A universal circuit part redundancy reinforcement method and system based on HITS

Abstract

The application discloses a general circuit part redundancy reinforcing method and system based on HITS, and the method comprises the following steps: generating a node relation matrix according to the connection relation between nodes; performing HITS iteration on the node relation matrix to obtain the HITS value of the node; sorting the nodes in descending order based on the HITS value, selecting the first number of nodes as the nodes to be reinforced; the first number is determined by the finally determined reinforcing proportion; and performing partial three-mode redundancy reinforcement on the nodes to be reinforced. Through the technical scheme, the additional resource consumption can be effectively reduced under the high reliability requirement.

Description

Technical Field

[0001] This invention relates to the field of redundancy hardening technology for general-purpose circuits, and specifically to a method and system for partial redundancy hardening of general-purpose circuits based on HITS. Background Technology

[0002] Single Event Upset (SEU) is one of the most common and typical single-event effects caused by the harsh radiation environment of space. This phenomenon causes the logic state of a device to flip, and it occurs because high-energy particles in space enter the sensitive areas of the device. Therefore, to reduce the impact of SEU, radiation hardening of on-orbit processors is usually required to improve the stability of aerospace electronic equipment. In recent years, the demand for high reliability in modern electronic systems has been increasing, and the increased integration and complexity have made VLSI circuits more sensitive to errors. To improve system reliability, effective fault-tolerant techniques are needed in the design.

[0003] In radiation hardening, triple-mode redundancy hardening is a common technique. Existing technologies incorporate full triple-mode redundancy hardening into ASIC designs to improve the system's soft error tolerance. While full triple-mode redundancy hardening effectively reduces the system's soft error rate, it leads to excessively large system area and high power consumption. Therefore, with the increasing scale and speed of aerospace processors, full triple-mode redundancy designs are increasingly unable to meet circuit design requirements. To further address these issues, researchers have proposed a new method for reducing precision redundancy (RPR) (offering a trade-off between computational accuracy and power consumption), as well as methods using error-correcting codes (ECCs) (suitable for linear operations such as FFT and adaptive filters).

[0004] Meanwhile, many scholars have begun to utilize partial redundancy hardening techniques to alleviate this problem. This hardening approach emphasizes the protection of critical components and offers greater flexibility than full tri-mode redundancy design. Theoretically, partial redundancy design can achieve a good balance between circuit size and stability. By simplifying the actual circuit to an approximate functional circuit and using partial redundancy techniques in the simplified circuit, a trade-off between system stability and area versus power consumption can be achieved. Although many researchers have proposed partial tri-mode redundancy hardening methods, there is no universally applicable method among the existing partial tri-mode redundancy hardening approaches, which hinders the widespread adoption of this approach.

[0005] Traditional fault-tolerance methods, such as full triple-modular redundancy radiation hardening, have high resource and power consumption, which limits their on-orbit application. For the main computational modules in on-orbit processing, improving fault tolerance while minimizing resource consumption is crucial.

[0006] It is evident that traditional full tri-mode redundancy hardening methods have excessively large areas and high energy consumption. While existing partial tri-mode redundancy hardening methods are more flexible than full tri-mode redundancy hardening methods and can achieve a good balance between circuit size and stability, there is no universal hardening method among the existing partial tri-mode redundancy hardening methods, nor is there a scientific evaluation of the effectiveness of circuit hardening, making it difficult to adapt to changes in harsh radiation environments. Summary of the Invention

[0007] To address the aforementioned issues, this invention provides a partial redundancy hardening method and system based on HITS, which can effectively reduce additional resource overhead under high reliability requirements.

[0008] To achieve the objectives of this invention, the following technical solution is adopted:

[0009] In a first aspect, the present invention provides a general circuit partial redundancy hardening method based on HITS, the method being applied to a dual logic cone model, comprising: S101: generating a node relationship matrix based on the connection relationships between nodes; S102: performing HITS iteration on the node relationship matrix to obtain the HITS values ​​of the nodes; S103: sorting the nodes in descending order based on the HITS values, and selecting the first number of nodes as nodes to be hardened; the first number is determined by a final hardening ratio; S104: performing partial tri-modular redundancy hardening on the nodes to be hardened.

[0010] Furthermore, the node is a central trigger.

