51 results about "Network congestion control" patented technology

Systems, computer program products, and methods are described for advanced congestion control using multiple congestion indicators in a networking environment. An example system may include an intelligent agent configured to learn congestion control policies. The agent may interact with real-world or simulated environments replicating real-world benchmarks. Congestion indicators such as telemetry information, packet drop metrics, congestion notification packet rate, pause frame rate, port utilization metrics, and / or the like form a comprehensive state representation of the network, enabling congestion state of the network environment. The intelligent agent evaluates these conditions using a reward function to optimize network performance. The intelligent agent may then implement a behavioral policy in response to the captured congestion indicators, thereby changing the congestion state of the network environment.

The invention provides a network traffic matrix prediction method based on multi-scale convolution and attention collaborative enhancement LSTM (Long Short Term Memory), and solves the problems that an existing traffic matrix prediction model is insufficient in multi-scale feature capture, weak in space-time correlation modeling and poor in generalization. A dynamic multi-scale convolution module (adopting a 3 * 3 / 5 * 5 / 7 * 7 two-dimensional convolution kernel) is used to extract local fine-grained association and global coarse-grained trend of a two-dimensional traffic matrix, deep fusion of LSTM time sequence features and two-dimensional multi-scale spatial features is realized in combination with a cross attention mechanism, and time-space cross-dimension association is enhanced. And meanwhile, the problem of gradient disappearance of the deep network is relieved by relying on residual connection, and topology independence is realized based on a flow-by-flow method so as to adapt to traffic matrixes of different dimensions. The method specifically comprises the following steps: performing stream-by-stream division and normalization preprocessing on an original two-dimensional traffic matrix, and after processing by modules such as dynamic multi-scale convolution, cross attention enhancement LSTM, residual block and gating screening and the like, outputting a single OD stream predicted value and reconstructing a complete traffic matrix; and integrating objective functions of the encoder and the decoder, taking the MSE as a loss function, and adopting an Adam optimizer to iteratively update model parameters until the model converges. Experiments show that the MAE and the MSE of the method on real data sets such as Ailene, GEANT and the like are better than those of mainstream models such as SVR, LSTM, MTGNN and the like, the reconstruction precision of a high-dimensional traffic matrix is particularly remarkable, and a reliable basis can be provided for network congestion control and resource optimization.

The invention relates to a network congestion control method, device and equipment, a medium and a program product, and relates to the technical field of communication. The method comprises the following steps: receiving a target message sent by a data sending end; acquiring historical congestion data of a target transmission path in the network equipment corresponding to a preset historical time period, and a target queue length of the target transmission path corresponding to a current moment; determining a reference queue length threshold according to the historical congestion data; under the condition that the target queue length is greater than the reference queue length threshold, adding speed regulation mark information into the target message to update the target message; the updated target message is sent to a data receiving end through the target transmission path, so that the data receiving end responds to the updated target message and sends feedback information to a data sending end, and the feedback information comprises speed regulation mark information. By adopting the method, the control effect of congestion control can be improved.

The invention provides a network congestion processing method and related equipment, and relates to the technical field of network communication. According to the method and the device, the to-be-discarded message which originally needs to execute the discarding strategy at the switch is converted into the target truncated message which has the ECN setting mark and is short in length, and the target truncated message is transmitted in time by scheduling the queue which is additionally configured on the basis of the configuration of the original message forwarding buffer queue and has a higher priority. A receiving end device timely notifies a sending end device to carry out message sending rate reduction processing and to-be-discarded message rapid retransmission processing based on a traditional ECN information feedback mechanism, so that in a network congestion control process based on an ECN technology, a network communication load is effectively reduced by using a message truncation operation; meanwhile, it is ensured that the sending end equipment can reduce the message sending rate in time, retransmission processing is carried out on the message to be discarded in time, and the zero packet loss requirement of the switch is met.

Mechanisms, including systems, methods, and media, for implementing network congestion control are provided, the methods including: imposing a delay on traffic on a connection between a sender and a receiver; determining a minimum round-trip time for traffic on the connection with the delay imposed; and adjusting the delay imposed on the traffic to an adjusted delay based on a minimum propagation delay goal and the round-trip time. In some of these embodiments, the methods further comprise setting an initial value of the delay to the propagation delay goal. In some of these embodiments, the delay is imposed on an acknowledgment message that is part of the traffic. In some of these embodiments, the adjusted delay is equal to the propagation delay goal minus the minimum round-trip time minus the delay.

