A multi-priority differential fault prediction and active prevention method for a vehicle-mounted fibre channel network

Abstract

The present application relates to a kind of multi-priority differential fault prediction and active prevention method for vehicle-mounted fiber channel network, belong to vehicle-mounted network communication technical field.The present application method includes: receiving the state report frame that each network node periodically reports, obtain the state information of each network node connected link from the state report frame, based on each state information, the network overall health index value corresponding to it, the bayesian probability anomaly value and the bayesian probability anomaly trend value of each network node are calculated;Based on each state information, network overall health index value, bayesian probability anomaly value and bayesian probability anomaly trend value, network fault risk prediction is carried out;Based on the predicted fault risk, the network optimization instruction frame carrying network optimization instruction is issued to optimize network.The present application method realizes the advance accurate prediction and active prevention of vehicle-mounted fiber channel network fault, effectively improves the reliability and security of network operation, improves the safety of vehicle operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle-mounted network communication, and in particular to a multi-priority differential fault prediction and active prevention method for a vehicle-mounted fiber channel network. BACKGROUND

[0002] With the rapid development of automobile intelligence and automatic driving technology, vehicle-mounted network communication is gradually developing towards high speed, low delay and high reliability. In recent years, fiber channel (FC for short) has gradually been applied in automotive electronic systems such as advanced driver assistance systems (ADAS), automatic driving systems and vehicle control systems due to its characteristics of high speed, low delay and high reliability.

[0003] However, the traditional fiber channel technology is originally designed for data center environment and does not fully consider the special requirements of vehicle-mounted application scenarios. The vehicle-mounted network environment is complex and dynamic, and any link or node failure, especially the abnormality of critical business links with high real-time and reliability, will seriously affect the safety of the vehicle. The existing FC network fault diagnosis technology mostly adopts a post-passive detection method, which cannot predict the fault risk in advance and lacks real-time performance. Moreover, it lacks precise analysis and active intervention mechanism for link abnormal trend, making it difficult to effectively guarantee the network safety and stability. SUMMARY

[0004] In view of the above analysis, the present application aims to disclose a multi-priority differential fault prediction and active prevention method for a vehicle-mounted fiber channel network, which can predict potential fault risks in advance and take active prevention measures, effectively improving the safety and stability of vehicle-mounted network operation and the safety of vehicle driving.

[0005] The method of the present application specifically comprises the following steps:

[0006] Receiving the state report frames periodically reported by each network node, obtaining the state information of the links connected by each network node from the state report frames, and calculating the corresponding network overall health index value, Bayesian probability anomaly value and Bayesian probability anomaly trend value of each network node based on each state information;

[0007] Performing network fault risk prediction based on each state information, network overall health index value, Bayesian probability anomaly value and Bayesian probability anomaly trend value;

[0008] Based on the predicted fault risk, a network optimization instruction frame carrying network optimization instructions is issued to optimize the network.

[0009] Further, the state information includes link frame loss rate, frame loss rate difference trend value, link round trip delay, high priority link round trip delay weighted moving average, high priority link round trip delay trend difference value and bit error rate;

[0010] The network node calculates the state information of the connected link based on the sent fault diagnosis frame and the received corresponding response frame, including:

[0011] The network node calculates the link frame loss rate of the connected link based on the total number of frames of the fault diagnosis frame sent in the current period and the total number of frames of the corresponding response frame received;

[0012] The network node calculates the frame loss rate difference trend value of the connected link based on the link frame loss rate of the current period and the last period;

[0013] The network node calculates the link round trip delay of the connected link based on the time of sending the fault diagnosis frame in the current period and the time of receiving the corresponding response frame;

[0014] The network node calculates the high priority link round trip delay weighted moving average of the connected high priority service link in the current period based on the link round trip delay of the high priority link in the current period and the high priority link round trip delay weighted moving average of the last period;

[0015] The network node calculates the high priority link round trip delay trend difference value of the connected link based on the high priority link round trip delay weighted moving average of the current period and the last period;

[0016] The network node determines the bit error rate of the connected link based on the cumulative bit error rate of the connected link per unit time or per unit bit transmission volume.

