A method and related system for trusted verification

By introducing an aggregation node into the computer system to aggregate measurement data from multiple measurement nodes, the capacity limitation problem of the trusted management center is solved, data compression and management are simplified, and it is suitable for third-party trusted management centers.

CN123116276APending Publication Date: 2026-11-17HUAWEI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510638967.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The Trusted Management Center supports a limited number of metric nodes, and the size of the data center exceeds the capacity of the Trusted Management Center, making it difficult to meet business needs.

Method used

By introducing aggregation nodes into the computer system, the measurement data from multiple measurement nodes are aggregated to obtain the aggregation result, which is then sent to the trusted management node, thus solving the problem of the limited number of measurement nodes supported by the trusted management center.

Benefits of technology

By compressing data volume, the storage requirements of trusted management nodes are reduced, the management of a large number of metric nodes is simplified, it is suitable for third-party trusted management centers, and the deployment cost is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116276A_ABST
    Figure CN123116276A_ABST
Patent Text Reader

Abstract

A trusted verification method is applied to a computer system, which includes a aggregation node and multiple measurement nodes. The measurement nodes are computing devices within the computer system that perform trusted measurements. The method includes: a first aggregation node in at least one aggregation node acquiring measurement data for trusted measurements of the multiple measurement nodes; the first aggregation node aggregating the measurement data from the multiple measurement nodes to obtain an aggregation result, the volume of which is less than the sum of the volume of the measurement data from the multiple measurement nodes; and the first aggregation node sending the aggregation result to a trusted management node, whereby the aggregation result is used to perform trusted verification of the multiple measurement nodes. This method, by introducing an aggregation node to aggregate the measurement data from multiple measurement nodes and reporting the aggregation result, solves the problem of a limited number of measurement nodes supported by a trusted management center and the problem of data center scale exceeding the capacity of the trusted management center, thus meeting business needs.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of trusted computing technology, and in particular to a trusted verification method, computer system, computer-readable storage medium, and computer program product. Background Technology

[0002] Trusted computing (TC) is a technology promoted by the Trusted Computing Group (TCG) to ensure the integrity of computer systems and software (such as applications). Trusted computing reduces the risk of attacks due to the use of unknown or tampered software by introducing a measurement-verification mechanism. Measurement involves collecting data on the state of the computer system or software being monitored. Verification compares the measurement results with a benchmark value. If they match, the verification passes, indicating that the computer system or software is intact; if they do not match, the verification fails, indicating that the computer system or software has been tampered with.

[0003] Currently, trusted computing is widely used in computer systems such as terminals and servers. Each computer system includes one or more measurement points, which are the basic measurement objects, including but not limited to various computing cards in the computer system, such as a single central processing unit (CPU), graphics processing unit (GPU), or neural network processing unit (NPU). The measurement content includes the firmware corresponding to the computing card and other auxiliary information. Among them, the auxiliary information may include at least one of the hardware's identity identifier or manufacturer identifier.

[0004] Data centers typically comprise a large number of computer systems. To facilitate management, current data centers employ trusted management centers (TLCs) to manage these multi-processor computer systems. In the trusted management process, each metric node initiates its own trusted startup and is then managed by the TLC one by one. However, the number of metric nodes that a TLC can manage is limited, and the deployment scale of data centers is rapidly increasing with the growth of computing services. Some data centers can even reach the scale of millions of cards, exceeding the capacity of the TLC and making it difficult to meet business demands. Summary of the Invention

[0005] This application provides a trusted verification method that introduces an aggregation node into a computer system to aggregate measurement data from multiple measurement nodes, thereby compressing the data volume. The aggregation result is then sent to a trusted management node (e.g., a trusted management center). This method solves the problems of limited support for the number of measurement nodes by the trusted management center and the data center's scale exceeding its capacity, thus meeting business needs. This application also provides a computer system, a computer-readable storage medium, and a computer program product.

[0006] Firstly, this application provides a method for trusted verification. This method is applied to a computer system. The computer system includes at least one aggregation node and multiple measurement nodes. Measurement nodes are computing devices within the computer system that perform trusted measurements. As a basic measurement object, a measurement node can be at the card level, such as a computing card within the computer system.

[0007] In practice, at least one aggregation node acquires metric data from multiple metric nodes for trusted metric measurements. This first aggregation node then aggregates the metric data from the multiple metric nodes to obtain an aggregation result. The size of the aggregation result is less than the sum of the metric data from the multiple metric nodes. In other words, the first aggregation node achieves metric data compression by aggregating the metric data. The first aggregation node can then send the aggregation result to the trusted management node, where it is used to verify the trustworthiness of the multiple metric nodes.

[0008] In this method, the trusted management node stores the aggregation results, which have a smaller data volume, instead of storing the larger metric data. Therefore, it can manage more metric nodes, solving the capacity limitation problem. Moreover, the aggregated metric nodes are treated as a single metric node (which can be called a virtual metric node or virtual node) by the trusted management node. The trusted management node processes the aggregated metric nodes as a single metric node, thereby simplifying the management of a large number of metric nodes in the system.

[0009] In some possible implementations, the measurement data includes the identity identifier and measurement value of the measurement node. The identity identifier uniquely identifies the measurement node, and the measurement value characterizes the state of the measurement node, such as its integrity status or whether it has been tampered with. Accordingly, the first aggregation node can use the identity identifier and measurement value of each of the multiple measurement nodes as the coordinates of each node, and fit the coordinates of the multiple measurement nodes to obtain a polynomial with the measurement values ​​as function values ​​and the identity identifier as function variables. The coefficients of the polynomial are then used as the aggregation result of the measurement data from the multiple measurement nodes. This aggregation result is used to construct a polynomial for trusted verification. The polynomial for trusted verification can be a polynomial obtained by restoring the coordinate fitting, used for trusted verification.

[0010] In this method, the first aggregation node converts identity identifiers and measurement values ​​into coordinates, aggregates the measurement data by coordinate fitting, and uses the coefficients of the fitted polynomial as the aggregation result. This allows the use of the coefficients of a polynomial with a smaller data volume to replace the measurement data for subsequent trust verification, reducing the demand of the trust management node on storage resources or memory.

[0011] In some possible implementations, the first aggregation node can be fitted based on the coordinates of multiple metric nodes using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermitian interpolation, finite field-based interpolation, or aggregate signatures of multiple samples.

[0012] Newton interpolation is an interpolation method that approximates known data points by constructing a Newton's difference quotient polynomial. Known data points can be data points formed by the measurement values ​​and identifiers of multiple measurement nodes, with each measurement value and identifier forming a data point. Newton interpolation is based on the concept of Newton's difference quotient, using the difference quotients of known data points to construct the interpolation polynomial. Lagrange interpolation is a method that directly constructs the interpolation polynomial using known data points. Lagrange interpolation is simple and intuitive in form, and convenient for theoretical analysis. Compared to Lagrange interpolation, Newton interpolation only requires adding one term to the original polynomial when adding new data points, resulting in relatively less computation and eliminating the need to recalculate all coefficients. Centroid Lagrange interpolation is an improved form of Lagrange interpolation. It simplifies calculations by introducing centroid weights, making the interpolation process more stable and efficient. During calculation, centroid Lagrange interpolation reduces the accumulation of numerical errors, especially noticeable in higher-order interpolations. Hermitian interpolation requires not only that the interpolating polynomial take known values ​​(equal function values) at given data points, but also that its derivatives at these points are equal. This allows Hermitian interpolation to better approximate functions with specific smoothness requirements. Finite field-based interpolation is an improved interpolation method. A finite field, also known as a Galois field, is a set of finite elements that satisfies all the axioms of a field: addition and multiplication are defined for each element, and the associative, commutative, and distributive laws of addition and multiplication apply. It also has an additive identity (usually denoted as 0) and a multiplicative identity (usually denoted as 1), and each element has an additive inverse, while non-zero elements have a multiplicative inverse. The use of additive or multiplicative inverses simplifies computation.

