A blockchain-based method for end-to-end traceability and access control of scientific research data.

By acquiring the execution time and latency of nodes, and comprehensively considering resource contention differences, permissions are dynamically adjusted to solve the data leakage problem caused by hardware resource contention in scientific research data management, thereby improving security and compliance.

CN122490563BActive Publication Date: 2026-11-17SHENZHEN KANGNING HOSPITAL (SHENZHEN MENTAL HEALTH INST SHENZHEN MENTAL HEALTH CENT)
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610894608.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-06-22
Publication Date
2026-11-17
Estimated Expiration
2046-06-22

AI Technical Summary

Technical Problem

Existing technologies cannot detect the dynamic contention of underlying hardware resources in scientific research data access control, leading to the risk of data leakage and rendering traditional static access control mechanisms ineffective.

Method used

By obtaining the execution time of a node in a state of exclusive resource access, calculating the latency multiplier and memory access latency multiplier, and combining the resource contention difference index, the isolation index of the node is determined, and the isolation index is stored on the blockchain for evidence storage, and permissions are dynamically adjusted.

Benefits of technology

It enables quantitative characterization of the node's operating environment, dynamically protects sensitive scientific research data, avoids the risk of data leakage due to environment drift, and improves the security and compliance of scientific research data management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122490563B_ABST
    Figure CN122490563B_ABST
Patent Text Reader

Abstract

This invention relates to the field of data management technology, specifically to a blockchain-based method for full-process traceability and access control of scientific research data. The method includes: obtaining the first exclusive execution time of a node executing a standard computing task under exclusive resource status, and the second exclusive execution time of executing a standard memory access task; determining resource contention difference indicators based on the preset workload of the node's standard computing task, the preset workload of the standard memory access task, the first exclusive execution time, the second exclusive execution time, the workload of business computing tasks during the execution of the scientific research task, the business execution time, the preset memory access task workload corresponding to the memory probe task, and the probe execution time; obtaining an isolation index based on all resource contention difference indicators within the current time period, and then determining whether to change the node's execution permissions, and storing the judgment basis and change record on the blockchain for evidence. This invention can avoid the risk of data leakage caused by node environment drift during scientific research data management.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically to a blockchain-based method for full-process traceability and access control of scientific research data. Background Technology

[0002] In distributed research collaboration and federated computing networks, computing tasks are typically distributed to computing nodes across participating institutions. To balance data privacy and computational efficiency, existing management architectures primarily rely on identity-based static access control and role-based logical access control. This architecture implicitly assumes that nodes holding legitimate credentials are trusted nodes, and that their operating environment conforms to pre-defined security standards. However, in actual engineering deployments, to reduce hardware ownership costs or improve resource utilization, legitimate computing nodes may migrate their computing environments from committed dedicated physical hardware to multi-tenant shared virtualization environments or container clusters without informing the task issuer. This migration of computing environments introduces specific side-channel risks. In a shared virtualization architecture, while different tenants on the same physical host are isolated in terms of logical instruction execution, they share processor cache, memory bandwidth, and bus interconnect resources at the microarchitecture level. When neighboring tenants perform high-frequency data throughput, they inevitably preempt shared microarchitecture resources, leading to unexpected execution delays in sensitive computing tasks on authorized nodes. This timing fluctuation caused by resource contention constitutes a side-channel characteristic, which not only leaks the non-isolated nature of the computing environment but also renders traditional static permission management mechanisms ineffective. Because the system is unable to perceive the dynamic contention state of the underlying hardware resources, it continues to distribute sensitive research data in plaintext to nodes in an uncontrollable environment even when physical isolation boundaries have disappeared. This undermines the physical security defenses of the research data and poses a risk of data leakage. Summary of the Invention