[0017] Further, the network node calculates the corresponding network overall health index value, the Bayesian probability anomaly value and the Bayesian probability anomaly trend value of each network node based on each state information, including:

[0018] The network node determines the state of each link based on each state information in the current period, and calculates the network overall health index value based on the state of each link and the priority weight of each link; the state of the link includes normal and abnormal;

[0019] The network node calculates the Bayesian probability anomaly value and the Bayesian probability anomaly trend value of each network node based on each state information in the current period and the corresponding historical data.

[0020] Further, the center management node determines the state of each link based on each state information in the current period, including:

[0021] When the link meets at least one of the following conditions, the link is determined to be abnormal, otherwise the link is normal, and the conditions include:

[0022] a link frame loss rate is greater than or equal to a frame loss rate threshold value;

[0023] a link round-trip delay exceeds a delay threshold value for N consecutive periods; N is a positive integer;

[0024] a fault diagnosis frame sent through the link fails to receive a response frame for two consecutive times within a limited time;

[0025] a bit error rate exceeds a preset multiple of a current link reference within a period in which two fault diagnosis frames are sent through the link and corresponding response frames are received.

[0026] Further, the network fault risk prediction based on the state information, the network overall health index value, the Bayesian probability anomaly value, and the Bayesian probability anomaly trend value comprises:

[0027] when the link frame loss rate is greater than or equal to the frame loss rate threshold value or a cumulative value of the frame loss rate difference trend value for a first set number of consecutive periods is greater than a frame loss rate difference trend threshold value, it is predicted that there is a single-point link intermittent fault risk;

[0028] when the link round-trip delay exceeds a link round-trip difference threshold value for a second set number of consecutive periods, it is predicted that there is a link transmission performance abnormality risk;

[0029] when the high-priority link round-trip weighted moving average is greater than or equal to a weighted moving average threshold value or a cumulative value of the high-priority link round-trip delay trend difference value for a third set number of consecutive periods is greater than a high-priority trend cumulative threshold value, it is predicted that there is a high-priority link performance decline risk;

[0030] when the network overall health index value is greater than or equal to a set health threshold value or a growth trend thereof is greater than an index growth rate threshold value, it is predicted that there is a network congestion risk;

[0031] when the Bayesian probability anomaly value of the network node is greater than or equal to an anomaly probability set threshold value or a cumulative value of the Bayesian probability anomaly trend value for a fourth set number of consecutive periods is greater than a set anomaly trend threshold value, it is predicted that there is an implicit or intermittent node fault risk.

[0032] Further, the network optimization instruction frame carrying a network optimization instruction issued based on the predicted fault risk comprises:

[0033] when it is predicted that there is a single-point link intermittent fault risk, the central management node issues a network optimization instruction frame carrying a link isolation or link activation instruction;

[0034] when it is predicted that there is a link transmission performance abnormality risk, the central management node issues a network optimization instruction frame carrying a link activation or topology reconstruction instruction;

[0035] When it is predicted that there is a high-priority link performance decline risk, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions;

[0036] When it is predicted that there is a network congestion risk, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions;

[0037] When it is predicted that there is a hidden or intermittent node failure risk, the central management node issues a network optimization instruction frame carrying link isolation or topology reconfiguration instructions.

[0038] Further, the central management node calculates the network overall health index value using the following calculation formula:

[0039]

[0040] wherein TSI(k) represents the network overall health index value of the kth period; N represents the number of network nodes; i and j represent the ith and jth network nodes, respectively; l ij (k) represents the link state of the link (i, j); p ij represents the priority weight of the link (i, j).

[0041] Further, the central management node calculates the Bayesian probability anomaly value of each network node using the following calculation formula:

[0042]

[0043] wherein X represents a random variable of a network node or a link, and its value is a feature vector composed of link frame loss rate, link round-trip delay, high-priority link round-trip weighted moving average, network overall health index value, bit error rate, node CPU load condition, and node temperature;

[0044] P(fault) is the prior probability, representing the estimated probability of the corresponding network node or link being in a fault state before the current data X is observed, and is valued based on historical statistics, empirical data, or failure rate estimation in system design;

[0045] P(X|fault) represents the likelihood of observing the current data X when the network node or link is actually in a fault state;

[0046] P(normal) represents the probability of estimating that the network node or link remains in normal operation without observing the current data X;

[0047] P(X|normal) represents the probability of observing X when the network node or link is normal;

[0048] The Bayesian probability abnormal trend value is calculated by using the following calculation formula:

[0049] Delta P (fault|X) k = P (fault|X) k -P (fault|X) k-1 .

