprises the following steps: deploying a lightweight flow detection module in a switch or a router, and after a controller receives a data request, automatically identifying a newly arrived unknown service flow by using the lightweight flow detection module, and judging whether the newly arrived unknown service flow is a periodic TT flow or an unpredictable burst flow; the controller performs classification management on the identified service flow types, collects topological information and flow requirements of the whole network and issues the topological information and the flow requirements to the terminal nodes through a network management interface; the controller constructs an intelligent queue scheduling task based on the collected network topology information and traffic demand and converts the task into a Markov decision process MDP, network resources, queue states and priorities are used as a state space, a scheduling strategy is used as an action space, a reward function is designed in combination with throughput and delay indexes, and an intelligent queue scheduling task is obtained. Driving a dynamic environment through real-time data and training a DRL model; an enhanced queue scheduling mechanism Pro-CQF is adopted, different priority labels are configured according to classified flow types, and then mixed flow scheduling is carried out; the controller periodically collects time delay, packet loss rate and end-to-end transmission delay indexes and feeds back the indexes to the DRL model, and a scheduling strategy is updated online. According to the method, the traffic sensing and scheduling efficiency is greatly improved in a network environment in which multiple service flows coexist and end-side equipment functions are different, and the reliability and the expandability of the TSN in industrial Internet of Things, edge computing and other high-real-time application scenes are remarkably enhanced.

The invention provides a PCIe-NTB (Peripheral Component Interconnect Express-Network Temporary Block)-based P2P (Peer-to-Peer) communication method and system between IO (Input / Output) devices, the system comprises a plurality of computing nodes, and the computing nodes comprise a PCIe Switch, an NTB and a plurality of PCIe IO devices; the plurality of PCIe IO devices are connected with the CPU through the PCIe Switch; the NTB is connected with the PCIe Switch; the PCIe IO equipment at least comprises two different types of IO equipment, namely an IO equipment type 1 and an IO equipment type 2; the plurality of computing nodes are connected through the NTB; the system supports P2P communication between IO device types 1 or between IO device types 1 and IO device types 2 of different computing nodes. In the scheme, P2P communication can be directly carried out among IO devices of a plurality of computing nodes on the basis of PCIe-NTB, protocol conversion overhead and main memory and IO device memory moving overhead required by traditional NIC RDMA communication are avoided, and the characteristics of low delay and high throughput of a PCIe network can be fully utilized.

The invention provides a stream computing resource scheduling method and system based on a dynamic time window, and relates to the field of stream computing resource scheduling, and the method comprises the following steps: constructing a master-slave event time monitoring network to obtain a task feature vector, and establishing an adaptive time window scheduling mechanism to dynamically adjust the window size; and constructing a multi-level task priority queue, carrying out state transition by adopting a remote direct memory access mechanism of incremental transmission, and finally coordinating computing nodes through a distributed task arrangement protocol to complete resource switching. According to the method, computing resource competition can be effectively reduced, the flow computing task processing efficiency is improved, dynamic optimal configuration of resources is realized, and the system throughput is enhanced.

The invention discloses a cross-fragmentation transaction and fragmentation migration method and system, and relates to the technical field of transaction methods, and the method comprises the following steps: for a transaction input into a network, if accounts of a payer and a payee are in one fragmentation, carrying out an on-chip transaction; if the account of the payer and the account of the payee do not belong to the same fragment, cross-fragment transaction processing is carried out, and the cross-fragment transaction processing comprises the steps of main fragment transaction distribution; verifying transactions in the fragments; carrying out cross-fragment transaction verification; and submitting the cross-fragment transaction. According to the invention, the introduction of the block chain fragmentation technology can enable the block chain network to maintain the decentration characteristic, and achieve higher throughput and lower transaction delay at the same time.

The invention discloses a multi-resonator filter collaborative optimization method and system based on cloud computing, particularly relates to the field of cloud computing task processing, is used for solving the problem of resource scheduling resources, and can realize intelligent grouping of tasks in a task input stage through a dynamic density clustering method. A dynamic weight coefficient is calculated in combination with the data association degree and the resource conflict probability, so that priority division of tasks is completed before scheduling, and the accuracy of scheduling decision making is improved; a capacity expansion threshold is adjusted in combination with a historical load trend prediction model and a dynamic weight, and a resource capacity expansion strategy based on benefit driving is realized, so that the energy consumption of the system is reduced, and meanwhile, resource expansion is ensured to accurately match actual requirements; task subblock disassembly and a greedy distribution strategy are carried out based on the real-time energy efficiency ratio and the network topology cost, so that the calculation load tends to the optimal node distribution all the time, the calculation throughput is improved, and the delay loss of cross-node data transmission is reduced.