[0013] Multi-sample aggregate signature is a technique that combines the signatures of multiple signers (such as metric nodes) on different messages (such as coordinates derived from metric values ​​and identity identifiers) into a single signature. This significantly reduces the storage space and transmission bandwidth required for signatures. Specifically, multiple signers each sign their own messages, and then these signatures are aggregated together using a specific algorithm to form an aggregate signature. A verifier can use this aggregate signature along with all the original messages to verify the validity of all the signers' signatures.

[0014] This method supports multiple ways to achieve coordinate fitting, which can meet different business needs and has high flexibility.

[0015] In some possible implementations, the trusted management node can receive a baseline value acquisition request, which is used to obtain aggregation results and send the aggregation results to the verification node. In this application, the aggregation results serve as the baseline value, or the aggregation results can be used to obtain the baseline value. The first aggregation node receives a truth value acquisition request, which is used to obtain the measurement data of at least one of the multiple measurement nodes during runtime. Then, the first aggregation node sends the measurement data of at least one measurement node during runtime to the verification node, so that the verification node can perform trusted verification based on the measurement data of at least one measurement node during runtime and the aggregation results.

[0016] When the aggregation results are obtained using coordinate fitting, the aggregation results may include polynomial coefficients. In this case, the aggregation results are used to construct the polynomial for trust verification. The verification node can convert the metric data from at least one metric node into coordinates and verify whether the coordinates satisfy the polynomial, i.e., whether the data point corresponding to the coordinate lies on the curve corresponding to the polynomial. If the coordinates satisfy the polynomial, or the data point corresponding to the coordinate lies on the curve corresponding to the polynomial, the trust verification passes. If the coordinates do not satisfy the polynomial, or the data point corresponding to the coordinate does not lie on the curve corresponding to the polynomial, the trust verification fails.

[0017] In this method, the trusted management node can save the aggregation results so that the verification node can perform trusted verification based on the aggregation results. Therefore, the method of this application does not require modification of the trusted management node's workflow, can be applied to third-party trusted management centers, and has high availability. Moreover, since the trusted management node saves the aggregation results rather than the signed policy file, even if the metric node in the computer system changes, the trusted management node does not need to re-sign or re-verify the signature, reducing the deployment cost of the trusted management node.

[0018] In some possible implementations, the first aggregation node corresponds to multiple second aggregation nodes, and the second aggregation nodes may include aggregation nodes directly connected to the metric node. The multiple metric nodes include multiple first metric nodes. The computer system also includes multiple second metric nodes. The first metric nodes may include the aforementioned second aggregation nodes.

[0019] Each of the multiple second aggregation nodes can also aggregate the measurement data of a portion of the multiple second measurement nodes to obtain a second aggregation result. Then, each of the multiple second aggregation nodes sends the second aggregation result to the first aggregation node. Correspondingly, the first aggregation node can aggregate the measurement data of multiple first measurement nodes to obtain a first aggregation result. The first aggregation node sends the first aggregation result and the second aggregation result to the trusted management node.

[0020] In this method, the first aggregation node can transmit the second aggregation result to the trusted management node. The trusted management node can store the full aggregation result. During trusted verification, the required aggregation result can be obtained from the trusted management node for trusted verification, simplifying the trusted verification process. Moreover, having the first aggregation node uniformly report the aggregation result can reduce the number of connections to the trusted management node and reduce network resource consumption.

[0021] In some possible implementations, the first aggregation node can merge the second aggregation result with its own metric data to obtain merged metric data. The first aggregation node then aggregates the merged metric data to obtain the first aggregation result. Specifically, the first aggregation node can use a hash method to extend the second aggregation result (or metric report) from the second aggregation node and embed it into its own metric value, thereby enabling the first aggregation result to be sent to the trusted management node.

[0022] When the first aggregation node's metric data (such as metric values) embeds the second aggregation node's report content (such as the second aggregation result), the first aggregation node does not need to send the second aggregation result separately. Thus, compared to the scheme where the first aggregation node reports both the first and second aggregation results, the trusted management node can use less memory to store the aggregation results.

[0023] In some possible implementations, a first aggregation node corresponds to multiple second aggregation nodes, and multiple metric nodes include multiple first metric nodes. The computer system also includes multiple second metric nodes. Each first metric node includes a second aggregation node. Accordingly, the first aggregation node can aggregate the metric data from multiple first metric nodes to obtain a first aggregation result; the first aggregation node sends the first aggregation result to the trusted management node. Each of the multiple second aggregation nodes aggregates the metric data from a subset of the multiple second metric nodes to obtain a second aggregation result. Each of the multiple second aggregation nodes sends the second aggregation result to the trusted management node.

[0024] In this method, each aggregation node can send its aggregation results to the trusted management node without having to report to the next level, thus shortening the reporting time and improving the reporting efficiency.

[0025] In some possible implementations, the aggregation node is a baseboard management controller in a computer system. A baseboard management controller is an embedded system primarily used for the management and monitoring of computer systems such as servers, workstations, and mainframes. The baseboard management controller operates independently of the host operating system, continuing to function even if the host system is shut down or crashes, enabling low-level control and status monitoring of the hardware.

[0026] This method introduces a baseboard management controller as an aggregation node, which can naturally aggregate measurement data from measurement nodes such as computing cards controlled by the baseboard management controller. It can continue to run even if the host system is shut down or crashes, thus having high reliability.

[0027] Secondly, this application provides a computer system. The computer system includes at least one aggregation node and multiple measurement nodes, wherein each measurement node is a computing device in the computer system that performs trust measurement, including:

[0028] The first aggregation node among the at least one aggregation node is used to acquire measurement data for trust measurement of the plurality of measurement nodes; aggregate the measurement data of the plurality of measurement nodes to obtain an aggregation result, wherein the data volume of the aggregation result is less than the sum of the data volumes of the measurement data of the plurality of measurement nodes; and send the aggregation result to the trust management node, wherein the aggregation result is used to perform trust verification on the plurality of measurement nodes.

[0029] In some possible implementations, the metric data includes the identity identifier and metric value of the metric node, whereby the metric value characterizes the state of the metric node, and the first aggregation node is specifically used for:

[0030] Using the identity identifier and metric value of each of the multiple metric nodes as the coordinates of each of the multiple metric nodes, a polynomial is obtained by fitting the coordinates of the multiple metric nodes, with the metric value as the function value and the identity identifier as the function variable. The coefficients of the polynomial are used as the aggregation result of the metric data of the multiple metric nodes. The aggregation result is used to construct a polynomial for trusted verification.

[0031] In some possible implementations, the first aggregation node is specifically used for:

[0032] Based on the coordinates of the multiple metric nodes, fitting is performed using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermitian interpolation, finite field-based interpolation, or aggregated signatures of multiple samples.

[0033] In some possible implementations, the trusted management node is used for:

[0034] Receive a baseline value acquisition request, the baseline value acquisition request being used to acquire the aggregation result, and send the aggregation result to the verification node;

[0035] The first aggregation node is also used for:

[0036] A real value acquisition request is received, which is used to acquire the metric data of at least one of the plurality of metric nodes during operation, and the metric data of the at least one metric node during operation is sent to the verification node so that the verification node performs trusted verification based on the metric data of the at least one metric node during operation and the aggregation result.

