An AI large model security all-in-one machine, a security channel establishment method and device

Abstract

The application discloses an AI large model security all-in-one machine, a security channel establishment method and device, relates to the technical field of multi-party secure computation, and comprises a security processor and a physical memory module; the security all-in-one machine has a plurality of trusted execution environments TEE, different TEEs have different keys for memory encryption and decryption, the memory key of the TEE is generated based on a hardware trusted root of the security processor for a memory encryption module of the security processor, and different TEEs are allocated with different memory spaces independently used on the physical memory module; an AI large model application is installed on each TEE, data generated by the AI large model application is encrypted and written and decrypted and read in the memory space allocated for the TEE by using the memory key corresponding to the TEE by the memory security module. By adopting the scheme, the data security of the AI large model local deployment is improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of multi-party secure computing, and particularly relates to an AI large model security all-in-one machine, a secure channel establishment method and device. BACKGROUND

[0002] Large language models (LLM) and their applications in the field of artificial intelligence have become a global research hotspot. A large language model is a deep learning model trained based on massive text data. It can not only generate natural language text, but also deeply understand the meaning of the text and process various natural language tasks such as text summarization, question answering and translation.

[0003] However, in the context of rapid evolution of large model technology and deep integration with industry, data security protection has become a core challenge restricting the development of artificial intelligence. Enterprises face serious data security risks when applying large models. For example, there are security risks of models and training data in the model localization deployment scenario.

[0004] In a private deployment environment, when enterprises use their own data to optimize model applications, they face the risk of potential unauthorized access to enterprise data by internal personnel, which makes it impossible to effectively protect the security of training data, model parameters and enterprise private knowledge bases. SUMMARY

[0005] The embodiments of the application provide an AI large model security all-in-one machine, a secure channel establishment method and device to solve the problem of poor data security for AI large model localization deployment in the prior art.

[0006] The embodiments of the application provide an AI large model security all-in-one machine, which comprises a secure processor and a physical memory module.

[0007] The secure all-in-one machine has a plurality of trusted execution environments (TEEs), different TEEs have different keys for memory encryption and decryption, and the memory keys of the TEEs are generated based on the hardware trusted root of the secure processor as memory keys for the memory encryption module of the secure processor. Different TEEs are allocated with different memory spaces for independent use on the physical memory module.

[0008] Each TEE is installed with an AI large model application, and the data generated by the AI large model application is encrypted and written and decrypted and read in the memory space allocated for the TEE by the memory security module using the memory key corresponding to the TEE.

[0009] Further, it further comprises a plurality of AI acceleration cards.

[0010] The AI acceleration card is connected with a device pass-through module of the secure processor;

[0011] The AI large model application of the TEE interacts with the AI acceleration card through the device pass-through module.

[0012] Further, the secure processor has a remote authentication module;

[0013] The remote authentication module realizes remote authentication of the TEE through message interaction with the client.

[0014] Further, a database is established in each TEE for storing identity authentication information;

[0015] In the process of establishing a secure channel with the client, the AI large model application authenticates the identity of the client based on the identity authentication information stored in the database of the TEE.

[0016] Further, the AI large model application has a public key for establishing a secure channel with the client;

[0017] The public key of the AI large model application is generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs are different.

[0018] The application also provides a secure channel establishment method, applied to the AI large model application in the TEE of the AI large model security all-in-one machine, comprising:

[0019] The AI large model application confirms the establishment information for establishing a secure channel through message interaction with the client;

[0020] Receive the identity certificate sent by the client, the identity certificate carrying identity authentication information for identity authentication;

[0021] Compare the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs, and authenticate the identity of the client;

[0022] After the identity authentication of the client is passed, send a token to the client, the token being carried in the data packet sent by the client to the AI large model application;

[0023] generate a session key through message interaction with the client based on the confirmed establishment information and the public key of the AI large model application, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

[0024] The embodiment of the application further provides a secure channel establishment method, applied to a client located outside the AI large model security all-in-one machine.