Disclosed is a method for transmitting and receiving a physical protocol data unit (PPDU) in a wireless communication system. A terminal receives a preamble of the PPDU from an access point (AP). The preamble includes a universal signal (U-SIG) field and an extremely high throughput (EHT)-SIG field, wherein the EHT-SIG field may be configured with at least one content channel on a frequency axis. Here, the number of the at least one content channel may be recognized on the basis of the U-SIG field. Subsequently, the terminal receives data of the PPDU on the basis of the U-SIG field and the EHT-SIG field.

Extended Reality (XR) applications may require high throughput and very low latency. New Radio (NR) Wireless receiver / transmitter units (WTRUs) running XR applications (XR-WTRUs) may be confronted with link failures when these applications are used in dense urban or indoor environments because of these devices operating in the millimeter band. Methods and devices are provided therefore for performing cell selection and beam management proactively to avoid link failure due to objects being in line of sight between the XR-WTRU and a network node with which it has established a connection. XR-WTRUs may therefore perform visual sensing that may enable visualizing current and future radio environment and using this information to prevent link failures by handing the XR-WTRU over to an alternative network node in line of sight of the XR-WTRU before the XR-WTRU is no longer in line of sight of the network node with which it has established a connection.

The invention discloses a CPE router, a data transmission method and device, a medium and a product, and relates to the technical field of network communication and data transmission optimizing.According to the scheme, the method comprises the steps of collecting and preprocessing inlet flow and wireless channel data of the CPE router, extracting transmission urgency and a channel quality index, constructing a transmission decision model and outputting service priority classification; and meanwhile, a dynamic transmission strategy and an early warning mechanism are formulated in combination with a channel prediction model result, and network resource allocation is optimized. According to the scheme, accurate matching and intelligent scheduling of network resources are achieved by dynamically sensing the channel state and the service characteristics, key service transmission delay is effectively reduced, the throughput of high-priority data is improved, meanwhile, transmission stability in a complex environment is enhanced, data retransmission caused by channel fluctuation is reduced, and the transmission efficiency is improved. Smoother multi-service concurrent experience is provided for users, potential transmission risks are warned in advance, and network service quality and equipment operation reliability are guaranteed.

The invention relates to the technical field of communication fault monitoring, in particular to a multi-core optical cable fault online monitoring method based on an optical fiber sensing technology. The method comprises the following steps: firstly, establishing a primary network topological graph according to a direct communication relationship of core nodes, selecting a key service optical cable according to a key service type of the primary network topological graph, and establishing a secondary network topological graph; the second-level link historical performance data of the second-level network topological graph is clustered, and the performance state of the second-level link is judged in real time; calculating a trunk index according to the bandwidth and the throughput and determining a trunk optical cable; performing fault monitoring on the trunk optical cable by using ROTDR equipment, completing super-resolution reconstruction and signal analysis through an SR-LSTM model, positioning a fault position and identifying a fault type; a state space and an action space are constructed through an Actor-Critic reinforcement learning framework in combination with a performance state, a fault position and a fault type, an intelligent agent learns an optimal operation and maintenance strategy, and online fault monitoring of a main optical cable is realized.

The invention relates to the technical field of low earth orbit satellite network communication, in particular to a low earth orbit satellite network load balancing intelligent routing algorithm based on a DQN, which adopts a time slicing method to learn node connection relation and load condition in each time slot static topology, according to the method, multiple targets such as the distance between target satellites, the node congestion degree and the link available bandwidth are jointly optimized, an optimized next-hop selection strategy is given for each node, a link connectivity matrix is updated for different time slots, training is continued to update network parameters, and the method has better performance in the aspects of time delay and network throughput, and can be applied to the field of wireless communication. The multiplexing of each satellite node in the network is reduced, the problem of uneven load in the network is relieved, and the network performance is improved.