[0011] Further, the step of performing HITS iteration on the node relationship matrix to obtain the HITS values ​​of the nodes includes: S102.1: Dividing the node relationship matrix G into a set of nodes V and a set of edges E, i.e., G = (V, E); S102.2: Each node set V includes a hub value and an authority value, and the hub value and authority value are initialized to 1; S102.3: Repeating S102.31-S102.33 until the hub value and the authority value converge: S102.31: Updating the authority value auth(v) of each node, calculated as follows: S102.32: Update the hub value (hub(v) of each node, calculated as follows: Where In(v) represents the set of nodes connected by edges pointing to node v, hub(u) represents the hub value of node u, Out(v) represents the set of nodes connected by edges originating from node v, and auth(u) represents the authority value of node u; S102.33: Normalize the authority value and hub value of each node to obtain the HITS value, calculated as follows: in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

[0012] Further, the step of sorting the nodes in descending order based on the HITS value and selecting the first number of nodes as nodes to be hardened includes: S103.1: Selecting a temporary hardening ratio according to monotonically increasing order, the number corresponding to the temporary hardening ratio is used as the number of temporary nodes to be hardened, setting the nodes of the number of temporary nodes to be hardened as nodes to be hardened, and replacing the nodes to be hardened with tri-mode trigger nodes; the temporary hardening ratio is a value between 10% and 100%; S103.2: Deriving the mean time between failures (MTBF) using a dual logic cone model, calculated as follows:

[0013]

[0014] Where Z is the number of center flip-flops, T is the continuous operating time of the circuit to be hardened, λ is the failure rate of the circuit to be hardened, P(T) is the error probability, and DC of the center flip-flops... i The number of source triggers is M i i is the center trigger number, and the center trigger DC i The number of target triggers is N i Mz and Nz are the number of source flip-flops and target flip-flops when i = Z, respectively, and the number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1; where the source trigger is the input node of the central trigger, and the target trigger is the output node of the central trigger; S103.3: Establish the relationship curve between the number of temporary node hardening and the mean time between failures (MTBF); S103.4: Obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio according to the temporary hardening ratio; S103.5: Compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple redundancy. When the values ​​of the two are similar or the same, the temporary hardening ratio is the final determined hardening ratio.

[0015] Further, the partial tri-mode redundancy reinforcement of the node to be reinforced includes: S104.1: replacing the first number of nodes with tri-mode trigger nodes; S104.2: comparing the first mean time between failures (MTBF) and the second mean time between failures (MTBF) again to obtain a comparison result; S104.3: if the comparison result is less than or equal to a threshold value, the reinforcement is confirmed to be effective, and the reinforcement process ends; the threshold value is less than 0.1%, and can be adjusted according to accuracy requirements.

[0016] A second aspect of the present invention provides a general circuit partial redundancy hardening system based on HITS, the system comprising: a generation module for generating a node relationship matrix based on the connection relationships between nodes; an iteration module connected to the generation module for performing HITS iteration on the node relationship matrix to obtain the HITS values ​​of the nodes; a selection module connected to the iteration module for sorting the nodes in descending order based on the HITS values ​​and selecting the first number of nodes as nodes to be hardened; the first number is determined by a final hardening ratio; and a hardening module connected to the selection module for performing partial tri-modular redundancy hardening on the nodes to be hardened.

[0017] Furthermore, the node is a central trigger.

[0018] Furthermore, the iterative module further includes: a partitioning subunit, used to partition the node relationship matrix, wherein the node matrix G includes a node set V and an edge set E, i.e., G = (V, E); an initialization subunit, connected to the partitioning subunit, used to initialize the hub value and authority value included in each node set V to 1; and a convergence subunit, connected to the initialization subunit, used to repeat the following steps until the hub value and authority value converge; S102.31: Update the authority value auth(v) of each node, calculated as follows: S102.32: Update the hub value (hub(v) of each node, calculated as follows: Where In(v) represents the set of all nodes connected by edges pointing to node v, hub(u) represents the hub value of node u, Out(v) represents the set of all nodes connected by edges originating from node v, and auth(u) represents the authority value of node u.