[0025] confirm establishment information for establishing a secure channel through message interaction with the AI large model application in the TEE;

[0026] send an identity certificate to the AI large model application, the identity certificate carrying identity authentication information used for identity authentication, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of the TEE to which the AI large model application belongs;

[0027] receive a token sent by the AI large model application to the client after the identity authentication of the client is passed, the token being carried in a data packet sent by the client to the AI large model application;

[0028] generate a session key through message interaction with the AI large model application based on the confirmed establishment information and the public key of the AI large model application, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

[0029] The embodiment of the application further provides a secure channel establishment device, applied to an AI large model application in a TEE of the AI large model security all-in-one machine.

[0030] a first establishment information confirmation module configured to confirm establishment information for establishing a secure channel through message interaction with a client;

[0031] a certificate receiving module configured to receive an identity certificate sent by the client, the identity certificate carrying identity authentication information used for identity authentication;

[0032] An identity authentication module is configured to authenticate the identity of the client by comparing identity authentication information carried in the identity certificate with identity authentication information stored in a database of a TEE to which the AI large model application belongs.

[0033] A token sending module is configured to send a token to the client after the identity authentication of the client is passed, the token being used to be carried in a data message sent by the client to the AI large model application.

[0034] A first session key generation module is configured to generate a session key by using a key exchange algorithm based on the confirmed establishment information and a public key of the AI large model application through message interaction between the client, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

[0035] Embodiments of the present application also provide a secure channel establishment device applied to a client located outside the AI large model security all-in-one machine, and the device comprises:

[0036] A second establishment information confirmation module is configured to confirm establishment information for establishing a secure channel through message interaction between the AI large model application in the TEE.

[0037] A certificate sending module is configured to send an identity certificate to the AI large model application, the identity certificate carrying identity authentication information used for identity authentication, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of a TEE to which the AI large model application belongs.

[0038] A token receiving module is configured to receive a token sent by the AI large model application to the client after the identity authentication of the client is passed, the token being used to be carried in a data message sent by the client to the AI large model application.

[0039] A second session key generation module is configured to generate a session key by using a key exchange algorithm based on the confirmed establishment information and a public key of the AI large model application through message interaction between the AI large model application, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

[0040] The application further provides an electronic device, comprising a processor and a machine readable storage medium, the machine readable storage medium stores machine executable instructions capable of being executed by the processor, and the processor is prompted by the machine executable instructions to implement any of the above-mentioned secure channel establishment methods applied to AI large model applications or to implement any of the above-mentioned secure channel establishment methods applied to clients.

[0041] The application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement any of the above-mentioned secure channel establishment methods applied to AI large model applications or to implement any of the above-mentioned secure channel establishment methods applied to clients.

[0042] The application further provides a computer program product containing instructions, which, when running on a computer, causes the computer to execute any of the above-mentioned secure channel establishment methods applied to AI large model applications or to execute any of the above-mentioned secure channel establishment methods applied to clients.

[0043] The application has the following beneficial effects:

[0044] In the scheme provided by the application, the AI large model secure all-in-one machine comprises a secure processor and a physical memory module, the secure all-in-one machine has a plurality of TEEs, different TEEs have different keys for memory encryption and decryption, as a memory key, the memory key of the TEE is generated by a memory encryption module of the secure processor based on a hardware trusted root of the secure processor, different TEEs are allocated with different memory spaces independently used on the physical memory module, and an AI large model application is installed on each TEE, data generated by the AI large model application is encrypted and written and decrypted and read in the memory space allocated for the TEE by the memory security module using the memory key corresponding to the TEE. By using the scheme, hardware encryption and decryption of data generated by the AI large model application in the TEE are realized, so that the host system, application programs outside the trusted execution environment cannot access the data generated by the AI large model application, and the operation and maintenance personnel of the host system cannot obtain the data generated by the AI large model application, and the memory keys of different TEEs are different, further realizing the isolation of data generated in different TEEs, and therefore, the data security of the AI large model local deployment is improved.

