Compute instance security domain full lifecycle binding method and system

By identifying and protecting the security domain identifier during the creation, operation, loading, and destruction of computing instances, the problem of variable security domain ownership for computing instances is solved, achieving the immutability of security domains and security assurance throughout the entire lifecycle.

CN123122906APending Publication Date: 2026-11-17BANGYAN TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611282170.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-24
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing technologies, the security domain affiliation of computing instances is variable, which poses a risk of illegal tampering and cannot meet the rigid requirement of maintaining the security domain affiliation of computing instances throughout their entire lifecycle.

Method used

When a compute instance is created, its security domain identifier is determined, and the binding relationship between the security domain identifier and the instance is written into the security domain binding record. Integrity protection is achieved through methods such as digital signatures. During runtime, resource access is constrained based on the security domain identifier, and online modification is rejected. Security domain attribute matching is verified during loading. Security operations are performed and audit records are generated during destruction.

Benefits of technology

This change transforms the security domain affiliation of computing instances from a configurable attribute to an inherent attribute, preventing unauthorized tampering, ensuring the structural integrity and durability of security domains, and guaranteeing the consistency of policy execution and traceability throughout the entire lifecycle.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure FT_1
    Figure FT_1
  • Figure FT_2
    Figure FT_2
  • Figure FT_3
    Figure FT_3
Patent Text Reader

Abstract

This application provides a method and system for binding security domains throughout the entire lifecycle of a computing instance, relating to the field of cloud computing technology. The method includes: when creating a computing instance, assigning it a security domain identifier and writing the binding relationship between the instance identifier and the security domain identifier into a security domain binding record with integrity protection; during the operation of the computing instance, restricting resource access according to the binding record and rejecting online modifications to the security domain identifier; when loading the computing instance into a target resource environment, verifying the binding record, and completing the loading only if the security domain attribute of the target environment matches the security domain identifier; when destroying the computing instance, performing security operations based on the security domain identifier and generating audit records. This invention elevates the security domain to an inherent attribute of the computing instance, effectively preventing the risk of unauthorized tampering, ensuring the consistency of policy execution, and achieving closed-loop security management throughout the entire lifecycle.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of cloud computing technology, specifically relating to a method and system for binding the security domain of a computing instance throughout its entire lifecycle. Background Technology

[0002] In existing cloud computing and centralized computing architectures, firewall policies, Virtual Local Area Networks (VLANs), Access Control Lists (ACLs), or security groups are commonly used to achieve network isolation between computing instances. However, a common drawback of these methods is that security domain affiliation is an attribute that can be configured and modified during runtime. This means that due to administrator errors, incorrect policy distribution, or attackers using privilege escalation, it is possible to illegally change the security affiliation of a computing instance during its operation, thereby introducing security risks.

[0003] For example, while some solutions statically divide resources into security zones, this division is essentially a reconfigurable management attribute rather than an inherent characteristic of the instance. Other solutions, while supporting the propagation of security classifications throughout the service lifecycle, allow classifications to evolve with deployment, migration, and other processes, making security attribution still variable.

[0004] In high-security application scenarios (such as government multi-network isolation and separation of financial production and office), regulations require that the security domain affiliation of computing instances must be structural, persistent, and immutable throughout the entire process of their creation, operation, migration, and destruction. Existing technologies cannot meet this rigid requirement of making security domains an inherent attribute of computing instances. They generally suffer from technical defects such as variable security domain affiliation, lack of unified resource constraint identifiers, easy loss or tampering of security attributes during migration, and inability to audit and trace the ownership of residual data after instance destruction. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for binding the security domain of a computing instance throughout its entire lifecycle, which aims to solve the technical problem that the security domain ownership of a computing instance is a variable configuration in the prior art, which poses a risk of being illegally tampered with and cannot meet the rigid security requirement of keeping the security domain ownership unchanged throughout the entire lifecycle of the computing instance.

[0006] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for binding a computing instance security domain throughout its entire lifecycle, including: When creating a computing instance, a corresponding security domain identifier is determined for the computing instance, and the binding relationship between the identifier of the computing instance and the security domain identifier is written into the security domain binding record to establish the correspondence between the computing instance and the security domain identifier. During the operation of the computing instance, resource access to the computing instance is restricted according to the security domain identifier constraint corresponding to the computing instance, and online modification of the security domain identifier corresponding to the computing instance is denied. The computing instance is loaded into the target computing node. The loading includes: verifying the security domain binding record, and only loading the computing instance is completed when the security domain attribute of the target computing node matches the security domain identifier corresponding to the computing instance. When the computing instance is destroyed, at least one security operation is performed on a set of predefined resources associated with the computing instance based on the security domain identifier corresponding to the computing instance, and an audit record containing the identifier of the computing instance and its corresponding security domain identifier is generated.

