A paperless meeting data encryption transmission method and system
By generating a dynamic root key and role weight allocation sub-key bound to the session, and combining a two-layer tunnel and verification frame chain mechanism, the security and real-time issues of key updates and data transmission in paperless conferencing systems are solved, achieving secure and reliable data transmission.
Patent Information
- Application Number
- CN202611343964.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-01
- Publication Date
- 2026-11-17
AI Technical Summary
In paperless conferencing systems, the risk of data leakage arises from the long-term immutability of keys. Fixed keys cannot be protected in a hierarchical manner, and the key update mechanism is out of sync with the conferencing process. Existing encrypted transmission methods are unable to meet the real-time and fault-tolerant requirements of different data types.
Generate a dynamic root key bound to the session, allocate sub-keys based on role weights and activate them in rotation, separate control signaling from data logic through a two-layer tunnel, and update keys by combining a verification frame chain and a historical key caching mechanism.
It enables dynamic, hierarchical, and synchronous updates of keys, ensuring secure and reliable transmission of conference data in complex network environments while accommodating the transmission needs of different data types.
Smart Images

Figure CN123122816A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of paperless meeting systems, specifically relating to a method and system for encrypted transmission of paperless meeting data. Background Technology
[0002] With the popularization of information-based office work, paperless meeting systems have been widely used in government agencies, enterprises, institutions and various organizations. Compared with traditional paper meetings, paperless meeting systems use electronic devices to distribute, display and interact with meeting materials, reducing resource consumption and improving meeting efficiency. However, when conference data is transmitted between terminals via wireless or wired networks, it faces security risks such as eavesdropping, tampering, and forgery; if the key remains unchanged for a long time and is cracked during the conference, all historical and subsequent data are at risk of being leaked; fixed keys cannot adapt to the dynamic changes of different speaking roles during the conference, making it difficult to achieve hierarchical protection of data for different roles; existing technologies have proposed schemes for periodic key updates, such as triggering key updates based on time periods or data volume thresholds; however, such schemes still have shortcomings: The key update mechanism is out of sync with the conference process, making it impossible to flexibly adjust the encryption strategy according to the actual data transmission status and the status of the participants; The lack of a reliable synchronization mechanism during key updates can easily lead to data decryption issues when switching between versions before and after an update. Meeting data typically includes various data types such as document sharing, screen projection, and audio / video streams. These different types of data have significantly different requirements for real-time transmission and fault tolerance, which existing unified encrypted transmission methods cannot adequately address. In summary, how to achieve dynamic, hierarchical, and synchronous updates of keys during paperless meetings, and ensure the secure and reliable transmission of various meeting data and control signaling in complex network environments, is a technical problem that urgently needs to be solved in this field. Summary of the Invention
[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a paperless method and system for encrypted transmission of meeting data.
[0004] The objective of this invention can be achieved through the following technical solutions: This application provides a method for encrypted transmission of paperless meeting data, including the following steps: S1: Collect the set of participant identities and meeting timestamps for the current meeting to generate a unique session ID; generate a dynamic root key for the session ID through a key negotiation protocol based on the hardware feature code of each participant's terminal and a real-time random number; and perform hierarchical expansion of the dynamic root key to form a dynamic session key matrix containing encryption keys, obfuscation keys, and update keys. S2: Parse the data type of the conference data to be transmitted, determine the block granularity based on the current network bandwidth fluctuation value; divide the conference data to be transmitted into multiple data segments according to the block granularity, extract the encryption key from the dynamic session key matrix and perform symmetric encryption on each data segment; use the confusion key to rearrange the encryption results of each data segment and fill in the redundancy to generate a heterogeneous encryption block sequence. S3: Establish a first-layer tunnel and a second-layer tunnel with the receiving end, wherein the first-layer tunnel is used to transmit the heterogeneous encrypted block sequence, and the second-layer tunnel is a control signaling logical channel containing a key index and a block mapping table; extract the update key from the dynamic session key matrix to independently encrypt the control signaling, and interleave the encrypted control signaling into the heterogeneous encrypted block sequence in time sequence to form a composite encrypted stream; S4: When the sending end transmits the composite encrypted stream to the receiving end, it embeds a verification frame carrying the current encryption round within a preset interval; the receiving end performs integrity verification based on the verification frame and confirms identity based on the authentication code generated based on the updated key in the verification frame. If the verification passes, it returns a synchronization confirmation frame to the sending end; after receiving the synchronization confirmation frame, the sending end triggers an irreversible transformation of the updated key in the dynamic session key matrix and applies the transformed new key parameters to the data encryption of the next round until the meeting ends.
[0005] This application provides a paperless conference data encryption transmission system, including the following modules: The session key initialization module is used to collect the set of participant identity identifiers and the meeting timestamp of the current meeting, and generate a unique session ID for the meeting; based on the hardware feature code of each participant's terminal and a real-time random number, a dynamic root key is generated for the session ID through a key negotiation protocol; the dynamic root key is hierarchically expanded to form a dynamic session key matrix containing an encryption key, an obfuscation key and an update key; The data block encryption module is used to parse the data type of the conference data to be transmitted and determine the block granularity according to the current network bandwidth fluctuation value; divide the conference data to be transmitted into multiple data segments according to the block granularity, and extract the encryption key from the dynamic session key matrix to perform symmetric encryption on each data segment; use the obfuscation key to rearrange the encryption results of each data segment and fill in redundancy to generate a heterogeneous encryption block sequence. A composite encrypted stream generation module is used to establish a first-layer tunnel and a second-layer tunnel with the receiving end, wherein the first-layer tunnel is used to transmit the heterogeneous encrypted block sequence, and the second-layer tunnel is a control signaling logical channel containing a key index and a block mapping table; the update key is extracted from the dynamic session key matrix to independently encrypt the control signaling, and the encrypted control signaling is interleaved into the heterogeneous encrypted block sequence in a time sequence to form a composite encrypted stream; The synchronization update module is used to embed a verification frame carrying the current encryption round and an authentication code generated based on the update key within a preset interval when the sending end transmits the composite encrypted stream to the receiving end. The receiving end confirms the identity based on the authentication code in the verification frame and performs integrity verification based on the verification frame. If the verification passes, it returns a synchronization confirmation frame to the sending end. After receiving the synchronization confirmation frame, the sending end triggers an irreversible transformation of the update key in the dynamic session key matrix and applies the transformed new key parameters to the data encryption of the next round until the meeting ends.