[0019] S102.33: Normalize the authority and hub values ​​of each node to obtain the HITS value, calculated as follows: in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

[0020] Furthermore, the selection module further includes: a selection subunit, used to select a temporary reinforcement ratio in monotonically increasing order, the number corresponding to the temporary reinforcement ratio being used as the number of temporary node reinforcements, setting the nodes of the number of temporary node reinforcements as nodes to be reinforced, and replacing the nodes to be reinforced with tri-mode trigger nodes; the temporary reinforcement ratio is a value between 10% and 100%; and an MTBF subunit, connected to the selection subunit, used to derive the mean time between failures (MTBF) using a dual logic cone model, calculated as follows:

[0021]

[0022] Where Z is the number of center flip-flops, T is the continuous operating time of the circuit to be hardened, λ is the failure rate of the circuit to be hardened, P(T) is the error probability, i is the center flip-flop number, and DC of the center flip-flop... i The number of source triggers is M i DC center trigger i The number of target triggers is N i The number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1, Mz and Nz are the corresponding values ​​when i = Z; where the source trigger is the input node of the central trigger, and the target trigger is the output node of the central trigger. A relational subunit, connected to the MTBF subunit, is used to establish a relationship curve between the number of temporary node hardenings and the mean time between failures (MTBF); an acquisition subunit, connected to the relational subunit, is used to obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio based on the temporary hardening ratio; a determination subunit, connected to the acquisition subunit, is used to compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple-modular redundancy, and when the two values ​​are similar or the same, the temporary hardening ratio is the finally determined hardening ratio.

[0023] Furthermore, the hardening module further includes: a replacement subunit, used to replace the first number of nodes with tri-mode trigger nodes; a comparison subunit, connected to the replacement subunit, used to compare the first mean time between failures (MTBF) and the second mean time between failures (MTBF) again to obtain a comparison result; and an end subunit, connected to the comparison subunit, which confirms the hardening is effective and ends the hardening process if the comparison result is less than or equal to a threshold value; the threshold value is less than 0.1% and can be adjusted according to accuracy requirements.

[0024] This invention proposes a new general solution to address the problems existing in current partial redundancy hardening techniques. After mapping the circuit connections, the HITS algorithm can rank the importance of circuit nodes based on their global connectivity. Furthermore, the introduction of the dual logic cone model establishes a circuit reliability evaluation criterion using MTBF, effectively solving the problem of hardening ratio selection. The algorithm used in this invention can be implemented based on the circuit's netlist signal connections, thus it is not limited by circuit functionality and can be directly used in most circuits requiring radiation hardening. Attached Figure Description

[0025] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0026] Figure 1 This is a structural diagram of the dual logic cone model of the present invention;

[0027] Figure 2 This is a schematic diagram illustrating the relationship between the authority value and the pivot value in this invention;

[0028] Figure 3 This is a flowchart of the method of the present invention;

[0029] Figure 4 This is a system structure diagram of the present invention;

[0030] Figure 5 This is a structural diagram of the iterative module of the present invention;

[0031] Figure 6 This is a structural diagram of the selected module in the present invention;

[0032] Figure 7 This is a structural diagram of the reinforcement module of the present invention. Detailed Implementation

[0033] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the invention. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the invention.

[0034] In the description of embodiments of the present invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0035] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and descriptions are considered exemplary in nature and not restrictive. There is no requirement for the order of the method steps described, as long as they are achievable, they are within the scope of protection of this invention.

[0036] To better describe this application, we first introduce two fundamental concepts: 1. The dual logic cone model, which is used to establish the relationship curve between hardening ratio and MTBF; 2. The HITS algorithm, which is used to rank the importance of circuit nodes. These are described below.

[0037] I. Dual Logic Cone Model

[0038] There are many methods for performing reliability analysis on circuits. The core idea is to find the relationships between circuit nodes and identify the critical nodes to analyze the system's reliability. (Combined with...) Figure 1 In circuit formal verification techniques, the logic cone model is used to analyze the correlation between circuit nodes. In the logic cone model, the flip-flop (defined as the center flip-flop), the combinational logic of the center flip-flop input, and the input flip-flop (defined as the source flip-flop) are regarded as a whole, and such a circuit is defined as the source logic cone.