[0007] Secondly, this application provides a computing instance security domain full lifecycle binding system for managing computing instances, including: The binding management module is used to determine the corresponding security domain identifier for the computing instance when the computing instance is created, and write the binding relationship between the identifier of the computing instance and the security domain identifier into the security domain binding record to establish the correspondence between the computing instance and the security domain identifier. The access constraint module is used to restrict resource access of the computing instance based on the security domain identifier constraint corresponding to the computing instance during the operation of the computing instance, and to refuse online modification of the security domain identifier corresponding to the computing instance. A loading verification module is used to control the loading of the computing instance to the target computing node. The loading includes: verifying the security domain binding record, and only completing the loading of the computing instance when the security domain attribute of the target computing node matches the security domain identifier corresponding to the computing instance. The destruction audit module is used to perform at least one security operation on a set of predefined resources associated with the computing instance when the computing instance is destroyed, based on the security domain identifier corresponding to the computing instance, and generate an audit record containing the identifier of the computing instance and its corresponding security domain identifier.

[0008] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the computing instance security domain full lifecycle binding method described in any one of the above.

[0009] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the computing instance security domain full lifecycle binding method described in any one of the above.

[0010] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the computing instance security domain full lifecycle binding method described above.

[0011] According to the specific embodiments provided in this application, the following technical effects are disclosed: 1. By determining the security domain identifier when a compute instance is created and forming a security domain binding record with integrity protection, the security domain is elevated from a runtime configurable attribute to an inherent affiliation throughout the entire lifecycle of the compute instance. This effectively prevents the risk of illegally upgrading or downgrading the security level through online configuration modifications, and enhances the structural and durable nature of business isolation.

[0012] 2. By using the same immutable binding relationship to perform verification and constraints in critical operations such as the running, loading, and migration of computing instances, security policy vulnerabilities caused by inconsistencies between security identifiers and underlying resource configurations are effectively avoided, ensuring the consistency of policy execution.

[0013] 3. By refusing online modification of the security domain identifier corresponding to the computing instance identifier and stipulating that security domain changes must be implemented by destroying the old instance and creating a new instance, the "intra-domain migration of the same instance" and "cross-domain change" are clearly separated technically, making the security management model more rigorous and clear.

[0014] 4. By performing security operations based on the security domain identifier bound to the computing instance when it is destroyed and generating an audit log containing the instance identifier and security domain identifier, it is ensured that the entire process of the computing instance from creation to destruction is traceable, forming a complete closed loop in management and security, and meeting the traceability requirements in high security and compliance scenarios. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the method for binding the security domain of a computing instance throughout its entire lifecycle, as described in one embodiment of this application.

[0017] Figure 2 This is a schematic diagram of the architecture of a computing instance security domain full lifecycle binding system in one embodiment of this application.

[0018] Figure 3This is a schematic diagram illustrating the process of securely migrating a computational instance in one embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] This invention provides a method and system for binding the security domain of a computing instance throughout its entire lifecycle, aiming to address the risk of unauthorized modification or misconfiguration of the security ownership of computing instances, which is a variable configuration item in existing technologies. The core idea of ​​this technical solution is to transform the security domain from a runtime configurable management attribute into an inherent attribute established from the creation of the computing instance and spanning its entire lifecycle, thereby achieving strong structural isolation.

[0022] The computing instance security domain full lifecycle binding method provided by this invention specifically includes: Step S110: When creating a computing instance, determine the corresponding security domain identifier for the computing instance, and write the binding relationship between the computing instance identifier and the security domain identifier into the security domain binding record to establish the correspondence between the computing instance and the security domain identifier.

[0023] Step S120: During the operation of the computing instance, resource access to the computing instance is restricted according to the security domain identifier constraint corresponding to the computing instance, and online modification of the security domain identifier corresponding to the computing instance is rejected.

[0024] Step S130: Load the computing instance to the target computing node. Loading includes: verifying the security domain binding record. The loading of the computing instance is only completed when the security domain attribute of the target computing node matches the security domain identifier corresponding to the computing instance.

[0025] Step S140: When destroying a computing instance, based on the security domain identifier corresponding to the computing instance, perform at least one security operation on a set of predefined resources associated with the computing instance, and generate an audit record containing the identifier of the computing instance and its corresponding security domain identifier.