[0006] Beneficial effects This invention solves the problems of data leakage due to long-term unchanged keys, the inability of fixed keys to provide hierarchical protection, and the disconnect between key updates and meeting processes by generating a dynamic root key bound to the session, allocating sub-keys based on role weights and activating them in rotation, and adaptively updating key strength according to the round sequence number. This invention solves the problem of difficulty in simultaneously meeting the different requirements of real-time transmission and fault tolerance for various types of conference data by parsing data types and dynamically adjusting the granularity of the blocks according to network bandwidth fluctuations, and by using different units of measurement and dynamically scaling for different types of data. This invention achieves separation of control signaling and data logic through a two-layer tunnel and time-interleaved composite transmission. Combined with a verification frame chain and historical key caching mechanism, it solves the problems of version switching failure and unusable retransmitted data due to the lack of a synchronization mechanism for key updates. Attached Figure Description
[0007] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.
[0008] Figure 1 A flowchart illustrating a paperless meeting data encryption transmission method provided in Embodiment 1 of this application; Figure 2 This is a schematic diagram of a paperless conference data encryption transmission system provided in Embodiment 1 of this application. Detailed Implementation
[0009] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.
[0010] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.
[0011] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.
[0012] Example 1 Please see Figure 1 This embodiment provides a paperless meeting data encryption transmission method, including the following steps: S1 Session Key Initialization The meeting initiator collects a set of identity identifiers for all participants in the current meeting. Each participant's identity identifier is a pre-assigned digital certificate serial number or a unified enterprise identity identifier, used to uniquely distinguish each participant. At the same time, the meeting initiator obtains the system time at which the current meeting started from the local clock module and generates a meeting timestamp.
[0013] The meeting initiator concatenates the participant identity set and the meeting timestamp into a binary string according to the lexicographical order of the participant identity. This string is then input into the SHA-256 hash function for hashing to obtain a fixed-length hash value, which is the unique session ID of the meeting. The session ID remains unchanged throughout the entire meeting lifecycle and is used to uniquely identify this paperless meeting session.
[0014] When the session starts, the initiator calls a local secure random number generator to generate a 256-bit one-time random key, which is generated by a cryptographically secure pseudo-random number generator. The initiator concatenates the session ID with the one-time random key, and then uses the public keys of each participating terminal to perform asymmetric encryption on the concatenated data. The encrypted ciphertext is then distributed to each participating terminal synchronously along with the session establishment signaling. Upon receiving the ciphertext, each participating terminal decrypts it using its own private key to obtain the session ID and the one-time random key.
[0015] After obtaining a one-time random key, each participating terminal reads its pre-stored hardware signature, which consists of the unique device serial number burned into the terminal at the factory, the network interface physical address, and the measurement value of the trusted platform module. Each participating terminal concatenates its local hardware signature, one-time random key, and session ID into an authentication data block, and then generates an authentication credential using an identity-based authentication mechanism. This credential is obtained by digitally signing the authentication data block using the participant's identity identifier and private key. Each participating terminal then returns this authentication credential to the conference initiator.
[0016] After receiving the authentication credentials returned by each participating terminal, the conference initiator performs two verifications simultaneously. The first verification is the validity of the one-time random key: the conference initiator extracts the one-time random key from the received authentication credentials, compares it with the original value stored locally, and marks the one-time random key as used to prevent replay attacks. The second verification is the consistency of the hardware signature: the conference initiator compares the received hardware signature with the device registration information pre-registered by the participating party. Only when both verifications pass can the participating terminal obtain access and enter the subsequent key negotiation protocol process; participating terminals that fail either verification will be refused access to the conference.
[0017] All verified participating terminals and the conference initiator jointly participate in the key negotiation protocol. This embodiment uses an authentication key negotiation algorithm based on elliptic curve cryptography, namely the SM2 key exchange protocol. During the key negotiation process, the conference initiator and each participating terminal generate a temporary elliptic curve public-private key pair. Each participating terminal concatenates its temporary public key with its own hardware signature, then encrypts it using the conference initiator's public key and sends it to the conference initiator. The conference initiator collects the temporary public keys of all participating terminals and verifies their hardware signature signatures. After successful verification, the conference initiator uses its own temporary public key and the temporary public keys of all participating terminals to form a public key pair. The public key set is broadcast to all participating terminals. Each participating terminal calculates its shared secret value with the conference initiator based on the received public key set, its own generated temporary private key, and a one-time random key. The conference initiator simultaneously calculates its shared secret value with each participating terminal. After all participants have completed the above calculations, all shared secret values are arranged in lexicographical order according to the participant's identity identifier and concatenated into a binary sequence. This sequence is then input into the key derivation function (KDF) to derive a 256-bit dynamic root key. This dynamic root key is bound to the current session ID, and any participating party destroys the dynamic root key after the conference ends.
[0018] The conference initiator uses a dynamic root key as a seed, inputting it into the key derivation function KDF. Three different fixed derivation strings are used as parameters for three derivations, resulting in three independent 2048-bit key components. The first derivation string is the ASCII encoding of ENCRYPT, yielding the encryption key; the second derivation string is the ASCII encoding of OBFUSCATE, yielding the obfuscation key; and the third derivation string is the ASCII encoding of UPDATE, yielding the update key. These encryption, obfuscation, and update keys together constitute the dynamic session key matrix. In subsequent steps involving symmetric encryption using the SM4 algorithm, the first 128 bits of the corresponding encryption key component are extracted as the SM4 encryption key. When using the MT19937 pseudo-random number generator, the lower 32 bits of the corresponding obfuscation key component are extracted as the random seed.
