Server BMC dynamic security authentication and firmware protection method and system based on hardware root of trust

Abstract

The application discloses a kind of based on hardware root of trust server BMC dynamic security authentication and firmware protection method and system, it is related to server remote management security technical field, the disclosed based on hardware root of trust server BMC dynamic security authentication and firmware protection method and system, realize dynamic authentication by the equipment fingerprint of security chip generation, combine fragment encryption, differential hash check and memory integrity verification to build multi-layer protection mechanism, simultaneously utilize block chain storage key hash value to enhance data credibility, effectively solve the problem of static authentication risk in traditional scheme, firmware tampering detection deficiency and trust chain loss, can improve the security protection capability of BMC system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of server remote management security, in particular to a server BMC dynamic security authentication and firmware protection method and system based on a hardware root of trust. BACKGROUND

[0002] With the growing demand for server remote management, BMC, as the core component carrying IPMI, Redfish and other management protocols, faces serious security threats. The traditional BMC security mechanism has serious defects: in the authentication layer, it relies on static password authentication or pre-set X.509 certificates, which not only has weak authentication risk, but also has high risk of key leakage due to the complex certificate update mechanism; in the firmware protection layer, it only supports static hash verification of the full firmware image, and has insufficient detection capability for differential tampering attacks, and lacks continuous protection in runtime; in the trust chain construction, it fails to establish a complete root of trust from hardware to software, making the BMC's own security configuration vulnerable to tampering. Existing technical solutions such as static certificate authentication, firmware hash verification and access control list cannot effectively solve the core problems of dynamic authentication, fine-grained firmware protection and complete trust chain construction. Especially in the face of advanced persistent threats, the protection effect of traditional solutions is insufficient.

[0003] The above content is only used to assist in understanding the technical solutions of the present application, and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0004] The main purpose of the present application is to provide a server BMC dynamic security authentication and firmware protection method and system based on a hardware root of trust, aiming to improve the security protection capability of BMC system.

[0005] To achieve the above purpose, the present application provides a server BMC dynamic security authentication and firmware protection method based on a hardware root of trust, which comprises:

[0006] Obtain the unique identifier data and hardware version data of the server motherboard through the security chip, generate the device fingerprint data based on the unique identifier data and hardware version data, and store the device fingerprint data in the tamper-proof storage area of the security chip;

[0007] Dynamic authentication processing based on device fingerprint data: generate a first random number, send the first random number to the management platform through the out-of-band channel of the management interface; receive the signature data returned by the management platform containing a second random number and timestamp data; use the security chip and device fingerprint data to verify the legality of the signature data; if the verification is successful, generate session key data based on the elliptic curve Diffie-Hellman algorithm; upload the hash value of the session key data to the blockchain network for storage;

[0008] The firmware protection processing based on the session key data: in the firmware transmission stage, the firmware update package data is fragmented and encrypted to obtain encrypted fragmented data; the check value data of the encrypted fragmented data is generated based on the SHA-3 hash algorithm; when the firmware update package data is received, the integrity of the check value data is verified piece by piece; in the firmware storage stage, the incremental update log data is generated based on the firmware update package data, and the incremental update log data is stored in the secure partition of the BMC flash memory, and the differential hash algorithm is used to monitor the abnormal change of the incremental update log data; in the firmware running stage, the trusted execution environment is established through the security chip and the reference hash value data is prestored, the periodic memory integrity check processing is performed on the key processes of the BMC kernel, and the memory check result data is obtained; if the memory check result data is abnormal, the management interface is frozen and the alarm notification is triggered;

[0009] The BMC user permission configuration data is digitally signed based on the security chip to obtain signed configuration data; when the permission change request data is received, the hardware key and the dynamic password are combined to form two-factor authentication data and authentication processing is performed; the security event log data is collected in real time, the security event log data is analyzed based on the anomaly detection model, and the automatic locking mechanism is triggered.

[0010] In an embodiment, the step of performing dynamic authentication processing based on the device fingerprint data comprises:

[0011] A first random number is generated by the security chip;

[0012] The first random number is sent to the management platform through an out-of-band channel;

[0013] The signature data returned by the management platform is received, which is generated by the management platform using its security chip to sign the first random number, the second random number and the timestamp data;

[0014] The legality of the signature data is verified using the security chip and the device fingerprint data;

[0015] If the verification is successful, the session key data is generated based on the first random number and the second random number through the elliptic curve Diffie-Hellman algorithm, and the hash value of the session key data is uploaded to the blockchain network.

[0016] In an embodiment, the step of generating the session key data based on the first random number and the second random number through the elliptic curve Diffie-Hellman algorithm, and uploading the hash value of the session key data to the blockchain network comprises:

[0017] Based on the elliptic curve Diffie-Hellman algorithm and the first random number, the second random number, the initial key data is generated;

[0018] The initial key data is processed by using a key derivation function to obtain derived key data;

[0019] The derived key data is output as session key data to a blockchain network.

[0020] In an embodiment, the steps of, in the firmware transmission phase, performing slice encryption on the firmware update package data to obtain encrypted slice data, generating check value data of the encrypted slice data based on a SHA-3 hash algorithm, and verifying the integrity of the check value data piece by piece when receiving the firmware update package data include:

[0021] The firmware update package data is encrypted slice by slice using the session key data to obtain encrypted slice data;

[0022] The check value data of each encrypted slice data is generated based on a SHA-3 hash algorithm, and a Merkle tree structure is constructed to store the check value data;

[0023] When receiving the firmware update package data, the encrypted slice data is extracted piece by piece and its check value data is verified to match the Merkle tree structure to verify the integrity of the check value data.