[0045] Other features and advantages of the application will be set forth in the following description, and in part will become apparent to those skilled in the art from the following or can be learned by practice of the application. The objects and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0046] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application, and explain the principles of the present application, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0047] Figure 1 A structural schematic diagram of an AI large model security all-in-one machine provided for an embodiment of the present application is shown in the figure;

[0048] Figure 2 A structural schematic diagram of an AI large model security all-in-one machine provided for another embodiment of the present application is shown in the figure;

[0049] Figure 3 A structural schematic diagram of an AI large model security all-in-one machine provided for another embodiment of the present application is shown in the figure;

[0050] Figure 4 A schematic diagram of interaction between a client and an AI large model application on an AI large model security all-in-one machine in an embodiment of the present application is shown in the figure;

[0051] Figure 5 A flowchart of a security channel establishment method applied to an AI large model application provided for an embodiment of the present application is shown in the figure;

[0052] Figure 6 A flowchart of a security channel establishment method applied to a client provided for an embodiment of the present application is shown in the figure;

[0053] Figure 7 A flowchart of a security channel establishment method provided for an embodiment of the present application is shown in the figure;

[0054] Figure 8 A structural schematic diagram of a security channel establishment device applied to an AI large model application provided for an embodiment of the present application is shown in the figure;

[0055] Figure 9 A structural schematic diagram of a security channel establishment device applied to a client provided for an embodiment of the present application is shown in the figure;

[0056] Figure 10 A structural schematic diagram of an electronic device provided for an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0057] In order to give an implementation scheme for improving the data security of AI large model localization deployment, the present application provides an AI large model security all-in-one machine, a security channel establishment method and device. The preferred embodiments of the present application are described below in conjunction with the drawings of the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. And in the case of no conflict, the embodiments and the features in the embodiments in the present application can be combined with each other.

[0058] An AI large model security all-in-one machine is provided, as shown in the accompanying drawings, comprising a secure processor and a physical memory module. Figure 1

[0059] The security all-in-one machine has a plurality of trusted execution environments (TEEs), and different TEEs have different memory encryption / decryption keys. The memory keys of the TEEs are generated by a memory encryption module of the secure processor based on a hardware trusted root of the secure processor. Different TEEs are allocated with different memory spaces for independent use on the physical memory module.

[0060] An AI large model application is installed on each TEE. Data generated by the AI large model application is encrypted and decrypted in the memory space allocated for the TEE by the memory security module using the memory key corresponding to the TEE.

[0061] The above AI large model security all-in-one machine provided by the embodiments of the present application realizes hardware encryption / decryption of data generated by AI large model applications in TEEs, so that the host system and applications outside the trusted execution environment cannot access the data generated by the AI large model applications, and the operation and maintenance personnel of the host system cannot obtain the data generated by the AI large model applications. In addition, the memory keys of different TEEs are different, further realizing the isolation of data generated in different TEEs, and thus improving the data security of the AI large model local deployment.

[0062] In an embodiment of the present application, the TEE can exist in the AI large model security all-in-one machine in the form of a confidential virtual machine or a confidential container. The TEE has an independent system kernel and is completely isolated from the host system. Therefore, the IT administrator / operation and maintenance personnel of the AI large model security all-in-one machine has no access to the data or code in the TEE, thereby preventing access and theft of business system data.

[0063] Different TEEs have different memory encryption / decryption keys, which are generated by the memory encryption module of the secure processor based on the hardware trusted root. For example, for a TEE, the memory key is generated based on the hardware trusted root and the attribute information of the TEE using a preset algorithm, and is stored. The generated memory key is not exposed to the host system or the hypervisor.

[0064] ​In the embodiments of the present application, different memory spaces are allocated on the physical memory module for different TEEs. For a TEE, the allocated memory space is only used by the TEE independently, and the memory encryption module encrypts and decrypts the data generated by the AI large model application in the TEE in the allocated memory space. The application program in the untrusted execution environment of the host machine cannot maliciously access and steal data from the memory space.

[0065] In an embodiment of the present application, as shown in Figure 2 The AI large model security all-in-one machine can further include a plurality of AI acceleration cards.

[0066] The AI acceleration card is connected to the device pass-through module of the security processor.

[0067] The AI large model application of the TEE and the AI acceleration card interact data through the device pass-through module.

[0068] In an embodiment of the present application, the device pass-through module can be a PCIe device pass-through module, which functions to unbind the AI acceleration card with a PCIe interface from the host machine at the device driver layer and pass it into the trusted execution environment TEE to pass through with the AI large model application in the TEE. After passing through, the host system has no access to the data in the AI acceleration card, thereby improving the security of the data in the AI acceleration card passed through with the TEE, and further improving the security of the data of the AI large model application in the TEE.

[0069] In an embodiment of the present application, as shown in Figure 3 The security processor of the AI large model security all-in-one machine can further have a remote authentication module.