[0019] The meeting initiator generates a list of speaking roles based on the meeting agenda. These roles include the meeting host, keynote speaker, discussion speaker, and observer. The meeting initiator assigns a role weight factor to each speaking role, which is determined by the role's preset speaking time percentage, speaking order priority coefficient, and data sensitivity level coefficient.
[0020] Let the total bit length of the dynamic root key be... In this embodiment =2048; for the first The weighting factor for each speaking role is calculated by obtaining the percentage of the preset speaking time for that role. Speaking order priority coefficient The speaking order priority coefficient The value of ᵢ ranges from 1 to 10, with larger values indicating earlier speaking order; the data sensitivity level coefficient The value of is a positive integer from 1 to 5, with a larger value indicating higher data sensitivity. After multiplying the three factors, normalization is performed so that the sum of the weight factors for all roles equals 1. The formula for calculating the weight factors is:
[0021] The preset speaking duration percentage Speaking order priority coefficient and data sensitivity level coefficient All weighting factors are set once by the meeting initiator according to the agenda at the start of the meeting; The weighting factor is calculated only once during this initialization phase and stored permanently in the metadata of the dynamic session key matrix, remaining unchanged throughout the entire lifecycle of the meeting. Any dynamic actions during the meeting, such as changes in the actual speaking time of any participant or temporary adjustments to the speaking order, will not affect the already calculated weighting factor. The allocation results of key fragments for each role, and the main function of the weighting factor is to distinguish the differences in security levels between roles. Deviations in the actual speaking time allocation do not affect the validity and correctness of the key derivation results; whereby... The denominator represents the total number of speaking roles. , , Summing the products.
[0022] Then allocate bit lengths according to the weight factors of each role. , which is the total length of the key fragment for that role; if any If the length is less than three times the preset security length threshold of 384 bits, it will be padded to 384 bits, with the insufficient portion filled using the hash extension value of the dynamic root key; the allocated lengths for each role... All are no less than 384 bits; the padded bits will be... As the key derivation seed corresponding to this role, this seed is combined with the role identifier and a fixed derivation string. After concatenation, the key is input into the key derivation function KDF, which generates encryption subkeys, confusion subkeys, and update subkeys, each with a length of 128 bits. The length of each subkey is fixed at 128 bits, satisfying the key length requirement of the SM4 algorithm. During the derivation process... The larger the value, the higher the input entropy of the KDF, and the key strength is still positively correlated with the role weight.
[0023] The conference initiator truncates the dynamic root key according to the bit length allocated above, obtaining an encryption subkey, a confusion subkey, and an update subkey that correspond one-to-one with each speaking role. For example, for the host role, a host encryption subkey, a host confusion subkey, and a host update subkey are allocated, and the total length of the three subkeys is the bit length allocated to that role. The subkeys of all roles are stored together in the dynamic session key matrix.
[0024] The dynamic session key matrix operates in a time-slice rotation manner. At any given moment, only one key combination matching the current speaking role is activated. This key combination includes the encryption subkey, obfuscation subkey, and update subkey corresponding to the role, which are used to encrypt the meeting data generated within the current time slice. The length of each time slice is dynamically determined based on the weight factor of the current speaking role. The larger the weight factor, the longer the time slice. The specific time slice length is a fixed base duration multiplied by the weight factor of the role. The fixed base duration is preset by the meeting initiator based on the total meeting duration and the total number of roles. For example, the fixed base duration can be set to be equal to twice the total meeting duration divided by the total number of roles. The specific value is configured by the meeting initiator when the meeting starts.
[0025] The time slice rotation is executed cyclically within the same encryption round. The encryption rounds are advanced round by round according to the synchronization confirmation frames in the subsequent S4. The irreversible key transformation only occurs when the encryption round is switched. The time slice switching within the same round does not trigger any key transformation.
[0026] When the current role is switched, the conference initiator generates a control signaling message containing the target role identifier and the corresponding key index and inserts it into the composite encrypted stream and sends it to the receiving end. After receiving the message, the receiving end returns a role switch confirmation frame. The conference initiator activates the key combination for the next time slice only after receiving the confirmation frame. If no confirmation frame is received within the preset timeout period, the control signaling message is retransmitted.
[0027] S2 Data Block Encryption The conference initiator parses the conference data to be transmitted and identifies its data type. Data types are divided into continuous data and discrete data. Continuous data refers to streaming data that is continuously generated at a fixed sampling rate or frame rate, such as screen-projected video streams and audio streams. This type of data has time-series dependence and real-time requirements, and the data volume continues to grow over time, making it impossible to predict the total data volume. Discrete data refers to non-streaming data that is transmitted completely at once, such as document files, images, and annotation information. This type of data has definite start and end boundaries and is required to arrive at the receiving end completely and without loss.
[0028] The conference initiator monitors the current network bandwidth in real time by periodically sending probe packets to the receiving end. These probe packets are User Datagram Protocol (UDP) data packets with a sending timestamp. Upon receiving the packet, the receiving end immediately returns it. The conference initiator calculates the current network bandwidth measurement based on the round-trip time and the number of bytes successfully received per unit time. The difference between the current measurement and the historical bandwidth average is divided by the historical bandwidth average to obtain the bandwidth fluctuation value. A positive bandwidth fluctuation value indicates an increase in network bandwidth, while a negative value indicates a decrease in network bandwidth. The historical bandwidth average is the moving average of the most recent ten measurements.
[0029] The conference initiator converts bandwidth fluctuation values into dynamic scaling factors. When the bandwidth fluctuation value is greater than 0, the scaling factor is equal to the sum of 1 and the bandwidth fluctuation value, that is, when the bandwidth increases by 20%, the scaling factor is 1.2. When the bandwidth fluctuation value is less than 0, the scaling factor is equal to the difference between 1 and the absolute value of the bandwidth fluctuation value. The lower limit of the scaling factor is 0.5, that is, when the bandwidth decreases by more than 50%, the scaling factor is still calculated as 0.5 to prevent excessive protocol overhead caused by excessively small blocks. When the bandwidth fluctuation value is 0, the scaling factor is equal to 1.