[0024] In an embodiment, the steps of, in the firmware storage phase, generating incremental update log data based on the firmware update package data, storing the incremental update log data in a secure partition of the BMC flash memory, and monitoring abnormal changes in the incremental update log data based on a differential hash algorithm include:

[0025] The incremental update log data is generated based on the firmware update package data;

[0026] A secure partition is divided in the BMC flash memory to store the incremental update log data;

[0027] The differential hash algorithm is applied to monitor changes in the incremental update log data in real time;

[0028] If the incremental update log data is detected to be abnormal, an alarm data is generated and the firmware loading is suspended.

[0029] In an embodiment, the steps of applying the differential hash algorithm to monitor changes in the incremental update log data in real time include:

[0030] The incremental update log data is processed based on a sliding window algorithm to generate current fingerprint data;

[0031] The current fingerprint data is compared with pre-stored reference fingerprint data;

[0032] If the differential value exceeds pre-set threshold rule data, it is determined that the incremental update log data is abnormal.

[0033] In an embodiment, the step of, during the firmware running stage, establishing a trusted execution environment by the security chip and pre-storing reference hash value data, periodically performing memory integrity check processing on key processes of the BMC kernel, and obtaining memory check result data comprises:

[0034] establishing a trusted execution environment by the security chip and loading the pre-stored reference hash value data;

[0035] periodically scanning key processes of the BMC kernel to obtain memory region data;

[0036] extracting key function segment data from the memory region data based on a preset key function segment address range;

[0037] calculating a hash value of the key function segment data, comparing the hash value with the reference hash value data, and generating memory check result data.

[0038] In an embodiment, the step of calculating a hash value of the key function segment data, comparing the hash value with the reference hash value data, and generating memory check result data comprises:

[0039] extracting the key function segment data;

[0040] applying a fast hash algorithm to calculate real-time hash value data of the key function segment data;

[0041] comparing the real-time hash value data with the pre-stored reference hash value data, and determining that the memory check result data is abnormal if the real-time hash value data is inconsistent with the reference hash value data.

[0042] In an embodiment, the step of real-time collecting security event log data, analyzing the security event log data based on an anomaly detection model, and triggering an automatic locking mechanism comprises:

[0043] real-time collecting security event log data from system logs;

[0044] analyzing the security event log data based on a Bayesian classifier model and preset threshold rule data to detect abnormal behavior;

[0045] if abnormal behavior is detected, generating locking instruction data to trigger the automatic locking mechanism.

[0046] Moreover, to achieve the above object, the application further provides a server BMC dynamic security authentication and firmware protection system based on a hardware root of trust, comprising a memory, a processor, and a server BMC dynamic security authentication and firmware protection program based on a hardware root of trust stored in the memory and capable of running on the processor, which is configured to implement the steps of the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust.

[0047] The server BMC dynamic security authentication and firmware protection method and system based on a hardware root of trust provided by the application achieve dynamic authentication through a device fingerprint generated by a security chip, and construct a multi-layer protection mechanism in combination with sharding encryption, differential hash verification, and memory integrity verification, while using a blockchain to store key hash values to enhance data credibility, effectively solving the problems of static authentication risks, insufficient firmware tampering detection, and missing trust chains in traditional solutions, and capable of improving the security protection capability of a BMC system. BRIEF DESCRIPTION OF DRAWINGS

[0048] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and serve to explain the principles of the application together with the specification.

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced hereinafter. Obviously, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative work.

[0050] Figure 1 A flowchart of an embodiment of the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust of the application is provided.

[0051] Figure 2 A structural diagram of an embodiment of the server BMC dynamic security authentication and firmware protection system based on a hardware root of trust of the application is provided.

[0052] DETAILED DESCRIPTION

[0053] 10, memory; 20, processor.

[0054] The object implementation, functional features, and advantages of the application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0055] The technical solutions in the present application will be described clearly and completely in the present application in combination with the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. The components of the present application described and shown in the drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work belong to the scope of protection of the present application.

[0056] It should be understood that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0057] In the prior art, the wide dependence on baseboard management controllers in the server remote management scene realizes device monitoring and maintenance. The traditional baseboard management controller adopts a static password authentication mechanism, which has the risk of default password leakage and cannot resist phishing attacks. The firmware integrity check only supports full image hash verification, and the detection capability for partial tampering attacks is insufficient. The access control policy is based on network address filtering, and lacks a dynamic identity verification mechanism, resulting in frequent unauthorized access vulnerabilities.

[0058] In order to solve the above problems, the research and development personnel found that the static authentication mechanism cannot adapt to the dynamic security environment, and needs to introduce an authentication factor based on the unique identifier of the device. The firmware protection needs to cover the whole life cycle of transmission, storage and running, and a single verification method is difficult to cope with complex attack scenarios. The lack of hardware trust root leads to easy tampering of security configuration, and an end-to-end trusted execution environment needs to be established.