[0070] The remote authentication module realizes remote authentication of the authenticity of the TEE through message interaction with the client.

[0071] In an embodiment of the present application, remote authentication can be initiated by a client located outside the AI large model security all-in-one machine, and the authenticity of the TEE of the AI large model security all-in-one machine is authenticated through message interaction with the remote authentication module. After authentication, a secure channel can be established between the client and the AI large model application in the TEE, and a data transmission operation can be initiated.

[0072] Further, in an embodiment of the present application, a database can be established in each TEE on the AI large model security all-in-one machine, for storing identity authentication information.

[0073] The AI large model application can authenticate the identity of the client based on the identity authentication information stored in the database of the TEE to which the AI large model application belongs in the process of establishing a secure channel with the client.

[0074] The identity authentication information used can be traditional username and password data, or biometric identity data information such as fingerprint, face, etc.

[0075] After the identity of the client is authenticated, a secure channel is established with the client, and data is transmitted through the established secure channel, for example, receiving a query request initiated by the client, returning a query result to the client, receiving data sent by the client for model training, etc.

[0076] Through the above identity authentication mechanism for the client, after the AI large model security all-in-one machine is deployed locally, the different TEEs on the AI large model security all-in-one machine can be accessed only by clients with specific identities, for example, the AI large model application in a TEE can be accessed only by specific organizations or specific department personnel of a company, thereby realizing the isolation of data between different organizations or departments in the AI large model service, and further improving the data security.

[0077] In actual applications, as shown in Figure 4 There are department A and department B (one is a finance department and the other is a marketing department) in a group, both of which have a demand for large model related reasoning business on the AI large model security all-in-one machine of the company, for example, department A uses a 32B model in TEE-1 on the AI large model security all-in-one machine, and department B uses a 70B model in TEE-2 on the AI large model security all-in-one machine, but at the same time, department A and department B have a demand for isolated and secure protection of their respective data requests, so the above AI large model security all-in-one machine provided by the embodiments of the present application can realize the isolation of data and further improve the data security.

[0078] In the embodiments of the present application, the AI large model application has a public key for establishing a secure channel with the client.

[0079] The public key of the AI large model application is generated by the secure processor based on the hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

[0080] Based on the above AI large model security all-in-one machine provided by the embodiments of the present application, the embodiments of the present application further provide a secure channel establishment method, which is applied to the AI large model application in the TEE of the above AI large model security all-in-one machine, as shown in Figure 5 , which comprises:

[0081] Step 51, the AI large model application confirms the establishment information for establishing the secure channel through message interaction with the client;

[0082] Step 52, the identity certificate sent by the client is received, and the identity authentication information for identity authentication is carried in the identity certificate;

[0083] Step 53, the identity of the client is authenticated by comparing the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs;

[0084] Step 54, after the identity authentication of the client is passed, a token is sent to the client, and the token is carried in the data message sent by the client to the AI large model application;

[0085] Step 55, through message interaction with the client, a session key is generated based on the confirmed establishment information and the public key of the AI large model application by using a key exchange algorithm, the session key is used for encrypted communication between the AI large model application and the client, the public key of the AI large model application is generated by the secure processor based on the hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

[0086] Correspondingly, the embodiment of the application also provides a secure channel establishment method, which is applied to a client outside the AI large model security all-in-one machine, such as Figure 6 As shown in the figure, the method comprises the steps of:

[0087] Step 61, the establishment information for establishing the secure channel is confirmed through message interaction between the AI large model application in the TEE and the client;

[0088] Step 62, the identity certificate is sent to the AI large model application, the identity authentication information for identity authentication is carried in the identity certificate, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs;

[0089] Step 63, the token sent by the AI large model application to the client after the identity authentication of the client is passed is received, and the token is carried in the data message sent by the client to the AI large model application;

[0090] Step 64, through message interaction with the AI large model application, a session key is generated based on the confirmed establishment information and the public key of the AI large model application by using a key exchange algorithm, the session key is used for encrypted communication between the AI large model application and the client, the public key of the AI large model application is generated by the secure processor based on the hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

[0091] With the above security channel establishment method provided by the embodiments of the present application, when a client located outside the AI large model security all-in-one machine needs to establish a security channel with an AI large model application in the TEE of the AI large model security all-in-one machine, the AI large model application will authenticate the client, and only after the authentication is passed, the security channel will be established, and the public key used by the AI large model application and the client to generate a session key is generated by the secure processor based on the hardware trusted root, and the public keys of the AI large model applications in different TEEs are different, so that the AI large model applications in different TEEs can be accessed only by clients with specific identities, thereby realizing the isolation of data between different identity clients in the AI large model service, and further improving the data security.