[0026] Specifically, when a computing instance is created, it is first assigned a unique and explicit security domain identifier. This aims to endow the computing instance with an immutable security foundation from its inception. Subsequently, the system writes the correspondence between the computing instance's identifier and the determined security domain identifier—the binding relationship—into a security domain binding record with integrity protection. Integrity protection is applied through methods such as digital signatures, message authentication codes, or measurement using trusted hardware (such as TPM), ensuring that the binding record itself cannot be tampered with, providing a reliable foundation for security decisions in all subsequent stages. This solves the fundamental problem in the background technology where the security policy baseline is easily tampered with.

[0027] During the runtime of a compute instance, all access permissions and isolation policies for that instance's critical resources, including network connectivity, storage read / write, and peripheral device usage, are uniquely derived and enforced based on the corresponding security domain identifier in the security domain binding record. More importantly, the system proactively rejects any requests attempting to modify the security domain identifier corresponding to the compute instance's identifier online. This design ensures that once an instance is assigned to a security domain, it cannot "drift" to another security domain during runtime through simple configuration changes. This achieves runtime security domain solidification, avoiding the risks of administrator misoperation or attackers gaining privileges and performing unauthorized cross-domain operations.

[0028] When a compute instance needs to be loaded, such as during initial startup, restoration from backup, or acceptance as a migration target, this method introduces a mandatory pre-load verification checkpoint. The target resource environment (such as the target compute node) performing the loading operation must first obtain and verify the integrity and authenticity of the compute instance's security domain binding record. After successful verification, the target resource environment's own security domain attributes must be rigorously compared with the security domain identifier declared in the binding record. Only when both match perfectly will the system approve and complete the loading of the compute instance. This "verify first, then load" mechanism physically or logically prevents compute instances from being incorrectly placed in mismatched security environments; for example, it effectively prevents a "production domain" instance from being started on a "workplace domain" server.

[0029] Finally, at the end of the compute instance's lifecycle, during the destruction phase, this method constructs a secure closed loop. Based on the instance's security domain identifier, the system performs cleanup or invalidation operations on all predefined resources associated with it within the boundary of its respective security domain. This includes completely erasing data volumes it has used, permanently invalidating its dedicated storage encryption keys, or removing its references to specific network or peripheral resources. After completing these security operations, the system generates a final audit log containing the compute instance identifier, its security domain identifier, and operation details. This ensures that sensitive data does not remain in the wrong domain after instance destruction and provides traceable evidence of security compliance throughout the entire lifecycle.

[0030] In some embodiments, the security domain binding record also includes a binding version, and when verifying the security domain binding record, it also includes verifying whether the binding version is not lower than the accepted binding version of the target compute node, and the target compute node rejects the security domain binding record with a binding version lower than its accepted binding version.

[0031] Specifically, to defend against replay attacks using old, but previously legitimate, binding records, the security domain binding record can also include a "binding version" field. Accordingly, when verifying the security domain binding record, the loading verification process adds an extra step: verifying the binding version number of the record to be loaded, ensuring it is not lower than the highest binding version that the target compute node has recorded and accepted in its trusted storage. By enforcing a rule of mandatory one-way version increment, any attempt to start an instance using an outdated configuration will be identified and rejected, significantly enhancing system security.

[0032] In some embodiments, the security domain binding record also includes a rollback protection value, and when verifying the security domain binding record, it also includes verifying whether the rollback protection value is not lower than the accepted rollback protection value of the target compute node. The target compute node rejects the security domain binding record whose rollback protection value is lower than its accepted rollback protection value.

[0033] Specifically, to achieve a similar security effect as binding versions, the security domain binding record can also include a "rollback protection value," such as a monotonically increasing counter or timestamp. Similarly, when verifying the binding record, the target compute node checks whether this rollback protection value is not lower than its locally accepted highest value. This mechanism also effectively prevents configuration rollback attacks, providing another layer of protection for the trustworthiness of the binding relationship. Binding versions and rollback protection values ​​can be used individually or in combination for stronger security.

[0034] In some embodiments, the compute instance security domain lifecycle binding method further includes: periodically verifying the security domain attributes of each resource in the target compute node connected to the compute instance during the operation of the compute instance; and performing at least one of the following actions when the security domain attribute of any resource is found to be inconsistent with the security domain identifier: disconnecting the corresponding resource, suspending the compute instance, or generating a security event.

[0035] Specifically, to address potential security policy deviations caused by changes in the external environment during runtime, this method may also include periodically and proactively verifying the security domain attributes of various resources (such as network ports and storage volumes) actually connected to the compute instance during runtime. The principle behind this is to extend security assurance from static pre-load checks to dynamic, continuous runtime monitoring. In this way, even if external factors (such as network administrator misoperation) change the security attributes of resources in the target compute node, the system can promptly detect this inconsistency.