[0039] In the source logic cone, i is the index of the central flip-flop, and the central flip-flop DC... i The number of source triggers is M i The source flip-flops are represented as DI(i, 0), DI(i, 1), ..., DI(i, M). i -1). In the target logic cone, the number of target flip-flops is N. i The target trigger can be represented as DO(i, 0), DO(i, 1), ..., DO(i, N). i -1), the number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1.

[0040] In a circuit with Z central flip-flops, the average frequency of spatial particles causing a single flip-flop to flip is L times / second. L is calculated based on long-term single-event upset statistics and the number of flip-flop nodes and the silicon wafer area. The probability P0 of a single flip-flop being directly hit by a single particle and causing it to flip is:

[0041] P0 = L / Z

[0042] Ignoring the impact of error propagation from other sources on the source flip-flop, what is the probability, p, that a source flip-flop error in the source logic cone causes a center flip-flop error? S (i) is:

[0043]

[0044] The probability P that a single-event upset causes the central trigger to malfunction c The sum of the probability of the source trigger causing its error and the probability of the center trigger itself failing:

[0045]

[0046] The combinational logic involved in the output of the central trigger and the output trigger (defined as the target trigger, where the number of target triggers for the i-th central trigger is N) are considered. i When considered together, they form another logic cone, called the target logic cone.

[0047] Based on the above analysis, the central trigger DC i The probability P of error propagation to the target trigger cd It can be calculated using the following formula:

[0048]

[0049] Therefore, we can establish an error propagation model where a single particle knocks over a source trigger or a central trigger. The structure of the dual logic cone is shown in the attached manual. Figure 1 As shown.

[0050] If the system has Z central triggers and a continuous operating time of T, and the error probability is equivalent to the failure rate λ, then the system's mean time between failures (MTBF) is:

[0051]

[0052] The above formula represents the average fault time interval obtained from the dual logic cone model, where Z is the number of center flip-flops, T is the continuous operating time of the circuit to be hardened, λ is the failure rate of the circuit to be hardened, P(T) is the error probability, and the center flip-flop DC... i The number of source triggers is M i DC center trigger i The number of target triggers is N i Where i is the center flip-flop index, Mz and Nz are the number of source flip-flops and target flip-flops when i = Z, and the number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1.

[0053] II. HITS Algorithm

[0054] The basic idea of ​​the HITS algorithm is that the importance of each webpage is characterized by two metrics: Authority and Hub. A webpage with a high Authority will be linked to by many other webpages, and a webpage with a high Hub will link to many other webpages, such as... Figure 2 As shown.

[0055] Authority value and hub value are interdependent and mutually influential, and they are calculated as follows:

[0056] The authority value of a webpage is equal to the sum of the hub values ​​of all webpages that point to it.

[0057]

[0058] The hub value of a webpage is equal to the sum of the authority values ​​of all the webpages it points to.

[0059]

[0060] The importance value of a webpage is calculated by the HITS algorithm based on the connection relationships between various nodes on the webpage. The process is as follows:

[0061] 1. Initialization: Set the authority and hub values ​​of each node to 1;

[0062] 2. Update the node's authority value;

[0063] 3. Update the node's hub value;

[0064] 4. Normalize the authority value and hub value;

[0065] 5. Repeat steps 2-4 until final convergence.

[0066] To address the problem of prioritizing the importance of circuit nodes, this invention creatively introduces the HITS algorithm, which is used for web page ranking, to solve the problems of difficulty in determining the redundancy reinforcement ratio of circuit nodes and the lack of a unified standard.

[0067] To implement this method, a mapping from the HITS algorithm to the circuit needs to be established. The HITS algorithm is based on two fundamental assumptions: a high-quality authority page will be pointed to by many high-quality hub pages; and a high-quality hub page will point to many high-quality authority pages. The quality of a page is determined by its hub value and authority value, calculated as follows: a page's hub value equals the sum of the authority values ​​of all pages it points to; and a page's authority value equals the sum of the hub values ​​of all pages pointing to it. In the HITS algorithm, based on these two fundamental assumptions, the importance of each webpage node is iteratively calculated using the reference relationships between webpages to obtain the page importance ranking. This approach is very similar to the redundancy hardening design of circuit components, and these two fundamental assumptions align with the requirements of this invention: flip-flops with more outputs need to be hardened; flip-flops with more inputs also need to be hardened.

