Congestion control method for electric power data field

Abstract

The invention discloses an electric power data field-oriented congestion control method, which comprises the following steps of: monitoring a broadband utilization rate, a data packet loss rate and a delay key index in real time in an electric power data field, constructing a network state real-time sensing system, calculating a link flow congestion condition by utilizing a DCQCN congestion control algorithm, and when the flow congestion exceeds a threshold value, executing the congestion control. The method comprises the following steps: compressing data packets at a sending end by adopting a compression algorithm, classifying the data packets, transmitting the data packets to a storage buffer area of the sending end by utilizing a scheduling strategy, carrying out emergency queuing processing on the data packets, adopting a token bucket scheduling algorithm, receiving the compressed data packets by a receiving end, and identifying the data packets subjected to coding compression. And decoding the data packet by using the same decompression algorithm, establishing a feedback mechanism, collecting congestion control effect data in real time, and adaptively adjusting a control strategy according to a feedback result, so that congestion control is continuously optimized. The method can effectively solve the problem that the data transmission rate is reduced due to network congestion and DCQCN failure of a large-scale high-concurrency power data field.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power communication, and in particular to a congestion control method for an electric power data field. Background Art

[0002] The power data network carries key business data such as monitoring data and control instructions. As the core area of ​​data transmission, it is crucial to ensure the efficiency and stability of transmission. With the development of Internet communication technology, the demand for resources in the network is gradually increasing, and the demand will exceed the available part that can be provided by bandwidth resources and switching equipment, resulting in network congestion.

[0003] The Chinese patent publication number: CN117061438A discloses a congestion control method and device. The invention targets the situation where high-concurrency controlled flows occur in network congestion control. The packet loss detection function of the RDMA protocol and the DCQCN algorithm are integrated to enhance the DCQCN algorithm's ability to perceive network packet loss. When packet loss is detected in the network, the controlled flow sending rate is quickly reduced to quickly alleviate the problem of large-scale network congestion. In the core solution, when the first network card determines that packet loss occurs when the first network card sends a service message to the second network card, the first network card switches the message sending rate of the first network card to the second network card to the second rate according to the first congestion control strategy. When the first network card determines that there is no packet loss when the first network card sends a service message to the second network card, the first network card increases or decreases the message sending rate of the first network card to the second network card according to the second congestion control strategy. However, it cannot solve the network congestion of large-scale high-concurrency data fields, and the market demand for reducing the data transmission rate due to the failure of DCQCN.

[0004] The present invention proposes a solution to the above shortcomings: by deploying monitoring equipment at key nodes to monitor key indicators of the power data field in real time, building a network real-time perception system, and using the DCQCN congestion control algorithm to calculate the current link traffic congestion. When the traffic congestion exceeds the threshold, the compression algorithm is used to compress the data and then transmit it to the sending end storage buffer through the priority scheduling strategy. After urgency processing, the token bucket scheduling algorithm is used to transmit the data compression package to the receiving end. Combined with real-time control effect data, the control strategy is adaptively adjusted to further continuously optimize congestion control. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present invention provides a congestion control method for a power data field.

[0006] The technical solution of the present invention is implemented as follows: a congestion control method for a power data field, comprising:

[0007] S1. Deploy monitoring equipment at key nodes of the power data field to monitor key indicators such as broadband utilization, packet loss rate, and latency in real time, build a real-time network status perception system, and reflect congestion conditions;

[0008] S2. Use the DCQCN congestion control algorithm to calculate the link traffic congestion and adjust the current traffic situation. When the traffic congestion exceeds the threshold, a compression algorithm is used at the sending end to compress the data packet for transmission.

[0009] S3, classifying the data compression packets according to the transport layer protocol, the packet destination, and the packet type, and transmitting them to the storage buffer using a scheduling strategy;

[0010] S4, the data compression packet is urgently queued for processing, and the receiving end receives the compressed data packet through the token bucket scheduling algorithm, and identifies the data packet, and after encoding and compression, uses the same decompression algorithm to decode the data packet;

[0011] S5. Establish a feedback mechanism to collect congestion control effect data in real time and adaptively adjust the control strategy based on the feedback results.

[0012] Furthermore, in the step S1, a real-time network status perception system is constructed:

[0013] Select key nodes and determine that the network topology of the power data field is a ring topology. The key nodes include switches and server access points.

[0014] Deploy monitoring equipment: Install monitoring equipment on selected key nodes. The equipment includes dedicated network monitoring hardware or monitoring software running on existing equipment.

[0015] Configure monitoring parameters to set the parameters that the monitoring device needs to monitor in real time, including bandwidth utilization, packet loss rate and delay; configure the monitoring device to collect data on bandwidth utilization, packet loss rate and delay at time intervals (every second or every minute);

[0016] The data collection and monitoring equipment starts to collect network data in real time, including:

[0017] Use SNMP (Simple Network Management Protocol) to collect status information of network devices; use NetFlow or sFlow technology to analyze network traffic and calculate bandwidth utilization; measure packet loss rate and network delay by sending test packets (ICMP ping);

[0018] Reflect the current congestion situation Based on the monitored data, analyze the network congestion situation. If the broadband utilization rate is high for a long time, or the packet loss rate and delay suddenly increase, it indicates that network congestion is occurring. Set thresholds. When the monitored indicators exceed these thresholds, the system automatically alarms.