[0050] Further, the calculation method of the link round-trip difference threshold is:

[0051] RTT threshold (k) = mu RTT (k) + alpha * sigma RTT (k) ;

[0052] Wherein, RTT threshold (k) is the link round-trip difference threshold of the kth period; W is the length of the history window, indicating the number of periods before the kth period; RTT(i) is the link round-trip delay of the ith period; alpha represents the difference sensitivity factor; sigma RTT (k) is the historical standard deviation,

[0053] Further, the calculation method of the high-priority link round-trip weighted moving average is:

[0054] RTT EWMA (k) = beta * RTT(k) + (1-beta) * RTT EWMA (k-1) ;

[0055] Wherein, RTT EWMA (k) is the high-priority link round-trip weighted moving average of the kth period; RTT(k) is the link round-trip delay of the kth period; beta represents the weight factor;

[0056] The calculation method of the high-priority link round-trip delay trend difference value is:

[0057] Delta RTT EWMA (k) = RTT EWMA (k) - RTT EWMA (k-1).

[0058] The present application can at least achieve one of the following beneficial effects:

[0059] By setting and monitoring multiple difference indicators to predict network failure risk, early signs of failure can be more accurately identified, and compared with traditional failure detection methods, potential problems can be found earlier, reducing the impact of network communication failure on vehicle operation.

[0060] By predicting the network failure risk, taking proactive preventive measures, issuing network optimization instruction frames carrying network optimization instructions to the vehicle-mounted network based on the predicted failure risk, the proactive preventive measures can reasonably allocate network resources according to the priority, and preferentially guarantee the transmission of critical data, thereby improving the operation efficiency and reliability of the entire vehicle-mounted network.

[0061] By taking proactive preventive measures, the frequency and severity of network failures can be reduced, and the overall reliability of the vehicle-mounted fiber channel network is improved, thereby avoiding vehicle operation failures caused by network communication failures and reducing the operation and maintenance cost of the vehicle.

[0062] By constructing special fault diagnosis frames, state report frames and network optimization instruction frames, the structure is compatible with the existing FC network standard frame structure, easy to implement, and suitable for wide industrial application.

[0063] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent from the description, or can be learned by practice of the present application. The objects and other advantages of the present application will be achieved and obtained by means of the instrumentalities and combinations pointed out in the following description, the claims, and the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0064] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the principles of the application.

[0065] Figure 1 The flowchart of the method of the present application. DETAILED DESCRIPTION

[0066] The preferred embodiments of the present application will be described in detail below with reference to the drawings, wherein the drawings constitute a part of this application and are used to explain the principles of the embodiments of the present application, but are not considered as limiting the scope of the present application.

[0067] One embodiment of the present application discloses a multi-priority differential fault prediction and proactive prevention method for a vehicle-mounted fiber channel network, specifically comprising steps S01-S03.

[0068] Step S01, the center management node receives the state report frame periodically reported by each network node, obtains the state information of the link connected by each network node from the state report frame, and calculates the corresponding network overall health index value, the Bayesian probability abnormal value and the Bayesian probability abnormal trend value of each network node based on each state information. It should be noted that the calculation period of the center management node based on each state information is consistent with the reporting period of the state report frame. Unless otherwise specified, the period described in the present application is the calculation period of the center management node (i.e. the reporting period of each node).

[0069] Specifically, the state information of the link connected by each network node includes the link frame loss rate, the frame loss rate difference trend value, the link round-trip delay, the high-priority link round-trip weighted moving average, the high-priority link round-trip delay trend difference value and the bit error rate.