[0037] In some possible implementations, the first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system further includes multiple second metric nodes, the first metric nodes including the second aggregation nodes;

[0038] Each of the plurality of second aggregation nodes is used for:

[0039] The measurement data of some of the multiple second measurement nodes are aggregated to obtain a second aggregation result, and the second aggregation result is sent to the first aggregation node;

[0040] The first aggregation node is specifically used for:

[0041] The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result;

[0042] Send the first aggregation result and the second aggregation result to the trusted management node.

[0043] In some possible implementations, the first aggregation node is specifically used for:

[0044] The second aggregation result is merged with the metric data of the second aggregation node to obtain the merged metric data;

[0045] The merged metric data is aggregated to obtain the first aggregation result.

[0046] In some possible implementations, the first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system further includes multiple second metric nodes, the first metric nodes including the second aggregation nodes;

[0047] The first aggregation node is specifically used for:

[0048] The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result;

[0049] Send the first aggregation result to the trusted management node;

[0050] Each of the plurality of second aggregation nodes is used for:

[0051] The measurement data of a portion of the multiple second measurement nodes are aggregated to obtain a second aggregation result;

[0052] The second aggregation result is sent to the trusted management node.

[0053] In some possible implementations, the aggregation node is a baseboard management controller in the computer system.

[0054] Thirdly, this application provides a computer-readable storage medium storing instructions that instruct a computer system to perform the trusted verification method described in the first aspect or any implementation thereof.

[0055] Fourthly, this application provides a computer program product containing instructions that, when run on a computer system, causes the computer system to perform the trusted verification method described in the first aspect or any implementation thereof.

[0056] Based on the implementation methods provided in the above aspects, this application can be further combined to provide more implementation methods. Attached Figure Description

[0057] To more clearly illustrate the technical methods of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below.

[0058] Figure 1 A schematic diagram illustrating the aggregation of metric values ​​from multiple metric nodes by a trusted management node, as provided in this application;

[0059] Figure 2A A schematic diagram of a server architecture provided for this application;

[0060] Figure 2B A schematic diagram of a data center architecture provided for this application;

[0061] Figure 3 A flowchart of a trusted verification method provided in this application;

[0062] Figure 4 A schematic diagram of a curve fitted by interpolation method provided in this application;

[0063] Figure 5 A schematic diagram illustrating how a single proxy node aggregates metric values ​​from multiple metric nodes, as provided in this application;

[0064] Figure 6 A schematic diagram of a hierarchical aggregation provided for this application;

[0065] Figure 7 This application provides a schematic diagram of the coordinates of a metric node and the aggregation results. Detailed Implementation

[0066] The terms "first" and "second" used in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first" and "second" may explicitly or implicitly include one or more of that feature.

[0067] First, some technical terms involved in the embodiments of this application will be introduced.

[0068] Trusted computing (TC), also known as reliable computing, is a security technology promoted and developed by the Trusted Computing Group (TCG) to ensure the reliability and integrity of computer systems. Trusted computing originates from the concept of a trusted system. Trustworthiness means that the system's behavior more fully follows its design, and the probability of it performing actions prohibited by the designers and software developers is very low. An entity (such as a computing component in a computer system) is considered trustworthy if it consistently achieves its intended goals in the expected manner.

[0069] Trusted computing introduces a measurement-verify mechanism to reduce the risk of attacks due to the use of unknown or tampered software. Measurement-verify specifically refers to collecting the state of the system (e.g., a computer system such as a terminal or server) or software being tested. In trusted computing, this is also called "trusted measurement." It typically combines cryptographic techniques, such as hash functions, to obtain a representation of the state. In other words, the state can be the hash value of a key component of the computer system or software (such as firmware, operating system kernel, application code, etc.). Verification, also known as trusted verification in trusted computing, usually involves comparing the measurement result (e.g., the measurement value) with a baseline value (or reference value). If the measurement value matches the baseline value, the verification passes, indicating that the system or software is intact; if the measurement value does not match the baseline value, the verification fails, indicating that the system or software has been tampered with.

[0070] Trusted boot, including trusted boot based on trusted computing hardware, includes, but is not limited to, trusted platform modules (TPMs). Trusted boot based on trusted computing hardware establishes a root of trust at the hardware level, sequentially measuring critical components during the boot process, such as the basic input / output system (BIOS) firmware, bootloader, and operating system kernel, and storing the measurement values ​​in the Platform Configuration Registers (PCR). Trusted boot based on trusted computing hardware can only record the integrity of components at the moment of measurement; it cannot prevent tampering by malicious software after measurement, nor can it guarantee that the measurement process itself is not interfered with. Therefore, it cannot guarantee the absolute trustworthiness of the boot process. To address this, measurement reports can be used to ensure the trustworthiness of the boot process. A measurement report refers to the secure reporting of measurement logs, including measurement values, to third-party challengers. Third-party challengers can be independent security audit organizations, remote servers, etc. Third-party challengers can perform secondary verification and evaluation of the system boot process, enhancing the reliability of the system's trustworthiness assessment.

[0071] Each computer system includes one or more measurement points, which are computing devices within the system that perform trust measurements. Measurement points serve as the basic measurement objects, and the granularity of these objects can be at the card level. For example, measurement objects can include, but are not limited to, various computing cards within the computer system. Specifically, measurement objects can include a single central processing unit (CPU), graphics processing unit (GPU), or neural network processing unit (NPU) within the computer system. The measurement content can include the firmware corresponding to the computing card and other auxiliary information. The auxiliary information can include at least one of the following: hardware identification or vendor identification.

[0072] Data centers typically comprise a large number of computer systems. These systems can include servers with heterogeneous processors (also known as heterogeneous servers) or servers with homogeneous processors (also known as homogeneous servers). For ease of management, current data centers utilize Trusted Management Centers (TMCs) to manage these computer systems. In the TMC process, each metric node is independently initiated with trust and then managed by the TMC one by one. The TMC is essentially a server that performs trusted management; it can only support a limited number of metric nodes. The deployment scale of data centers is rapidly increasing with the growth of computing services, and data centers can reach the million-card level, exceeding the capacity of the TMC.

[0073] To address this, the industry has developed a policy-based mechanism that centralizes the measurement data from all measurement nodes on a computer system (such as a terminal or a server) and submits it to a trusted management center on the control plane for signing. The remote verifier only needs to verify once, eliminating the need to verify every measurement node on the computer system, thus reducing computational and communication overhead. The remote verifier can be a user-side trusted verification node or a third-party challenger such as a security auditing agency.

[0074] Specifically, the Trusted Management Center (TMC) on the control plane can derive a remote proof policy file based on the pre-installed software and hardware model of the computer system. The hardware model includes the organizational relationships of the hardware within the computer system. For example, when the computer system is a server, the hardware model can include the organizational relationships of hardware such as the CPU and NPU. The remote proof policy file can serve as a benchmark for trusted verification by remote provers. The TMC can sign the remote proof policy file to obtain a signed policy file. The TMC can store the signed policy file in the corresponding computer system. During task distribution, remote provers verify the status of the computer system based on the signed policy file and the measurement values ​​of each metric node in the computer system. For example, a remote prover can first request the TMC to verify the signed policy file. Signature verification is a core technology in the field of information security, typically employing cryptographic methods to ensure the integrity, authenticity, and non-repudiation of data. When signature verification passes, it indicates that the policy file is complete, authentic, and non-repudiable. Remote provers can then perform trusted verification based on the measurement values ​​of each metric node and the policy file (as a benchmark). The remote verifier processes the metrics from each metric node to obtain the overall metric of the computer system, and then compares this overall metric with a benchmark value. If they match, the computer system's trust verification passes. If they don't match, the trust verification fails. Only when the trust verification passes will the task be sent to the computer system for execution.