The application relates to the field of 5G uplink, and discloses a self-adaptive network congestion control system and method, wherein the self-adaptive network congestion control method comprises the following steps: forming an uplink occupation evolution record according to the sending time sequence and feedback indication of a physical uplink shared channel and a physical random access channel; introducing cross-terminal access track alignment processing around the dense overlapping section appearing in the uplink occupation evolution record to generate a heterogeneous congestion indication; embedding an access mode distinguishing identifier in the heterogeneous congestion indication to build the distribution form of the interference degree between different access modes; introducing a variable-granularity scheduling adjustment unit based on the phased change characteristics of the distribution form to form a sending behavior feedback result; and combining the sending behavior feedback result after adjustment to jointly evaluate the resource acquisition efficiency and data discharge rate of each terminal in a continuous scheduling period and dynamically correct a congestion judgment boundary. The application has the advantage of improving the network resource utilization efficiency.

The application discloses a mass data network congestion control method and device, electronic equipment and storage medium, belongs to the computer network communication technical field, method includes: according to the congestion mark byte number of the confirmation data packet, total confirmation byte number and historical congestion ratio, determine the congestion ratio estimation value, and combine the sampling data packet sending time and confirmation data packet receiving time to determine the round trip communication time and communication time gradient;According to the confirmation byte number, the sending byte number, the round trip communication time and the historical bottleneck bandwidth to determine the bandwidth utilization rate;When the round trip communication time does not exceed the minimum communication time, maintain the sending end control parameter;When the round trip communication time exceeds the minimum communication time, according to the congestion ratio estimation value, the communication time gradient, the round trip communication time and the bandwidth utilization rate to determine the congestion score, and accordingly adjust the sending end control parameter. Through the implementation of the application, the problem of inaccurate congestion control of mass data network in the prior art can be solved.

The present application relates to the technical field of network communication congestion control, in particular to an end-network cooperative ubiquitous network congestion control method. The method is applied to a ubiquitous network comprising end nodes, intermediate nodes and content production nodes. The end nodes generate and send interest packets. The intermediate nodes obtain queue usage state, pending interest table occupation state, available bandwidth state and interest packet satisfaction state when the data packets return, generate a first congestion score or a second congestion score, and generate a congestion marking probability in combination with the occupation contribution of the target interest flow in the pending interest table to mark the return data packets with congestion. The intermediate nodes perform interest flow splitting or rate shaping according to the congestion level, and the end nodes adjust the subsequent interest packet sending window according to the congestion score information and the congestion marking result. The present application is suitable for high dynamic and multi-path ubiquitous network congestion control.

Provided in the embodiments of the present disclosure are a network congestion control method, a communication apparatus, a storage medium and a program product. The network congestion control method comprises: acquiring a service message, wherein the service message comprises a first flag bit and at least one second flag bit, the first flag bit is used for indicating the type of at least one type of congestion control signal carried in the service message, the at least one second flag bit corresponds to at least one type of congestion control signal, and one of the at least one second flag bit is used for indicating a sub-type of the type of congestion control signal corresponding to the second flag bit; and on the basis of the first flag bit and the at least one second flag bit, processing the service message.

The application discloses a routing optimization method based on a graph neural network and deep reinforcement learning, which is improved from two aspects of network topology perception and intelligent decision-making by combining GNN and DRL technologies, starting from the actual needs of modern communication networks. Under the condition of ensuring network stability and user experience, the algorithm has been significantly improved in network resource allocation, routing path calculation and network congestion control, and the communication network is promoted to develop in a more intelligent and greener direction.

The invention discloses a network congestion control method, a network congestion control device and electronic equipment. The method comprises the following steps: acquiring network state information respectively corresponding to a plurality of agents in a data center network; determining initial adjustment strategies corresponding to the plurality of agents according to the network state information, and determining response results obtained after the initial adjustment strategies are executed; updating network parameters of the global network according to the response result to obtain a target global network; and updating the network parameters of the local networks corresponding to the plurality of agents by adopting the target global network to obtain a target local network. The technical problem that in a distributed network control scene, related technologies lack an effective mechanism to ensure coordination of network congestion control decisions among a plurality of agents, and consequently, overall performance deterioration is likely to occur is solved.