[0068] Comparative analysis of webpage links and circuit dual logic cone models reveals that the basic assumptions of the HITS algorithm are consistent with the circuit node importance evaluation criteria. Therefore, the HITS algorithm can be applied to the importance ranking of circuit flip-flop nodes to improve the effectiveness of redundancy hardening.

[0069] One aspect of the present invention provides a partial redundancy hardening method based on HITS, such as... Figure 3 As shown, it specifically includes:

[0070] Step 1: Generate a node relationship matrix based on the connection relationships between nodes.

[0071] In some embodiments, a node can be a trigger or a central trigger.

[0072] This step can be implemented in the following ways:

[0073] Step 1.1: Use chip logic synthesis tools to determine the timing information between the central trigger nodes and generate a report file to store the timing information.

[0074] Step 1.2: Analyze the report file to obtain the node relationship matrix G.

[0075] Step 2: Perform HITS iteration on the node relationship matrix to obtain the HITS value of the node.

[0076] This step can be implemented in the following ways:

[0077] Step 2.1: Divide the node relationship matrix G into a set of nodes V and a set of edges E, i.e., G = (V, E);

[0078] Step 2.2: Each node set V includes a hub value and an authority value, and the hub value and authority value are initialized to 1;

[0079] Step 2.3: Repeat the following steps until the pivot value and the authority value converge:

[0080] Step 2.31: Update the authority value auth(v) of each node, calculated as follows:

[0081]

[0082] Step 2.32: Update the hub value (hub(v) of each node, calculated as follows:

[0083]

[0084] Where In(v) represents the set of all nodes connected by edges pointing to node v, hub(u) represents the hub value of node u, Out(v) represents the set of all nodes connected by edges originating from node v, and auth(u) represents the authority value of node u.

[0085] Step 2.33: Normalize the authority value and hub value of each node to obtain the HITS value, calculated as follows:

[0086]

[0087] in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

[0088] By applying mean square normalization, the data can be observed on the same scale. Quantizing the authority and pivot values ​​to the [0,1] interval makes it easier to statistically rank the HITS values ​​in the [0,2] interval, effectively improving accuracy.

[0089] Step 3: Based on the HITS values, sort the nodes in descending order and select the first number of nodes as nodes to be reinforced; the first number is determined by the final reinforcement ratio.

[0090] This step can be achieved in the following way:

[0091] Step 3.1: Select a temporary reinforcement ratio in monotonically increasing order. The number corresponding to the temporary reinforcement ratio is used as the number of temporary nodes to be reinforced. Set the nodes of the number of temporary nodes to be reinforced as nodes to be reinforced. Replace the nodes to be reinforced with tri-mode trigger nodes. The temporary reinforcement ratio is a value between 10% and 100%.

[0092] Step 3.2: Derive the Mean Time Between Failures (MTBF) using the dual logic cone model, as follows:

[0093]

[0094] Where Z is the number of center flip-flops, T is the continuous operating time of the circuit to be hardened, λ is the failure rate of the circuit to be hardened, P(T) is the error probability, i is the center flip-flop number, and DC of the center flip-flop... i The number of source triggers is M i DC center trigger i The number of target triggers is N i The number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1, Mz and Nz are the number of source triggers and target triggers when i = Z; wherein, the source triggers are the input nodes of the central triggers and the target triggers are the output nodes of the central triggers.

[0095] Step 3.3: Establish the relationship curve between the number of temporary node reinforcements and the mean time between failures (MTBF).

[0096] By substituting the number of temporary node reinforcements and related parameters into the formula of S103.2, a relationship curve between the temporary node reinforcement number and the mean time between failures (MTBF) is gradually established.

[0097] Step 3.4: Based on the temporary hardening ratio, obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio;

[0098] Step 3.5: Compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple redundancy. When the two values ​​are similar or the same, the temporary reinforcement ratio is the final determined reinforcement ratio.

[0099] Step 4: Perform partial tri-modal redundancy reinforcement on the node to be reinforced.

[0100] This step can be achieved in the following way:

[0101] Step 4.1: Replace the first number node with a three-mode trigger node;

[0102] Step 4.2: Compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) again to obtain the comparison result;

[0103] Step 4.3: If the comparison result is less than or equal to the threshold value, the reinforcement is confirmed to be effective and the reinforcement process ends; if the comparison result is greater than the threshold value, steps 3.1-3.5 and steps 4.1-4.3 are repeated; the threshold value is less than 0.1% and can be adjusted according to the accuracy requirements.