[0075] The above solution merely collects and signs the baseline values ​​of each metric node (such as each computing component) in the computer system to form a policy file. During trusted verification, verification only needs to be performed once, eliminating the need to verify each metric node separately. While this solution reduces the computational and communication overhead of trusted verification, it does not address the capacity limitations of the trusted management center. Furthermore, the above solution requires modification of the entire workflow, such as changing the trusted management center's processing flow for baseline values, making it difficult to apply to third-party trusted management centers.

[0076] In view of this, this application provides a method for trusted verification. This method is applied to trusted computing management operations. For a computer system comprising multiple measurement nodes, a convergence node is introduced to aggregate the measurement data from the multiple measurement nodes, obtaining an aggregation result. The size of the aggregation result is smaller than the sum of the size of the measurement data from the multiple measurement nodes. This achieves measurement data compression. The convergence node then sends the aggregation result to a trusted management node (e.g., a trusted management center). This aggregation result is used for trusted verification of the multiple measurement nodes. By storing the smaller aggregation result, the trusted management node does not need to store the larger measurement data, thus managing a larger number of measurement nodes and solving the capacity limitation problem.

[0077] Moreover, the multiple aggregated metric nodes are treated as a single metric node by the trusted management node. The trusted management node processes the aggregated metric nodes as a single metric node. For example, the trusted management node can save the aggregation results so that the verification node can perform trusted verification based on the aggregation results. Therefore, the method of this application does not require modification of the trusted management node's workflow, can be applied to third-party trusted management centers, and has high availability.

[0078] like Figure 1 As shown, from the perspective of the trusted management node 100, multiple measurement nodes 202 are interconnected through a convergence algorithm, forming a unified "virtual node 200". The virtual node 200 includes one or more convergence nodes 204, typically an onboard baseboard management controller (BMC) or a rack-mounted BMC, responsible for performing the convergence algorithm on the measurement data of some or all of the measurement nodes 202 within the virtual node 200 to obtain the convergence result. The convergence node 204 can also report the convergence result to the trusted management node 100. This overcomes the capacity limitations of the trusted management node 100, supporting scalable management services and remote authentication for measurement nodes. Furthermore, the converged content can also include an identifier (ID), which uniquely identifies the measurement node 202 and can be used to locate the measurement node during error tracing. It should be noted that in this application, the convergence node 204 can act as a proxy for the measurement node 202 to interact with the trusted management node 100; therefore, the convergence node 204 can also be called a proxy or proxy node.

[0079] The trusted verification method provided in this application can be applied to a single computer system, such as a single server or a single terminal. Alternatively, the trusted verification method provided in this application can also be applied to a data center or computing device cluster formed by multiple computer systems.

[0080] First, see Figure 2A The diagram illustrates the architecture of a server 20, which includes a baseboard management controller 22 and various heterogeneous processors, including a CPU 24, a GPU 26, and an NPU 28. Heterogeneity refers to the different types of computing cores within the processors; these different types of cores have unique architectures, functionalities, and performance characteristics, and can work together to complete different types of tasks. Software can also be deployed or run on the processors. For example, the CPU 24 can also deploy or run BIOS firmware, an operating system (OS), and applications.

[0081] Each processor (such as CPU24, GPU26, and NPU28) can act as a measurement node. Upon trusted startup, each measurement node measures its own state and obtains measurement data. This measurement data can be stored in a measurement report. The measurement report can also record the identity of the measurement node. Furthermore, the measurement report can also include the measurement time.

[0082] The substrate management controller 22 can act as a convergence node. Metric nodes managed or controlled by the substrate management controller 22 can centralize metric reports, including metric data, at this convergence node. The convergence node executes a convergence algorithm on the metric reports, including metric data, and obtains the convergence result. The convergence node can verify the convergence result, thereby maintaining the metric nodes based on the verification result and providing services externally. It should be noted that if the substrate management controller 22 is attached to another substrate management controller, that substrate management controller 22 can also be considered a metric node, and the metric data of that metric node can be converged by the other basic management controller mentioned above.

[0083] It should be noted that the aggregation node can send aggregation results to the trusted management node, and the aggregation results are used to perform trusted verification on multiple metric nodes.

[0084] Next, see Figure 2B The diagram shows an architecture of a data center, which includes multiple servers 20, each server 20 including one or more baseboard management controllers 22. Figure 2B Each server in the example includes a baseboard management controller 22. For ease of management, the data center also introduces a trusted management node 100 to perform trusted computing management on multiple servers 20 and support external remote attestation services.

[0085] Remote proof is a technique used to verify the trustworthiness of the hardware and software status of a remote computer system (the prover). The specific verification process may include the following steps: 1) The prover (i.e., the challenger, such as a cloud service customer, regulator, trusted management node, or other system administrator) sends a proof request to the prover (such as a server, IoT terminal, or other computer system), specifying the component to be verified (such as a computing card, firmware, or operating system); 2) The prover obtains the metric values, i.e., the status, of the components in the computer system through trusted computing hardware (such as a TPM) and generates a proof report; 3) The prover sends the signed report to the prover; 4) The prover verifies the signature and the proof report to determine whether the component to be verified is trustworthy.

[0086] In this application, the trusted management node 100 can directly communicate with the baseboard management controller 22 to manage the aggregation results uploaded by the baseboard management controller 22. In addition, the organization methods between different baseboard management controllers 22 are not entirely the same. They can communicate directly with the trusted management node 100, or multiple baseboard management controllers 22 can select an aggregation node for aggregation, and the aggregation node communicates with the trusted management node 100.

[0087] It should be noted that, Figure 2A , Figure 2B The example of a computer system including heterogeneous processors is used for illustration. In practical applications, a computer system may also include homogeneous processors, and this application does not impose any restrictions on this.

[0088] Figure 2A and Figure 2B The system architectures of computer systems such as servers and data centers are described separately. The trusted verification method of this application is then described below with reference to the accompanying drawings.

[0089] See Figure 3 The flowchart illustrates a trusted verification method applied to a computer system. The computer system includes at least one aggregation node and multiple measurement nodes. Measurement nodes are computing devices within the computer system that perform trusted measurements, and aggregation nodes are nodes that aggregate measurement data from multiple measurement nodes. Measurement nodes are typically at the card level. For example, the aggregation node includes... Figure 2A or Figure 2B The substrate management controller 22 shown includes a measurement node. Figure 2A or Figure 2B The processors shown are, for example, CPU24, GPU26, and NPU28. The method includes the following steps:

[0090] S302, The first aggregation node in at least one aggregation node acquires the measurement data for the trusted measurement of multiple measurement nodes.

[0091] A measurement node is a computing device in a computer system that performs trust measurements. As a basic measurement object, a measurement node includes, but is not limited to, various computing cards in a computer system, such as CPU24, GPU26, and NPU28 as shown in Figure 2. Measurement data includes data obtained by performing trust measurements on the measurement node. For example, measurement data may include measurement values. Measurement values ​​are used to represent the state of the measurement node. The state of the measurement node can be an integrity state, which can be used to determine whether the measurement node is intact or has been tampered with. The measurement value can be obtained by hashing the measurement content. The measurement content may include the firmware corresponding to the computing card and other auxiliary information. Auxiliary information may include at least one of the following: hardware-borne identity identifier or vendor identifier. Furthermore, the measurement content may also include one or more of the following: operating system (or operating system kernel), and applications.

[0092] In practice, the TPM can measure the content during the trusted startup process to obtain measurement data, and then report the measurement data. The aggregation node can receive the measurement data reported by the TPM. Alternatively, the aggregation node can proactively obtain measurement data from the TPM. For example, the aggregation node can periodically access the TPM to obtain measurement data.

[0093] S304. The first aggregation node aggregates the measurement data from multiple measurement nodes to obtain the aggregation result.