The application relates to a network congestion control method and device, computer equipment, a storage medium and a product, and relates to the technical field of communication. The method comprises the following steps: obtaining a response message for a sent message; when it is determined that a data flow where the sent message is located exists congestion based on the response message, identifying a congestion occurrence position according to congestion information carried by the response message; if the congestion occurrence position is located at a message forwarding port, obtaining path mapping information; wherein the path mapping information is used for recording the corresponding relationship between message transmission parameter values and message forwarding paths; the original message transmission parameter value of a to-be-sent message in the data flow is adjusted to a first transmission parameter; the first transmission parameter refers to other transmission parameters in each message transmission parameter, except the original message transmission parameter of the to-be-sent message; and the next to-be-sent message of the sent message in the data flow is transmitted through a message forwarding path indicated by the first transmission parameter. The method can improve network bandwidth utilization.

The application discloses a network congestion control method and device, equipment, a storage medium and a program product, and relates to the technical field of communication, so as to realize accurate network congestion control. The method comprises the following steps: a first message is sent to a second device; wherein the first message comprises the identification of a first data flow in which network congestion occurs, the information of a first queue to which the first data flow belongs, and first information used for representing congestion processing action information. The embodiment of the application can realize accurate network congestion control.

The invention provides a congestion control selection method and system based on reinforcement learning, and relates to the technical field of network congestion control. The method comprises the following steps: firstly, in an offline preparation state, dividing a network state into a plurality of grades, and matching a candidate congestion control algorithm pool for each network grade; after entering an online running state, collecting currently running network performance data, determining the grade of the currently running network performance data, loading a corresponding candidate congestion control algorithm pool, and monitoring the network dynamic state so as to feed back and adjust a network grade division basis of an offline stage; and a reinforcement learning agent is introduced, the selection of the congestion control algorithm is used as the action of the reinforcement learning agent, the reinforcement learning agent is trained, and the trained reinforcement learning agent is utilized to decide and output the congestion control algorithm. The preparation of the offline stage accelerates the convergence speed of training, can quickly and dynamically switch to the current optimal congestion control algorithm, and ensures the stability and high efficiency of network transmission.

The embodiment of the present application provides a congestion control method and device for real-time communication. The technical scheme provided by the embodiment of the present application enters a code rate restriction state according to a set code rate restriction period when network congestion is detected, determines a current restriction code rate according to a historical restriction code rate and a maximum receiving code rate of a data receiving end when network congestion occurs, and performs network congestion control based on the current restriction code rate in the code rate restriction state, wherein the detection code rate enters a detection state after the duration of network congestion reaches a set detection waiting time or after each code rate restriction period, and the detection code rate is obtained by upward detection of the code rate in the detection state. The network congestion control is performed based on the current restriction code rate in the code rate restriction period. While ensuring that the code rate is close to the code rate restriction under the current network environment, the number of upward detection of the code rate is effectively reduced, the code rate fluctuation caused by frequent upward detection of the code rate is reduced, and the quality of experience of real-time communication of a user is improved.

The invention discloses a cross-data center RDMA network congestion control method and device based on in-band telemetering, and belongs to the technical field of network transmission control. The method comprises the following steps: sending a first cross-data center flow to an upstream DCI switch at an initial rate; according to a Switch-INT message returned by an upstream DCI switch, a first fair rate is calculated, a second cross-data center flow is sent to the upstream DCI switch at the first fair rate, an ACK message of a downstream data center responding to the first cross-data center flow is obtained, and the ACK message comprises a suggested rate; and generating a final rate in combination with the first fair rate and the suggested rate, and sending a third cross-data center flow to an upstream DCI switch at the final rate. According to the method, the network sensing capability of the cross-data center flow can be improved, the fairness of the flow in the data center and the cross-data center flow is improved, and the transmission delay of the datagram is optimized.

The application discloses a dual-time-scale congestion control method of an RDMA lossless data center network and belongs to the technical field of data center network congestion control. The application constructs a dual-time-scale congestion control framework. The method introduces a reinforcement learning model in a coarse-grained layer, realizes the control of data sending rate and a sending window by multi-dimensional perception of global network state, traffic distribution characteristics and historical performance indexes, and thus achieves an optimization decision from a global perspective. In a fine-grained layer, the instantaneous link state is perceived in real time based on a fast feedback congestion mechanism, and fine adjustment of traffic and fast convergence of a micro-time scale are completed. By the optimization capability of the reinforcement learning in dual-dimensional scheduling of the rate and the window, and in combination with the low-delay response characteristics of the fast feedback congestion mechanism, the application can effectively inhibit the accumulation of switch queues, block the risk of congestion diffusion, and thus achieve a dynamic balance among high network throughput, multi-flow fairness and ultra-low queuing delay.