[0019] Furthermore, in step S2, the DCQCN congestion control algorithm is used to calculate the link traffic congestion:

[0020] Integrated DCQCN congestion control algorithm Integrate the DCQCN algorithm in network devices (switches) and monitoring systems. It is used as a congestion control algorithm for data center networks and manages network congestion by increasing and decreasing the sending rate.

[0021] Calculate the link traffic congestion using the network data collected in real time by the monitoring equipment, including bandwidth utilization and packet loss rate, and apply the DCQCN algorithm to calculate the congestion level of each link, including the following steps:

[0022] S201. Monitor the queue length or delay of each link. If the delay continues to increase while the queue length does not change significantly, it indicates that the bandwidth utilization is high and the queue is too long, resulting in data packet loss:

[0023] Calculate the current link traffic congestion index CI. The calculation formula is:

[0024]

[0025] Among them, CI represents the link traffic congestion index;

[0026] U represents the bandwidth utilization, ranging from 0 to 1;

[0027] D represents the packet loss rate, ranging from 0 to 1;

[0028] λ and γ are weight factors used to adjust the relative importance of bandwidth utilization and data packet loss rate in the calculation, and λ>γ;

[0029] S202, adjusting the size of the sending window according to the state of the link, thereby controlling the flow;

[0030] S203, adjusting the current flow situation:

[0031] When the DCQCN algorithm detects link congestion, it reduces the size of the sending window, thereby reducing the transmission rate of the sender;

[0032] The DCQCND speed reduction algorithm is calculated through the following three parallel processes:

[0033] α update is used to reflect the degree of network congestion. A timer is set at the RP. If the RP does not receive the CNP sent by the NP after K time units, the RP updates the a value. The calculation formula for α update is:

[0034] new_α=g*old_α+(1-g)*CNP_arrived

[0035] Wherein, g represents a constant parameter between 0 and 1, and CNP_arrived represents a one-bit field indicating whether a CNP has arrived in the previous time slot;

[0036] When the DCQCN algorithm detects link congestion, it reduces the size of the sending window, thereby reducing the transmission rate of the sender. The speed reduction calculation formula is:

[0037]

[0038] Among them, R T represents the current sending rate, R C represents the old sending rate, and α represents the degree of network congestion;

[0039] If link congestion eases, the algorithm gradually increases the send window size:

[0040] The speed-up logic is divided into three sequential stages: fast recovery, aggressive increase (keep detection), and high aggressive increase (keep detection);

[0041] After moving to the next stage to speed up in one stage, before slowing down, the current speed will be saved in the target_rate parameter. In the fast recovery stage, the speed increases by half the distance to the target_rate (logarithmically approximate). The rate adjustment formula is:

[0042] CR = (CR + TR) / 2;

[0043] Among them, CR is the current rate and TR is the target rate;

[0044] The formula for actively increasing the rate adjustment is: TR = TR + Rai;

[0045] Among them, TR is the target rate, Rai is the step size of rate increase;

[0046] Super aggressive rate adjustment formula, the rate increase is greater than the step size of Rai or less than the adjustment period of Rai;

[0047] S204, setting a congestion threshold: setting a congestion threshold in the system, and when the monitored link traffic congestion exceeds the threshold, triggering a compression processing mechanism;

[0048] Compression processing mechanism: When the link congestion exceeds the set threshold, the sender receives a congestion notification and uses a data compression algorithm including LZ77, LZ78, and Deflate to compress the data packet to be sent, thereby reducing the amount of data transmitted.

[0049] S205, transmitting the compressed data packet, the compressed data packet is sent to the network, and a compression flag, a compression method identifier, and a context identifier are set in the data packet header;

[0050] At the same time, ensure that the receiving end can correctly decompress the compressed data packets. For compressed data packets, the compression algorithm used, dictionary synchronization information (including the latest information of the dictionary, newly added entries), serial number or timestamp are clearly indicated in the header of each compressed data packet.

[0051] Furthermore, in the step S3, the data compression packets are classified according to the transport layer protocol, the data packet destination, and the data packet type, and are transmitted to the storage buffer using a scheduling strategy:

[0052] Packet classification:

[0053] Before entering the storage buffer, the packets are classified according to the following criteria:

[0054] Transport layer protocol: Identify the transport layer protocols TCP and UDP used by the data packet, because different protocols require different processing methods;

[0055] Packet destination: Classify packets into corresponding destination queues according to broadcast packets, unicast packets and multicast packets;

[0056] Data packet type: Classify data packets according to their control packets and data packets;

[0057] Through decision tree algorithms including ID3, C5, and CART, a priority prediction model for data packet processes is established. Information gain, information gain rate, and Gini index indicators are used for feature selection. A decision tree is generated from top to bottom in a recursive manner, and optimized by pre-pruning or post-pruning to improve the generalization ability of the model. The generated decision tree predicts the priority of the data packet based on the feature value of the input data packet. The generated decision tree is as follows:

[0058] Control packet, UDP, unicast packet, process P1, priority 1;

[0059] Control packets, UDP, multicast packets, process P2, priority 2;

[0060] Control packet, UDP, broadcast packet, process P3, priority 3;

[0061] Control packet, TCP, unicast packet, process P4, priority 4;

[0062] Control packets, TCP, multicast packets, process P5, priority 5;

[0063] Control packets, TCP, multicast packets, process P6, priority 6;

[0064] Data packets, UDP, unicast packets, process P7, priority 7;

[0065] Data packets, UDP, multicast packets, process P8, priority 8;

[0066] Data packets, UDP, broadcast packets, process P9, priority 9;

[0067] Data packet, TCP, unicast packet, process P10, priority 10;

[0068] Data packets, TCP, multicast packets, process P11, priority 11;

[0069] Data packets, TCP, multicast packets, process P12, priority 12;

[0070] The process priority is represented by a number, the smaller the number, the higher the process level;

[0071] At the packet scheduling layer, the packets are reordered and rate-limited, and the scheduler checks all processes in the ready state, selects the process with the highest priority to execute, and transmits the compressed packets to the storage buffer at the sender according to the priority scheduling strategy.

[0072] Furthermore, in the step S4, the data compression packets are queued for urgent processing, and the receiving end receives the compressed data packets through the token bucket scheduling algorithm:

[0073] Urgent queuing processing allocates compressed data packets to different queues according to the urgency of the data packets (determined by the type and priority of the data packets), ensuring that the data packets with high urgency can be processed and transmitted faster;

[0074] Token Bucket Scheduling Algorithm The token bucket scheduling algorithm is implemented at the sending end, which controls the sending rate of data packets;

[0075] The token bucket algorithm works by following these steps:

[0076] Tokens are added to the token bucket at a fixed rate. When a data packet arrives,

[0077] If there are enough tokens in the token bucket, the data packet is allowed to be sent;

[0078] If there are not enough tokens, the data packet needs to wait until there are enough tokens in the token bucket;

[0079] Sending compressed data packets: Sending compressed data packets from the storage buffer to the network according to the urgency and scheduling of the token bucket algorithm;

[0080] The receiving end receives the compressed data packet, and the receiving end receives the compressed data packet through a network interface;

[0081] Identify the data packet encoding compression. The receiving end parses the header information of the data packet to identify whether the data packet has been encoded and compressed;

[0082] Decode the data packet using the same decompression algorithm. According to the compression algorithm indicated in the header information of the data packet, the receiving end uses the same decompression algorithm to decode the data packet;

[0083] The decoding process consists of the following steps:

[0084] Extract the compressed data in the data packet;

[0085] Apply decompression algorithms to restore original data;

[0086] Verify the integrity of the decompressed data.

[0087] Furthermore, in step S5, congestion control effect data is collected in real time, and the control strategy is adaptively adjusted according to the feedback result:

[0088] Establishing a feedback channel A feedback channel is established between the sender and the receiver to transmit information related to congestion control.

[0089] Define feedback data to determine the congestion control effect data that needs to be collected, including network delay, packet loss rate, bandwidth utilization, and queue length;

[0090] The sender and receiver use fixed data formats and transmission protocols to facilitate the exchange of this information between the sender and the receiver;

[0091] Collect congestion control effect data in real time:

[0092] At the receiving end, the congestion control effect data defined above is monitored and recorded in real time;

[0093] Periodically or when a congestion event is detected, this data is sent back to the sender through a feedback channel;

[0094] Analysis feedback results:

[0095] At the sending end, feedback data from the receiving end is received and analyzed;

[0096] Observe the fluctuations of network performance parameters (network delay, packet loss rate, bandwidth utilization, queue length), evaluate whether the congestion control strategy can keep the network in a relatively stable state, and identify existing problems or potential improvement points;

[0097] Adaptive adjustment of control strategy:

[0098] Dynamically adjust congestion control parameters based on feedback results, including token generation rate of token bucket, queue management strategy, and compression algorithm threshold;

[0099] Dynamic adjustments include:

[0100] Change the parameters of the compression algorithm to balance the compression ratio and network load;

[0101] By adding coding redundancy to the transmitted data, the strength of the FEC (Forward Error Correction) coding is adjusted to improve error resilience.

[0102] Beneficial Effects

[0103] The present invention aims at the shortcomings of the prior art that the network congestion of large-scale high-concurrency data fields cannot be solved, and the failure of DCQCN leads to a reduction in data transmission rate. By deploying detection equipment monitoring at key nodes of the power data field, the broadband utilization rate, data packet loss rate, and delay key indicators are monitored in real time, and a real-time perception system for network status is constructed to reflect the current network congestion situation. In combination with the DCQCN congestion control algorithm, the link traffic congestion situation is calculated, the current traffic situation is adjusted, and when the link traffic congestion exceeds the set threshold, a compression algorithm is used to compress the data packets sent by the metering. The compressed data reduces the amount of transmitted data and alleviates the network congestion situation. After the compressed and data packets are classified, they are transmitted to the storage buffer of the sending end through a priority scheduling strategy, and the compressed data packets are transmitted to the receiving end through a scheduling algorithm after emergency queuing processing. The problem of DCQCN failure in the prior art that reduces the data transmission rate is solved. In combination with the effect data of real-time congestion control, an adaptive adjustment control strategy is adopted to continuously optimize the congestion control strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 It is a structural block diagram of a congestion control method for a power data field in an embodiment of the present invention;