[0036] Based on the aforementioned periodic checks, once a discrepancy is found between the actual resource security domain attributes and the security domain identifier declared by the computing instance in its binding record, the system will immediately execute preset actions. These actions include at least: immediately disconnecting the connection to the mismatched resource to cut off potential risk paths; or suspending the entire computing instance, freezing all its activities pending manual intervention; or at least generating a high-priority security event alert to notify the security operations team. This automated, rapid response mechanism significantly shortens the risk exposure window and effectively curbs security deviations.

[0037] In some embodiments, when a compute instance is received as a migration target, the method further includes: before initiating the migration, selecting a target compute node only from compute nodes whose security domain attributes match the security domain identifier corresponding to the compute instance; and / or, releasing the compute instance from the compute node it was on before the migration only after the target compute node has been verified and returned a successful reception result.

[0038] Specifically, the above embodiments are enhanced for the migration of computing instances. Before initiating a migration operation, the management platform first performs a pre-screening, selecting target computing nodes only from those computing nodes whose own security domain attributes match the security domain identifier of the computing instance to be migrated. This "pre-filtering" design avoids initiating invalid migration attempts to a large number of unqualified nodes, improves migration efficiency, and reduces the possibility of errors from the source.

[0039] In the aforementioned secure migration process, to ensure the atomicity and data consistency of the migration process, the above embodiments also introduce a "confirm first, release later" mechanism. Specifically, the management platform will only issue an instruction to the source compute node to release the original compute instance and its occupied resources after the target compute node has completed all pre-load verifications of the compute instance and explicitly returned a confirmation result of successful reception to the management platform. If the target node returns a failure during the verification process, the entire migration is aborted, and the source instance remains unchanged or is restored, ensuring that there will be no instance loss or inconsistency under any circumstances.

[0040] Furthermore, in the implementation combining the two migration enhancement features mentioned above, a complete and highly secure migration process is achieved: First, the management platform filters qualified candidate target nodes based on the instance's security domain identifier; then, only after the target node completes all verifications and confirms successful reception is the instance on the source node finally released. This combined scheme, through the two checkpoints of "pre-screening" and "post-confirmation," together constitutes a robust migration security system, effectively preventing various risks during the migration process.

[0041] In some embodiments, audit records are written to an audit domain independent of system administration privileges. The audit domain employs at least one of the following techniques to ensure the immutability of audit records: append-only logging, write-unalterable storage, hash chains, or trusted timestamps.

[0042] Specifically, to meet high-level compliance and auditability requirements, audit logs generated during the destruction phase are written to an audit domain independent of regular system management permissions. The principle behind this design is to achieve "separation of rights and responsibilities," meaning that administrator accounts responsible for daily system maintenance do not have permission to access or modify log data used for auditing. By storing audit data in a dedicated, access-restricted area, it effectively prevents internal personnel (whether malicious or unintentional) from concealing traces of violations.

[0043] Building upon the aforementioned independent audit domain, at least one of several technical methods can be employed to ensure the immutability of audit records. For example, an append-only log format can be used, or the log can be stored on a write-once, read-many (WORM) physical medium. Hash chaining technology can also be used to link successive log entries with cryptographic digests, ensuring that any modification to the historical record will break the integrity of the entire chain. Furthermore, a trusted timestamp service can be combined to attach a legally valid trusted timestamp to each log entry. The application of these technologies collectively guarantees the absolute credibility of audit records as final evidence.

[0044] In some embodiments, accessing resources of a computing instance based on the security domain identifier constraint corresponding to the computing instance includes: determining the security domain to which the computing instance belongs based on the security domain identifier constraint corresponding to the computing instance; configuring the computing instance to only use the storage key corresponding to its security domain for resource access; wherein, different security domains are configured with different storage keys that are independent of each other.

[0045] Specifically, to achieve strong isolation, independent storage keys can be configured for different security domains. When a compute instance is created and bound to a security domain, it is authorized to use only the storage key corresponding to that security domain. Instance data is encrypted using this key when written to backend storage. Therefore, even if an instance in another security domain gains underlying access to the storage area due to misconfiguration, it will only read unparseable ciphertext because it lacks the correct decryption key. This cryptographically based isolation is far stronger than traditional logical access control list-based isolation.

[0046] In some embodiments, changing the security domain identifier of a computing instance is achieved by destroying the computing instance and creating a new computing instance with a new identifier and a new security domain binding record.