[0003] To address the problem that existing methods fail to detect the dynamic contention of underlying hardware resources during scientific research data access management, leading to the risk of data leakage, this invention aims to provide a blockchain-based method for end-to-end traceability and access management of scientific research data. The specific technical solution adopted is as follows: This invention provides a blockchain-based method for full-process traceability and access control of scientific research data, which includes the following steps: Get the first exclusive execution time of the node executing the standard computing task issued by the on-chain standard operator library under exclusive resource state, and the second exclusive execution time of executing the standard memory access task; The computation latency ratio and memory access latency ratio are obtained based on the preset workload of the standard computation task of the node, the preset workload of the standard memory access task, the first exclusive execution time, the second exclusive execution time, the workload of the business computation task during the execution of the scientific research task, the business execution time, the preset memory access task and probe execution time corresponding to the memory access probe task. By combining the calculated latency multiplier and the memory access latency multiplier, a resource contention difference index is determined; based on all resource contention difference indices within the current time period, an isolation index is obtained. The system determines whether to change permissions on a node based on the isolation level index, and records the determination criteria and changes on the blockchain for evidence storage.

[0004] Preferably, obtaining the calculation delay ratio includes: Based on the preset workload of the standard computation task of the candidate node and the first exclusive execution duration, the reference computation time rate of the candidate node is obtained. The computation latency ratio is obtained by combining the workload of the business computation task, the execution duration of the business, and the reference computation time consumption rate; the workload of the business computation task and the execution duration of the business are obtained based on the main task mapping; The candidate node can be any node.

[0005] Preferably, obtaining the memory access latency ratio includes: Based on the preset workload of the standard memory access task of the candidate node and the second exclusive execution duration, the reference memory access time rate of the candidate node is obtained. The memory access latency ratio is obtained by combining the preset memory access task volume, the probe execution time, and the reference memory access time rate.

[0006] Preferably, obtaining the computation latency multiplier by combining the workload of the business computation task, the execution duration of the business, and the reference computation time consumption rate includes: Calculate the first product of the business computation task volume and the reference computation time consumption rate; The ratio of the service execution time to the first product is determined as the calculation delay ratio.

[0007] Preferably, the step of obtaining the memory access latency ratio by combining the preset memory access task volume, the probe execution duration, and the reference memory access time rate includes: Calculate the second product of the preset memory access task volume and the reference memory access time rate; The ratio of the probe execution time to the second product is determined as the memory access latency multiplier.

[0008] Preferably, the memory access probe task is forcibly bound to the processor core where the main task resides, and is executed concurrently and alternately with the main task.

[0009] Preferably, the step of determining the resource contention difference index by combining the computation latency multiplier and the memory access latency multiplier includes: The computational latency multiplier and the memory access latency multiplier are respectively subjected to nonlinear transformations; The difference between the transformed computation latency multiplier and the transformed memory access latency multiplier is used as an indicator of resource contention difference.

[0010] Preferably, obtaining the isolation index based on all resource contention difference indicators within the current time period includes: Calculate the arithmetic mean of all resource contention difference indicators for each node within the current time period, and use it as the isolation index for each node within the current time period.

[0011] Preferably, the step of determining whether to perform permission changes on a node based on the isolation index includes: If the isolation index is greater than the preset isolation compliance threshold, the corresponding node is determined to be in an uncontrolled sharing state, triggering a circuit breaker operation for smart contract execution permissions. If the isolation index is less than or equal to the preset isolation compliance threshold, the corresponding node is determined to be in an exclusive resource state, and the node's execution permissions are not changed.

[0012] Preferably, the permission circuit breaker operation includes: revoking the data decryption key of the node and downgrading the node to a restricted mode that only receives encrypted computation tasks.