A method for optimizing service-level-policy-based network performance management in a 5G wireless network slice includes: analyzing, by a slice analytics module, a target slice throughput of the slice in comparison to an observed slice throughput of the slice to compute a slice-throughput error function; and updating, by the slice analytics module, based on at least the slice-throughput error function, at least one of the following: i) radio resource management policy maximum ratio for the slice and a corresponding shared pool of resource blocks (RBs) for the slice; ii) radio resource management policy minimum ratio for the slice and a corresponding prioritized pool of RBs for the slice; and iii) radio resource management policy dedicated ratio for the slice and a corresponding dedicated pool of RBs for the slice.

Methods, systems, and devices for wireless communications are described. A user equipment (UE) may receive, at a first modem and from a first device, one or more data packets. Based on an explicit congestion notification (ECN) capable transport (ECT) codepoint indication in respective header portions of the data packets, the UE may route the data packets from the first modem to two or more queues associated with a second modem of the UE. Based on receiving buffer and link state information associated with the second modem, the UE may perform a congestion indication procedure that is associated with congestion of a communication link associated with the UE. The UE may thereafter transmit, to a second device, packet data from the two or more queues associated with the second modem of the UE based on a respective priority associated with the two or more queues.

The provided is an anti-radiation low-latency chip for neural network inference acceleration. The following steps are included: deploying a neural network inference acceleration chip at a frontend of a detector, using a streaming architecture to correspond each pipeline stage to each major neural network layer, balancing a limited on-chip memory resource and support for a large-sized input, and constructing layer parallelism, channel parallelism, and convolution kernel parallelism. This application proposes to deploy an anti-radiation, low-latency, and efficient chip for convolutional neural network (CNN) inference acceleration is deployed at the frontend of the detector to improve an intelligence level of future detector hardware. To this end, a fine-grained streaming architecture, and fine-grained storage management, flexible compression and quantization, and anti-radiation digital chip design technologies are proposed to achieve a high throughput and a low on-chip memory consumption while achieving anti-radiation and low-latency inference.

The invention discloses a multi-band network resource allocation method based on ML-MADDPG, and belongs to the technical field of communication. According to the technical scheme, a basic framework of a communication system is established, and the environment and condition of multi-band resource allocation are defined; an integrated MBN network is adopted, and specific modeling is carried out by considering the characteristics of each frequency band; a meta learning method is fused into an MADDPG algorithm, and a meta learning MADDPG framework is constructed; according to task release of a multi-band network, each task corresponds to one network setting, specifically, the whole design is divided into a meta-training stage and a meta-adaptation stage; and target imbalance caused by direct weighted summation is avoided through a layered reward mechanism. The method has the beneficial effects that the packet loss rate is remarkably reduced, the network throughput is improved, the dynamic environment adaptability and sample efficiency are enhanced, the multi-agent cooperation and global coordination capability is optimized, the training stability and generalization capability are improved, efficient resource scheduling under complex constraints is supported, and the actual deployment cost and complexity are reduced.

The invention relates to the technical field of network communication, and discloses a ubiquitous network dynamic fair rate cooperative control method, which aims at link time varying, topology reconstruction and multi-source and multi-path concurrency characteristics of an air-space-ground-sea fusion network. A routing node periodically collects indexes such as a request satisfaction rate, a queue length and a multi-hop queuing-free proportion to obtain link transmission quality and dynamically update an interest packet forwarding table; deducing a queue upper limit, calculating a rate shaping value in combination with an interest packet forwarding proportion, and carrying out delayed transmission on an over-limit interest packet; when the interest queue exceeds a threshold value, a congestion feedback value is generated and goes down along with a data packet, and an uncongested interface obtains a release bandwidth through probability compensation; and if the whole path is still congested, the consumer multiplicatively shrinks the sending window according to the feedback proportion. According to the method, the throughput is remarkably improved and the queuing delay is reduced while fair allocation of bandwidth resources is ensured.

The invention discloses a communication method and system of a computer system, equipment, a product and a storage medium, relates to the field of computers, and aims to solve the problem that a single-channel architecture can become a bottleneck of system performance, the method comprises the following steps: identifying a plurality of direct memory access controllers and a plurality of non-transparent bridging equipment in a first computer system; determining a current target non-transparent bridging device corresponding to each direct memory access controller in the plurality of non-transparent bridging devices; and for each direct memory access controller, carrying out binding operation on the direct memory access controller and the corresponding current target non-transparent bridging equipment, so that the direct memory access controller accesses the second computer system through the corresponding current target non-transparent bridging equipment. According to the invention, the data transmission load can be dispersed to a plurality of communication links, and all data are prevented from being transmitted through the same communication link, so that the throughput of the system is improved, the possibility of bottleneck is reduced, and the data processing capability and performance of the whole system are improved.