[0059] Based on this, the embodiments of the present application provide a server BMC dynamic security authentication and firmware protection method based on a hardware trust root, referring to Figure 1 , the server BMC dynamic security authentication and firmware protection method based on the hardware trust root comprises steps S100-S400, wherein:

[0060] Step S100, obtaining the unique identifier data and hardware version data of the server mainboard through the security chip, generating device fingerprint data based on the unique identifier data and the hardware version data, and storing the device fingerprint data in the tamper-proof storage area of the security chip;

[0061] Step S200, dynamic authentication processing based on device fingerprint data: generating a first random number, sending the first random number to the management platform through the out-of-band channel of the management interface; receiving the signature data returned by the management platform containing the second random number and the timestamp data; verifying the legality of the signature data using the security chip and the device fingerprint data; if the verification is successful, generating session key data based on the elliptic curve Diffie-Hellman algorithm; uploading the hash value of the session key data to the blockchain network for storage;

[0062] Step S300, firmware protection processing based on session key data: in the firmware transmission stage, performing fragmentation encryption processing on the firmware update package data to obtain encrypted fragment data; generating check value data of the encrypted fragment data based on the SHA-3 hash algorithm; when receiving the firmware update package data, verifying the integrity of the check value data piece by piece; in the firmware storage stage, generating incremental update log data based on the firmware update package data, storing the incremental update log data in the secure partition of the BMC flash, and monitoring the abnormal changes of the incremental update log data based on the differential hash algorithm; in the firmware running stage, creating a trusted execution environment through the security chip and pre-storing reference hash value data, performing periodic memory integrity verification processing on the key processes of the BMC kernel to obtain memory verification result data; if the memory verification result data is abnormal, freezing the management interface and triggering an alarm notification;

[0063] Step S400, based on the security chip, digitally signing the BMC user permission configuration data to obtain signed configuration data; when receiving the permission change request data, obtaining the hardware key and the dynamic password to form two-factor authentication data and performing authentication processing; real-time collection of security event log data, analysis of security event log data based on an abnormal detection model, and triggering of an automatic locking mechanism.

[0064] In this embodiment, the device fingerprint data refers to a unique identity credential generated by combining the server motherboard physical identifier and the hardware version, which can be implemented by performing hash operation on the motherboard serial number and version number using the encryption engine of the security chip, and is used to establish an unforgeable device identity. The dynamic authentication processing refers to dynamic identity confirmation through two-way random number exchange and timestamp verification, which can specifically use the elliptic curve digital signature algorithm to verify the identity of the management platform, and eliminate the risk of static credential leakage. The fragmentation encryption processing refers to dividing the firmware data into multiple encrypted units, which can specifically use the session key to implement AES-GCM encryption on each data block to prevent whole data leakage during transmission. The incremental update log data refers to the audit information recording the version difference changes, which can specifically use the binary difference algorithm to generate version change records for easy detection of illegal tampering. The trusted execution environment refers to an isolated secure computing area, which can specifically use the TEE technology of the security chip to create a protected memory space to ensure that the key processes are not disturbed.

[0065] In this embodiment, the security chip collects the mainboard unique identification information, generates a device fingerprint and stores it in a tamper-proof area, forming a hardware-level trust anchor. In the dynamic authentication process, a random number is generated and sent to the management platform, and a signed response containing a timestamp is received. The platform identity is confirmed by verifying the legality of the signature. After successful authentication, the session key is negotiated based on the elliptic curve algorithm, and the key hash value is written to the blockchain to achieve an unalterable record. When updating the firmware, the data packets are encrypted in blocks, each encrypted block generates an independent check value to build a Merkle tree, and the receiving end verifies the hash value block by block to ensure the integrity of the transmission. In the storage stage, the firmware version difference log is recorded, and the sliding window algorithm is used to calculate the log fingerprint in real time, and the abnormal changes are found by comparing with the reference value. In the running stage, the reference hash is loaded in the trusted environment, the memory area of the key process is scanned periodically, and the code tampering behavior is detected by the fast hash algorithm. When the authority is changed, two-factor authentication is performed by combining the hardware key and the dynamic password, and the automatic locking mechanism is triggered by real-time analysis of the security log.

[0066] Compared with the prior art, the dynamic authentication mechanism replaces the static password authentication, and the anti-replay attack capability is realized through random number exchange and blockchain storage. Compared with the Merkle tree verification, the fragmented encryption can accurately locate the tampered data block. The incremental log monitoring technology significantly reduces the storage overhead and improves the detection sensitivity compared with the traditional full comparison method. The establishment of the trusted execution environment breaks through the limitations of software layer protection and realizes hardware-level memory protection. The two-factor authentication combined with the anomaly detection model forms a multi-dimensional defense system. Through the above technical solutions, the problem that the traditional authentication mechanism is easy to be cracked is effectively solved, and dynamic identity verification and secure negotiation of session keys are realized. The firmware protection covers the whole process of transmission encryption, storage monitoring and running verification, which significantly improves the ability to resist complex attacks. The introduction of the hardware trust root ensures that the security configuration is unalterable, the anomaly detection mechanism blocks potential threats in time, and a complete server management controller security protection system is formed.

[0067] In a feasible implementation, the step of performing dynamic authentication based on the device fingerprint data includes: generating a first random number by the security chip; sending the first random number to the management platform through an out-of-band channel; receiving signature data returned by the management platform, the signature data being generated by the management platform using its security chip to sign the first random number, a second random number and timestamp data; verifying the legality of the signature data using the security chip and the device fingerprint data; if the verification is successful, generating session key data based on the first random number and the second random number through an elliptic curve Diffie-Hellman algorithm, and uploading a hash value of the session key data to a blockchain network.

[0068] In this embodiment, the out-of-band channel refers to a management communication channel independent of the server service network, which can be implemented by a dedicated network interface of a baseboard management controller, and is used to avoid interference of the data transmission process by the service network attack. The elliptic curve Diffie-Hellman algorithm refers to a key exchange protocol based on elliptic curve cryptography, which can be implemented by NIST P-256 curve parameters, and the generation security of the temporary session key is ensured by the discrete logarithm problem. The blockchain network refers to a distributed ledger storage system, which can be implemented by a consortium chain architecture, and the non-tamperability of the key hash value is ensured by a multi-node consensus mechanism.