[0013] The present invention has at least the following beneficial effects: This invention obtains the first and second exclusive execution times of a node executing standard computing and memory access tasks under exclusive resource conditions, providing an accurate and reproducible hardware baseline for subsequent performance evaluation and effectively avoiding misjudgments of permissions due to differences in node hardware. Combining the workload and execution time of business computing tasks during research task execution, as well as the preset memory access workload and probe execution time corresponding to memory access probe tasks, the computation latency multiplier and memory access latency multiplier are calculated respectively. This process quantitatively compares the actual performance of the node in the real operating environment with the hardware baseline, accurately identifying the degree of performance degradation caused by environmental changes, and providing reliable data support for dynamic permission adjustment. By comprehensively calculating the latency multiplier and memory access latency multiplier, the invention further determines the performance degradation of the node in the real operating environment. By defining resource contention difference indicators and obtaining isolation indicators based on all resource contention difference indicators within the current time period, a quantitative representation of the isolation status of the node's operating environment is achieved. This indicator can objectively reflect whether a node is in a controlled, exclusive resource state, thus providing a scientific basis for permission changes. The mechanism determines whether to perform permission changes on a node based on the isolation indicator and records the judgment criteria and changes on the blockchain for evidence storage. This mechanism binds permission control to the real-time operating environment status, achieving dynamic protection of sensitive scientific research data and effectively avoiding the risk of data leakage due to node environment drift. Simultaneously, all judgment criteria and change records are stored on the blockchain for evidence storage, ensuring the entire permission management process is traceable and tamper-proof, thus improving the security and compliance of scientific research data management. Attached Figure Description

[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating a blockchain-based method for full-process traceability and access control of scientific research data, as provided in an embodiment of the present invention. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of the blockchain-based scientific research data full-process traceability and access management method proposed in accordance with the present invention is provided in conjunction with the accompanying drawings and preferred embodiments.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] The following description, in conjunction with the accompanying drawings, details the specific solution of the blockchain-based scientific research data full-process traceability and access management method provided by this invention.

[0019] Example of a blockchain-based method for full-process traceability and access control of scientific research data: This embodiment proposes a blockchain-based method for full-process traceability and access control of scientific research data, such as... Figure 1 As shown, the blockchain-based scientific research data full-process traceability and access management method in this embodiment includes the following steps: Step S1: Obtain the first exclusive execution time of the standard computing task issued by the on-chain standard operator library and the second exclusive execution time of the standard memory access task when the node is in exclusive resource state.

[0020] First, an on-chain standard operator library is deployed. This library is pre-installed in the blockchain's smart contracts and contains two types of rigorously audited baseline code segments as the source of the entire computational workload. The first type consists of standard computational tasks issued by the on-chain standard operator library, executed by nodes under exclusive resource conditions. These tasks employ pure floating-point arithmetic logic, such as fixed-dimensional dense matrix multiplication. The working set of this task is designed to be smaller than the L2 cache capacity of mainstream processors to isolate the impact of main memory bandwidth. The preset workload of this task is denoted as... Its physical unit is defined as one billion floating-point operations (GFLOPs). The second type involves nodes executing standard memory access tasks issued by the on-chain standard operator library under exclusive resource status. This task employs a large-scale random pointer chasing logic. The task is forced to have a working set far exceeding the capacity of the final level cache (LLC) of mainstream processors, forcing the processor to frequently access main memory, thereby reaching the physical bandwidth limit of the memory controller. The preset workload of this task is denoted as... Its physical unit is defined as gigabyte (GB).

[0021] Because competition for hardware resources in a multi-tasking concurrent environment can cause measurement data to deviate from the hardware's inherent capabilities, benchmarking needs to be conducted under strict isolation conditions. When a compute node initiates a network access request or responds to a periodic baseline calibration command, the smart contract issues a test package containing the aforementioned standard tasks. The compute node first suspends other background processes and enters a state of exclusive resource access. In this state, the node executes standard computation tasks and standard memory access tasks sequentially.

[0022] The node calls the underlying operating system's hardware counters (such as the CPU's timestamp counter TSC or the high-precision event timer HPET) to record the start and end time difference of the execution of the core code segment of the standard computing task, thus obtaining the exclusive execution duration, which is recorded as the first exclusive execution duration. The unit is milliseconds. Similarly, the node records the exclusive execution time of the standard memory access task, and records this exclusive execution time as the second exclusive execution time. The unit is milliseconds. This type of duration data only includes core computing and memory transfer processes, excluding network I / O and disk read / write latency.