[0094] Specifically, the first aggregation node can use an aggregation algorithm to aggregate the measurement data from multiple measurement nodes to obtain an aggregation result. The amount of data in the aggregation result is less than the sum of the amounts of measurement data from the multiple measurement nodes. For example, if the amount of measurement data from each of the multiple measurement nodes is m bytes (B), the number of measurement nodes aggregated by the aggregation node is q, and the amount of data in the aggregation result is n bytes, then n is less than q × m.

[0095] In some possible implementations, the core idea of ​​the aggregation algorithm is to use a curve to represent the metric data (such as metric values) of multiple metric nodes, thereby achieving management and maintenance. Here, multiple metric nodes can be a series of metric nodes within a virtual node. Metric data can include the identity ID and metric value of the metric node. The identity ID is used to uniquely identify the metric node, such as as its name or code. The metric value represents the state of the metric node. The first aggregation node can obtain the identity IDs of multiple metric nodes, for example, by obtaining the identity IDs along with the metric values. Accordingly, the first aggregation node can aggregate the data based on the identity IDs and metric values ​​of the multiple metric nodes to obtain the aggregation result.

[0096] The first aggregation node can use the identity ID and metric value of each of the multiple metric nodes, along with the coordinates of each metric node, to fit a polynomial to the coordinates of the multiple metric nodes. This polynomial, with the metric value as the function value and the identity ID as the function variable, has its coefficients used as the aggregation result of the metric data from the multiple metric nodes. In other words, the aggregation result includes the coefficients of the polynomial expression of the curve fitted based on the coordinates of the multiple metric nodes, also known as aggregation coefficients. Furthermore, when the identity ID or metric value is a large integer, the first aggregation node can process the identity ID or metric value, for example, by performing modulo operations, to obtain a smaller integer to facilitate fitting.

[0097] For example, the metric value and identity ID of metric node i can be obtained through the following coordinates (x i ,,y i ) represents. Where x i y represents the identity ID of metric node i (or the result of processing the identity ID), i This represents the metric value of metric node i (or the result of processing the identity ID). The first aggregation node can fit the coordinates of multiple metric nodes using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermite interpolation, finite field-based interpolation, or multi-sample aggregated signatures to obtain the aggregation result. The aggregation result can be represented by the coefficients of the polynomial expression f(x) of the curve.

[0098] Next, we will explain in detail the various implementation methods of fitting the first aggregation node with the coordinates of multiple metric nodes.

[0099] Newton interpolation is an interpolation method that approximates known data points by constructing a Newton's difference quotient polynomial. In this application, the known data points can be data points formed by the metric values ​​and identity IDs of multiple metric nodes. Each metric value and identity ID can form a data point, represented as (x... i ,,y i Newton interpolation is based on the concept of Newton's difference quotient, using the difference quotient of known data points to construct an interpolation polynomial.

[0100] Lagrange interpolation is a method that directly constructs an interpolation polynomial using known data points. Lagrange interpolation is simple and intuitive in form, and convenient for theoretical analysis. Compared to Lagrange interpolation, Newton interpolation only requires adding one term to the existing polynomial when adding new data points, resulting in relatively less computation and eliminating the need to recalculate all coefficients.

[0101] Centroid Lagrange interpolation is an improved form of Lagrange interpolation. It simplifies calculations by introducing a centroid weight, making the interpolation process more stable and efficient. During calculation, centroid Lagrange interpolation reduces the accumulation of numerical errors, especially at higher-order interpolation levels.

[0102] Hermitian interpolation requires not only that the interpolating polynomial take known values ​​at given data points (equal function values), but also that its derivative values ​​at these points are equal. This allows Hermitian interpolation to better approximate functions with specific smoothness requirements. Figure 4 A schematic diagram of a curve fitted using interpolation is shown. The fitted curve may pass through or approach the aforementioned metric nodes. It should be noted that... Figure 4 The curves obtained by fitting the data points formed by the metric values ​​and identity IDs of the respective convergence nodes are shown.

[0103] Finite field-based interpolation is an improved interpolation method. A finite field, also known as a Galois field, is a set of finite elements that satisfies all the axioms of a field. Specifically, for each element in the set, addition and multiplication are defined, and their associative, commutative, and distributive laws apply. There exist additive and multiplicative identity elements (usually denoted as 0) and multiplicative identity elements (usually denoted as 1). Every element has an additive inverse, and all non-zero elements have a multiplicative inverse. The use of additive or multiplicative inverses simplifies computation.

[0104] To facilitate understanding, the following example uses Newton interpolation based on finite fields. Finite field-based Newton interpolation can include several stages such as parameter configuration, coordinate generation, and aggregation. In the parameter configuration stage, the aggregation algorithm can configure 8 metric nodes, and the polynomial dimension of the aggregation result can be 8. The aggregation algorithm operates within a finite field, and the modulus q is a prime number. Let the modulus q be 983041, or 15 × 2^16 + 1. The modulus q has 20 bits and can be represented by a 32-bit integer (int) data type. In the coordinate generation stage, the metric nodes perform numerical transformations based on their own identity IDs, such as using a hash operation to obtain a corresponding hash value. The hash value can be a large integer, and then the large integer is modulo q to obtain the x-coordinate. i Each time a new x-coordinate is generated, it is compared with existing x-coordinates. If a conflict is found, auxiliary information (such as date, manufacturer identifier, etc.) is added, and the x-coordinate is regenerated. Similarly, the metric node can convert the metric value into another large integer, and then take the modulus q to obtain the y-coordinate of each metric node. i During the convergence phase, the first convergence node converges the coordinates (x, y) of each metric node using Newton's interpolation algorithm over a finite field. i ,y iThe calculation process is similar to Newton's interpolation algorithm. In a finite field (Galois Field, GF), elements are usually represented as an nth-degree polynomial. Addition in a finite field is similar to addition in a regular polynomial, adding corresponding coefficients and following the rules of the modulus q. After obtaining a polynomial in a finite field through addition and other arithmetic operations, the first pool node can convert it into a regular polynomial. The regular polynomial can be a polynomial over a non-finite field, and its coefficients can be represented as {a0, a1, ..., a7}.

[0105] Multi-sample aggregate signature is a technique that combines the signatures of multiple signers (such as metric nodes) on different messages (such as coordinates formed by metric values ​​and identity IDs) into a single signature. This significantly reduces the storage space and transmission bandwidth required for signatures. Specifically, multiple signers each sign their own messages, and then these signatures are aggregated together using a specific algorithm to form an aggregate signature. A verifier can use this aggregate signature along with all the original messages to verify the validity of all the signers' signatures.

[0106] S306. The first aggregation node sends the aggregation result to the trusted management node.

[0107] The convergence results are used to verify the trustworthiness of multiple metric nodes. Trusted management nodes can include a trusted management center, such as a server used for trusted management. When managing trusted computing, the trusted management center only needs to manage and maintain the curve's expression, without needing to manage each metric node individually. During remote proof, challengers (such as cloud service customers, regulators, trusted management nodes, or other system administrators) determine the prover's trustworthiness based on whether the coordinates sent by the prover (such as a computer system or metric nodes / convergence nodes within a computer system) conform to the curve's expression (i.e., whether they lie on the curve).

[0108] In some possible implementations, the trusted management node can receive a baseline value acquisition request. This request is used to obtain the aggregation result, which is then sent to the verification node. The aggregation result, obtained by aggregating the metric data from initialization, can serve as the baseline value for trusted verification. The first aggregation node can also receive a real value acquisition request. This request is used to obtain the runtime metric data (also called the real value or real metric value) of at least one of the multiple metric nodes. This real value acquisition request is then sent to the verification node, enabling the verification node to perform trusted verification based on the runtime metric data of at least one metric node and the aggregation result.