[0105] Figure 2 It is a flowchart of the steps of a congestion control method for a power data field in an embodiment of the present invention;

[0106] Figure 3 The present invention is a flowchart of the steps of calculating the congestion degree of each link in a congestion control method for a power data field in an embodiment of the present invention. DETAILED DESCRIPTION

[0107] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0108] The preferred implementation methods of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these implementation methods are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0109] When used herein, the singular forms "a", "an", and "said / the" may also include plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "include / comprise" or "have" and the like specify the presence of stated features, wholes, steps, operations, components, parts, or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof. At the same time, the term "and / or" used in this specification includes any and all combinations of the relevant listed items.

[0110] See also Figure 1-Figure 3 As shown, a congestion control method for a power data field includes:

[0111] S1. Deploy monitoring equipment at key nodes of the power data field to monitor key indicators such as broadband utilization, packet loss rate, and latency in real time, build a real-time network status perception system, and reflect congestion conditions;

[0112] S2, using the DCQCN congestion control algorithm, calculate the link traffic congestion and adjust the traffic situation. When the traffic congestion exceeds the threshold, the compression algorithm is used at the sending end to compress the data packet for transmission;

[0113] S3, classifying the data compression packets according to the transport layer protocol, the packet destination, and the packet type, and transmitting them to the storage buffer using a scheduling strategy;

[0114] S4, the data compression packet is urgently queued for processing, and the receiving end receives the compressed data packet through the token bucket scheduling algorithm, and identifies the data packet, and after encoding and compression, uses the same decompression algorithm to decode the data packet;

[0115] S5. Establish a feedback mechanism to collect congestion control effect data in real time and adaptively adjust the control strategy based on the feedback results.

[0116] like Figure 1-Figure 2 As shown, in the step S1, monitoring equipment is deployed at key nodes of the power data field to monitor key indicators such as broadband utilization, packet loss rate, and delay in real time, and a real-time network status perception system is built to reflect the current congestion situation;

[0117] Specifically, in step S1, a real-time network status perception system is constructed:

[0118] Select key nodes and determine that the network topology of the power data field is a ring topology. The key nodes include switches and server access points.

[0119] Deploy monitoring equipment: Install monitoring equipment on selected key nodes. The equipment includes dedicated network monitoring hardware or monitoring software running on the equipment.

[0120] Configure monitoring parameters to set the parameters that the monitoring device needs to monitor in real time, including bandwidth utilization, packet loss rate and delay; configure the monitoring device to collect data on bandwidth utilization, packet loss rate and delay at time intervals (every second or every minute);

[0121] The data collection and monitoring equipment starts to collect network data in real time, including:

[0122] Use SNMP (Simple Network Management Protocol) to collect status information of network devices; use NetFlow or sFlow technology to analyze network traffic and calculate bandwidth utilization; measure packet loss rate and network delay by sending test packets (ICMP ping);

[0123] Reflect the current congestion situation Based on the monitored data, analyze the network congestion situation. If the broadband utilization rate is high for a long time, or the packet loss rate and delay suddenly increase, it indicates that network congestion is occurring. Set thresholds. When the monitored indicators exceed these thresholds, the system automatically alarms.

[0124] like Figure 1-Figure 3 As shown, in the step S2, the DCQCN congestion control algorithm is used to calculate the link traffic congestion and adjust the current traffic situation. When the traffic congestion exceeds the threshold, a compression algorithm is used at the sending end to compress the data packet for transmission;

[0125] Specifically, in step S2, the DCQCN congestion control algorithm is used to calculate the link traffic congestion:

[0126] Integrated DCQCN congestion control algorithm Integrate the DCQCN algorithm in network devices (switches) and monitoring systems. It is used as a congestion control algorithm for data center networks and manages network congestion by increasing and decreasing the sending rate.

[0127] Calculate the link traffic congestion using the network data collected in real time by the monitoring equipment, including bandwidth utilization and packet loss rate, and apply the DCQCN algorithm to calculate the congestion level of each link, including the following steps:

[0128] S201. Monitor the queue length or delay of each link. If the delay continues to increase while the queue length does not change significantly, it indicates that the bandwidth utilization is high and the queue is too long, resulting in data packet loss:

[0129] Calculate the current link traffic congestion index CI. The calculation formula is:

[0130]

[0131] Among them, CI represents the link traffic congestion index;

[0132] U represents the bandwidth utilization, ranging from 0 to 1;

[0133] D represents the packet loss rate, ranging from 0 to 1;

[0134] λ and γ are weight factors used to adjust the relative importance of bandwidth utilization and data packet loss rate in the calculation, and λ>γ;

[0135] S202, adjusting the size of the sending window according to the state of the link, thereby controlling the flow;