[0023] Step S2: Based on the preset workload of the standard computing task of the node, the preset workload of the standard memory access task, the first exclusive execution duration, the second exclusive execution duration, the workload of the business computing task during the execution of the scientific research task, the business execution duration, the preset memory access task workload and probe execution duration corresponding to the memory access probe task, the computing latency ratio and the memory access latency ratio are obtained respectively.

[0024] In order to convert absolute physical time into a relative capability index that characterizes hardware performance, the collected duration data needs to be processed to obtain the time cost required to process a unit of workload. The lower the time cost, the stronger the node hardware performance.

[0025] This embodiment uses one node as an example for illustration. The method provided in this embodiment can be used to process other nodes.

[0026] Each node is designated as a candidate node. Based on the preset workload of the candidate node's standard computation task and the first exclusive execution duration, a reference computation time rate is obtained for the candidate node. Specifically, the ratio of the first exclusive execution duration to the preset workload of the candidate node's standard computation task is used as the reference computation time rate. This reference computation time rate characterizes the time required for the processor pipeline to process a unit of standard computing power task under ideal, interference-free conditions. Similarly, based on the preset workload of the candidate node's standard memory access task and the second exclusive execution duration, a reference memory access time rate is obtained for the candidate node. Specifically, the ratio of the second exclusive execution duration to the preset workload of the candidate node's standard memory access task is used as the reference memory access time rate. This reference memory access time rate characterizes the time required for the memory subsystem to transfer a unit of standard memory access data under ideal, interference-free conditions.

[0027] Considering that existing access control mechanisms cannot accurately distinguish between overall fluctuations in hardware physical performance and side-channel interference caused by competition for virtualization resources, including fluctuations in overall hardware physical performance such as dynamic frequency adjustments due to heat dissipation, this embodiment will next compare the latency ratio differences between the compute channel and the memory access channel within the same spatiotemporal window to extract nonlinear characteristics caused by multi-tenant resource contention from the overall performance data, thereby accurately locating the risk of failure of the isolation attributes of the physical environment.

[0028] In order to accurately quantify the degree of attenuation of the execution efficiency of the current scientific research task in the processor pipeline dimension relative to the inherent clustering standard in S1, and to prevent computing nodes from covering up performance defects by exaggerating the task size, this embodiment will adopt a mandatory homogeneous workload mapping mechanism.

[0029] Specifically, firstly, the transaction metadata of the research task to be executed is parsed to extract the input parameter vector used to describe the task scale. The system does not accept any workload values ​​directly reported by nodes, but instead calls the on-chain standard operator library based on the vector. The pre-defined workload mapping function in ) This function is based on a deterministic algorithm complexity model (such as matrix multiplication). (where O represents the algorithm's time complexity symbol, and n is the number of rows or columns of the input matrix) to calculate the workload of this task. This value, used as the theoretical denominator for this monitoring, is unique.

[0030] Then, the candidate nodes are used to initiate the computation process for this research task. The system calls the underlying operating system's process-level timer (such as CLOCK_PROCESS_CPUTIME_ID) to collect the execution time of the core computation segment of the task. This duration only counts the effective clock cycles of the CPU in user mode and kernel mode, excluding suspension time caused by disk I / O wait or network congestion, to ensure that the measurement data only reflects the processor's computing power.

[0031] Furthermore, by combining the business computing task volume, business execution time, and reference computing time consumption rate obtained above, the computing latency ratio is obtained.

[0032] Specifically, the product of the business computing task volume and the reference computing time consumption rate is recorded as the first product; the ratio of the business execution time to the first product is determined as the computing latency multiplier.

[0033] As a concrete example, the specific formula for calculating the delay ratio is given. The delay ratio can be expressed as:

[0034] in, This indicates the calculation of the delay ratio. Indicates the duration of the business execution. This indicates the amount of business calculation tasks. This indicates the reference calculation time rate.