[0030] The conference initiator presets a basic block size for the data to be transmitted. For continuous data, the basic block size is a fixed time window length of 100 milliseconds. This time window length is determined based on the frame interval of the audio and video codecs and matches the common video frame rate of 25 frames per second, so that each block contains approximately two to three video frames. For discrete data, the basic block size is 512 kilobytes. This value is determined based on an integer multiple of the maximum transmission unit (MTU) of the network transmission path to avoid data packets being fragmented and reassembled at the network layer.
[0031] The conference initiator multiplies the base chunk size by the aforementioned dynamic scaling factor to obtain the actual chunk size. When bandwidth decreases, the scaling factor is less than 1, and the actual chunk size is reduced accordingly to decrease the load of a single transmission and reduce retransmission overhead caused by packet loss. When bandwidth increases, the scaling factor is greater than 1, and the actual chunk size is increased accordingly to reduce data packet header overhead and make full use of bandwidth resources. Continuous data types are chunked based on time length, while discrete data types are chunked based on byte length. Both are adjusted by the same scaling factor at the same ratio to ensure that the scaling ratio remains consistent between the two data types.
[0032] The conference initiator divides the conference data to be transmitted into multiple data segments according to the actual block size. For continuous data, a time window is defined as the time length corresponding to the actual block size, starting from the current moment. All data generated within each time window constitutes a data segment. If the last window is less than the actual block size, it is processed according to the actual data volume. This segmentation strategy ensures that continuous data is segmented in a time-aligned manner, which is convenient for the receiving end to reassemble and play according to the timestamp. For discrete data, it is segmented sequentially according to the number of bytes of the actual block size, starting from the beginning of the data. The number of bytes in each data segment is equal to the actual block size. If the last data segment is less than the actual block size, it is processed according to the remaining number of bytes.
[0033] After the data is partitioned, the conference initiator extracts the currently active encryption subkey from the dynamic session key matrix. The first 128 bits of this key component are used as the SM4 encryption key to perform symmetric encryption on each data segment. The encryption algorithm uses the SM4 block cipher algorithm, and the working mode is CBC mode, i.e., cipher block chaining mode. The initial vector of each data segment is obtained by concatenating the segment's sequence number in the original data and the current encryption round number, and then inputting it into the SHA-256 hash function to obtain the hash value. The first 128 bits of this hash value are used as the initial vector. This generation method ensures that each segment uses a different initial vector in different rounds, and the same segment generates different initial vectors in different encryption rounds due to different round numbers, which meets the basic security requirement of cryptography that the initial vector cannot be reused.
[0034] After encryption, the conference initiator extracts the currently active obfuscation subkey from the dynamic session key matrix and inputs the lower 32 bits of the obfuscation subkey as a seed into the pseudo-random number generator. The pseudo-random number generator uses the MT19937 version of the Mersenne rotation algorithm. The initialization of the MT19937 pseudo-random number generator follows a standard initialization procedure, where the 32-bit obfuscation seed is expanded into 624 32-bit state variables using the standard initialization function of the Mersenne rotation algorithm. The specific implementation is consistent with the MT19937 standard reference implementation. The sending and receiving ends use the same initialization function to ensure that the random permutation sequences generated by both ends are completely identical. If a non-standard simplified initialization is used, an initialization mode identifier must be additionally carried in the control signaling, and the receiving end selects the corresponding initialization method based on this identifier. The period of the MT19937 algorithm is... The output is a 32-bit unsigned integer. The pseudo-random number generator outputs a random permutation sequence with the same number of encrypted data fragments. This random permutation sequence is an arrangement of integers from zero to the total number of fragments minus one, with each integer appearing exactly once. This sequence defines a global permutation order for all encrypted data fragments. The conference initiator reads the encrypted data fragments sequentially according to this random permutation sequence and rearranges them. That is, the fragment at position i in the original sequence is moved to the target position indicated by i in the permutation sequence. Fragments that were originally adjacent are no longer adjacent in the shuffled sequence, and there is at least one other fragment between any two originally adjacent fragments. Through global position shuffling, even if an attacker intercepts some continuous data, they cannot infer the structural information of the original data based on the packet sequence number.
[0035] The conferencing initiator dynamically determines the redundancy padding length based on the current network packet loss rate. The current network packet loss rate is obtained from the transmission statistics returned by the receiver after each round of transmission. Specifically, it is the ratio obtained by dividing the difference between the expected total number of data packets received in this round and the actual total number of data packets received by the receiver, by the expected total number of data packets received. The padding length is calculated using the following formula:
[0036] in The final padding length, in bytes, indicates the number of padding bytes required for each data segment. The preset basic padding length is set to 32 bytes in this embodiment. This value is determined based on the difference between the maximum transmission unit of common networks and the average length of data segments, which can provide sufficient obfuscation space without introducing too much redundancy. As a preset scaling factor, in this embodiment the scaling factor is set to 256 bytes. This value makes the padding length linearly increase from 32 bytes to 160 bytes when the packet loss rate increases from zero to 50%, with an increment range of 128 bytes. This represents the current network packet loss rate, a decimal between 0 and 1, for example, when the packet loss rate is 50%. Take 0.5; As can be seen from the formula, the higher the network packet loss rate, the larger the padding length. When the packet loss rate is zero, only the basic length is filled. When the packet loss rate is 50%, the padding length reaches the maximum value of 160 bytes. This maximum value is set based on 30% of the actual block size to ensure that the redundancy overhead does not exceed 30% of the transmission load and to ensure transmission efficiency.