[0069] In this embodiment, in the dynamic authentication process, the security chip first generates a random number with cryptographic strength as a temporary session parameter, and transmits it to the remote management platform through the physically isolated out-of-band communication link. After receiving the random number, the management platform combines the second random number and the precise timestamp generated by itself, and performs a digital signature operation using the hardware security module. The security chip on the server side verifies the validity of the signature by presetting the device fingerprint, confirms the legal identity of the communication parties, and performs key negotiation operation based on the elliptic curve cryptography principle, taking the two random numbers as input parameters, to generate an encryption key valid only for this session. Finally, the key hash value is written into the blockchain node to form an audit record with timestamp and multi-party verification features.

[0070] In this embodiment, by dynamically generating a random number as a session base parameter and combining a timestamp signature verification mechanism, replay attacks and man-in-the-middle attacks can be effectively resisted. The one-time pad key negotiation method is used to replace the fixed certificate, which eliminates the security risks caused by long-term storage of the key, and further enhances the traceability and anti-repudiation of the key usage record through the blockchain, thereby realizing a strong identity authentication mechanism based on dynamic random parameters and solving the security risks caused by easy forgery and long-term validity of traditional static certificates.

[0071] In a feasible implementation, the step of generating session key data based on the first random number and the second random number through the elliptic curve Diffie-Hellman algorithm and uploading the hash value of the session key data to the blockchain network includes: generating initial key data based on the elliptic curve Diffie-Hellman algorithm and the first random number and the second random number; applying a key derivation function to process the initial key data to obtain derived key data; and outputting the derived key data to the blockchain network as session key data.

[0072] In this embodiment, the elliptic curve Diffie-Hellman algorithm refers to a key exchange protocol based on the elliptic curve discrete logarithm problem, which can be implemented using secp256k1 curve parameters, and is used to securely negotiate a shared key in an insecure channel. The key derivation function refers to an algorithm that expands an initial key into a more complex key sequence, which can be implemented using the HKDF standard, and enhances the entropy value of the derived key through a salt value and context information. The derived key data refers to the session key generated after key expansion processing, which can use the AES-256 algorithm as the encryption basis, and the length can be set to 256 bits to meet the high-strength encryption requirement. The blockchain network refers to a distributed ledger storage system, which can be implemented using the Hyperledger Fabric framework, and ensures the non-tamperability of the key hash value through a consensus mechanism.

[0073] In this embodiment, in the dynamic authentication process, the first random number and the second random number are taken as input parameters by the elliptic curve Diffie-Hellman algorithm to calculate an initial shared key. Then, the key derivation function is applied to iteratively process the initial key, and a pseudo-random function is used to generate a derived key with higher complexity. The derived key is output as the final session key to the blockchain node, and its hash value is written into the immutable ledger of the blockchain through the smart contract. For example, the number of iterations can be set to 10,000 in the key derivation stage, and a 512-bit intermediate key is generated through the HMAC-SHA512 function, and the first 256 bits are finally taken as the effective session key.

[0074] In this embodiment, by dynamically generating an initial key and superimposing key derivation processing, the key of each session has uniqueness and unpredictability, and the blockchain storage mechanism eliminates the single point failure risk of centralized storage, ensuring that the key verification record cannot be maliciously tampered with. In this way, the application solves the technical defect that the traditional static key is easy to be cracked, significantly improves the randomness and complexity of the key generation process through the dynamic key derivation mechanism, and forms a complete key life cycle management chain by combining the distributed storage characteristics of the blockchain, ensuring the independence and traceability of each session key.

[0075] In a feasible implementation, in the firmware transmission phase, the firmware update package data is subjected to a fragmentation encryption process to obtain encrypted fragmented data; check value data of the encrypted fragmented data is generated based on a SHA-3 hash algorithm; when the firmware update package data is received, the step of verifying the integrity of the check value data piece by piece includes: the firmware update package data is subjected to a fragmentation encryption process using session key data to obtain encrypted fragmented data; check value data is generated for each encrypted fragmented data based on a SHA-3 hash algorithm, and a Merkle tree structure is constructed to store the check value data; when the firmware update package data is received, the encrypted fragmented data is extracted piece by piece and it is verified whether the check value data thereof matches the Merkle tree structure, so as to verify the integrity of the check value data.

[0076] In this embodiment, the fragmentation encryption process refers to independent encryption of the firmware update package by dividing the firmware update package into multiple data blocks, which can be specifically implemented by using an AES-GCM algorithm in combination with a dynamically generated session key. Independent encryption of each fragment can prevent an intermediate person from obtaining a complete firmware image through an attack. The SHA-3 hash algorithm refers to a third-generation secure hash algorithm standard, and a SHA3-512 variant can be specifically used to generate a 512-bit check value. The anti-collision property of the SHA3-512 variant can ensure the unique identification of the fragmented data. The Merkle tree structure refers to a binary tree-shaped hash value aggregation structure, and a root hash value can be specifically generated by layer-by-layer hash merging of sub-node hash values. This structure allows the integrity of a specific fragment to be verified through a small number of hash values.