[0035] Represents the first product. Calculate the delay ratio. The closer the value is to 1, the closer the processor performance is to the ideal baseline state; calculate the latency multiplier. When the value is significantly greater than 1, it indicates that the processor pipeline has stalled or reduced its frequency.

[0036] Furthermore, considering that the noisy neighbor effect in virtualized environments typically manifests as pollution of the last-level cache (LLC) and preemption of memory bandwidth, this interference may occur independently of CPU scheduling. Therefore, it is necessary to introduce independent memory access probes for orthogonal probing.

[0037] To ensure that the probe and the main task experience the exact same hardware resource contention, the probe is bound to the CPU core affinity context of the main task and configured to execute concurrently and alternately with the main task. The probe directly calls the on-chain standard operator library (…). The standard memory access operator in the probe memory access algorithm has a fixed workload, denoted as the preset memory access workload. The node records the probe execution time of this probe task. .

[0038] Furthermore, the product of the preset memory access task volume and the reference memory access time rate is recorded as the second product; the ratio of the probe execution time to the second product is determined as the memory access latency ratio.

[0039] As a concrete example, the specific formula for calculating the memory access latency ratio is given. The memory access latency ratio can be expressed as:

[0040] in, Indicates the memory access latency ratio. Indicates the probe execution time. This indicates the preset memory access task volume. This indicates the reference memory access time rate.

[0041] This represents the second product. The memory access latency multiplier obtained in the above manner is used to characterize the congestion multiple of the memory subsystem relative to the inherent baseline.

[0042] Thus, the computation latency ratio and memory access latency ratio have been obtained through the above methods.

[0043] Step S3: Combine the calculated latency multiplier and the memory access latency multiplier to determine the resource contention difference index; obtain the isolation index based on all resource contention difference indices within the current time period.

[0044] Hardware aging or thermal throttling typically causes CPU and memory bus frequencies to decrease synchronously, resulting in a proportional increase in both compute latency and memory access latency. However, in shared virtualization environments, external tenants often preempt memory bandwidth far more than CPU time slices, leading to uncorrelated orthogonal changes in the two latency ratios. Therefore, this embodiment will eliminate hardware common-mode interference by performing mathematical transformations on the compute latency and memory access latency ratios, extracting features solely caused by incomplete virtualization isolation.

[0045] Considering that performance degradation physically manifests as a multiple relationship, direct linear subtraction cannot accurately measure the difference across different orders of magnitude. Therefore, this embodiment employs a logarithmic domain difference model to calculate the resource contention difference index.

[0046] Specifically, nonlinear transformations are performed on the computation latency multiplier and the memory access latency multiplier respectively; the difference between the transformed computation latency multiplier and the transformed memory access latency multiplier is used as an indicator of resource contention difference.

[0047] As a concrete example, the specific formula for calculating the resource contention difference index is given. The resource contention difference index can be expressed as:

[0048] in, Indicators representing differences in resource competition. Indicates the memory access latency ratio. This indicates the calculation of the delay ratio. This represents the logarithmic function with the natural constant as the base. Indicates the absolute value sign.

[0049] This represents the transformed memory access latency multiplier. This represents the calculation delay factor after the transformation. This represents the difference between the transformed computation latency multiplier and the transformed memory access latency multiplier, which is also a resource contention difference index.

[0050] When the resource contention difference index approaches 0, it indicates that the memory access latency multiplier is approximately equal to the compute latency multiplier. This means that compute and memory access performance have experienced proportional degradation, consistent with the normal thermal throttling characteristics of a dedicated physical machine under load. When the resource contention difference index is significantly greater than 0, it indicates a significant imbalance in the performance degradation of the two channels. This corresponds to the decoupling characteristics of resource contention in a virtualization environment. For example, if there is sufficient CPU time slice but the cache is exhausted by neighboring tenants, it leads to a surge in memory access latency multiplier while the compute latency multiplier changes relatively little.

[0051] By using the above methods, we can obtain indicators of resource competition differences at different times during the execution of scientific research tasks.