[0092] The above security channel establishment method provided by the embodiments of the present application is described in detail below. Based on the above AI large model security all-in-one machine, as shown in Figure 7 , it includes:

[0093] Step 71, the AI large model application and the client confirm the establishment information for establishing a security channel through message interaction.

[0094] In this step, the client can initiate the establishment of the security channel. First, the client generates a random number C, and generates a security channel establishment request containing the random number C, the client TLS version number, the cipher suite list, and the compression algorithm list, and sends it to the AI large model application.

[0095] Then, after receiving the security channel establishment request, the AI large model application generates a random number S, and based on the client TLS version number, the cipher suite list and the compression algorithm list carried in the security channel establishment request, generates a security channel establishment response containing the random number S, the confirmed TLS version number, the selected cipher suite and the compression algorithm, and returns it to the client.

[0096] In this step, the random number C, the random number S, the confirmed TLS version number, the selected cipher suite and the compression algorithm belong to the establishment information confirmed between the server and the client for establishing a security channel.

[0097] Step 72, the client sends an identity certificate to the AI large model application, and the identity certificate carries identity authentication information for identity authentication.

[0098] Step 73, after receiving the identity certificate sent by the client, the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs.

[0099] Step 74, after the identity authentication of the client is passed, the AI large model application sends a token to the client, which is used to carry in the data message sent by the client to the AI large model application subsequently.

[0100] Step 75, the AI large model application generates a private key and a public key, and sends the generated public key to the client.

[0101] For the sake of distinction, they can be called server private key and server public key.

[0102] In this step, the server private key and the server public key of the AI large model application are generated based on the hardware trusted root of the security processor of the AI large model security all-in-one machine, and the server public keys of the AI large model applications in different TEEs are different.

[0103] Specifically, the server public key of the AI large model application in the TEE can be obtained by confusing the ID of the hardware trusted root and the TEE, and the derived key is used as the server public key of the AI large model application in the TEE.

[0104] Step 76, the client generates a private key and a public key, and sends the generated public key to the AI large model application.

[0105] For the sake of distinction, they can be called client private key and client public key.

[0106] Step 77, the client generates a session key by using a key exchange algorithm.

[0107] In this step, the client can generate a session key by using a key exchange algorithm based on the random number C, the random number S, the client private key and the client public key, and the received server public key.

[0108] Step 78, after receiving the client public key sent by the client, the AI large model application generates a session key by using a key exchange algorithm.

[0109] In this step, the AI large model application can generate a session key by using a key exchange algorithm based on the random number C, the random number S, the server private key and the server public key, and the received client public key.

[0110] In the embodiments of the present application, various feasible key exchange algorithms can be used, which will not be described in detail here.

[0111] After the client and the AI large model application generate the session key respectively, the generated session key is used for encrypted communication between the AI large model application and the client, and the session key is only valid in this connection and is secret to the third party except the communication parties.

[0112] The AI large model security all-in-one machine and the corresponding secure channel establishment method provided by the embodiment of the application can realize the integrated scheme based on the cooperation of software and hardware, and realize the end-to-end security protection of the large model training, reasoning and deployment, in view of the need for data security of the localized deployment of the AI large model.

[0113] Based on the same inventive concept, according to the secure channel establishment method applied to the AI large model application provided by the above-mentioned embodiments of the application, correspondingly, another embodiment of the application also provides a secure channel establishment device, which is applied to the AI large model application in the TEE of the AI large model security all-in-one machine, and a structure diagram is as shown in Figure 8 The specific structure includes:

[0114] The first establishment information confirmation module 81 is configured to confirm the establishment information for establishing the secure channel through message interaction with the client.

[0115] The certificate receiving module 82 is configured to receive the identity certificate sent by the client, and the identity certificate carries the identity authentication information for identity authentication.

[0116] The identity authentication module 83 is configured to authenticate the identity of the client by comparing the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs.