[0037] The padding content is generated by a random byte sequence driven by a random number generator using a confusion subkey as a seed. Each byte in this sequence is a pseudo-random number. Two key pieces of information are embedded in each padding byte block: the first is the cyclic redundancy check value of the data segment, which is calculated using the CRC32 algorithm to obtain the original data before encryption and is used by the receiver to verify data integrity; the second is the sequence number of the segment in the original order, which starts from 0 and increments to the total number of segments minus 1 and is used by the receiver to restore the original order.
[0038] The padding positions are distributed among the shuffled segments in reverse order of the random permutation sequence. The reverse order of the random permutation sequence refers to the new sequence obtained by swapping the positions and values of the elements in the permutation sequence, i.e., the value of the i-th position in the reverse order is equal to the position index of the i-th position in the permutation sequence. For each shuffled segment position, the conferencing initiator inserts a padding block between the segment and the next segment. The insertion offset is obtained by multiplying the position index of the segment in the reverse sequence by a preset base offset step of eight bytes. The offset is calculated from the beginning of the current gap, and the offset of each segment increases with the reverse index, so that the padding blocks are evenly distributed throughout the transmission sequence rather than being piled up in one place. Piled up blocks are easy for attackers to identify and peel off, while evenly distributed blocks among the segments enhance the obfuscation effect of the data structure, making it impossible for attackers to distinguish the padding bytes from the real encrypted data.
[0039] After receiving the heterogeneous encrypted block sequence that has been rearranged and padded with redundancy, the receiving end performs the reverse operation to recover the original data. The receiving end first uses the confusion subkey as a seed to generate the same random permutation sequence as the sending end, and then locates the position of each padding block according to the reverse order of the random permutation sequence. Specifically, the receiving end traverses the entire sequence, checking for the presence of a padding block at each gap position. The header of each padding block contains a preset padding identifier, which is 4 fixed bytes long. The fixed value of this 4-byte padding identifier is 0xA5A5A5A5, which has a bit alternation pattern and has an extremely low probability of appearing in the binary data stream. This effectively avoids accidental collisions with encrypted data. The identifier is generated by the conference initiator at the start of the session and distributed synchronously to each participant's terminal as a session parameter along with the session ID. After identifying the padding block, the receiving end extracts the original sequence number from the padding content and reads the cyclic redundancy check value. After removing all padding bytes, the receiving end reorders the data segments according to the extracted original sequence number to obtain encrypted data arranged in the original order. Finally, the receiving end uses the decryption key to decrypt the data segments one by one and verifies the integrity of each segment through the cyclic redundancy check value in the padding. If the verification fails, it requests the sending end to retransmit the corresponding segment.
[0040] S3 Composite Encrypted Stream Generation Two logical tunnels, referred to as the first-layer tunnel and the second-layer tunnel, are established between the conference initiator and receiver. The two tunnels share the same transmission link at the physical layer, meaning they both use the same network interface and physical medium for data transmission and reception. However, they are independently addressed at the logical layer, with each tunnel having a unique port number or protocol identifier. The operating system protocol stack distinguishes data ownership through the port number. The first-layer tunnel is configured as the main data channel for transmitting the heterogeneous encrypted block sequence generated in step S2. The second-layer tunnel is configured as the control signaling logical channel for transmitting control signaling containing the key index and block mapping table. During normal conference proceedings, the second-layer tunnel is dedicated to carrying control signaling containing the key index and block mapping table. It is also used for key synchronization during new participant access or terminal reconnection. In the distribution scenario, the second-layer tunnel can temporarily serve as a secure key distribution channel to carry key materials that have been asymmetrically encrypted with the receiver's public key. At this time, the data types transmitted within the tunnel are not limited by the aforementioned control signaling range, and the key materials are encapsulated in an independent encryption format (distinguished from the format of conventional control signaling) within the tunnel, ensuring that the core key materials are logically strictly isolated from ordinary control signaling. The control signaling consists of two components: a key index and a block mapping table. The key index records the identification information of each key component in the current dynamic session key matrix and the version number corresponding to the current encryption round. The block mapping table records the correspondence between the original order of each data segment in the heterogeneous encryption block sequence and the current transmission order, as well as the padding length and padding position of each segment.
[0041] The first and second layer tunnels each employ independent transport protocol stacks. The first layer tunnel's transport protocol stack is configured with a first-level check strategy prioritizing throughput. This strategy uses a lightweight check method, performing checksum verification only on the packet header, with no checksum verification or only simple cyclic redundancy check on the data payload. This allows for a small number of packets to be lost during transmission without triggering retransmission, ensuring a high transmission rate. The first layer tunnel's transport protocol stack uses User Datagram Protocol (UDP) as the transport layer protocol. This protocol is connectionless and has no retransmission mechanism, making it suitable for the rapid transmission of large amounts of data. The second layer tunnel's transport protocol stack is configured with a second-level check strategy prioritizing integrity. This strategy uses forward error correction coding and Automatic Repeat Request (ARQ) mechanism, performing a complete checksum verification on each packet. Once packet loss or errors are detected, retransmission is immediately triggered to ensure the complete and reliable arrival of control signaling. The second layer tunnel's transport protocol stack uses Transmission Control Protocol (TCP) as the transport layer protocol. This protocol is connection-oriented and has an acknowledgment and retransmission mechanism, making it suitable for signaling transmission with small data volumes and high reliability requirements.
[0042] When encrypted control signaling in the second-layer tunnel needs to be interspersed within the heterogeneous encrypted block sequence of the first-layer tunnel, the control signaling temporarily borrows the transmission channel of the first-layer tunnel, that is, it sends data through the port of the first-layer tunnel at the physical layer. The verification method of this control signaling still follows the second verification strategy of the second-layer tunnel, that is, appending a cyclic redundancy check value as verification information to the end of the control signaling data packet. The receiving end independently verifies this verification information upon receipt, instead of using the lightweight verification of the first-layer tunnel. After the control signaling completes its transmission, the independent logical channel state of the second-layer tunnel is restored, and subsequent control signaling is still sent through the port of the second-layer tunnel itself.