[0052] To eliminate sporadic noise caused by operating system process scheduling jitter in a single measurement and to avoid frequent flickering of permission status, this embodiment will perform time-domain integration on the resource contention difference index.

[0053] Specifically, the arithmetic mean of all resource contention difference indices corresponding to candidate nodes within the current time period is calculated as the isolation index of candidate nodes in the current time period. The isolation index is used to characterize the average degree of resource decoupling of nodes within the current time period. Here, the current time period is the time period consisting of all historical moments whose time interval with the current moment is less than or equal to a preset duration. In this embodiment, the preset duration is 10 seconds. In specific applications, the implementer can set it according to the specific situation.

[0054] Step S4: Determine whether to perform permission changes on the node based on the isolation index, and upload the judgment basis and change record to the blockchain for evidence storage.

[0055] In this embodiment, the isolation index is determined in step S3. Next, the relationship between the isolation index and the preset isolation compliance threshold will be compared to determine whether to perform permission changes on the node.

[0056] Specifically, if the isolation index is greater than the preset isolation compliance threshold, the significant decoupling characteristic indicates that the node is experiencing uncontrollable side-channel resource contention, physical isolation has failed, and the corresponding node is determined to be in an uncontrolled sharing state. If the isolation index is less than or equal to the preset isolation compliance threshold, the computation and memory access performance remain linearly coupled, the physical environment meets the security baseline, and the corresponding node is determined to be in an exclusive resource state, without changing the node's execution permissions. The preset isolation compliance threshold is determined based on the statistical distribution of background noise of trusted nodes across the network under no-load baseline conditions, for example, using the 99th percentile.

[0057] When a node is in an uncontrolled shared state, a circuit breaker operation is triggered to suspend the execution permissions of the smart contract. Specifically, the smart contract sends a command to the Key Management Center (KMS) to trigger a multicast key update process, removing the node from the active key distribution list and preventing it from decrypting subsequently distributed raw data blocks. The smart contract also modifies the node's attribute label in the task scheduler, downgrading it to a restricted node. In subsequent task distributions, the scheduler will only distribute anonymized computation tasks marked as "Homomorphic Encryption (HE)" or "Differential Privacy (DP)" to it. This method ensures that when a node's physical environment becomes untrusted, its access to sensitive plaintext can be forcibly stripped through technical means, while preserving its channel for contributing computing power, thus achieving a balance between security and efficiency.

[0058] To meet the auditing requirements of scientific research data management, the criteria for judgment can be solidified to form a complete chain of evidence. Specifically, a governance transaction is constructed, whose data payload includes: the target node ID, the timestamp that triggered the circuit breaker, a snapshot of the isolation index value as the direct basis for the penalty, and the type of permission change action executed. This transaction is broadcast and packaged into the blockchain ledger.

[0059] When subsequent security audits are conducted or nodes file appeals, the auditors can retrieve the transaction and trace the physical environment fingerprint of the node when it was blocked. The isolation metric objectively reflects the resource contention situation at the time, eliminating interference factors such as hardware failure or aging, thus providing irrefutable objective attribution evidence for changes in data permissions.

[0060] Thus, the method provided in this embodiment has enabled full-process traceability and management of scientific research data.

[0061] This embodiment obtains the first and second exclusive execution times of a node executing standard computing and memory access tasks under exclusive resource status, providing an accurate and reproducible hardware baseline for subsequent performance evaluation and effectively avoiding misjudgments of permissions due to differences in node hardware. Combining the workload and execution time of business computing tasks during research task execution, as well as the preset memory access workload and probe execution time corresponding to the memory probe task, the computation latency multiplier and memory access latency multiplier are calculated respectively. This process quantitatively compares the actual performance of the node in the real operating environment with the hardware baseline, accurately identifying the degree of performance degradation caused by environmental changes, and providing reliable data support for dynamic permission adjustments. By comprehensively calculating the latency multiplier and memory access latency multiplier... By identifying resource contention difference indicators and obtaining isolation indicators based on all resource contention difference indicators within the current time period, a quantitative representation of the isolation status of the node's operating environment is achieved. This indicator objectively reflects whether a node is in a controlled, exclusive resource state, thus providing a scientific basis for permission changes. The mechanism determines whether to perform permission changes on a node based on the isolation indicator, and records the determination criteria and changes on the blockchain for evidence storage. This mechanism binds permission control to the real-time operating environment status, achieving dynamic protection of sensitive scientific research data and effectively avoiding the risk of data leakage due to node environment drift. Simultaneously, all determination criteria and change records are stored on the blockchain for evidence storage, ensuring the entire permission management process is traceable and tamper-proof, thus improving the security and compliance of scientific research data management.