[0117] The token sending module 84 is configured to send a token to the client after the identity authentication of the client is passed, and the token is carried in the data packet sent by the client to the AI large model application.

[0118] The first session key generation module 85 is configured to generate a session key based on the confirmed establishment information and the public key of the AI large model application by using a key exchange algorithm through message interaction with the client, and the session key is used for encrypted communication between the AI large model application and the client. The public key of the AI large model application is generated by the secure processor based on the hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

[0119] Based on the same inventive concept, according to the secure channel establishment method applied to the client provided by the above-mentioned embodiments of the application, correspondingly, another embodiment of the application also provides a secure channel establishment device, which is applied to the client located outside the AI large model security all-in-one machine, and a structure diagram is as shown in Figure 9 The specific structure includes:

[0120] The second establishment information confirmation module 91 is configured to confirm establishment information for establishing a secure channel through message interaction with the AI large model application in the TEE.

[0121] The certificate sending module 92 is configured to send an identity certificate to the AI large model application, the identity certificate carrying identity authentication information for identity authentication, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of the TEE to which the AI large model application belongs.

[0122] The token receiving module 93 is configured to receive a token sent by the AI large model application to the client after the identity authentication of the client is passed, the token being carried in a data packet sent by the client to the AI large model application.

[0123] The second session key generation module 94 is configured to generate a session key through message interaction with the AI large model application based on the confirmed establishment information and a public key of the AI large model application, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

[0124] The functions of the above modules can correspond to Figures 1 to 7 the corresponding processing steps in the flowchart, and will not be described here again.

[0125] The secure channel establishment apparatus provided by the embodiments of the present application can be implemented by a computer program. Those skilled in the art should understand that the above-mentioned module division manner is only one of many module division manners, and as long as the secure channel establishment apparatus has the above-mentioned functions, the secure channel establishment apparatus divided into other modules or without module division should be within the protection scope of the present application.

[0126] The embodiments of the present application further provide an electronic device, such as Figure 10 as shown, comprising a processor 101 and a machine readable storage medium 102, the machine readable storage medium 102 storing machine executable instructions capable of being executed by the processor 101, and the processor 101 is prompted by the machine executable instructions to implement any of the above-mentioned secure channel establishment methods applied to the AI large model application or to implement any of the above-mentioned secure channel establishment methods applied to the client.

[0127] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement any of the above-mentioned secure channel establishment methods applied to AI large model applications, or to implement any of the above-mentioned secure channel establishment methods applied to clients.

[0128] The embodiments of the present application further provide a computer program product containing instructions, which, when executed on a computer, cause the computer to perform any of the above-mentioned secure channel establishment methods applied to AI large model applications, or to perform any of the above-mentioned secure channel establishment methods applied to clients.

[0129] The machine readable storage medium in the above-mentioned electronic device can include a random access memory (RAM) and can also include a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0130] The processor described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0131] Each of the embodiments in the specification is described in a relevant manner, and the same and similar parts between each of the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, for the device, the electronic device, the computer readable storage medium, and the computer program product embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0132] It is to be understood that the terms "including", "comprising", or any other variation thereof, are intended to cover the contents "open", such that a process, a method, an article, or an apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or even include elements that are inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the stated element.

[0133] The present application is described with reference to the flowchart illustrations and / or block diagrams according to the embodiments of the application. It is understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0134] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device that implements the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0135] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0136] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. An AI large model security all-in-one machine, characterized in that, Comprise: A secure processor and a physical memory module; The secure all-in-one machine has a plurality of trusted execution environments (TEEs), and different TEEs have different keys for memory encryption and decryption. The memory keys of the TEEs are generated by a hardware trusted root of the secure processor based on a memory encryption module of the secure processor, and different TEEs are allocated with different memory spaces for independent use on the physical memory module; Each of the TEEs is installed with an AI large model application, and data generated by the AI large model application is encrypted and written in the memory space allocated for the TEE by the memory encryption module using the memory key corresponding to the TEE.

2. The AI large model security all-in-one machine of claim 1, wherein, Further comprise: A plurality of AI acceleration cards; The AI acceleration cards are connected to a device pass-through module of the secure processor; The AI large model application of the TEE and the AI acceleration card interact data through the device pass-through module.