[0047] Furthermore, to thoroughly implement the core principle of "immutable security domains" at the management strategy level, the above embodiments explicitly stipulate that any requirement to change the security domain identifier of a computing instance must be achieved by destroying the original computing instance and creating a new computing instance with a completely new computing instance identifier and a completely new security domain binding record in the target security domain. This rigid strategy of "preferring reconstruction to modification" effectively prevents any attempt to "take shortcuts" by modifying instance attributes to cross security domain boundaries, fundamentally ensuring the complete realization of the design intent of this invention.

[0048] Reference Figure 2 This invention also provides a computing instance security domain full lifecycle binding system. This system manages a computing instance, and its structure corresponds one-to-one with the steps of the aforementioned method. The system includes a binding management module, an access constraint module, a load verification module, and a destruction audit module. These modules can be logically separate software services or functional components, working together to achieve complete lifecycle binding.

[0049] Specifically, the core responsibility of the binding management module is to determine a security domain identifier for a compute instance when it is created, and to write the binding relationship between the compute instance identifier and the security domain identifier into a security domain binding record with integrity protection applied.

[0050] The access constraint module is responsible for continuously performing its function during the runtime of the compute instance. It reads the security domain binding records generated by the binding management module and, based on the security domain identifiers recorded therein, uniformly constrains and configures access to various resources of the compute instance. These resources include, but are not limited to, host domain attributes, network domain attributes, storage domain attributes, and peripheral domain attributes. Simultaneously, this module also undertakes the responsibility of proactive defense, explicitly rejecting any requests attempting to modify the security domain identifier corresponding to the compute instance identifier online.

[0051] The load verification module is a crucial step in ensuring the effective implementation of security domain policies, and it is typically deployed on each compute node. Its function is to forcibly verify the security domain binding records of a compute instance before it is loaded (e.g., during startup, recovery, or migration) to a target resource environment, and to rigorously compare these records with the security domain attributes of the target resource environment. Only if the verification passes and the attributes match will the module approve the load operation; otherwise, it will reject the request.

[0052] The destruction audit module is responsible for the end-of-lifecycle closure. When it receives an instruction to destroy a compute instance, this module controls and supervises the execution of a series of security operations, ensuring that these operations (such as data cleanup and key expiration) strictly occur within the security domain bound to the instance. After the operations are completed, the module generates a final audit log containing the compute instance identifier and security domain identifier for compliance checks and post-event traceability.

[0053] The present invention also provides a computing node device, which is a physical entity that carries computing instances and executes the core security mechanism of the present invention. The device internally includes at least one binding verification unit and one domain constraint enforcement unit.

[0054] The binding verification unit is specifically responsible for performing critical verification tasks upon receiving a load request for a compute instance (whether it's the initial startup or a migration). It reads the security domain binding record of the compute instance and uses cryptographic methods or trusted computing technology to verify the integrity and authenticity of the record, ensuring it has not been tampered with.

[0055] The domain constraint enforcement unit then follows the binding verification unit. It is responsible for acquiring and comparing the resource security domain attributes of the compute node itself (e.g., the physical cluster partition to which the node belongs, the VLAN connected to the network port, the accessible storage LUNs, etc.) with the security domain identifier declared in the security domain binding record. If the two do not match, the unit will immediately take action to refuse to load the compute instance, thus forming a robust defense against instances "accidentally" entering an incompatible environment.

[0056] Furthermore, the domain constraint enforcement unit within the compute node device not only performs a one-time check before loading but is also responsible for continuous monitoring after the compute instance is successfully loaded and enters the running state. It periodically re-verifies the actual security domain attributes of various resources connected to the compute instance. If any mismatch is detected during operation, the unit immediately executes pre-defined actions, such as disconnecting the mismatched resource connection, suspending the potentially at-risk compute instance, or generating a security event alarm, thereby extending security protection from static configuration to the entire dynamic operation process.

[0057] The following detailed embodiments illustrate the computing instance security domain full lifecycle binding method and system provided by the present invention.

[0058] Reference Figure 1 , Figure 2 and Figure 3 In a basic embodiment, the basic process for implementing full lifecycle binding of computing instance security domains is demonstrated, covering four core stages: "binding at creation time, runtime constraints, pre-load verification, and closed-loop at destruction time".

[0059] In the computing instance security domain lifecycle binding system of this embodiment, there is a DMS management and control platform that serves as the decision center and management entry point, as well as binding management modules, access constraint modules, loading verification modules, and destruction audit modules that respectively implement specific functions. Among them, the loading verification module, as a security gatekeeper, is embedded in each computing node device.

[0060] During the creation phase, when a user requests the creation of a new computing instance through the DMS management and control platform and specifies a security domain (e.g., "production domain"), the DMS management and control platform instructs the binding management module to perform a binding operation. The binding management module generates a unique instance identifier for this new instance and writes it, along with the user-specified "production domain" security domain identifier, into a new security domain binding record. To prevent tampering, this record is digitally signed using the platform's private key during generation.