[0109] When the aggregation results are obtained using coordinate fitting, the aggregation results may include the coefficients of a polynomial. In this case, the aggregation results are used to construct a polynomial for trusted verification. In other words, the aggregation results are used to recover the polynomial obtained from coordinate fitting of multiple metric nodes. Accordingly, the verification node can construct a trusted verification polynomial based on the coefficients of the polynomial. For example, the verification node can agree or negotiate the number of terms and degree (e.g., the highest power exponent) of the polynomial with the trusted management node, and then assign the coefficients of the polynomial to each term of the polynomial to construct the trusted verification polynomial. Specifically, the verification node can sequentially assign the degree of the polynomial to each term to obtain the trusted verification polynomial. Then, the verification node can convert the metric data from at least one metric node into coordinates and verify whether the coordinates satisfy the polynomial, i.e., whether the data point corresponding to the coordinate lies on the curve corresponding to the polynomial. If the coordinates satisfy the polynomial, or the data point corresponding to the coordinate lies on the curve corresponding to the polynomial, the trusted verification passes. If the coordinates do not satisfy the polynomial, or the data point corresponding to the coordinate does not lie on the curve corresponding to the polynomial, the trusted verification fails.

[0110] To facilitate understanding, this application also provides an example using a task scheduling scenario. In a task scheduling scenario, the client can send a task scheduling request to the data center's management platform. The data center's management platform is used to manage the data center, such as performing resource scheduling and trusted management. Specifically, the management platform includes scheduling nodes and trusted management nodes. Upon receiving a task scheduling request, the scheduling node can allocate computing resources from the data center to the task. These computing resources can include a computer system or computing cards such as CPUs, NPUs, or GPUs within the computer system. The scheduling node can then return the resource information of the computing resources to the client. This resource information can include resource identifiers, such as the identifier of the computer system or the rank ID of the computing card. The client can send the resource information to a verification node, requesting the verification node to perform trusted verification of the computing resources. The verification node can be a third-party verification node or a node on the user side responsible for verification or security.

[0111] Verification nodes can send baseline value retrieval requests to the trusted management node and real value retrieval requests to the first aggregation node. Both baseline and real value retrieval requests can carry resource identifiers. It should be noted that when the computer system includes multiple aggregation nodes, verification nodes can send real value retrieval requests via polling. Accordingly, the verification node receives aggregation data (baseline values) from the trusted management node and runtime metric data (real values) from the metric node identified by the resource identifier sent by the first aggregation node, and then performs trusted verification based on the aggregation data and the runtime metric data. The specific verification process can be found in the relevant description above and will not be repeated here.

[0112] The above is just one example of a task scheduling scenario. In practical applications, the scheduling node can directly send the resource identifier to the verification node without returning it to the client, which then sends it to the verification node. Alternatively, the scheduling node can send the resource identifier to the trusted management node. The trusted management node obtains the aggregated data and the runtime measurement data from the computer system, and then returns both the aggregated data and the runtime measurement data to the verification node for trusted verification.

[0113] It should be noted that, in scenarios involving a single computer system (such as a single server), the aggregation node itself can also act as a trusted management node, managing and maintaining multiple metric nodes through the expression of the curve. For example, if the aggregation node detects that the coefficients of the polynomial expression of the curve are inconsistent with the preset coefficients, it indicates that at least one metric node among the multiple metric nodes represented by the curve has been tampered with, and tasks can be refused to be sent to that virtual node.

[0114] Based on the above description, this application provides a method for trusted verification. This method introduces a convergence node into a computer system, such as a baseboard management controller. The convergence node can aggregate measurement data from multiple measurement nodes it manages or controls, for example, by using curve fitting to aggregate the measurement data from multiple measurement nodes to obtain an aggregation result. The data volume of the convergence node is less than the sum of the measurement data volumes of the multiple measurement nodes (or the total data volume of the measurement data from the multiple measurement nodes). Thus, the trusted management node can manage a larger number of measurement nodes by storing the smaller aggregation result, without needing to store the larger measurement data volume, thereby solving the capacity limitation problem and supporting scalable management services and remote verification of measurement nodes.

[0115] Next, with reference to the accompanying diagram, we will explain in detail the specific implementation method of the aggregation node aggregating measurement data from multiple measurement nodes.

[0116] First, we will explain the basic aggregation scheme, which is the case where there is only one aggregation node (e.g., the first aggregation node).

[0117] See Figure 5 The diagram illustrates a method for aggregating metric values ​​from multiple metric nodes using a single aggregation node, and specifically includes the following steps:

[0118] Step 1: The aggregation node performs coordinate transformation on multiple metric nodes.

[0119] Specifically, the aggregation node can represent the coordinates (x, y) of each metric node using (identity ID, metric value). i ,y i Assume there are a total of k metric nodes.

[0120] Step 2: The convergence node uses interpolation to fit a curve represented by a polynomial expression based on the coordinates of multiple metric nodes.

[0121] For example, a convergence node can be fitted with a polynomial curve using Lagrange interpolation or Newton interpolation, expressed as:

[0122] f(x) = a0 + a1x + ... + a k-1 x k-1 (1)

[0123] Among them, a0, a1...a k-1 The coefficients of the polynomial are the coefficients of the polynomial. The coefficients of the polynomial are the convergence result.

[0124] Thus, during trusted management, the aggregation node / trusted management node only needs to manage the aggregation result, without needing to manage each metric node individually. Similarly, during remote proof, the challenger determines whether the coordinates sent by the prover (computer system or aggregation node or metric node within the computer system) satisfy the expression f(x) = a0 + a1x + ... + a k- 1x k-1 This allows us to determine whether the prover has tampered with the coordinates. If the coordinates sent by the prover satisfy the expression, it means that the prover has not tampered with the coordinates; if the coordinates sent by the prover do not satisfy the expression, it means that the prover has tampered with the coordinates.

[0125] This application supports scaling the number of aggregation nodes, enabling larger-scale server deployments. In this case, each aggregation node still uses basic aggregation capabilities, but a sufficient number of aggregation nodes can employ... Figure 6 The diagram illustrates a hierarchical aggregation scheme. Specifically, a first aggregation node corresponds to multiple second aggregation nodes. A second aggregation node can be a aggregation node directly connected to a metric node. Each second aggregation node is responsible for aggregating information from various metric nodes downwards and reporting specified information upwards to higher-level aggregation nodes, such as the first aggregation node. Figure 6 In the example, the first aggregation node can be a level+1 aggregation node, and the second aggregation node can be a level-1 aggregation node. The multiple metric nodes aggregated by the first aggregation node include multiple first metric nodes. The computer system also includes multiple second metric nodes, which are metric nodes aggregated by the second aggregation nodes. First metric nodes can include second aggregation nodes. It should be noted that first metric nodes can also include metric nodes other than second aggregation nodes. Figure 6 (Not shown in the image). During hierarchical aggregation, the second aggregation node can report the aggregation results to the first aggregation node, which in turn reports them to the trusted management node.

[0126] In some possible implementations, the second aggregation node can directly report the second aggregation result to the first aggregation node. Specifically, each of the multiple second aggregation nodes aggregates the measurement data of a portion of the multiple second measurement nodes to obtain a second aggregation result. Each of the multiple second aggregation nodes can also send the second aggregation result to the first aggregation node. The first aggregation node aggregates the measurement data of multiple first measurement nodes to obtain a first aggregation result. The first aggregation node can send the first and second aggregation results to the trusted management node. The trusted management node can store the full set of verification results. In this way, verification nodes (such as remote provers) can directly obtain the required aggregation results from the trusted management center when performing verification.