The invention provides a network congestion control method and system based on multi-agent reinforcement learning, computer equipment and a storage medium, and relates to the technical field of network congestion control, a communication network comprising a plurality of communication nodes is constructed, the communication nodes are mapped into independent agents, and the independent agents are used for controlling the network congestion. Designing a weighted composite reward function containing indexes such as throughput and delay by taking a flow-level index as a local state and a data sending rate adjustment coefficient as an action; the intelligent agent selects action execution according to a current state based on a strategy network, calculates an environment reward based on the current state, the action, a next state after the action is executed and a weighted composite reward function, and stores the environment reward in a shared experience playback pool; training a value network and a strategy network by using the data in the pool; the trained strategy network is deployed at each node, distributed congestion control is realized, the stability of a congestion control decision in a complex network environment is improved, and bandwidth resources are distributed fairly while the throughput is ensured.

Aspects of the present disclosure improve wireless networking performance by reducing or limiting performance issues caused by undetected or unknown interference sources and / or attenuation sources. Feedback from client devices can be utilized to reduce or limit wireless network performance issues caused by varying interference parameters and / or attenuation parameters observed by each client device. The feedback, associated with wireless communication conditions of corresponding client devices, in conjunction with a machine learning algorithm can be used to enable a dynamic selection of one or more particular operating bands, channels, and / or channel widths to avoid or limit signal interference and / or signal attenuation observed by each client device. Feedback from a plurality of client devices can be used to recommend one or more channels and / or one or more channel widths of one or more operating bands to use that reduces latency, CCI, and / or ACI and improves speed, throughput, and / or QoS.

A UE may be configured with an enhanced network throughput for data transmission with a small burst or a low burst. In one aspect, for the network throughput estimation for an UL transmission with a small burst, the UE may calculate the burst duration by excluding the SR lead time from the burst period. In another aspect, for the network throughput estimation for a window with the low burst duration, the UE may calculate a scaling factor between the channel bit rate and the estimated network throughput of a time window with a sufficient burst duration and estimate the network throughput of the low burst duration based on the estimated channel bit rate and the scaling factor. In another aspect, to improve the overall process of the network throughput estimation, the UE may dynamically adjust the associated parameters by comparing the estimated network throughput with a base throughput.

The invention discloses a LoRa-based parallel backscatter communication method and system. The method comprises the following steps: transmitting a single-tone radio frequency source signal as a signal excitation source; designing a backscattering label, and performing backscattering modulation processing on the environmental perception data in combination with the signal excitation source to obtain a modulated standard LoRa data packet; and the receiver presets the center frequency of a communication link, and performs frequency band distinguishing and demodulation processing on the modulated standard LoRa data packet to complete parallel backscatter communication of the LoRa data. According to the invention, while a parallel backscattering LoRa communication system compatible with commercial LoRa equipment is realized, the throughput of the system is improved, and frequency division multiple access parallel transmission is realized. The LoRa-based parallel backscatter communication method and system can be widely applied to the technical field of parallel backscatter communication.

The invention discloses a GPU Box hardware decoupling system based on a Retimer card and a PCIe Switch chip, and belongs to the technical field of computer hardware architecture and high-speed interconnection. According to the system, a Retimer card and a PCIe Switch chip are integrated in an independent GPU Box, and a decoupling link of a CPU server and a GPU acceleration card is constructed; the Retimer card realizes 30-meter long-distance PCIe signal transmission and breaks through physical distance limitation; the PCIe Switch chip pools GPU resources through a dynamic routing and MRIOV technology, supports flexible allocation of computing power by multiple servers, and realizes Peer-to-Peer direct connection communication between GPUs. Aiming at a large model reasoning scene, the system optimizes KV cache bandwidth allocation and video memory and memory cooperative scheduling, so that the 100B parameter model reasoning throughput is greatly improved; and meanwhile, the usability of the system is greatly improved through fault isolation and hot plug design. According to the method, the problems of physical binding of the CPU and the GPU, limited transmission distance, rigid resource allocation and the like in a traditional architecture are solved, and the method is suitable for large-scale AI calculation and distributed GPU cluster deployment.