[0136] S203, adjusting the current flow situation:

[0137] When the DCQCN algorithm detects link congestion, it reduces the size of the sending window, thereby reducing the transmission rate of the sender;

[0138] The DCQCND speed reduction algorithm is calculated through the following three parallel processes:

[0139] α update is used to reflect the degree of network congestion. A timer is set at the RP. If the RP does not receive the CNP sent by the NP after K time units, the RP updates the a value. The calculation formula for α update is:

[0140] new_α=g*old_α+(1-g)*CNP_arrived

[0141] Wherein, g represents a constant parameter between 0 and 1, and CNP_arrived represents a one-bit field indicating whether a CNP has arrived in the previous time slot;

[0142] When the DCQCN algorithm detects link congestion, it reduces the size of the sending window, thereby reducing the transmission rate of the sender. The speed reduction calculation formula is:

[0143]

[0144] Among them, R T represents the current sending rate, R C represents the old sending rate, and α represents the degree of network congestion;

[0145] If link congestion eases, the algorithm gradually increases the send window size:

[0146] The speed-up logic is divided into three sequential stages: fast recovery, aggressive increase (keep detection), and high aggressive increase (keep detection);

[0147] After moving to the next stage to speed up in one stage, before slowing down, the current speed will be saved in the target_rate parameter. In the fast recovery stage, the speed increases by half the distance to the target_rate (logarithmically approximate). The rate adjustment formula is:

[0148] CR = (CR + TR) / 2;

[0149] Among them, CR is the current rate and TR is the target rate;

[0150] The formula for actively increasing the rate adjustment is: TR = TR + Rai;

[0151] Among them, TR is the target rate, Rai is the step size of rate increase;

[0152] Super aggressive increase rate adjustment formula, involving a larger step size or shorter adjustment period than Rai;

[0153] S204, setting a congestion threshold: A congestion threshold is set in the system. When the monitored link traffic congestion exceeds this threshold, the compression processing mechanism is triggered:

[0154] After receiving the ECN-marked data packet, the receiving network card NP converts the ECN data packet into a congestion notification packet CNP and sends it to the sending network card RP. The sending network card adjusts the sending rate according to the received CNP to reduce the link traffic and alleviate congestion. If the congestion cannot be effectively alleviated by adjusting the sending rate, a further compression processing mechanism is implemented;

[0155] Compression processing mechanism: When the link congestion exceeds the set threshold, the sender receives a congestion notification and uses a data compression algorithm including LZ77, LZ78, and Deflate to compress the data packet to be sent, thereby reducing the amount of data transmitted.

[0156] Transmitting the compressed data packet, the compressed data packet is sent to the network, and a compression flag, a compression method identifier, and a context identifier are set in the data packet header;

[0157] At the same time, ensure that the receiving end can correctly decompress the compressed data packets. For compressed data packets, the compression algorithm used, dictionary synchronization information (including the latest information of the dictionary, newly added entries), serial number or timestamp are clearly indicated in the header of each compressed data packet.

[0158] like Figure 1-Figure 3 As shown, in the step S3, the data compression packets are classified according to the transport layer protocol, the data packet destination, and the data packet type, and are transmitted to the storage buffer using a scheduling strategy;

[0159] Specifically, in step S3, the data compression packets are classified according to the transport layer protocol, packet destination, and packet type, and transmitted to the storage buffer using the scheduling strategy:

[0160] Packet classification:

[0161] Before entering the storage buffer, the packets are classified according to the following criteria:

[0162] Transport layer protocol: Identify the transport layer protocols TCP and UDP used by the data packet, because different protocols require different processing methods;

[0163] Packet destination: Classify packets into corresponding destination queues according to broadcast packets, unicast packets and multicast packets;

[0164] Data packet type: Classify data packets according to their control packets and data packets;

[0165] Through decision tree algorithms including ID3, C5, and CART, a priority prediction model for data packet processes is established. Information gain, information gain rate, and Gini index indicators are used for feature selection. A decision tree is generated from top to bottom in a recursive manner, and optimized by pre-pruning or post-pruning to improve the generalization ability of the model. The generated decision tree predicts the priority of the data packet based on the feature value of the input data packet. The generated decision tree is as follows:

[0166] Control packet, UDP, unicast packet, process P1, priority 1;

[0167] Control packets, UDP, multicast packets, process P2, priority 2;

[0168] Control packet, UDP, broadcast packet, process P3, priority 3;

[0169] Control packet, TCP, unicast packet, process P4, priority 4;

[0170] Control packets, TCP, multicast packets, process P5, priority 5;

[0171] Control packets, TCP, multicast packets, process P6, priority 6;

[0172] Data packets, UDP, unicast packets, process P7, priority 7;

[0173] Data packets, UDP, multicast packets, process P8, priority 8;

[0174] Data packets, UDP, broadcast packets, process P9, priority 9;

[0175] Data packet, TCP, unicast packet, process P10, priority 10;

[0176] Data packets, TCP, multicast packets, process P11, priority 11;

[0177] Data packets, TCP, multicast packets, process P12, priority 12;

[0178] The process priority is represented by a number, the smaller the number, the higher the process level;

[0179] At the packet scheduling layer, the packets are reordered and rate-limited, and the scheduler checks all processes in the ready state, selects the process with the highest priority to execute, and transmits the compressed packets to the storage buffer at the sender according to the priority scheduling strategy.