[0061] Once in operation, the access constraint module reads and verifies the newly created security domain binding record. Based on the "production domain" identifier, the module automatically configures the resources required by the instance. For example, it connects the virtual network interface card of the compute instance to the virtual switch corresponding to the network domain attribute marked "production domain," restricts the access permissions of its data volume to the storage pool with the "production domain" storage domain attribute, and configures its peripheral redirection policy to comply with the requirements of the peripheral domain attribute. During this period, any request attempting to modify the instance's security domain identifier via API or other means will be identified and resolutely rejected by the access constraint module, triggering an alarm.

[0062] When a compute instance needs to be started or migrated from one compute node to another target compute node, the security mechanism during the loading phase is activated. The loading verification module on the target compute node executes a rigorous joint verification process. First, it reads the security domain binding record of the compute instance to be loaded. Next, it uses the public key of the DMS management platform to verify the digital signature of the record, ensuring its integrity and authenticity. After successful verification, it sequentially obtains and compares the actual environment attributes (security domain attributes of various resources in the target compute node) such as the host domain attribute, network port domain attribute, and available storage resource domain attribute of the target compute node to see if they all match the "production domain" identifier declared in the record. Only when all verification items match will the loading verification module allow the instance to be loaded and run on the target compute node. If any mismatch is found, such as the target compute node itself belonging to the "office domain," loading will be immediately aborted, and a security violation event will be reported to the DMS management control platform. Through this process, mandatory constraints on the instance's operating environment are implemented.

[0063] Finally, during the destruction phase, when the instance's lifecycle ends, the DMS management and control platform issues a destruction command to the destruction audit module. The destruction audit module ensures that all cleanup operations are performed within the boundaries of the "production domain," such as overwriting data on storage space used by the instance. After cleanup, the module generates an audit log containing the instance identifier, its original security domain ("production domain"), destruction time, and operation result, and archives it in the audit system. This process ensures that the security domain binding remains consistent and traceable from instance creation to destruction, thus achieving a basic full lifecycle security loop.

[0064] Building upon the basic implementation described above, enhanced protection and continuous verification mechanisms can be introduced to address more complex security threats. In a preferred embodiment, the data structure of the security domain binding record is expanded to include "binding version" and "anti-rollback value" fields. Simultaneously, a periodic verifier is added within the domain constraint enforcement unit of each compute node device. When the DMS management platform needs to update a non-core attribute of an instance for compliance procedures (without changing the security domain identifier), it generates a new security domain binding record with an incrementing version number. The compute node securely records the highest binding version number it has loaded for a given instance in its local trusted storage (e.g., within a TPM chip). During pre-load verification, the loading verification module, in addition to verifying the signature, also compares the "binding version" of the record to be loaded with the highest version number recorded locally. If a record with a lower version number is received, the system determines this to be a replay attack attempting to use an outdated configuration and immediately refuses to load it. This mechanism effectively defends against attacks that exploit old, obsolete, legitimate configurations to bypass current security policies.

[0065] In this preferred embodiment, after the instance is successfully loaded and running, the periodic verifier on the compute node automatically rechecks the domain attributes of the resources currently actually connected to the instance at regular intervals (e.g., every 30 seconds). For example, it queries the current VLAN configuration of the switch port or the access domain settings of the volume mounted by the instance on the storage array. Once it is found that the actual attribute of a resource is no longer consistent with the security domain identifier fixed in the instance binding record (e.g., a network administrator mistakenly changed the port from the "production domain" VLAN to the "office domain" VLAN on the switch), the domain constraint enforcement unit will immediately trigger a preset emergency response. The response measures may be to immediately disconnect the mismatched network connection to prevent potential data leakage; or to suspend the entire compute instance for investigation by security personnel; and at the same time generate a critical security alert. This continuous runtime verification extends security protection from static "access control" to dynamic "process monitoring," enabling timely detection and containment of security risks caused by environmental changes or management errors, significantly improving the robustness of the overall solution.

[0066] Furthermore, a preferred embodiment details the interactive process of secure migration and demonstrates an enhanced resource constraint implementation using a storage encryption domain as an example. This process involves three parties: the DMS management and control platform, the source compute node, and the target compute node. In this implementation, the storage system configures independent storage encryption domains for different security domains (e.g., "production domain," "office domain"), each using a completely different root key. A compute instance can only obtain the key of its own security domain.