[0127] In other possible implementations, the second aggregation result of the second aggregation node is embedded into the report content reported by the first aggregation node. For example, each aggregation node can compress the information of the lower-level metric nodes and embed it into its own metric result. Specifically, the first aggregation node merges the second aggregation result with the metric data of the second aggregation node to obtain merged metric data. Then, the first aggregation node aggregates the merged metric data to obtain the first aggregation result. The first aggregation node can use a hash method to expand the second aggregation result (or metric report) of the second aggregation node and embed it into its own metric value, thereby embedding the second aggregation result into the first aggregation result. The first aggregation node can send the first aggregation result to the trusted management node.

[0128] It should be noted that when the metric data (such as metric values) of the first aggregation node embeds the report content (such as the second aggregation result) of the second aggregation node, the first aggregation node may not need to send the second aggregation result separately. Thus, compared to the scheme where the first aggregation node reports both the first and second aggregation results, the trusted management node can use less memory to store the aggregation results. In this scheme, when performing trusted verification, the verification node can communicate with the lower-level metric node to obtain the metric data from the metric node's runtime, and then combine this metric data for trusted verification.

[0129] Still with Figure 6This example illustrates that the Level layer aggregation node is responsible for aggregating the metric values ​​from various metric nodes and reporting specified information upwards. Depending on the specific needs, the Level layer aggregation node can report information in two ways: 1) It can also report the aggregation results from lower layers (e.g., the aggregation results obtained by aggregating metric data from the metric node layer) to the trusted management node. In this way, remote provers can directly obtain the aggregated curves from the trusted management node; 2) It can compress and embed the lower-layer metric node information into the report content, such as using a hash method to expand the lower-layer metric report and embed it into its own metric value. In this case, remote provers need to increase communication with the Level layer metric nodes to obtain metric data. The benefits of the second reporting method become more pronounced as the node scale increases.

[0130] This method organizes large-scale metric nodes through a combination of aggregation and hierarchy, addressing the capacity limitations of trusted management centers and the difficulty of traceability, thus meeting business needs. Traceability can be achieved by using coordinates and polynomial expressions provided by metric or aggregation nodes.

[0131] The following is combined with Figure 7 An example illustrating the source tracing process. Figure 7 In the example, each coordinate value is an integer of type int, including 32 bits. The challenger (i.e., the validator) obtains the coordinates of the metric nodes from the prover, including the x-coordinate and y-coordinate of each metric node. The challenger also obtains the convergence coefficient 'a' from the Trust Management Center. The challenger then sequentially sets the x-coordinate... i Substituting into the polynomial P(x) = a0 + a1x + ... + a7x 7 In the middle, compare y i With P(x) i If y i With P(x) i If y is inconsistent, it indicates that the i-th metric node is faulty, for example, the i-th metric node has been tampered with or is untrustworthy. i With P(x) i If the values ​​are consistent, it means that the i-th metric node is not faulty, for example, the i-th metric node has not been tampered with or the i-th metric node is trustworthy.

[0132] It should be noted that in the scheme of this application, the trusted management node stores the aggregation result obtained by aggregating the measurement data, not the signed policy file. Even if the measurement node in the computer system changes, such as being tampered with, the trusted management node does not need to re-sign. Therefore, the trusted management node does not need to frequently perform signing and verification operations, reducing deployment costs.

[0133] Based on the trusted verification method provided in the foregoing embodiments, this application also provides a computer system. An example of the computer system may be as follows: Figure 2A or Figure 2B As shown. The computer system includes at least one aggregation node and multiple measurement nodes, where each measurement node is a computing device within the computer system that performs trust measurements.

[0134] A first aggregation node in at least one aggregation node is used to acquire measurement data for trust measurement of the plurality of measurement nodes; aggregate the measurement data of the plurality of measurement nodes to obtain an aggregation result, wherein the data volume of the aggregation result is less than the sum of the data volumes of the measurement data of the plurality of measurement nodes; and send the aggregation result to a trust management node, wherein the aggregation result is used to perform trust verification on the plurality of measurement nodes.

[0135] The first aggregation node can be used to execute the content described in S302 to S306 of the aforementioned embodiments, which will not be repeated here.

[0136] In some possible implementations, the metric data includes the identity identifier and metric value of the metric node, whereby the metric value characterizes the state of the metric node, and the first aggregation node is specifically used for:

[0137] Using the identity identifier and metric value of each of the multiple metric nodes as the coordinates of each of the multiple metric nodes, a polynomial is obtained by fitting the coordinates of the multiple metric nodes, with the metric value as the function value and the identity identifier as the function variable. The coefficients of the polynomial are used as the aggregation result of the metric data of the multiple metric nodes. The aggregation result is used to construct a polynomial for trusted verification.

[0138] The specific implementation of the first aggregation node in aggregating metric data can be found in S304, and will not be elaborated here.

[0139] In some possible implementations, the first aggregation node is specifically used for:

[0140] Based on the coordinates of the multiple metric nodes, fitting is performed using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermitian interpolation, finite field-based interpolation, or aggregated signatures of multiple samples.

[0141] The specific implementation of the first aggregation node when aggregating measurement data using interpolation or multi-sample aggregation signatures can be found in S304, and will not be elaborated here.

[0142] In some possible implementations, the trusted management node is used for:

[0143] Receive a baseline value acquisition request, the baseline value acquisition request being used to acquire the aggregation result, and send the aggregation result to the verification node;

[0144] The first aggregation node is also used for:

[0145] A real value acquisition request is received, which is used to acquire the metric data of at least one of the plurality of metric nodes during operation, and the metric data of the at least one metric node during operation is sent to the verification node so that the verification node performs trusted verification based on the metric data of the at least one metric node during operation and the aggregation result.

[0146] The specific implementations of the trusted management node and the first aggregation node can be found in [reference needed]. Figure 3 The relevant descriptions of the embodiments shown will not be repeated here.

[0147] In some possible implementations, the first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system further includes multiple second metric nodes, the first metric nodes including the second aggregation nodes;

[0148] Each of the plurality of second aggregation nodes is used for:

[0149] The measurement data of some of the multiple second measurement nodes are aggregated to obtain a second aggregation result, and the second aggregation result is sent to the first aggregation node;

[0150] The first aggregation node is specifically used for:

[0151] The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result;

[0152] Send the first aggregation result and the second aggregation result to the trusted management node.

[0153] In some possible implementations, the first aggregation node is specifically used for:

[0154] The second aggregation result is merged with the metric data of the second aggregation node to obtain the merged metric data;

[0155] The merged metric data is aggregated to obtain the first aggregation result.

[0156] The specific implementation of hierarchical aggregation between the first and second aggregation nodes can be found in [reference needed]. Figure 6 Description of relevant content in the illustrated embodiment.

[0157] In some possible implementations, the first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system further includes multiple second metric nodes, the first metric nodes including the second aggregation nodes;

[0158] The first aggregation node is specifically used for:

[0159] The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result;

[0160] Send the first aggregation result to the trusted management node;

[0161] Each of the plurality of second aggregation nodes is used for:

[0162] The measurement data of a portion of the multiple second measurement nodes are aggregated to obtain a second aggregation result;

[0163] The second aggregation result is sent to the trusted management node.

[0164] The specific implementation of hierarchical aggregation between the first and second aggregation nodes can be found in [reference needed]. Figure 6 Description of relevant content in the illustrated embodiment.

[0165] In some possible implementations, the aggregation node is a baseboard management controller in the computer system.

[0166] Furthermore, this application also provides a trust measurement apparatus. The trust measurement apparatus can be a software apparatus. This software apparatus can be deployed in a distributed manner on a computer system, for example, distributedly deployed on a first aggregation node and a second aggregation node. The computer system executes the program code of the trust measurement apparatus, thereby performing the aforementioned trust verification method.