[0180] like Figure 1-Figure 3 As shown, in the step S4, the data compression packets are urgently queued for processing, and the receiving end receives the compressed data packets through the token bucket scheduling algorithm, and identifies the data packets. After encoding and compression, the data packets are decoded using the same decompression algorithm;

[0181] Specifically, in step S4, the data compression packets are queued for urgent processing, and the receiving end receives the compressed data packets through the token bucket scheduling algorithm:

[0182] Urgent queuing processing allocates compressed data packets to different queues according to the urgency of the data packets (determined by the type and priority of the data packets), ensuring that the data packets with high urgency can be processed and transmitted faster;

[0183] Token Bucket Scheduling Algorithm The token bucket scheduling algorithm is implemented at the sending end, which controls the sending rate of data packets;

[0184] The token bucket algorithm works by following these steps:

[0185] Tokens are added to the token bucket at a fixed rate. When a data packet arrives,

[0186] If there are enough tokens in the token bucket, the data packet is allowed to be sent;

[0187] If there are not enough tokens, the data packet needs to wait until there are enough tokens in the token bucket;

[0188] Sending compressed data packets: Sending compressed data packets from the storage buffer to the network according to the urgency and scheduling of the token bucket algorithm;

[0189] The receiving end receives the compressed data packet, and the receiving end receives the compressed data packet through a network interface;

[0190] Identify the data packet coding compression. The receiving end parses the header information of the data packet and identifies whether the data packet has been coded and compressed through the compression flag, compression method identifier, and context identifier in the data packet header information.

[0191] Decode the data packet using the same decompression algorithm. According to the compression algorithm indicated in the header information of the data packet, the receiving end uses the same decompression algorithm to decode the data packet;

[0192] The decoding process consists of the following steps:

[0193] Extract the compressed data in the data packet;

[0194] Apply decompression algorithms to restore original data;

[0195] Verify the integrity of the decompressed data.

[0196] like Figure 1-Figure 3 As shown, in the step S5, a feedback mechanism is established to collect congestion control effect data in real time, and the control strategy is adaptively adjusted according to the feedback results;

[0197] Specifically, in step S5, congestion control effect data is collected in real time, and the control strategy is adaptively adjusted according to the feedback results:

[0198] Establishing a feedback channel A feedback channel is established between the sender and the receiver to transmit information related to congestion control.

[0199] Define feedback data to determine the congestion control effect data that needs to be collected, including network delay, packet loss rate, bandwidth utilization, and queue length;

[0200] The sender and receiver use fixed data formats and transmission protocols to facilitate the exchange of this information between the sender and the receiver;

[0201] Collect congestion control effect data in real time:

[0202] At the receiving end, the congestion control effect data defined above is monitored and recorded in real time;

[0203] Periodically or when a congestion event is detected, this data is sent back to the sender through a feedback channel;

[0204] Analysis feedback results:

[0205] At the sending end, feedback data from the receiving end is received and analyzed;

[0206] Observe the fluctuations of network performance parameters (network delay, packet loss rate, bandwidth utilization, queue length), evaluate whether the congestion control strategy can keep the network in a relatively stable state, and identify existing problems or potential improvement points;

[0207] Adaptive adjustment of control strategy:

[0208] Dynamically adjust congestion control parameters based on feedback results, including token generation rate of token bucket, queue management strategy, and compression algorithm threshold;

[0209] Dynamic adjustments include:

[0210] Change the parameters of the compression algorithm to balance the compression ratio and network load;

[0211] By adding coding redundancy to the transmitted data, the strength of the FEC (Forward Error Correction) coding is adjusted to improve error resilience.

[0212] like Figure 1-Figure 3 As shown, the present invention works as follows to solve the shortcomings of the prior art that the network congestion of large-scale high-concurrency data fields cannot be solved, and the DCQCN failure leads to the market demand for reducing the data transmission rate:

[0213] Deploy monitoring equipment at key nodes of the power data field to monitor key indicators such as broadband utilization, packet loss rate, and latency in real time, build a real-time network status perception system, and reflect the current congestion situation;

[0214] The DCQCN congestion control algorithm is used to calculate the link traffic congestion and adjust the current traffic situation. When the traffic congestion exceeds the threshold, the compression algorithm is used at the sending end to compress the data packet for transmission;

[0215] Classify the data compression packets according to the transport layer protocol, packet destination, and packet type, and transmit them to the storage buffer using a scheduling strategy;

[0216] The data compression packets are queued for urgent processing, and the receiving end receives the compressed data packets through the token bucket scheduling algorithm, and recognizes that the data packets have been encoded and compressed, and decodes the data packets using the same decompression algorithm;

[0217] Establish a feedback mechanism to collect congestion control effect data in real time and adaptively adjust the control strategy based on the feedback results.

[0218] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0219] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and rules of the present invention shall be included in the protection scope of the present invention.