[0067] When the DMS management and control platform needs to migrate an instance in the "production domain," it first queries the resource pool, selecting only compute nodes belonging to the "production domain" as candidate target compute nodes. After selecting a target, the platform sends a migration preparation command to the source compute node, which then begins transferring the instance's memory and disk data to the target compute node. During or after data transfer, the load verification module on the target compute node performs a complete pre-load verification. If verification fails (e.g., a peripheral device on the target node is found to be non-compliant with "production domain" requirements), the target compute node immediately clears all received temporary data and returns a "verification failed" result to the DMS management and control platform. Upon receiving this result, the platform aborts the entire migration process and ensures the instance on the source compute node resumes operation. Only after the target compute node passes all verifications and returns a "received successfully" feedback to the platform will the DMS management and control platform issue the final "release instance" command to the source compute node. This atomic operation process of "verify first, confirm then release" ensures the atomicity of the migration process, avoiding the risk of the instance entering an insecure environment or being lost due to migration failure.

[0068] In this implementation, the constraints of the storage encryption domain play a crucial role. Since data in the "production domain" instance is encrypted using the "production domain" key during writing, even in an extreme scenario where an instance in the "office domain" gains unauthorized read / write access to this storage, it cannot obtain the "production domain" key from the key management system. Therefore, the data it reads is simply ciphertext that cannot be decrypted. This cryptographically based isolation mechanism penetrates traditional logical access control layers, providing a more fundamental and reliable security guarantee.

[0069] To meet the high-level compliance requirements of sectors such as finance and government, other embodiments also introduce an independent audit mechanism separate from management permissions and a multi-role approval emergency response process for disaster scenarios. In this embodiment, the system establishes an independent audit domain with independent storage resources and dedicated service accounts. Regular system administrator accounts do not have the authority to modify or delete any log data within this domain. All critical events related to security domain binding, such as the creation of binding records, alarms for failed verification, records of rejected migrations, and confirmation of instance destruction, are forcibly and in real-time sent to this independent audit domain. To ensure the immutability of the logs, the audit domain can employ hash chain technology, binding each new log entry to the hash value of the previous log entry, forming an immutable cryptographic chain. Any tampering with historical logs will cause the chain verification to fail. This design provides credible and non-repudiable audit evidence to meet high-level regulatory requirements.

[0070] This embodiment also defines three separate management roles: "System Administrator," "Security Administrator," and "Security Auditor," to handle extreme and exceptional situations. For example, in the event of a disaster such as a crash of the DMS management and control platform or loss of keys, any emergency operation requiring bypassing regular security mechanisms (such as forcibly destroying an out-of-control instance) must be jointly approved by the administrators of these three roles to generate a one-time authorization token with strict time limits and operational restrictions. The entire emergency response process is also independently audited, thus providing necessary "escape pod" functionality while also helping to prevent the abuse of highest privileges.

[0071] The above is a detailed description of the computing instance security domain full lifecycle binding method and system provided by the present invention.

[0072] The technical solution of this invention has significant application value in multiple industries.

[0073] In the financial industry, core applications such as trading systems face rigid regulatory requirements of "no downgrading." Using this invention, a computing instance bound to the "trading production domain" will be forced by the system to be loaded only into a resource pool belonging to the same "trading production domain" level under any circumstances, including routine resource scheduling, failover, or disaster recovery. Any request attempting to restore or migrate it to the "office domain" or "test domain" environment will be directly rejected by the verification module during the access phase, triggering a security event report, thus structurally ensuring compliance.

[0074] In disaster recovery drills, enterprises can pre-define and register security domain resources of various levels at a remote disaster recovery site, using the same standards as the primary site. During a disaster recovery drill switchover, compute instances from the primary site are migrated to the disaster recovery site in batches. The mechanism of this invention automatically ensures that each instance is accurately restored within its corresponding security domain partition; for example, instances in the "core business domain" will only reside on the "core business domain" resources at the disaster recovery site. All domain compliance verification results throughout the entire drill process are recorded in detail in the audit log, serving as strong evidence to regulatory agencies or internal audit departments demonstrating the true effectiveness of their disaster recovery capabilities.

[0075] In government or classified application scenarios, information systems typically require strict hierarchical protection. Using this invention, the binding relationship of a computing instance processing "classified" information can be established at creation time through separate approvals by the system administrator, security administrator, and security auditor. Throughout its lifecycle, this instance will be strictly confined to a "classified" network and storage environment. When the project ends and the instance needs to be destroyed, the destruction process automatically triggers a media-level data erasure operation within the "classified domain" that complies with national security standards, and invalidates its encryption key. The destruction audit report, including the erasure scope checksum and the approval records of the three personnel, will be archived in an independent audit domain strictly isolated from ordinary management systems for long-term review.