[0070] Further, the network node calculates the state information of the connected link based on the sent fault diagnosis frame (FDF) and the received corresponding response frame. Wherein, the fault diagnosis frame is defined based on the standard FC frame format, including: defining the frame header TYPE field as a special type identifier based on the standard FC frame format, and marking it as a fault diagnosis frame type; using part of the bits of the Parameter field in the standard FC frame to define the priority identifier; adding a high-precision timestamp field before Payload. The fault diagnosis frame format is shown in Table 1:

[0071] Table 1, fault diagnosis frame format

[0072] Field name Length (bytes) Description (compared with standard FC frame format) SOF 4 Standard R_CTL 1 Standard DID 3 Standard SID 3 Standard TYPE 1 Define special type (0xFE) to identify FDF special frame F_CTL 3 Standard SEQ_ID 1 Standard DF_CTL 1 Standard SEQ_CNT 2 Standard, can be used as sequence number directly OX_ID 2 Standard RX_ID 2 Standard Parameter 4 Standard, can be used for frame extension parameter Priority 1 New field, use part of Parameter field space to realize Timestamp 8 Payload pre-field (new) Payload Optional Extension data field, store link information CRC 4 Standard, unchanged EOF 4 Standard, unchanged

[0073] Further, the FDF frame adopts an active interaction mode. The network node actively sends the FDF frame to the adjacent node, and can send one or more FDF frames to other nodes in one reporting period, that is, one reporting period includes one or more FDF dynamic sending periods; the node receiving the FDF frame must immediately return a FDF response frame (ACK frame). The ACK frame has the same frame structure as the original FDF frame, but the ACK frame is marked as a response frame type (0XEF) in the TYPE field, and carries the corresponding receiving and sending time stamps (merged and compressed encoding in the Payload front field). The period of active interaction is differentially set according to different business priorities, and the high-priority link uses a shorter sending period to improve the real-time diagnosis of critical business links.

[0074] Further, the network node calculates the link frame loss rate of the connected link based on the total number of fault diagnosis frames FDF sent in the current period and the total number of corresponding response frames ACK received;

[0075] (k) is calculated based on the link frame loss rate of the current period and the last period;

[0076] The link round-trip time of the connected link is calculated based on the time of sending the fault diagnosis frame in the current period and the time of receiving the corresponding response frame; if multiple fault diagnosis frames are sent in the current period, the link round-trip time is the average of the round-trip times of all fault diagnosis frames and the corresponding response frames;

[0077] The high-priority link round-trip weighted moving average of the connected link in the current period is calculated based on the link round-trip time of the high-priority service link in the current period and the high-priority link round-trip weighted moving average of the last period;

[0078] The high-priority link round-trip delay trend difference value of the connected link is calculated based on the high-priority link round-trip weighted moving average of the current period and the last period.

[0079] The bit error rate of the connected link is determined based on the accumulated bit error rate of the connected link per unit time or per unit bit transmission volume.

[0080] Further, the calculation formula of the link frame loss rate is:

[0081] FLR(k) = 1 - [N recv (k) / N send (k)].

[0082] Wherein, FLR(k) represents the link frame loss rate of the kth monitoring period; N send (k) is the total number of FDF frames sent in the period; N recv (k) is the total number of ACK frames received in the period.

[0083] Further, the calculation formula of the frame loss rate difference trend value ΔFLR(k) is:

[0084] ΔFLR(k) = FLR(k) - FLR(k-1).

[0085] Further, the calculation formula of the high-priority link round-trip weighted moving average is:

[0086] RTT EWMA (k) = β · RTT(k) + (1-β) · RTT EWMA (k-1).

[0087] Wherein, RTT EWMA (k) is the high-priority link round-trip weighted moving average of the kth monitoring period; β is the weight factor, and the preferred value is 0.6.

[0088] Further, the high-priority link round-trip delay trend difference value ΔRTTEWMA The calculation formula of (k) is:

[0089] ΔRTT EWMA (k) = RTT EWMA (k) = RTT EWMA (k-1).

[0090] Further, the error rate is based on the link physical layer or MAC layer built-in error code counter statistics.