[0104] Another aspect of the present invention provides a partially redundant hardening system based on HITS, such as Figure 4 As shown, the system includes a generation module, an iteration module, a selection module, and a reinforcement module. Each module is described in detail below.

[0105] The generation module is used to generate a node relationship matrix based on the connection relationships between nodes.

[0106] In some embodiments, the node may be a central trigger.

[0107] The generation module is also used to determine the timing information between the central trigger nodes using chip logic synthesis tools, and to generate a report file for storing the timing information. The report file is analyzed to obtain the node relationship matrix G.

[0108] An iteration module, connected to the generation module, is used to perform HITS iteration on the node relationship matrix to obtain the HITS value of the node.

[0109] like Figure 5 As shown, the iteration module further includes:

[0110] The sub-unit is used to divide the node relationship matrix. The node matrix G includes a set of nodes V and a set of edges E, i.e., G = (V, E).

[0111] Initialize the subunit and connect it to the partitioning subunit to initialize the hub value and authority value of each node set V to 1;

[0112] A convergence subunit, connected to the initialization subunit, is used to repeatedly perform the following steps until the pivot value and authority value converge:

[0113] Step 2.31: Update the authority value auth(v) of each node, calculated as follows:

[0114]

[0115] Step 2.32: Update the hub value (hub(v) of each node, calculated as follows:

[0116]

[0117] Where In(v) represents the set of all nodes connected by edges pointing to node v, hub(u) represents the hub value of node u, Out(v) represents the set of all nodes connected by edges originating from node v, and auth(u) represents the authority value of node u.

[0118] Step 2.33: Normalize the authority value and hub value of each node to obtain the HITS value, calculated as follows:

[0119]

[0120] in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

[0121] The selection module, connected to the iteration module, is used to sort the nodes in descending order based on the HITS value and select the first number of nodes as nodes to be reinforced; the first number is determined by the final reinforcement ratio.

[0122] like Figure 6 As shown, the selection module further includes:

[0123] Select a sub-unit to select a temporary reinforcement ratio in a monotonically increasing order. The number corresponding to the temporary reinforcement ratio is used as the number of temporary nodes to be reinforced. Set the nodes of the number of temporary nodes to be reinforced as nodes to be reinforced, and replace the nodes to be reinforced with tri-mode trigger nodes. The temporary reinforcement ratio is a value between 10% and 100%.

[0124] The MTBF sub-unit, connected to the selected sub-unit, is used to derive the Mean Time Between Failures (MTBF) using a dual logic cone model, calculated as follows:

[0125]

[0126] Where Z is the number of center flip-flops, T is the continuous operating time of the circuit to be hardened, λ is the failure rate of the circuit to be hardened, P(T) is the error probability, i is the center flip-flop number, and DC of the center flip-flop... i The number of source triggers is M i DC center trigger i The number of target triggers is N i The number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j = 0, 1, ..., N i -1, Mz and Nz are the number of source triggers and target triggers when i = Z; wherein, the source triggers are the input nodes of the central triggers and the target triggers are the output nodes of the central triggers.

[0127] The relational subunit, connected to the MTBF subunit, is used to establish a relationship curve between the number of temporary node reinforcements and the mean time between failures (MTBF).

[0128] The acquisition sub-unit is connected to the relational sub-unit and is used to obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio according to the temporary hardening ratio.

[0129] The determined subunit, connected to the obtained subunit, is used to compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple redundancy. When the two values ​​are similar or the same, the temporary reinforcement ratio is the final determined reinforcement ratio.

[0130] The reinforcement module, connected to the selection module, is used to perform partial tri-modal redundancy reinforcement on the node to be reinforced.

[0131] like Figure 7 As shown, the reinforcement module also includes:

[0132] Replacement subunit, used to replace the first number of nodes with a three-mode trigger node;

[0133] A comparison subunit, connected to the replacement subunit, is used to compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) again to obtain a comparison result.