[0076] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0077] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0078] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0079] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0080] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0081] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for binding security domains throughout the entire lifecycle of a computing instance, characterized in that, include: When creating a computing instance, a corresponding security domain identifier is determined for the computing instance, and the binding relationship between the identifier of the computing instance and the security domain identifier is written into the security domain binding record to establish the correspondence between the computing instance and the security domain identifier. During the operation of the computing instance, resource access to the computing instance is restricted according to the security domain identifier constraint corresponding to the computing instance, and online modification of the security domain identifier corresponding to the computing instance is denied. The computing instance is loaded into the target computing node. The loading includes: verifying the security domain binding record, and only loading the computing instance is completed when the security domain attribute of the target computing node matches the security domain identifier corresponding to the computing instance. When the computing instance is destroyed, at least one security operation is performed on a set of predefined resources associated with the computing instance based on the security domain identifier corresponding to the computing instance, and an audit record containing the identifier of the computing instance and its corresponding security domain identifier is generated.

2. The computing instance security domain full lifecycle binding method according to claim 1, characterized in that, The security domain binding record also includes a binding version, and when verifying the security domain binding record, it also includes verifying whether the binding version is not lower than the accepted binding version of the target computing node. The target computing node rejects the binding record with a version lower than its accepted binding version.

3. The method for binding the entire lifecycle of a computing instance security domain according to claim 1, characterized in that, The security domain binding record also includes a rollback prevention value. When verifying the security domain binding record, it also includes verifying whether the rollback prevention value is not lower than the accepted rollback prevention value of the target computing node. The target computing node rejects security domain binding records where the rollback prevention value is lower than its accepted rollback prevention value.

4. The method for binding the security domain of a computing instance throughout its entire lifecycle according to claim 1, characterized in that, Also includes: During the operation of the computing instance, the security domain attributes of each resource in the target computing node connected to the computing instance are periodically verified. When the security domain attribute of any resource is found to be inconsistent with the security domain identifier, at least one of the following actions is performed: disconnecting the corresponding resource, suspending the computing instance, or generating a security event.

5. The method for binding the entire lifecycle of a computing instance security domain according to claim 1, characterized in that, When the compute instance is loaded and received as a migration target, the method further includes: Before initiating the migration, the target compute node is selected only from compute nodes whose security domain attributes match the security domain identifier corresponding to the compute instance; And / or, the computing instance is released only on the computing node where it was located before the migration after the target computing node has passed the verification and returned a successful receipt result.

6. The method for binding the security domain of a computing instance throughout its entire lifecycle according to claim 1, characterized in that, The audit logs are written to an audit domain that is independent of system administration privileges.

7. The method for binding the entire lifecycle of a computing instance security domain according to claim 6, characterized in that, The audit domain employs at least one of the following techniques to ensure the immutability of the audit records: append-only logging, write-unchangeable storage, hash chain, or trusted timestamp.

8. The method for binding the security domain of a computing instance throughout its entire lifecycle according to claim 1, characterized in that, Access to the resources of the computing instance based on the security domain identifier constraints corresponding to the computing instance includes: The security domain to which the computation instance belongs is determined based on the security domain identifier constraint corresponding to the computation instance. The computing instance is configured to access resources only using the storage key corresponding to its security domain; Different security domains are configured with different storage keys that are independent of each other.

9. The method for binding the security domain of a computing instance throughout its entire lifecycle according to claim 1, characterized in that, Changing the security domain identifier of the computing instance is achieved by destroying the computing instance and creating a new computing instance with a new identifier and a new security domain binding record.

10. A computing instance security domain full lifecycle binding system for managing computing instances, characterized in that, include: The binding management module is used to determine the corresponding security domain identifier for the computing instance when the computing instance is created, and write the binding relationship between the identifier of the computing instance and the security domain identifier into the security domain binding record to establish the correspondence between the computing instance and the security domain identifier. The access constraint module is used to restrict resource access of the computing instance based on the security domain identifier constraint corresponding to the computing instance during the operation of the computing instance, and to refuse online modification of the security domain identifier corresponding to the computing instance. A loading verification module is used to control the loading of the computing instance to the target computing node. The loading includes: verifying the security domain binding record, and only completing the loading of the computing instance when the security domain attribute of the target computing node matches the security domain identifier corresponding to the computing instance. The destruction audit module is used to perform at least one security operation on a set of predefined resources associated with the computing instance when the computing instance is destroyed, based on the security domain identifier corresponding to the computing instance, and generate an audit record containing the identifier of the computing instance and its corresponding security domain identifier.