[0077] In this embodiment, in the firmware update package transmission process, the firmware data is first subjected to block encryption using a dynamically negotiated session key, for example, 1 GB of firmware is divided into 1024 1 MB fragments and is encrypted respectively. Each encrypted fragment generates an independent check value through a SHA-3 algorithm, all check values are used to construct a Merkle tree according to a binary tree structure, and finally a root hash value is stored in a secure chip. When the encrypted fragments are obtained at the receiving end, the hash chain of the fragment hash value and the corresponding node of the Merkle tree is compared, for example, when the 512th fragment is verified, only the fragment hash and the hash path of the sibling node to the root node need to be provided, so that it can be confirmed that the fragment has not been tampered with in the transmission process.

[0078] It can be understood that the traditional firmware hash check only performs single hash verification on a complete image, and when a differential tampering attack is encountered, the modified fragment cannot be located. According to the present scheme, through fragmentation encryption and a Merkle tree structure, not only is the integrity of each fragment verified, but also the tampered data block can be accurately located, for example, when it is detected that a certain fragment hash value does not match the Merkle tree, the decryption process of the fragment can be terminated immediately. The encrypted fragments are encrypted using a dynamic session key, which is more effective in resisting a replay attack than a static certificate encryption scheme.

[0079] In an implementation, the step of storing the incremental update log data in a secure partition of the BMC flash memory and monitoring abnormal changes in the incremental update log data based on a differential hash algorithm during the firmware storage phase includes: generating the incremental update log data based on the firmware update package data; dividing a secure partition in the BMC flash memory and storing the incremental update log data; applying the differential hash algorithm to monitor changes in the incremental update log data in real time; and if the incremental update log data is detected to be abnormal, generating alarm data and pausing the firmware loading.

[0080] In the embodiment, the incremental update log data refers to a data structure that records only differences between a firmware update package and a previous version, and can be generated by using a binary difference comparison algorithm to reduce storage space occupation and improve tampering detection efficiency. The secure partition of the BMC flash memory refers to a protected storage area divided by a hardware isolation mechanism, and can use an access control policy implemented by a flash memory controller to prevent unauthorized processes from modifying data. The differential hash algorithm refers to an algorithm that calculates hash differences of data blocks by using a sliding window mechanism, and can use a fingerprint generation method based on rolling hash to achieve fast detection of local data tampering.

[0081] In the embodiment, after the firmware update package completes transmission verification, the system generates incremental update log data containing only modified parts by using a binary difference comparison tool. The log data is written into a secure partition of the BMC flash memory that is isolated by a hardware access control policy, for example, an independent storage block is set and read-only permission is configured. The system continuously runs a differential hash monitoring process, for example, a sliding window scan of the log data is performed every 5 seconds, a current data fingerprint is generated and compared with a reference fingerprint. When an unexpected hash difference value exceeding a preset threshold is detected, for example, the hash values of three consecutive data blocks do not match, the system immediately interrupts the firmware loading process and triggers an audible and light alarm signal.

[0082] In the embodiment, by using the incremental log monitoring combined with the differential hash algorithm, local data tampering behavior can be accurately identified, and the technical problems of the update log being easily tampered with and low detection efficiency during the firmware storage phase are solved. By using the incremental storage mechanism, storage pressure is reduced, the differential hash algorithm is used to improve tampering detection sensitivity, the hardware protection mechanism of the secure partition blocks unauthorized modification, and a multi-level firmware storage protection system is formed.

[0083] In an implementation, the step of applying the differential hash algorithm to monitor changes in the incremental update log data in real time includes: processing the incremental update log data based on a sliding window algorithm to generate current fingerprint data; performing differential comparison between the current fingerprint data and pre-stored reference fingerprint data; and if a differential value exceeds a preset threshold value rule data, determining that the incremental update log data is abnormal.

[0084] In this embodiment, the sliding window algorithm refers to a sliding window technology for processing incremental update log data in fixed length blocks. Specifically, a window with a length of 512 bytes can be used for data slicing, and continuous coverage can be achieved through a sliding step. This technology can effectively capture local data change characteristics. The differential hash algorithm refers to a method for calculating the difference in hash values of adjacent data blocks. Specifically, SHA-256 algorithm can be used to generate hash values of each data block and perform XOR operation. By quantifying the difference value, abnormal modification can be identified. The preset threshold rule data refers to the pre-set difference tolerance range. Specifically, a dynamic threshold mechanism can be used, for example, based on historical data statistics, the mean value plus three times the standard deviation is set as the threshold. This mechanism can adapt to normal fluctuations in different scenarios.

[0085] In this embodiment, the incremental update log data is divided into continuous data blocks, and each data block generates a corresponding hash value as the current fingerprint data through the sliding window algorithm. The system compares the current fingerprint with the reference fingerprint block by block, and calculates the difference between the hash values. When the difference exceeds the preset threshold, it is determined that the log data has unauthorized modification. For example, during the firmware storage stage, if the hash difference value of a certain data block in the incremental log reaches the threshold, the system will immediately suspend the firmware loading process and generate an alarm. This process realizes fine-grained monitoring through differential comparison mechanism, which significantly reduces the computational overhead compared with the traditional full-check method.

[0086] In this embodiment, the incremental update log is monitored in real time through the differential hash algorithm, which can capture small data changes. For example, detecting abnormal offset of a single function address in the firmware log can trigger the protection mechanism, which can effectively identify hidden tampering attacks during the firmware storage stage, such as malicious code injection behavior against security patch update logs.