[0134] The end sub-unit is connected to the comparison sub-unit. If the comparison result is less than or equal to the threshold value, the reinforcement is confirmed to be effective and the reinforcement process ends. If the comparison result is greater than the threshold value, the functions of selecting sub-units, MTBF sub-units, relational sub-units, obtaining sub-units, determining sub-units, replacing sub-units, comparing sub-units, and comparing sub-units are repeated. The threshold value is less than 0.1% and can be adjusted according to the accuracy requirements.

[0135] Compared with the prior art, the beneficial effects of the present invention are:

[0136] 1) The HITS algorithm-based sorting method ranks the nodes of a circuit by importance, effectively strengthening the circuit with partial redundancy. This method is applicable to general-purpose circuits. Compared with traditional fault-tolerant techniques such as triple modular redundancy, the method adopted in this invention can reduce additional resource overhead while ensuring reliability, thus playing an important role in circuit reliability protection and other fields.

[0137] 2) After mapping the circuit connections, the HITS algorithm can rank the importance of circuit nodes based on their global connectivity. The introduction of the dual logic cone model establishes a circuit reliability evaluation criterion using MTBF, effectively solving the problem of hardening ratio selection. Because this algorithm is based on the circuit's netlist signal connections, it is not limited by circuit functionality and can be directly used in most circuits requiring radiation hardening.

[0138] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the scope of the technology disclosed in the present invention, or make equivalent substitutions for some of the technical features. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A general circuit redundancy hardening method based on HITS, characterized in that, The method is applied to a dual logic cone model, and the method includes: S101: Generate a node relationship matrix based on the connection relationships between nodes; S102: Perform HITS iteration on the node relationship matrix to obtain the HITS value of the node; S103: Based on the HITS value, sort the nodes in descending order and select the first number of nodes as nodes to be reinforced; the first number is determined by the final reinforcement ratio. This step specifically includes: S103.1: Select a temporary reinforcement ratio in monotonically increasing order, and use the number corresponding to the temporary reinforcement ratio as the number of temporary nodes to be reinforced. Set the nodes of the number of temporary nodes to be reinforced as nodes to be reinforced, and replace the nodes to be reinforced with tri-mode trigger nodes. The temporary reinforcement ratio is a value between 10% and 100%. S103.2: The Mean Time Between Failures (MTBF) is derived using the dual logic cone model and calculated as follows: Where Z represents the number of center flip-flops, and T represents the continuous operating time of the circuit to be hardened. Let P(T) be the failure rate of the circuit to be hardened, P(T) be the error probability, and DC be the center trigger. i The number of source triggers is i is the center trigger number, and the center trigger DC i The number of target triggers is Mz and Nz are the number of source flip-flops and target flip-flops when i=Z, respectively, and the number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j=0,1,…,N i -1; Wherein, the source trigger is the input node of the central trigger, and the target trigger is the output node of the central trigger; S103.3: Establish the relationship curve between the number of temporary node reinforcements and the mean time between failures (MTBF); S103.4: Based on the temporary hardening ratio, obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio; S103.5: Compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple redundancy. When the values ​​of the two are similar or the same, the temporary reinforcement ratio is the final determined reinforcement ratio. S104: Perform partial tri-modal redundancy reinforcement on the node to be reinforced.

2. The method as described in claim 1, characterized in that, The node is a central trigger.

3. The method as described in claim 1, characterized in that, The step of performing HITS iteration on the node relationship matrix to obtain the HITS value of the node includes: S102.1: Divide the node relationship matrix G into a set of nodes V and a set of edges E, i.e.: G=(V,E); S102.2: Each node set V includes a hub value and an authority value, and the hub value and authority value are initialized to 1; S102.3: Repeat the following steps until the pivot value and the authority value converge: S102.31: Update the authority value of each node. The calculation is as follows: S102.32: Update the hub value for each node. The calculation is as follows: Among them, I n(v) Represents all pointer nodes v The set of nodes connected by the edges. hub(u) Represents a node u The pivot value, Out(v) Represents all nodes v The set of nodes connected by the starting edge. auth(u) Represents a node u Authority value; S102.33: Normalize the authority and hub values ​​of each node to obtain the HITS value, calculated as follows: in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

4. The method as described in claim 1, characterized in that, The partial tri-modal redundancy reinforcement of the node to be reinforced includes: S104.1: Replace the first number node with a three-mode trigger node; S104.2: Compare the first mean time between failures (MTBF) and the second mean time between failures (MTBF) again to obtain the comparison result; S104.3: If the comparison result is less than or equal to the threshold value, the reinforcement is confirmed to be effective and the reinforcement process ends; the threshold value is less than 0.1% and can be adjusted according to the accuracy requirements.