[0043] The conferencing initiator interleaves the encrypted control signaling into the heterogeneous encrypted block sequence in a time-series manner. The specific interleaving rule uses a fixed number of heterogeneous encrypted blocks as an interleaving cycle. This fixed number is preset to ten data blocks. This value is determined based on the ratio between the control signaling update frequency and the data block arrival rate. This ensures that the control signaling is inserted in a timely manner to avoid the receiving end being unable to parse the data due to a lack of control information, while also preventing the data channel from being crowded due to excessive interleaving. Within each interleaving cycle, a fixed time slot is reserved for the encrypted control signaling. The length of this fixed time slot is equal to the average transmission time of a data block. The specific value is dynamically calculated based on the current network bandwidth and the average size of the data block, i.e., the data block transmission time is obtained by dividing the average size of the data block by the current network bandwidth.
[0044] Each interleaving cycle reserves a fixed time slot for encrypted control signaling. This time slot is located between the third and fourth data blocks within the cycle, which is slightly earlier in the middle of the cycle. This ensures that the receiver receives control information early in the cycle to prepare for processing subsequent data blocks, while also preventing data blocks from being unparseable due to the control information not yet arriving at the very beginning of the cycle. When this fixed time slot is exhausted, regardless of whether the current control signaling has been completely sent, the interleaving of control signaling in the current cycle is forcibly terminated, and the remaining control signaling that has not yet been sent is postponed to the next interleaving cycle. This mechanism prevents control signaling from occupying the data channel indefinitely and affecting data throughput.
[0045] At the start of each cycle, the conference initiator checks the queue of control signaling to be sent and determines the interleaving priority based on the urgency level of the control signaling. The urgency level is jointly determined by the update frequency of the key index and the change range of the block mapping table. The update frequency of the key index is the number of times the key index changes per unit time, measured in changes per second. This value reflects the activity level of key updates; a higher update frequency indicates more frequent key changes, and the receiving end needs to obtain the new key index in a timely manner to correctly decrypt the data. The change range of the block mapping table is defined as the Hamming distance between the current mapping table and the mapping table of the previous cycle, i.e., the number of different corresponding bits in the two mapping tables. This value reflects the degree of change in the data block structure; a larger change range indicates more drastic changes in the data reassembly method, and the receiving end needs to obtain the new mapping table in a timely manner to correctly reassemble the data.
[0046] The formula for calculating the emergency level is:
[0047] This represents the update frequency of the key index. Its statistical time window is a 1-second sliding interval prior to the current moment, and the total number of key index changes within this time window is counted in units of changes per second. The change range of the block mapping table is calculated by comparing the binary sequence of the original sequential array of segments in the current period's block mapping table and the previous period's block mapping table. If the current period is the first period, meaning there is no previous period mapping table, then... The value is 0, To determine the urgency level, for non-first cycles, the comparison is performed bitwise XORing based on the shorter of the two binary sequences. Any excess is not included in the difference, but this length difference is added as an additional difference term. , The final value is equal to the number of bits that result in 1 in the effective XOR bits plus the absolute value of the difference between the lengths of the two sequences. Both of the above parameters are dimensionless integers and are calculated once in real time at the beginning of each interleaving period. The character switching control command belongs to the highest emergency level and is not subject to the calculation limitations of the above formula. Processing based on the maximum value ensures priority insertion into the reserved time slot of the current cycle; in the formula This is a dimensionless numerical value representing the level of urgency; the higher the value, the greater the urgency. The weighting coefficient for the key index update frequency is set to 0.6 in this embodiment. This value is determined based on the degree of influence of the key index update frequency on the correctness of data decryption. A key index error will cause all subsequent data to be unable to be decrypted, so it is given a high weight. The key index update frequency is the number of times the key index changes per unit of time, expressed in terms of changes per second. The weighting coefficient for the magnitude of changes in the block mapping table is set to 0.4 in this embodiment. This value is determined based on the degree of impact of changes in the block mapping table on data reorganization. Mapping table errors only affect data reorganization within the current period and do not affect subsequent periods, so they are given a low weight. The change range of the block mapping table is defined as the Hamming distance between the current mapping table and the mapping table of the previous cycle. Control signaling with higher urgency level takes priority in occupying the reserved time slot of the current cycle. That is, when multiple control signaling is queued and waiting to be sent, the control signaling with the highest urgency level is inserted into the reserved time slot first, and the control signaling with the second highest urgency level is sent in the next available time slot.
[0048] The key index contains the version number corresponding to the current encryption round. The version number starts from 0 and increments by 1 with each irreversible transformation. The key index data structure contains four fields: the first field is the version number of the current encryption round, used to identify which round of key parameters is being used; the second field is the position identifier of the encryption key in the dynamic session key matrix; the third field is the position identifier of the obfuscation key in the dynamic session key matrix; and the fourth field is the position identifier of the update key in the dynamic session key matrix. After the receiving end extracts the control signaling from the composite encrypted stream, it reads the version number field in the key index, extracts the matching historical key parameter from its local buffer based on the version number, and uses this key parameter to decrypt the control signaling, ensuring that the receiving end always uses the same key version as the sending end for decryption. If the receiving end's buffer does not contain the key parameter corresponding to the version number, it requests the sending end to retransmit the key parameter of that version before decryption.
[0049] S4 Synchronous Update and Retransmission Mechanism When the conferencing initiator transmits the composite encrypted stream to the receiving end, it embeds a verification frame within a preset interval. The length of the preset interval is jointly determined by the data throughput of the current composite encrypted stream and the historical packet loss rate. The data throughput rate is the amount of data successfully transmitted per unit time, measured in megabits per second, and is obtained by dividing the number of data bytes actually sent in the most recent time window by the length of the time window. The historical packet loss rate is the ratio of the average number of data packets lost over the past ten preset intervals to the total number of packets sent, and this value is calculated from the cumulative packet loss count returned by the receiving end in each acknowledgment feedback.