[0091] It should be noted that in the vehicle-mounted fibre channel network, the state report frame periodically reported by each network node is received by the center management node. The frame format of the state report frame of the application is shown in Table 2:

[0092] Table 2, state report frame format

[0093]

[0094] Further, the center management node determines the link state based on the state information of the current period, and calculates the network overall health index value based on the link state and the link priority weight; the link state includes normal and abnormal;

[0095] The Bayesian probability anomaly value and the Bayesian probability anomaly trend value of each network node are calculated based on the state information of the current period and the corresponding historical data.

[0096] Further, the center management node determines the link state based on the state information of the current period, and calculates the network overall health index value based on the link state and the link priority weight; the link state includes normal and abnormal;

[0097] When the link meets at least one of the following conditions, it is determined that the link is abnormal (represented as 1), otherwise the link is normal (represented as 0), and the conditions include:

[0098] The link frame loss rate FLR(k) is greater than or equal to the frame loss rate threshold FLR threshold , for example, 5%;

[0099] The link round-trip time exceeds the time delay threshold for N consecutive periods; N is a positive integer, preferably 3, and the time delay threshold is preferably three times the standard deviation of the historical time delay average; it should be noted that in one reporting period, as long as the link round-trip time exceeds the time delay threshold once, the link round-trip time exceeding the time delay threshold in the period is recorded;

[0100] The fault diagnosis frame sent through the link does not receive a response frame within a limited time for two consecutive times; specifically, when the fault diagnosis frame sent through the link does not receive a response frame within a limited time, it is characterized by ACK_Missed_Counts and saved in the link state field of the state report frame;

[0101] In the period of sending two fault diagnosis frames through the link and receiving the corresponding response frames, the error rate exceeds the preset multiple of the current link reference, and the preset multiple can be 100% for example.

[0102] Further, the center management node calculates the network overall health index value using the following calculation formula:

[0103]

[0104] Wherein, TSI(k) represents the network overall health index value of the kth period; N represents the number of network nodes; i and j represent the ith and jth network nodes respectively; l ij (k) represents the link state of the link (i, j); p ij represents the priority weight of the link (i, j).

[0105] Further, the center management node calculates the Bayesian probability anomaly value of each network node using the following calculation formula in each period:

[0106]

[0107] Wherein, X is a feature vector composed of link frame loss rate, link round-trip delay, high-priority link round-trip weighted moving average, network overall health index value, error rate, node CPU load condition, and node temperature;

[0108] P(fault) is the prior probability, which represents the estimated probability of the corresponding network node or link being in a fault state before the current data X is observed, and is valued based on historical statistics, empirical data or failure rate estimation in system design;

[0109] It should be noted that when calculating the probability of the observed current data X, it means that the probability distribution model of each index value in X is calculated jointly, that is, the joint probability of observing the feature combination under the fault state is calculated; Even if the conditional probability of each feature can be small, but through Gaussian distribution, the joint probability of them can effectively capture the overall fault risk; For example, assume that a network node connected link has a small performance problem, but can still transmit and receive data, in this case, the link round-trip delay and link frame loss rate can still be in the normal range, but if combined with other multiple feature values (such as network overall health index value, etc.), the combination of the entire feature vector X can still effectively capture the performance deterioration trend, thus giving a higher "fault probability" evaluation in Bayesian inference.

[0110] P(X|fault) represents the possibility of observing the current data X when the network node or link is actually in a fault state; the value is determined based on online learning or adaptive updating of the central management node;

[0111] In the above formula, P(fault) and P(X|fault) are multiplied to obtain the weighted possibility of observing X in a fault state, that is, assuming that the current node is actually in a fault state, and how much the prior belief in the fault state itself is, how great the comprehensive possibility of observing X is;

[0112] P(normal) represents the probability of estimating that the network node or link remains normal when the current data X is not observed;

[0113] P(X|normal) represents the probability of observing X when the network node or link is normal; for example, if the value of P(X|normal) is large, it means that "even if it looks like a sign of failure, it may often appear in a normal state", so the final fault probability will be reduced; the value is determined based on online learning or adaptive updating of the central management node;

[0114] The meaning of the denominator in the above formula is all possible sources of observing X, that is, to ensure that the final result is a normalized probability value in the interval [0, 1].