[0062] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A blockchain-based method for end-to-end traceability and access control of scientific research data, characterized in that, The method includes the following steps: Get the first exclusive execution time of the node executing the standard computing task issued by the on-chain standard operator library under exclusive resource state, and the second exclusive execution time of executing the standard memory access task; Based on the preset workload of the standard computation task, the preset workload of the standard memory access task, the first exclusive execution duration, the second exclusive execution duration, the workload of the business computation task during the execution of the research task, the business execution duration, and the preset memory access task and probe execution duration corresponding to the memory probe task, the computation latency ratio and the memory access latency ratio are obtained respectively. This includes: obtaining a reference computation time consumption rate for the candidate node based on the preset workload of the standard computation task and the first exclusive execution duration of the candidate node; calculating a first product of the business computation task workload and the reference computation time consumption rate; and determining the computation latency ratio as the ratio of the business execution duration to the first product. The business computation task workload and business execution duration are obtained based on the main task mapping; the candidate node can be any node; obtaining a reference memory access time consumption rate for the candidate node based on the preset workload of the standard memory access task and the second exclusive execution duration of the candidate node; calculating a second product of the preset memory access task workload and the reference memory access time consumption rate; and determining the memory access latency ratio as the ratio of the probe execution duration to the second product. By combining the calculated latency multiplier and the memory access latency multiplier, a resource contention difference index is determined; based on all resource contention difference indices in the current time period, an isolation index is obtained, including: calculating the arithmetic mean of all resource contention difference indices corresponding to each node in the current time period, which is used as the isolation index of each node in the current time period. The system determines whether to change permissions on a node based on the isolation level index, and records the determination criteria and changes on the blockchain for evidence storage.

2. The method for full-process traceability and access control of scientific research data based on blockchain according to claim 1, characterized in that, The memory access probe task is forcibly bound to the processor core where the main task resides, and is executed concurrently and alternately with the main task.

3. The method for full-process traceability and access control of scientific research data based on blockchain according to claim 1, characterized in that, The method of combining the calculated latency multiplier and the memory access latency multiplier to determine the resource contention difference index includes: The computational latency multiplier and the memory access latency multiplier are respectively subjected to nonlinear transformations; The difference between the transformed computation latency multiplier and the transformed memory access latency multiplier is used as an indicator of resource contention difference.

4. The blockchain-based method for full-process traceability and access control of scientific research data as described in claim 1, characterized in that, The method of determining whether to perform permission changes on nodes based on isolation index includes: If the isolation index is greater than the preset isolation compliance threshold, the corresponding node is determined to be in an uncontrolled sharing state, triggering a circuit breaker operation for smart contract execution permissions. If the isolation index is less than or equal to the preset isolation compliance threshold, the corresponding node is determined to be in an exclusive resource state, and the node's execution permissions are not changed.

5. The blockchain-based method for full-process traceability and access control of scientific research data according to claim 4, characterized in that, The permission circuit breaker operation includes: revoking the data decryption key of the node and downgrading the node to a restricted mode for receiving encrypted computation tasks.

Citation Information

Patent Citations

  • Data classification processing method and system for parallel computing of multi-core processor

    CN120508936A

  • Distributed password service management and monitoring platform software and implementation method

    CN121907430A