[0050] The formula for calculating the preset interval is:
[0051] in The final preset interval length, in milliseconds; The preset basic interval length is set to 1000 milliseconds, or 1 second, in this embodiment. This value is determined based on the balance between the real-time requirements of conference data and network overhead. The 1-second interval can ensure that the verification frame arrives in time without occupying too much transmission bandwidth. When the verification frame is sent once per second, the length of each verification frame is about 100 bytes, and the verification frame traffic accounts for less than 0.1% of the total bandwidth. The throughput impact factor is set to 50 milliseconds per megabit per second in this embodiment. This value ensures that the preset interval is shortened by 50 milliseconds for every megabit per second increase in throughput. This factor is set based on the actual situation that the data throughput can reach 50 megabits per second in a typical conference scenario. When the throughput is in the range of 1 to 16 megabits per second, the preset interval is linearly reduced from 950 milliseconds to the lower limit of 200 milliseconds. When the throughput exceeds 16 megabits per second, the preset interval remains unchanged at 200 milliseconds to prevent excessively dense check frames from crowding out the effective data transmission bandwidth. The lower limit of the preset interval is 200 milliseconds. The current data throughput of the complex encrypted stream, in megabits per second; The packet loss rate impact coefficient is set to 20 milliseconds per 1% in this embodiment. This value means that the preset interval is shortened by 20 milliseconds for every 1% increase in packet loss rate. This coefficient is determined based on the degree of impact of packet loss rate on data retransmission. Every 1% increase in packet loss rate leads to an increase of about 2% in the amount of data that needs to be retransmitted. By shortening the preset interval, the density of the verification frame is increased to detect packet loss in advance and trigger retransmission. This represents the historical packet loss rate, expressed as a percentage; the preset interval has a lower limit of 200 milliseconds, which is the calculated value. If the time is less than 200 milliseconds, it will be executed at 200 milliseconds. This lower limit is determined based on the balance between the processing overhead of the verification frame and the real-time requirements. If the time is less than 200 milliseconds, the frequency of sending the verification frame exceeds 5 times per second. The protocol overhead of the verification frame itself will occupy more than 5% of the transmission bandwidth, affecting the effective transmission rate of conference data.
[0052] The verification frame structure contains three fields: The first field is the round number of the current encryption round, which increments by 1 after each irreversible transformation starting from zero; the second field is the authentication code generated based on the current update key, which is generated using the HMAC-SM3 algorithm. The current update key is used as the key for HMAC operation, and the binary sequence of the current round number and the complete content of the previous round verification frame is used as the message input. The HMAC-SM3 operation outputs a 32-byte hash value as the authentication code. The receiving end recalculates the HMAC-SM3 value using the locally stored current update key and the same message input, and compares the calculation result with the received authentication code byte by byte. If they match completely, the sender's identity is confirmed as legitimate, preventing third parties from forging verification frames; the third field is the hash value of the previous round verification frame, which is the SHA-256 hash value of the entire content of the previous verification frame. Through the third field, a chain structure is formed between the verification frames, that is, each verification frame points to the hash value of the previous verification frame, which is called the verification frame chain.
[0053] After parsing the verification frame from the composite encrypted stream, the receiving end performs integrity verification and identity confirmation. The specific process of identity confirmation is as follows: the receiving end recalculates the authentication code using the locally stored current update key and compares it with the received authentication code; integrity verification is achieved through the continuity verification of the verification frame chain: after receiving each verification frame, the receiving end extracts the hash value field of the previous round verification frame in the verification frame and compares this value with the SHA-256 hash value of the complete content of the previous round verification frame in the local cache; if the two match, the continuity verification of the verification frame chain passes, and the data in this round is considered continuous and complete; if the two do not match, the verification frame chain is broken at this point, and it is determined that there is data loss between this round and the previous round; the continuity verification of the round-level verification frame chain takes precedence over the fragment-level CRC32 verification; after the round-level verification passes, the fragment-level CRC32 verification is performed on each data fragment in the receiving end's buffer to locate the specific damaged fragment and trigger the corresponding retransmission; if both identity confirmation and integrity verification pass, the receiving end returns a synchronization confirmation frame to the sending end.
[0054] When multiple receivers exist, the sender maintains a receiver acknowledgment status table. This table uses the receiver identifier as the key and records the current round number that each receiver has acknowledged. The sender only triggers the irreversible transformation of the dynamic session key matrix and enters the next round after all active receivers have returned the synchronization acknowledgment frame for the current round. If any receiver fails to return a synchronization acknowledgment frame within a preset timeout period, the sender retransmits all data for the current round to that receiver. If the number of timeout receivers exceeds a preset fault tolerance threshold, such as one-third of the total number of participants, the key update process is paused and a network status alarm is triggered. Synchronization will resume from the current round after the network recovers. The synchronization confirmation frame consists of three fields: the first field is the current confirmation round number, which identifies the encryption round corresponding to this confirmation; the second field is the current timestamp of the receiver, which is used in conjunction with the round number to form the transformation seed for subsequent key transformations; the third field is the reception status bitmap, which is a binary sequence with a sequence length equal to the total number of data segments in the current round. Each bit corresponds to the reception status of a data segment, with the first bit indicating that it has been received correctly and zero bits indicating that it has not been received or that it was received incorrectly.
[0055] After receiving the synchronization confirmation frame, the sending end triggers an irreversible transformation of the update key in the dynamic session key matrix. The irreversible transformation uses the transformation seed as input for hash iteration. The transformation seed is constructed by converting the round number into a 32-bit unsigned integer and the timestamp in the synchronization confirmation frame into a 64-bit unsigned integer. The two are concatenated in order to form a 96-bit binary sequence. This sequence is then input into the SHA-256 hash function to obtain a 32-byte hash value as the final transformation seed. This transformation seed generation method ensures that the transformation seed is completely different when either the round number or the timestamp changes, and the one-way nature of the hash function guarantees that the original update key cannot be deduced from the transformation result.