[0087] In a feasible implementation, during the firmware running stage, a trusted execution environment is established through a security chip and the reference hash value data is pre-stored, the key processes of the BMC kernel are periodically subjected to memory integrity verification processing, and memory verification result data is obtained; if the memory verification result data is abnormal, the steps of freezing the management interface and triggering the alarm notification include: establishing a trusted execution environment through a security chip and loading the pre-stored reference hash value data; periodically scanning the key processes of the BMC kernel to obtain memory region data; based on the pre-set key function segment address range, extracting key function segment data from the memory region data; calculating the hash value of the key function segment data, and comparing it with the reference hash value data to generate memory verification result data.

[0088] In this embodiment, the trusted execution environment refers to a secure execution space created by the hardware isolation mechanism of the security chip, which can be implemented by ARM TrustZone or Intel SGX technology, and is used to isolate critical code from external malicious code interference. The reference hash value data refers to the hash value of the legal process memory region pre-calculated and stored, which can be generated by using the SHA-256 algorithm and used as a reference for integrity verification. The periodic memory integrity verification process refers to scanning the memory data at a predetermined time interval, which can be implemented by using a timer triggering mechanism, and is used to detect runtime tampering behavior in real time. The critical process refers to the core function module of the BMC kernel related to system management and network communication, which can include the IPMI protocol processing module or the Redfish interface service process.

[0089] In this embodiment, in the firmware running phase, the trusted execution environment is divided by the hardware isolation function of the security chip, and the reference hash value of the legal memory region of the critical process pre-calculated is loaded into the environment. The system periodically triggers the memory scanning operation, extracts the critical function segment data from the memory region of the BMC kernel, for example, determines the function address range by analyzing the process memory mapping table. The extracted data is subjected to hash calculation to generate real-time hash value, which is compared with the pre-stored reference hash value. If the hash values do not match, it is determined that there is memory tampering behavior, and the network communication function of the management interface is immediately frozen, and an encrypted alarm notification is sent to the management platform through the out-of-band channel.

[0090] In this embodiment, by periodically verifying the memory integrity of the critical process, combined with the hardware protection mechanism of the trusted execution environment, malicious code injection behavior can be identified in real time, the critical function segment can be accurately located and high-frequency verification can be implemented, the interception efficiency of runtime attack is improved, the dynamic protection of BMC runtime memory data is realized, and the residence and spread of malicious code in the system running phase are effectively prevented. Through the cooperative mechanism of the trusted execution environment and the periodic verification, the technical defects of the traditional scheme that cannot detect partial tampering and lack of real-time protection are solved, and the security of the verification process is ensured by using the hardware root of trust.

[0091] In a feasible implementation, the step of calculating the hash value of the critical function segment data and comparing it with the reference hash value data to generate memory verification result data includes: extracting the critical function segment data; applying a fast hash algorithm to calculate the real-time hash value data of the critical function segment data; comparing the real-time hash value data with the pre-stored reference hash value data, and if they are inconsistent, determining that the memory verification result data is abnormal.

[0092] In this embodiment, the key function segment data refers to the specific memory region data of the key process in the BMC kernel, which can be implemented by extracting the function entry address and code segment length range data of the process using the memory address mapping technology, and is used to locate the core code region that needs to be checked. The fast hash algorithm refers to a hash function based on low computational complexity, which can be implemented by using the XXH64 or FarmHash algorithm, and is used to quickly generate the hash value of the memory data under limited computing resources. The real-time hash value data refers to the immediate hash calculation result of the memory content of the key function segment in the current running state, which can be implemented by obtaining the real-time memory mapping data of the process through the memory image snapshot technology and performing hash operation, and is used to reflect the real-time integrity state of the memory code. The reference hash value data refers to the legal function segment hash value stored in the trusted execution environment in advance, which can be implemented by performing hash calculation on the verified original code in the trusted startup phase through the security chip and storing it encrypted, and is used as a reference standard for integrity check.

[0093] In this embodiment, in the firmware running phase, the reference hash value data is loaded by the trusted execution environment established by the security chip. The system periodically scans the key processes of the BMC kernel, obtains the memory region data thereof, and extracts the corresponding code segment data based on the preset key function segment address range. After extracting the key function segment data, the real-time hash value thereof is calculated by using the fast hash algorithm, and is compared with the reference hash value. If the two are inconsistent, it is determined that the memory check result is abnormal, and the management interface freezing and alarm mechanism is triggered. This process processes the key function segment data in stages, realizes high-frequency checking by combining the fast hash algorithm, and relies on the trusted execution environment to ensure the non-tamperability of the reference data.

[0094] In this embodiment, by dividing the key function segment and implementing periodic fast hash calculation, the abnormal code region can be accurately located. For example, when an attacker injects a malicious code fragment, even if the overall hash of the firmware does not change, the real-time hash value of the specific function segment will deviate, thereby solving the problem of low efficiency of runtime memory tampering detection, and realizing fine-grained real-time monitoring of the key process. By combining the fast hash algorithm and the key function segment positioning technology, the malicious behaviors such as injection attack and code hijacking can be quickly identified while maintaining the stability of the system performance. Based on the reference data stored in the trusted execution environment, the safety of the check reference system is ensured, and the attacker is prevented from evading detection by tampering with the reference value. After abnormality determination, the automatic freezing mechanism is triggered, which can timely block the spread of potential attack chain, and improves the security of the BMC running environment.

[0095] In an implementable embodiment, the step of collecting security event log data in real time, analyzing the security event log data based on an anomaly detection model, and triggering an automatic locking mechanism comprises: collecting security event log data in real time from system logs; analyzing the security event log data based on a Bayesian classifier model and preset threshold rule data to detect abnormal behavior; and generating locking instruction data to trigger the automatic locking mechanism if abnormal behavior is detected.