[0115] Further, the Bayesian probability abnormal trend value of the network node in the kth period is calculated using the following calculation formula:

[0116] ΔP(fault|X) k =P(fault|X) k -P(fault|X) k-1 ;

[0117] It should be noted that the trend monitoring of ΔP(fault|X) k indicates that if the posterior probability continuously increases and approaches the threshold value, it means that the node fault risk is likely to break out in the short term.

[0118] Step S02, the central management node performs network fault risk prediction based on each of the state information, the overall network health index value, the Bayesian probability abnormal value and the Bayesian probability abnormal trend value.

[0119] Specifically, it includes:

[0120] When the link frame loss rate FLR(k) is greater than or equal to the frame loss rate threshold FLR threshold (represented as FLR(k)≥FLR threshold ) or the frame loss rate difference trend value ΔFLR thresholdWithin a consecutive first set number of periods, the cumulative value is greater than the cumulative threshold ΔFLR of the frame drop rate differential trend. threshold Time (represented as) M is the first set number, which is generally in the range of 3-5), predicting the risk of intermittent single-point link failure;

[0121] When the link round-trip time (RTT(k)) exceeds the link round-trip differential threshold (RTT) for a consecutive second set number of periods (preferably 3), threshold When (k) is reached, it is predicted that there is a risk of abnormal link transmission performance; among which, RTT threshold The calculation method for (k) is: RTT threshold (k)=μ RTT (k)+α·σ RTT (k), μ RTT (k) is the historical statistical mean. σ RTT (k) represents the historical standard deviation. W represents the historical window length, which refers to the number of periods before the k-th period, typically 20 periods. α represents the differential sensitivity factor, with a preferred value of 3.

[0122] When the round-trip weighted moving average (RTT) of high-priority links EWMA (k) is greater than or equal to the weighted moving average threshold RTT EWMA,threshold (represented as RTT) EWMA (k)≥RTT EWMA,threshold ), or the round-trip delay trend difference value ΔRTT for high-priority links. EWMA (k) The cumulative value within the third consecutive set period is greater than the high-priority trend cumulative threshold ΔRTT EWMA,threshold (represented as) When Q is the third set number, it is predicted that there is a risk of performance degradation of high-priority links;

[0123] When the overall network health index value TSI(k) is greater than or equal to the set health threshold TSI threshold (represented as TSI(k)≥TSI) threshold If the growth trend of a network is greater than the exponential growth rate threshold γ, a network congestion risk is predicted; the formula for calculating the growth trend is... The preferred value for γ is 0.1;

[0124] When the Bayesian probability of a network node is an outlier P(fault|X) k Greater than or equal to the abnormal probability set threshold P threshold (represented as P(fault|X)) k ≥P threshold ( ), or the cumulative value of the Bayesian probability anomaly trend value within the fourth consecutive set number of periods is greater than the set anomaly trend threshold ΔP.threshold (represented as) Where R is the fourth set number), it predicts the risk of hidden or intermittent node failures.

[0125] Step S03: Based on the predicted fault risk, issue a network optimization instruction frame carrying network optimization instructions to optimize the network.

[0126] Specifically, as soon as the central management node predicts any of the fault risks mentioned in step S02, it immediately triggers proactive optimization measures, including:

[0127] When a single point of failure is predicted to be intermittent, the central management node issues a network optimization instruction frame carrying link isolation or link activation instructions.

[0128] When a risk of abnormal link transmission performance is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions.

[0129] When a risk of performance degradation of a high-priority link is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions.

[0130] When a network congestion risk is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions.

[0131] When a hidden or intermittent node failure risk is predicted, the central management node issues a network optimization instruction frame carrying link isolation or topology reconfiguration instructions.