A method for congestion control in a data communication protocol employing acknowledged communication may include measuring a flight size. A transmission rate may be measured. A trend of the flight size may be determined. A trend of the transmission rate may be determined, and the trend may be derived from a transmission rate gradient calculation, in which either the transmission rate measurements or the transmission rate gradient calculations or both may be filtered to reduce their temporal variability. Whether there is a congestion may be detected according to the determined trend of the transmission rate and the trend of the flight size. Upon detection of the congestion, a change may be made from a current congestion control state to a new congestion control state. Data may be transmitted while respecting a maximum amount of unacknowledged data which the transmitting node may transmit. An apparatus is also disclosed.

The present disclosure relates to a network congestion control method, apparatus, electronic device, and storage medium. The method includes: obtaining target delay information of a target data packet, the cumulative number of data packets with continuously increasing delay, and a target quantity threshold; determining a data reception rate corresponding to a receiving terminal of the target data packet within a preset period when the target delay information is greater than a preset delay threshold and the cumulative number of data packets is greater than the target quantity threshold; the end time of the preset period is the reception time of the target data packet; determining a data transmission rate corresponding to a sending terminal of the target data packet based on the data reception rate, so that the sending terminal sends data packets at the data transmission rate. According to the technical solution provided by the present disclosure, congestion control can be performed efficiently and accurately.

The invention discloses a low-invasion congestion control method based on RTT and AoI fusion perception, and belongs to the technical field of computer network congestion control. The method comprises the following steps: generating a timestamp and a receiving timestamp at a receiving end to calculate an information age; fusing the normalized historical minimum round-trip delay with the weight of the historical minimum information age to generate a comprehensive congestion index; comparing the comprehensive congestion index with a historical reference index and a previous measurement period index to determine a network state; and selecting a window increasing or decreasing strategy according to the network state. According to the method, the AoI is sensitive to forward path congestion, so that misjudgment caused by reverse path interference is avoided, and an algorithm can recognize a real congestion state earlier and more accurately in a complex network environment; according to the method, different network phases such as congestion deterioration, remission, idle and the like can be clearly distinguished, and differentiated increasing and reducing strategies are driven, so that dynamic optimization is realized among low invasion, high utilization rate and quick response.

An RoCE network congestion control method includes a first network device that sends a RoCE protocol packet to a second network device. The first network device receives an acknowledgment packet from the second network device that includes indication information and acknowledgment information in response to the ROCE protocol packet. The indication information indicates whether a network path between the first network device and the second network device is congested. The first network device performs congestion control based on the acknowledgment packet.

The invention provides a congestion control method and device, electronic equipment and a storage medium, and belongs to the technical field of networks. The method comprises the steps that in each preset period, the number of clock periods in a normal scheduling state and the number of clock periods in a congestion waiting state of switching equipment are determined, the preset period comprises a plurality of clock periods, and the preset period is a congestion signal detection period; determining a candidate congestion signal based on the number of the clock cycles in the normal scheduling state and the number of the clock cycles in the congestion waiting state, the candidate congestion signal indicating a congestion degree of the switching device in a preset period; and based on the candidate congestion signal, determining a target congestion signal to adjust a data transmission rate in a network link where the switching device is located, the target congestion signal indicating a congestion degree of the network link. According to the method, the congestion signal capable of accurately reflecting the congestion degree is provided, and low-overhead network congestion control is realized.

A network congestion control method includes, during transmission of a first service flow, a first network device that obtains a measurement packet including groups of measurement results corresponding to the first service flow, and each group includes an identifier of a forwarding device on a forwarding path of the first service flow and a measurement value corresponding to the forwarding device. The first network device determines, based on the groups, a congested second network device on the forwarding path. The first network device determines a control policy based on location information of the second network device, where the control policy resolves congestion of the first service flow.