[0096] In this embodiment, the security event log data refers to record data generated from the running process of the BMC system, including user login, permission change, and firmware operation behavior. Specifically, the syslog protocol can be used to collect system kernel logs and audit logs in real time to achieve the implementation, which is used to build the data basis for anomaly detection. The Bayesian classifier model refers to a classification algorithm model constructed based on probability statistics theory. Specifically, the Naive Bayes algorithm can be combined with historical security event training data sets to achieve the implementation, which is used to dynamically identify abnormal behavior patterns in logs. The threshold rule data refers to a set of preset abnormal behavior judgment conditions. Specifically, multi-dimensional rule combination can be used to achieve the implementation, such as a threshold of the number of permission changes within a single hour and a threshold of cross-region login attempt frequency, which is used to quantify the judgment standard of abnormal behavior. The automatic locking mechanism refers to a security control policy for access restriction of the management interface. Specifically, the firewall policy dynamic update or session connection forced termination technology can be used to achieve the implementation, which is used to immediately block potential attacks when abnormality is detected.

[0097] In this embodiment, the system continuously captures BMC kernel events and user operation records through the log collection module to form a structured security event log data stream. The data stream is input to the Bayesian classifier model for real-time analysis. The model calculates the deviation degree of the event feature vector from the normal behavior probability distribution, and makes a comprehensive judgment in combination with the preset threshold rule. For example, when it is detected that a same user account triggers 5 permission change operations within 10 minutes and the operation IP address is abnormal, the system will determine that it is abnormal behavior. At this time, the security control module will generate a locking instruction to immediately freeze the corresponding account permission and close the related management port by calling the BMC underlying interface.

[0098] In some specific embodiments, the log collection module can be configured to collect hardware sensor data and firmware operation logs simultaneously, and the Bayesian classifier model can load different versions of training data sets to adapt to the security policy of a specific environment. The threshold rule data can use a dynamic adjustment mechanism, such as automatically adjusting the alarm sensitivity according to the system load state. The automatic locking mechanism can be set to a hierarchical response mode, such as temporary locking for the first abnormal trigger and permanent locking for repeated abnormal triggers.

[0099] In this embodiment, the dynamic behavior modeling is realized through the Bayesian classifier model, which can adaptively learn the normal operation mode and significantly improve the detection capability of unknown threats. At the same time, the threshold rule data and the automatic locking mechanism work together to solve the problem of response lag in the traditional scheme, and realize the closed-loop control from anomaly detection to security protection. In this way, the application realizes intelligent monitoring and real-time response to the security events of the BMC system, effectively identifies malicious behaviors disguised as normal operations, timely blocks unauthorized access and configuration tampering, and improves the security protection level of the server management control unit.

[0100] In the embodiments of the application, the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust realizes dynamic authentication through a security chip to generate a device fingerprint, constructs a multi-layer protection mechanism by combining sharding encryption, differential hash verification and memory integrity verification, and uses a blockchain to store key hash values to enhance data credibility, effectively solving the problems of static authentication risk, insufficient firmware tampering detection and missing trust chain in the traditional scheme, and being capable of improving the security protection capability of the BMC system.

[0101] It should be noted that the above examples are only used for understanding the application and do not constitute a limitation on the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust of the application. More forms of simple transformation based on this technical concept are within the protection scope of the application.

[0102] The application also provides a server BMC dynamic security authentication and firmware protection system based on a hardware root of trust, which refers to Figure 2 The server BMC dynamic security authentication and firmware protection system based on a hardware root of trust comprises a memory 10, a processor 20 and a server BMC dynamic security authentication and firmware protection program based on a hardware root of trust stored on the memory 10 and capable of running on the processor 20, and the server BMC dynamic security authentication and firmware protection program based on a hardware root of trust is configured to implement the steps of the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust.

[0103] The server BMC dynamic security authentication and firmware protection system based on a hardware root of trust provided by the application adopts the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust in the above embodiments, and can improve the security protection capability of the BMC system. Compared with the prior art, the server BMC dynamic security authentication and firmware protection system based on a hardware root of trust provided by the application has the same beneficial effects as the server BMC dynamic security authentication and firmware protection method based on a hardware root of trust provided by the above embodiments, and other technical features in the server BMC dynamic security authentication and firmware protection system based on a hardware root of trust are the same as the features disclosed in the above method embodiments, which will not be repeated here.

[0104] It should be understood that portions of the application disclosed herein can be implemented with hardware, software, firmware or combinations thereof. In the description of the embodiments above, specific features, structures, materials or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0105] The above description is merely illustrative of the application and not restrictive. The scope of the application should be determined by the appended claims, along with equivalents thereof.