3. The AI large model security all-in-one machine of claim 1, wherein, The secure processor has a remote authentication module; The remote authentication module realizes remote authentication of the TEE through message interaction with the client.

4. The AI large model security all-in-one machine of claim 1, wherein, A database is established in each of the TEEs for storing identity authentication information; In the process of establishing a secure channel with the client, the AI large model application authenticates the identity of the client based on the identity authentication information stored in the database of the TEE to which the AI large model application belongs.

5. The AI large model security all-in-one machine of claim 4, wherein, The AI large model application has a public key for establishing a secure channel with the client; The public key of the AI large model application is generated by the secure processor based on a hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

6. A secure channel establishment method, characterized by, The AI large model application in the TEE of the AI large model secure all-in-one machine of any one of claims 1-5 comprises: The AI large model application confirms the establishment information for establishing a secure channel through message interaction with the client; Receive the identity certificate sent by the client, the identity certificate carries the identity authentication information for identity authentication; By comparing the identity authentication information carried in the identity certificate with the identity authentication information stored in the database of the TEE to which the AI large model application belongs, the identity of the client is authenticated; After the identity authentication of the client is passed, a token is sent to the client, the token is carried in the data packet sent by the client to the AI large model application; Through message interaction with the client, a session key is generated based on the confirmed establishment information and the public key of the AI large model application using a key exchange algorithm, the session key is used for encrypted communication between the AI large model application and the client, and the public key of the AI large model application is generated by the secure processor based on a hardware trusted root, and the public keys of the AI large model applications in different TEEs are different.

7. A secure channel establishment method, characterized by, Applied to the client outside the AI large model secure all-in-one machine of any one of claims 1-5, comprising: Confirm the establishment information for establishing a secure channel through message interaction with the AI large model application in the TEE; An identity certificate is sent to the AI large model application, and the identity certificate carries identity authentication information for identity authentication, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of a TEE to which the AI large model application belongs; A token is received, which is sent by the AI large model application to the client after the identity authentication of the client is passed, and the token is carried in a data message sent by the client to the AI large model application; Through message interaction between the AI large model application and the client, a session key is generated based on the confirmed establishment information and a public key of the AI large model application by using a key exchange algorithm, the session key is used for encrypted communication between the AI large model application and the client, and the public key of the AI large model application is generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs are different.

8. A secure channel establishment device, characterized in that, The AI large model application in the TEE of the AI large model security all-in-one machine of any one of claims 1-5 comprises: A first establishment information confirmation module is configured to confirm establishment information for establishing a secure channel through message interaction with a client. A certificate receiving module is configured to receive an identity certificate sent by the client, and the identity certificate carries identity authentication information for identity authentication. An identity authentication module is configured to authenticate the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of a TEE to which the AI large model application belongs. A token sending module is configured to send a token to the client after the identity authentication of the client is passed, and the token is carried in a data message sent by the client to the AI large model application. A first session key generation module is configured to generate a session key based on the confirmed establishment information and a public key of the AI large model application by using a key exchange algorithm through message interaction with the client, the session key is used for encrypted communication between the AI large model application and the client, and the public key of the AI large model application is generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs are different.

9. A secure channel establishment device, characterized in that, The client located outside the AI large model security all-in-one machine of any one of claims 1-5 comprises: A second establishment information confirmation module is configured to confirm establishment information for establishing a secure channel through message interaction with the AI large model application in the TEE. A certificate sending module is configured to send an identity certificate to the AI large model application, and the identity certificate carries identity authentication information for identity authentication, so that the AI large model application authenticates the identity of the client by comparing the identity authentication information carried in the identity certificate with identity authentication information stored in a database of a TEE to which the AI large model application belongs. The token receiving module is configured to receive a token sent by the AI large model application to the client after the identity authentication of the client is passed, the token being used for carrying in a data message sent by the client to the AI large model application. The second session key generating module is configured to generate a session key by message interaction between the AI large model application and the client based on the confirmed establishment information and a public key of the AI large model application, the session key being used for encrypted communication between the AI large model application and the client, the public key of the AI large model application being generated by the secure processor based on a hardware trusted root, and the public keys of AI large model applications in different TEEs being different.

10. An electronic device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method in claim 6 or the method in claim 7.

11. A computer readable storage medium characterized by, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method in claim 6 or the method in claim 7.

Images