[0167] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computer system can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct a computer system to perform a method for trusted verification.

[0168] This application also provides a computer program product containing instructions. The computer program product may be a software or program product containing instructions, capable of running on a computer system or stored on any usable medium. When the computer program product is run on a computer system, it causes at least one computer system to perform the aforementioned trusted verification method.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for trusted verification, characterized in that, Applied to a computer system, the computer system including at least one aggregation node and multiple measurement nodes, wherein the measurement nodes are computing devices in the computer system that perform trust measurements, the method includes: The first aggregation node in the at least one aggregation node acquires the measurement data for the trusted measurement of the plurality of measurement nodes; The first aggregation node aggregates the measurement data of the multiple measurement nodes to obtain an aggregation result. The amount of data in the aggregation result is less than the sum of the amount of data in the multiple measurement nodes. The first aggregation node sends the aggregation result to the trusted management node, and the aggregation result is used to perform trusted verification on the multiple metric nodes.

2. The method according to claim 1, characterized in that, The measurement data includes the identity identifier and measurement value of the measurement node. The measurement value is used to characterize the state of the measurement node. The first aggregation node aggregates the measurement data of the multiple measurement nodes to obtain an aggregation result, including: The first aggregation node uses the identity and metric value of each of the plurality of metric nodes as the coordinates of each of the plurality of metric nodes. It fits the coordinates of the plurality of metric nodes to obtain a polynomial with the metric value as the function value and the identity as the function variable. The coefficients of the polynomial are used as the aggregation result of the metric data of the plurality of metric nodes. The aggregation result is used to construct a polynomial for trusted verification.

3. The method according to claim 2, characterized in that, The first convergence node fits the coordinates of the plurality of metric nodes, including: The first aggregation node fits the coordinates of the multiple metric nodes using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermitian interpolation, finite field-based interpolation, or aggregated signatures of multiple samples.

4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: The trusted management node receives a baseline value acquisition request, which is used to obtain the aggregation result and send the aggregation result to the verification node. The first aggregation node receives a truth value acquisition request, which is used to acquire the metric data of at least one of the plurality of metric nodes during operation, and sends the metric data of the at least one metric node during operation to the verification node so that the verification node performs trusted verification based on the metric data of the at least one metric node during operation and the aggregation result.

5. The method according to any one of claims 1 to 4, characterized in that, The first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system also includes multiple second metric nodes, the first metric node including the second aggregation node; The method further includes: Each of the plurality of second aggregation nodes aggregates the measurement data of a portion of the plurality of second measurement nodes to obtain a second aggregation result; Each of the plurality of second aggregation nodes sends the second aggregation result to the first aggregation node; The first aggregation node aggregates the measurement data from the multiple measurement nodes to obtain an aggregation result, including: The first aggregation node aggregates the measurement data of the plurality of first measurement nodes to obtain a first aggregation result; The first aggregation node sends the aggregation result to the trusted management node, including: The first aggregation node sends the first aggregation result and the second aggregation result to the trusted management node.

6. The method according to claim 5, characterized in that, The first aggregation node aggregates the measurement data from the plurality of first measurement nodes to obtain a first aggregation result, including: The first aggregation node merges the second aggregation result with the measurement data of the second aggregation node to obtain the merged measurement data; The first aggregation node aggregates the merged metric data to obtain a first aggregation result.

7. The method according to any one of claims 1 to 4, characterized in that, The first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system also includes multiple second metric nodes, the first metric node including the second aggregation node; The first aggregation node aggregates the measurement data from the multiple measurement nodes to obtain an aggregation result, including: The first aggregation node aggregates the measurement data of the plurality of first measurement nodes to obtain a first aggregation result; The first aggregation node sends the aggregation result to the trusted management node, including: The first aggregation node sends the first aggregation result to the trusted management node; The method further includes: Each of the plurality of second aggregation nodes aggregates the measurement data of a portion of the plurality of second measurement nodes to obtain a second aggregation result; Each of the plurality of second aggregation nodes sends the second aggregation result to the trusted management node.

8. The method according to any one of claims 1 to 7, characterized in that, The aggregation node is the baseboard management controller in the computer system.

9. A computer system, characterized in that, The computer system includes at least one aggregation node and multiple measurement nodes, wherein the measurement nodes are computing devices in the computer system that perform trust measurements. The first aggregation node among the at least one aggregation node is used to obtain measurement data for the trusted measurement of the plurality of measurement nodes; The measurement data of the multiple measurement nodes are aggregated to obtain an aggregation result, and the amount of data in the aggregation result is less than the sum of the amount of data of the multiple measurement nodes. The aggregation result is sent to the trusted management node, and the aggregation result is used to perform trusted verification on the multiple metric nodes.

10. The system according to claim 9, characterized in that, The measurement data includes the identity identifier and measurement value of the measurement node. The measurement value is used to characterize the state of the measurement node. The first aggregation node is specifically used for: Using the identity identifier and metric value of each of the multiple metric nodes as the coordinates of each of the multiple metric nodes, a polynomial is obtained by fitting the coordinates of the multiple metric nodes, with the metric value as the function value and the identity identifier as the function variable. The coefficients of the polynomial are used as the aggregation result of the metric data of the multiple metric nodes. The aggregation result is used to construct a polynomial for trusted verification.

11. The system according to claim 10, characterized in that, The first aggregation node is specifically used for: Based on the coordinates of the multiple metric nodes, fitting is performed using Newton interpolation, Lagrange interpolation, centroid Lagrange interpolation, Hermitian interpolation, finite field-based interpolation, or aggregated signatures of multiple samples.

12. The system according to any one of claims 9 to 11, characterized in that, The trusted management node is used for: Receive a baseline value acquisition request, the baseline value acquisition request being used to acquire the aggregation result, and send the aggregation result to the verification node; The first aggregation node is also used for: A real value acquisition request is received, which is used to acquire the metric data of at least one of the plurality of metric nodes during operation, and the metric data of the at least one metric node during operation is sent to the verification node so that the verification node performs trusted verification based on the metric data of the at least one metric node during operation and the aggregation result.

13. The system according to any one of claims 9 to 12, characterized in that, The first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system also includes multiple second metric nodes, the first metric node including the second aggregation node; Each of the plurality of second aggregation nodes is used for: The measurement data of some of the multiple second measurement nodes are aggregated to obtain a second aggregation result, and the second aggregation result is sent to the first aggregation node; The first aggregation node is specifically used for: The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result; Send the first aggregation result and the second aggregation result to the trusted management node.

14. The system according to claim 13, characterized in that, The first aggregation node is specifically used for: The second aggregation result is merged with the metric data of the second aggregation node to obtain the merged metric data; The merged metric data is aggregated to obtain the first aggregation result.

15. The system according to any one of claims 9 to 12, characterized in that, The first aggregation node corresponds to multiple second aggregation nodes, the multiple metric nodes include multiple first metric nodes, and the computer system also includes multiple second metric nodes, the first metric node including the second aggregation node; The first aggregation node is specifically used for: The measurement data of the multiple first measurement nodes are aggregated to obtain a first aggregation result; Send the first aggregation result to the trusted management node; Each of the plurality of second aggregation nodes is used for: The measurement data of a portion of the multiple second measurement nodes are aggregated to obtain a second aggregation result; The second aggregation result is sent to the trusted management node.

16. The system according to any one of claims 9 to 15, characterized in that, The aggregation node is the baseboard management controller in the computer system.

17. A computer-readable storage medium, characterized in that, Includes computer-readable instructions; the computer-readable instructions are used to implement the trusted verification method according to any one of claims 1 to 8.

18. A computer program product, characterized in that, Includes computer-readable instructions; the computer-readable instructions are used to implement the trusted verification method according to any one of claims 1 to 8.