[0132] Furthermore, this invention defines a network optimization instruction frame sent by the central management node, which adopts an FC-2 layer frame structure. The specific frame format definition is shown in Table 3:

[0133] Table 3. Network optimization instruction frame format

[0134]

[0135] For example, the detailed correspondence between the predicted fault risks and the issued network optimization command frames is shown in Table 4:

[0136] Table 4. Correspondence between predicted fault risks and issued network optimization command frames

[0137]

[0138]

[0139] For example, the central management node sets extended parameters in the network optimization instruction frame to generate specific instructions. Each network node extracts specific instructions from the network optimization instruction frame to optimize the network. Table 5 shows examples of sub-instructions for optimization:

[0140] Table 5. Examples of Sub-Instructions

[0141]

[0142]

[0143] This embodiment discloses a multi-priority differential fault prediction and proactive prevention method for vehicular fiber channel networks. By setting and monitoring multiple differential indicators to predict network fault risks, it can more accurately identify early signs of faults. Compared with traditional fault detection methods, it can discover potential problems earlier and reduce the impact of network communication failures on vehicle operation.

[0144] By predicting network failure risks and taking proactive preventative measures, network optimization command frames carrying network optimization instructions are issued based on the predicted failure risks to optimize the vehicular network. Based on failure prediction, proactive preventative measures can rationally allocate network resources according to priority, prioritizing the transmission of critical data, thereby improving the overall operational efficiency and reliability of the vehicular network.

[0145] By taking proactive preventative measures, the frequency and severity of network failures can be reduced, improving the overall reliability of the vehicle fiber optic channel network, thereby avoiding vehicle malfunctions caused by network communication failures and reducing vehicle operation and maintenance costs.

[0146] By constructing dedicated fault diagnosis frames, status report frames, and network optimization instruction frames, the structure is highly compatible with the existing FC network standard frame structure, easy to implement, and suitable for a wide range of industrial applications.

[0147] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for multi-priority differential fault prediction and proactive prevention in vehicular fiber channel networks, characterized in that, Includes the following steps: The central management node receives status report frames periodically reported by each network node, obtains the status information of the links connected to each network node from the status report frames, and calculates the corresponding overall network health index value, Bayesian probability outlier value and Bayesian probability outlier trend value of each network node based on the status information. The status information includes link frame loss rate, frame loss rate differential trend value, link round-trip time, weighted moving average of high-priority link round-trip time, trend differential value of high-priority link round-trip time, and bit error rate. The central management node predicts network failure risks based on the aforementioned status information, the overall network health index value, Bayesian probability outliers, and Bayesian probability outlier trend values. Based on the predicted failure risk, a network optimization instruction frame carrying network optimization instructions is issued to optimize the network.

2. The multi-priority differential fault prediction and active prevention method according to claim 1, characterized in that, Network nodes calculate the status information of the connected links based on the sent fault diagnosis frames and the received corresponding response frames, including: The network node calculates the link frame loss rate of the connected link based on the total number of fault diagnosis frames sent in the current period and the total number of corresponding response frames received. The frame loss rate difference trend value of the connected links is calculated based on the link frame loss rate of the current period and the previous period. The round-trip delay of the connected link is calculated based on the time it takes to send the fault diagnosis frame and receive the corresponding response frame in the current cycle. The weighted moving average of the high-priority links in the current period is calculated based on the link round-trip delay in the current period and the weighted moving average of the high-priority links in the previous period. The round-trip delay trend difference value of the high-priority link is calculated based on the round-trip weighted moving average of the high-priority link in the current period and the previous period. The bit error rate of the connected link is determined based on the cumulative bit error rate per unit time or unit bit transmission.

3. The multi-priority differential fault prediction and active prevention method according to claim 2, characterized in that, The overall network health index value, the Bayesian probability outlier value of each network node, and the Bayesian probability outlier trend value calculated based on each state information include: The status of each link is determined based on the status information of the current period, and the overall network health index value is calculated based on the status of each link and the priority weight of each link; the link status includes normal and abnormal. Based on the current state information and corresponding historical data, the Bayesian probability outlier value and Bayesian probability outlier trend value of each network node are calculated.

4. The multi-priority differential fault prediction and active prevention method according to claim 3, characterized in that, The central management node determines the status of each link based on the status information of the current period, including: A link is considered abnormal if it meets at least one of the following conditions; otherwise, the link is considered normal. These conditions include: The link frame loss rate is greater than or equal to the frame loss rate threshold. The link round-trip latency exceeds the latency threshold for N consecutive cycles; N is a positive integer. The fault diagnosis frames sent through this link did not receive a response frame twice within the specified time. During the period when two fault diagnosis frames are sent and corresponding response frames are received through this link, the bit error rate exceeds the preset multiple of the current link benchmark.