Claims

1. A method for dynamic security authentication and firmware protection of server BMC based on hardware root of trust, characterized in that, The method includes: The unique identifier data and hardware version data of the server motherboard are obtained through the security chip. Device fingerprint data is generated based on the unique identifier data and hardware version data, and the device fingerprint data is stored in the tamper-proof storage area of ​​the security chip. Dynamic authentication processing based on device fingerprint data: Generate a first random number and send it to the management platform through the out-of-band channel of the management interface; receive signature data containing a second random number and timestamp data returned by the management platform; verify the legality of the signature data using the security chip and device fingerprint data; if the verification is successful, generate session key data based on the elliptic curve Diffie-Hellman algorithm; upload the hash value of the session key data to the blockchain network for storage. Firmware protection based on session key data: During firmware transmission, the firmware update package data is fragmented and encrypted to obtain encrypted fragment data; checksum data of the encrypted fragment data is generated based on the SHA-3 hash algorithm; when receiving firmware update package data, the integrity of the checksum data is verified piece by piece; during firmware storage, incremental update log data is generated based on the firmware update package data and stored in the secure partition of the BMC flash memory; abnormal changes in the incremental update log data are monitored based on the differential hash algorithm; during firmware operation, a trusted execution environment is established through a security chip and a base hash value is pre-stored; periodic memory integrity checks are performed on critical processes of the BMC kernel to obtain memory checksum data; if the memory checksum data is abnormal, the management interface is frozen and an alarm notification is triggered. The security chip performs digital signature processing on BMC user permission configuration data to obtain the signature configuration data; when a permission change request is received, the hardware key and dynamic password are used to form two-factor authentication data and perform authentication processing; security event log data is collected in real time, and the security event log data is analyzed based on the anomaly detection model to trigger an automatic locking mechanism.

2. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 1, characterized in that, The steps for dynamic authentication based on device fingerprint data include: The first random number is generated using the security chip; The first random number is sent to the management platform via an out-of-band channel; Receive signature data returned by the management platform. This signature data is generated by the management platform using its security chip to sign the first random number, the second random number, and the timestamp data. The signature data is verified using a security chip and device fingerprint data. If the verification is successful, the session key data is generated using the elliptic curve Diffie-Hellman algorithm based on the first and second random numbers, and the hash value of the session key data is uploaded to the blockchain network.

3. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 2, characterized in that, The steps of generating session key data based on a first random number and a second random number using the elliptic curve Diffie-Hellman algorithm, and uploading the hash value of the session key data to the blockchain network, include: Initial key data is generated based on the elliptic curve Diffie-Hellman algorithm and the first and second random numbers. The initial key data is processed using the key derivation function to obtain the derived key data; The derived key data is output to the blockchain network as session key data.

4. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 1, characterized in that, During the firmware transmission phase, the firmware update package data is fragmented and encrypted to obtain encrypted fragment data. Generate verification data for encrypted fragmented data based on the SHA-3 hash algorithm; When receiving firmware update package data, the steps for verifying the integrity of the checksum data piece by piece include: Use session key data to perform fragmented encryption on firmware update package data, and obtain encrypted fragmented data; Verification value data is generated for each encrypted data fragment based on the SHA-3 hash algorithm, and a Merkle tree structure is constructed to store the verification value data; When receiving firmware update package data, the encrypted fragment data is extracted piece by piece and its checksum data is verified to match the Merkle tree structure in order to verify the integrity of the checksum data.

5. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 1, characterized in that, The steps of generating incremental update log data based on firmware update package data during the firmware storage phase, storing the incremental update log data in the security partition of BMC flash memory, and monitoring abnormal changes in the incremental update log data based on a differential hash algorithm include: Incremental update log data is generated based on firmware update package data; Divide the BMC flash memory into a security partition to store incremental update log data; The differential hashing algorithm is used to monitor changes in incremental update log data in real time. If an anomaly is detected in the incremental update log data, an alert is generated and firmware loading is paused.

6. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 5, characterized in that, The steps for using the differential hash algorithm to monitor changes in incremental update log data in real time include: The incremental update log data is processed based on the sliding window algorithm to generate the current fingerprint data. Perform a differential comparison between the current fingerprint data and the pre-stored baseline fingerprint data; If the difference value exceeds the preset threshold rule data, the incremental update log data is determined to be abnormal.

7. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 1, characterized in that, During the firmware execution phase, a trusted execution environment is established through a security chip and baseline hash value data is pre-stored. Periodic memory integrity verification is performed on key processes of the BMC kernel to obtain memory verification result data. If the memory verification result is abnormal, the steps to freeze the management interface and trigger an alarm notification include: A trusted execution environment is established using a secure chip, and pre-stored baseline hash value data is loaded. Periodically scan key processes in the BMC kernel to obtain memory region data; Based on a preset range of key function segment addresses, extract key function segment data from memory region data; Calculate the hash value of the key function segment data and compare it with the baseline hash value data to generate memory verification result data.

8. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 7, characterized in that, The step of calculating the hash value of the key function segment data and comparing it with the baseline hash value data to generate memory verification result data includes: Extract key function segment data; The fast hash algorithm is applied to calculate the real-time hash value of key function segment data; The real-time hash value data is compared with the pre-stored baseline hash value data. If they are inconsistent, the memory verification result data is determined to be abnormal.

9. The server BMC dynamic security authentication and firmware protection method based on hardware root of trust as described in claim 1, characterized in that, The steps of real-time collection of security event log data, analysis of the security event log data based on the anomaly detection model, and triggering of the automatic locking mechanism include: Collect security event log data in real time from the system log; Security event log data is analyzed based on a Bayesian classifier model and preset threshold rules to detect abnormal behavior. If abnormal behavior is detected, a lock command is generated to trigger the automatic locking mechanism.

10. A server BMC dynamic security authentication and firmware protection system based on hardware root of trust, characterized in that, The hardware root-based server BMC dynamic security authentication and firmware protection system includes: a memory, a processor, and a hardware root-based server BMC dynamic security authentication and firmware protection program stored on the memory and executable on the processor. The hardware root-based server BMC dynamic security authentication and firmware protection program is configured to implement the steps of the hardware root-based server BMC dynamic security authentication and firmware protection method as described in any one of claims 1 to 9.

Images