Some embodiments described herein provide a method for reducing congestion in a network using per-hop telemetry data and network adapters implementing such a method. As compared to conventional congestion control mechanisms, the method may include sharing telemetry data between flows on a per-hop basis to accelerate the convergence of congestion control efforts in a multi-flow network environment. In some embodiments, a destination network adapter may be configured to generate a response packet that includes all of the telemetry data for a packet without combining it, such that the response packet includes distinct telemetry data for each hop in the flow. A sender network adapter may be configured to parse the telemetry data in the response packet to determine the queue length and the link utilization for each hop in the flow through the network, which it then uses to determine a hop rate for each hop in the flow.

In various embodiments, a congestion control module within a transport stack limits the rate at which packets are transmitted from a server to a client device based on a percentage of the available capacity of a network path through which the packets are transmitted. In some embodiments, the available network path capacity can be determined by first performing a linear regression using (1) send durations over which packets associated with encoded frames are transmitted, and (2) corresponding reception durations over which the packets associated with the encoded frames are received, in order to determine a line that relates send duration and reception duration. After the line is determined, the available network path capacity can be computed as an estimated intersection between the determined line and the line y=x, with the intersection being approached as a limit.

The invention provides a network-on-chip congestion control method and system, and belongs to the technical field of network-on-chip, and the method comprises the steps: responding to any memory reading request, and obtaining current response data; when it is judged that the current response data is approximable data according to the current approximation flag bit of the current response data, any target entry data in a preset output cache table is compared with the current response data, and a precision difference is obtained; when it is judged that the approximate entry data exists, setting a target bit position in the flag bit field, and discarding the current response data; when it is judged that the approximate entry data does not exist, the current response data serves as new target entry data to be stored in an output cache table; and packaging the target entry data into a data packet, and sending the data packet to one or more streaming multiprocessors. According to the method and the device, a plurality of independent memory reading responses aiming at approximate data are converted into single data responses sent through multicast, so that the overall performance of the general parallel computing processor is improved.

A system and method for reducing congestion in onboarding networks before onboarding and provisioning connections are made by user equipment (UE), and for reducing congestion in onboarding networks after attempts are made by UEs to make connections to onboarding networks for onboarding and provisioning which result in network congestion. Broadcast information (referred to herein as “onboarding access control information”) from a radio access network (RAN) to user equipment (UE) is used in order to guide onboarding UEs to detect and access the right onboarding network to prevent unnecessary network access attempts by UEs which are not authorized to access an onboarding network for onboarding.

The invention discloses an electric power emergency network congestion control method based on a low-orbit broadband satellite, which belongs to the technical field of electric power, and comprises the following steps: judging a network state as a non-uniform flow distribution or malicious node attack state according to congestion inducements; executing a flow scheduling strategy based on trust evaluation aiming at a flow distribution non-uniform state: calculating a link trust value to identify a key link, screening a key flow, and calculating a rerouting path by optimizing an objective function to perform flow migration; and aiming at a malicious node attack state, executing a congestion relieving strategy based on spatial-temporal characteristics and reinforcement learning: evaluating a node trust value and isolating malicious nodes based on the spatial-temporal characteristics, and dynamically adjusting a packet loss strategy of satellite nodes by utilizing a deep reinforcement learning model which is executed in a centralized training and distribution manner. According to the method, the problems of satellite link congestion misjudgment, malicious attack defense deficiency, slow sudden service response and the like in an electric power emergency scene are solved, and reliable and low-delay transmission of key electric power services in an extreme environment is remarkably guaranteed.

A switch dynamic ECN configuration method and system based on reinforcement learning and a switch belong to the field of network congestion control and AI optimization, and the method comprises the following steps: training a reinforcement learning model by using historical data, deploying the model, collecting multi-dimensional state data S of a switch port in real time, and outputting an ECN configuration action value A; and calculating a reward value R, updating an experience pool in combination with state transition data {S, A, R, S '}, and optimizing a model by adopting a double-inference mechanism, a gradual change epsilon-greedy strategy and historical optimal action backtracking. By dynamically optimizing the ECN threshold, balancing the throughput rate and the queue stability, reducing the packet loss rate and delay jitter and improving the network performance and the model reliability, the method is suitable for a real-time congestion control scene of a large-scale data center switch.

A wireless device receives, from a network node of a first network via a packet data unit (PDU) session of a second network, a congestion notification of the first network. The wireless device sends, to the second network, the congestion notification of the first network.