5. A general-purpose circuit redundancy hardening system based on HITS, characterized in that, The system includes: The generation module is used to generate a node relationship matrix based on the connection relationships between nodes; An iteration module, connected to the generation module, is used to perform HITS iteration on the node relationship matrix to obtain the HITS value of the node; A selection module, connected to the iteration module, is used to sort the nodes in descending order based on the HITS value and select the first number of nodes as nodes to be reinforced; the first number is determined by the final reinforcement ratio. The selection module further includes: Select a sub-unit to select a temporary reinforcement ratio in a monotonically increasing order. The number corresponding to the temporary reinforcement ratio is used as the number of temporary nodes to be reinforced. Set the nodes of the number of temporary nodes to be reinforced as nodes to be reinforced. Replace the nodes to be reinforced with tri-mode trigger nodes. The temporary reinforcement ratio is a value between 10% and 100%. The MTBF sub-unit, connected to the selected sub-unit, is used to derive the Mean Time Between Failures (MTBF) using a dual logic cone model, calculated as follows: Where Z represents the number of center flip-flops, and T represents the continuous operating time of the circuit to be hardened. Let P(T) be the failure rate of the circuit to be hardened, and P(T) be the error probability. The DC-DC trigger is a center-triggered flip-flop. i The number of source triggers is DC center trigger i The number of target triggers is Where i is the center flip-flop index, Mz and Nz are the number of source flip-flops and target flip-flops when i=Z, and the number of combinational logic input signals corresponding to the j-th target flip-flop is K. j j=0,1,…,N i -1; where the source trigger is the input node of the center trigger, and the target trigger is the output node of the center trigger; The relational subunit, connected to the MTBF subunit, is used to establish a relationship curve between the number of temporary node reinforcements and the mean time between failures (MTBF). The acquisition sub-unit is connected to the relational sub-unit and is used to obtain the first mean time between failures (MTBF) of the system under the temporary hardening ratio according to the temporary hardening ratio. A determination subunit is connected to the acquisition subunit and is used to compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) of full triple redundancy. When the values ​​of the two are similar or the same, the temporary reinforcement ratio is the final determined reinforcement ratio. The reinforcement module, connected to the selection module, is used to perform partial tri-modal redundancy reinforcement on the node to be reinforced.

6. The system as described in claim 5, characterized in that, The node is a central trigger.

7. The system as described in claim 5, characterized in that, The iteration module also includes: The sub-unit is used to divide the node relationship matrix. The node matrix G includes a set of nodes V and a set of edges E, i.e., G=(V,E). Initialize the subunit and connect it to the partitioning subunit to initialize the hub value and authority value of each node set V to 1; A convergence subunit, connected to the initialization subunit, is used to repeatedly perform the following steps until the pivot value and authority value converge: S102.31: Update the authority value of each node. The calculation is as follows: S102.32: Update the hub value for each node. The calculation is as follows: Among them, I n(v) Represents all pointer nodes v The set of nodes connected by the edges. hub(u) Represents a node u The pivot value, Out(v) Represents all nodes v The set of nodes connected by the starting edge. auth(u) Represents a node u Authority value; S102.33: Normalize the authority and hub values ​​of each node to obtain the HITS value, calculated as follows: in, This represents the result of mean square normalization of the authority value; This represents the result of mean square normalization of the pivot value.

8. The system as described in claim 5, characterized in that, The reinforcement module also includes: Replacement subunit, used to replace the first number of nodes with a three-mode trigger node; A comparison subunit, connected to the replacement subunit, is used to compare the first mean time between failures (MTBF) with the second mean time between failures (MTBF) again to obtain a comparison result. The end subunit is connected to the comparison subunit. If the comparison result is less than or equal to the threshold value, the reinforcement is confirmed to be effective, and the reinforcement process ends. The threshold value is less than 0.1% and can be adjusted according to the accuracy requirements.

Images