5. The multi-priority differential fault prediction and active prevention method according to claim 3, characterized in that, Network failure risk prediction based on the aforementioned state information, overall network health index value, Bayesian probability outliers, and Bayesian probability anomaly trend value includes: When the link frame loss rate is greater than or equal to the frame loss rate threshold, or when the cumulative value of the frame loss rate differential trend value within a consecutive first set number of periods is greater than the frame loss rate differential trend threshold, it is predicted that there is a risk of intermittent single-point link failure. When the link round-trip delay exceeds the link round-trip differential threshold within a second consecutive set number of periods, it is predicted that there is a risk of abnormal link transmission performance. When the weighted moving average of round-trip delays of high-priority links is greater than or equal to the weighted moving average threshold, or when the cumulative value of the round-trip delay trend difference of high-priority links within the third consecutive set number of periods is greater than the cumulative threshold of high-priority trend, it is predicted that there is a risk of performance degradation of high-priority links. When the overall network health index value is greater than or equal to the set health threshold, or its growth trend is greater than the exponential growth rate threshold, network congestion risk is predicted. When the outlier value of the Bayesian probability of a network node is greater than or equal to the set threshold for outlier probability, or when the cumulative value of the outlier trend value of the Bayesian probability is greater than the set threshold for outlier trend in the fourth consecutive set number of periods, a hidden or intermittent node failure risk is predicted.

6. The multi-priority differential fault prediction and active prevention method according to claim 5, characterized in that, The network optimization instruction frame, which carries network optimization instructions and is issued based on predicted fault risk, includes: When a single point of failure is predicted to be intermittent, the central management node issues a network optimization instruction frame carrying link isolation or link activation instructions. When a risk of abnormal link transmission performance is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions. When a risk of performance degradation of a high-priority link is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions. When a network congestion risk is predicted, the central management node issues a network optimization instruction frame carrying link activation or topology reconfiguration instructions. When a hidden or intermittent node failure risk is predicted, the central management node issues a network optimization instruction frame carrying link isolation or topology reconfiguration instructions.

7. The multi-priority differential fault prediction and active prevention method according to claim 3, characterized in that, The central management node calculates the overall network health index using the following formula: ； in, Indicates the first The overall network health index value for each period; N represents the number of network nodes; , They represent the first one. , Network nodes; Indicates link The link status; Indicates link Priority weights.

8. The multi-priority differential fault prediction and active prevention method according to claim 3, characterized in that, The central management node calculates the Bayesian probability outliers of each network node using the following formula: ； in, To represent a random variable of a network node or link, its value is a feature vector consisting of link frame loss rate, link round-trip time, weighted moving average of round-trip time for high-priority links, overall network health index, bit error rate, node CPU load, and node temperature. The prior probability represents the probability without observing the current data. Previously, the central management node assigned values ​​to the estimated probability of a corresponding network node or link being in a faulty state based on historical statistics, empirical data, or failure rate estimates in the system design. This indicates that the current data was observed when a network node or link was actually in a faulty state. The possibility; This indicates that the current data was not observed. At that time, estimate the probability that the network node or link remains in normal operation; This indicates that the network node or link is functioning normally. The probability of; The Bayesian probability outlier value is calculated using the following formula: ； in, For network node number One cycle.

9. The multi-priority differential fault prediction and active prevention method according to claim 5, characterized in that, The method for calculating the link round-trip differential threshold is as follows: ； in, For the first Periodic link round-trip differential threshold; ; The historical window length refers to the number of periods before the k-th period; For the first Periodic link round-trip latency; Indicates the differential sensitivity factor; For historical standard deviation, .

10. The multi-priority differential fault prediction and active prevention method according to claim 2, characterized in that, The method for calculating the round-trip weighted moving average of the high-priority link is as follows: ； in, For the first Periodic high-priority link round-trip weighted moving average; For the first Periodic link round-trip delay; β represents the weighting factor; The method for calculating the round-trip delay trend difference value of the high-priority link is as follows: 。

Images