Claims

1. A congestion control method for a power data field, characterized in that: S1. Deploy monitoring equipment at key nodes of the power data field to monitor key indicators such as broadband utilization, packet loss rate, and latency in real time, build a real-time network status perception system, and reflect congestion conditions; S2. Use the DCQCN congestion control algorithm to calculate the link traffic congestion and adjust the current traffic situation. When the traffic congestion exceeds the threshold, a compression algorithm is used at the sending end to compress the data packet for transmission. S3, classifying the data compression packets according to the transport layer protocol, the packet destination, and the packet type, and transmitting them to the storage buffer using a scheduling strategy; S4, the data compression packet is urgently queued for processing, and the receiving end receives the compressed data packet through the token bucket scheduling algorithm, and identifies the data packet, and after encoding and compression, uses the same decompression algorithm to decode the data packet; S5. Establish a feedback mechanism to collect congestion control effect data in real time and adaptively adjust the control strategy based on the feedback results.

2. The congestion control method for a power data field according to claim 1, characterized in that: The construction of a real-time network status perception system: Select key nodes and determine that the network topology of the power data field is a ring topology. The key nodes include switches and server access points. Deploy monitoring equipment: Install monitoring equipment on selected key nodes. The equipment includes dedicated network monitoring hardware or monitoring software running on existing equipment. Configure monitoring parameters to set the parameters that the monitoring device needs to monitor in real time, including broadband utilization, packet loss rate and delay; configure the monitoring device to collect data on broadband utilization, packet loss rate and delay every second or every minute; The data collection and monitoring equipment starts to collect network data in real time, including: Use SNMP to collect status information of network devices; use NetFlow or sFlow technology to analyze network traffic and calculate bandwidth utilization; measure packet loss rate and network delay by sending test packets ICMP ping; Reflect the current congestion situation Based on the monitored data, analyze the network congestion situation. If the broadband utilization rate is high for a long time, or the packet loss rate and delay suddenly increase, it indicates that network congestion is occurring. Set the threshold. When the monitored indicator exceeds the threshold, the system automatically alarms.

3. The congestion control method for a power data field according to claim 1, characterized in that: The DCQCN congestion control algorithm is used to calculate the link traffic congestion: Integrated DCQCN congestion control algorithm Integrate the DCQCN algorithm in the switch and monitoring system. It is used as a congestion control algorithm for data center networks and manages network congestion by increasing and decreasing the sending rate. Calculate the link traffic congestion using the network data collected in real time by the monitoring equipment, including bandwidth utilization and packet loss rate, and apply the DCQCN algorithm to calculate the congestion level of each link, including the following steps: S201. Monitor the queue length or delay of each link. If the delay continues to increase while the queue length does not change, it indicates that the bandwidth utilization is high and the queue is too long, resulting in data packet loss: Calculate the current link traffic congestion index CI. The calculation formula is: Among them, CI represents the link traffic congestion index; U represents the bandwidth utilization, ranging from 0 to 1; D represents the packet loss rate, ranging from 0 to 1; λ and γ are weight factors used to adjust the relative importance of bandwidth utilization and data packet loss rate in the calculation, and λ>γ; S202, adjusting the size of the sending window according to the state of the link to control the flow; S203, adjusting the current flow situation: When the DCQCN algorithm detects link congestion, it reduces the size of the sending window and lowers the transmission rate of the sender; The DCQCND speed reduction algorithm is calculated through the following three parallel processes: α update reflects the degree of network congestion. A timer is set at the RP. If the RP does not receive the CNP sent by the NP after K time units, the RP updates the a value. The calculation formula for α update is: new_α=g*old_α+(1-g)*CNP_arrived Wherein, g represents a constant parameter between 0 and 1, and CNP_arrived represents a one-bit field indicating whether a CNP has arrived in the previous time slot; When the DCQCN algorithm detects link congestion, it reduces the size of the sending window, thereby reducing the transmission rate of the sender. The speed reduction calculation formula is: Among them, R T represents the current sending rate, R C represents the old sending rate, and α represents the degree of network congestion; If link congestion eases, the algorithm gradually increases the send window size: The speed-up logic is divided into three sequential stages: fast recovery, aggressive increase, and high aggressive increase; After moving to the next stage to speed up in one stage, before slowing down, the current speed will be saved in the target_rate parameter. In the fast recovery stage, the speed increases logarithmically according to half the distance to the target_rate. The rate adjustment formula is: CR = (CR + TR) / 2; Among them, CR is the current rate and TR is the target rate; The formula for actively increasing the rate adjustment is: TR = TR + Rai; Among them, TR is the target rate, Rai is the step size of rate increase; In the super aggressive rate increase adjustment formula, the rate increase is a step greater than Rai or an adjustment period less than Rai; S204, setting a congestion threshold: setting a congestion threshold in the system, and when the monitored link traffic congestion exceeds the threshold, triggering a compression processing mechanism; Compression processing mechanism: When the link congestion exceeds the set threshold, the sender receives the congestion notification and uses data compression algorithms including LZ77, LZ78, and Deflate to compress the data packets to be sent. S205, transmitting the compressed data packet, the compressed data packet is sent to the network, and a compression flag, a compression method identifier, and a context identifier are set in the data packet header; At the same time, it is ensured that the receiving end can correctly decompress the compressed data packets. For compressed data packets, the compression algorithm used is clearly indicated in the header of each compressed data packet. The dictionary synchronization information contains the latest information of the dictionary, newly added entries, serial numbers or timestamps.

4. The congestion control method for a power data field according to claim 1, characterized in that: The data compression packets are classified according to the transport layer protocol, the data packet destination, and the data packet type, and transmitted to the storage buffer using a scheduling strategy: Packet classification: Before entering the storage buffer, the packets are classified according to the following criteria: Transport layer protocol: Identify the transport layer protocols TCP and UDP used by the data packet. Different protocols use different processing methods; Packet destination: Classify packets into destination queues based on broadcast, unicast, and multicast packets; Data packet type: Classify data packets according to their control packets and data packets; Through decision tree algorithms including ID3, C5, and CART, a priority prediction model for data packet processes is established. Information gain, information gain rate, and Gini index indicators are used for feature selection. A decision tree is generated from top to bottom in a recursive manner, and optimized by pre-pruning or post-pruning. The generated decision tree predicts the priority of the data packet based on the feature value of the input data packet. The generated decision tree is as follows: Control packet, UDP, unicast packet, process P1, priority 1; Control packets, UDP, multicast packets, process P2, priority 2; Control packet, UDP, broadcast packet, process P3, priority 3; Control packet, TCP, unicast packet, process P4, priority 4; Control packets, TCP, multicast packets, process P5, priority 5; Control packets, TCP, multicast packets, process P6, priority 6; Data packets, UDP, unicast packets, process P7, priority 7; Data packets, UDP, multicast packets, process P8, priority 8; Data packets, UDP, broadcast packets, process P9, priority 9; Data packet, TCP, unicast packet, process P10, priority 10; Data packets, TCP, multicast packets, process P11, priority 11; Data packets, TCP, multicast packets, process P12, priority 12; The process priority is represented by numbers, with the number 1 representing the highest process level; At the packet scheduling layer, the packets are reordered and rate-limited, and the scheduler checks all processes in the ready state, selects the process with the highest priority to execute, and transmits the compressed packets to the storage buffer at the sender according to the priority scheduling strategy.

5. The congestion control method for a power data field according to claim 1, characterized in that: The data compression packets are queued for urgent processing, and the receiving end receives the compressed data packets through the token bucket scheduling algorithm: Urgent queuing processing allocates compressed data packets to different queues according to the urgency of the data packets, which is determined by the type and priority of the data packets, ensuring that the data packets with high urgency can be processed and transmitted quickly; Token Bucket Scheduling Algorithm The token bucket scheduling algorithm is implemented at the sending end, which controls the sending rate of data packets; The token bucket algorithm works by following these steps: Tokens are added to the token bucket at a fixed rate. When a data packet arrives, If there are enough tokens in the token bucket, the data packet is allowed to be sent; If there are not enough tokens, the data packet needs to wait until there are enough tokens in the token bucket; Sending compressed data packets: Sending compressed data packets from the storage buffer to the network according to the urgency and scheduling of the token bucket algorithm; The receiving end receives the compressed data packet, and the receiving end receives the compressed data packet through a network interface; Identify the data packet encoding compression. The receiving end parses the header information of the data packet to identify whether the data packet has been encoded and compressed; Decode the data packet using the same decompression algorithm. According to the compression algorithm indicated in the header information of the data packet, the receiving end uses the same decompression algorithm to decode the data packet; The decoding process consists of the following steps: Extract the compressed data in the data packet; Apply decompression algorithms to restore original data; Verify the integrity of the decompressed data.

6. The congestion control method for a power data field according to claim 1, characterized in that: The real-time collection of congestion control effect data and the adaptive adjustment of the control strategy based on the feedback results: Establishing a feedback channel A feedback channel is established between the sender and the receiver to transmit information related to congestion control. Define feedback data to determine the congestion control effect data that needs to be collected, including network delay, packet loss rate, bandwidth utilization, and queue length; The sender and receiver use fixed data formats and transmission protocols to facilitate the exchange of data information between the sender and the receiver; Collect congestion control effect data in real time: At the receiving end, congestion control effect data is monitored and recorded in real time; Periodically or when a congestion event is detected, data is sent back to the sender through a feedback channel; Analysis feedback results: At the sending end, feedback data from the receiving end is received and analyzed; Observe network performance parameters, including fluctuations in network latency, packet loss rate, bandwidth utilization, and queue length, identify network stability as reflected by congestion control policies, and identify existing problems or potential areas for improvement; Adaptive adjustment of control strategy: Dynamically adjust congestion control parameters based on feedback results, including token generation rate of token bucket, queue management strategy, and compression algorithm threshold; Dynamic adjustments include: Change the parameters of the compression algorithm to balance the compression ratio and network load; By increasing coding redundancy in the transmitted data and adjusting the strength of the FEC (forward error correction) coding, the error recovery capability is improved.

Images