[0056] The specific implementation of the hash iteration operation is as follows: Using the 32 bytes of data from the current updated key as the initial value, the initial value is concatenated with the transformation seed and input into the SHA-256 hash function to obtain the first iteration result; then, the first iteration result is concatenated with the transformation seed again and input into the SHA-256 hash function to obtain the second iteration result; this process is repeated until the specified number of iterations is reached. The input concatenation order for each iteration is fixed as follows: the 32-byte hash value output from the previous round is used as the left-hand data, and the transformation seed is used as the right-hand data. These are concatenated in sequence and input into the SHA-256 hash function. The iteration formula is: This is equal to the 32 bytes of data in the current update key. Equal to SHA-256 pairs The hash operation is performed on the concatenated value with the transformation seed, where the binary concatenation order is represented. The transformation seed is 32 bytes long, and after concatenation with the output of the previous round, a 64-byte input is formed. Both the sending and receiving ends strictly follow this concatenation order to perform key transformation to ensure that the transformation results at both ends are consistent.
[0057] The mapping relationship between the number of iterations and the current round number is as follows: the number of iterations equals the sum of the preset base number of iterations and the current round number. The base number of iterations is set to 256, a value determined based on the balance between security strength and computational overhead. The computational delay generated by 256 hash iterations is approximately 0.1 milliseconds, which does not affect real-time transmission performance. The larger the round number, the more iterations are performed, allowing the key to continuously strengthen its security as the conference progresses. The maximum number of iterations is set to 1000. When the current round number exceeds 744, the number of iterations is fixed at 1000. This maximum value is determined based on the ratio between the duration of a single encryption round and the hash operation time. The computational delay of 1000 hash iterations is approximately 0.4 milliseconds, which ensures that the iteration calculation is completed within the time budget allocated in each round, avoiding impact on real-time transmission performance. In each iteration, the output result of the previous iteration is concatenated with the transformation seed and input into the hash function. This process is repeated until the specified number of iterations is reached, and a new 32-byte value is output as the new updated key after transformation.
[0058] The encryption key and obfuscation key are synchronously transformed using the same transformation seed and the same number of iterations. That is, the three key components are independent but their transformation parameters are completely identical. The new key parameters generated after the transformation replace all key components of the previous round. The hierarchical relationship between the replaced key components is maintained. The hierarchical relationship refers to the relative hierarchical structure formed by the encryption key, obfuscation key and update key being derived from the same source of the dynamic root key. Specifically, the encryption key is always the first set of outputs derived from the dynamic root key, the obfuscation key is always the second set of outputs, and the update key is always the third set of outputs. After the overall replacement, the three key components are synchronously transformed using the same transformation seed and the same number of iterations. The derivation order and relative hierarchical relationship between the three remain unchanged, thus maintaining the original hierarchical relationship.
[0059] The sending end stores the historical key parameters from the previous round of replacement into a buffer. The buffer uses a circular queue data structure, with each buffer unit storing one historical version. Each version contains five fields: round number, the encryption key for that round (32 bytes), the obfuscation key for that round (32 bytes), the update key for that round (32 bytes), and the receive acknowledgment flag for that round. The buffer capacity is preset to five units, a value determined based on a balance between storage overhead and retransmission coverage. Each key version contains three 32-byte key components, a 4-byte round number, and a 1-byte flag, totaling 101 bytes of storage space. Five versions require only 505 bytes. The buffer has a storage space of byte, and the retransmission range covering five rounds can cope with continuous packet loss scenarios caused by sudden network congestion. The buffer management rules are as follows: the historical version that has been confirmed to have been received is immediately removed from the buffer and destroyed, releasing the storage unit; the buffer always prioritizes ensuring that the latest round key parameters can be stored. When the buffer is full and a new round key needs to be stored, the oldest unconfirmed version with the flag bit set to 0 is forcibly overwritten, and a retransmission request is triggered on the receiving end before the overwrite; if all five buffer units are filled with unconfirmed versions with the flag bit set to 0, they are all marked as overwriteable, and the storage of the current round key is prioritized to avoid system stagnation.
[0060] During transmission, the receiving end continuously verifies the continuity of the check frame chain. When the receiving end finds that the hash value of a check frame is inconsistent with the hash value of the previous round check frame in the local cache, that is, the check frame chain is broken at this point, it is determined that the round data corresponding to the check frame has been lost or corrupted. The receiving end identifies the break position, determines the round number of the missing round, and returns a retransmission request to the sending end. The retransmission request carries the missing round information, which includes the missing start round number and the missing end round number.
[0061] Upon receiving a retransmission request, the sending end first determines the data type of the data in that round. For continuous data, the sending end compares the deviation between the timestamp of the data in that round and the current playback progress to see if it exceeds a preset timeout threshold. If the deviation exceeds the threshold, for example, 200 milliseconds, the retransmission is skipped, and a skipped retransmission status code is returned to the receiving end. The receiving end then uses forward error correction or frame loss concealment strategies to continue playing subsequent data. If the deviation does not exceed the threshold, normal retransmission is performed. For discrete data, the above timeout judgment is not performed, and the retransmission process is directly initiated. If retransmission is required, the sending end pauses the key update process for the current round, i.e., stops triggering irreversible transformations to prevent the retransmitted data from becoming undecryptable due to continued key transformations.
[0062] The sending end determines whether the missing round is still within the buffer's retention range. Specifically, it checks if the missing round's sequence number is greater than or equal to the current round's sequence number minus the number retained in the buffer, and less than the current round's sequence number. If it is within the range, it extracts the historical key parameters corresponding to that round from the buffer, including the encryption key and update key for that round. Using these historical key parameters, it retransmits the composite encrypted stream from the beginning of that round. During retransmission, when constructing the initial vector for each data segment, the sending end concatenates the current retransmission count for that segment with the original segment sequence number and round sequence number, incrementing from 1. Generating factor, i.e. To ensure that the same data segment is transmitted in different retransmissions The retransmission data packets are all different; the retransmission packet header carries the round number of the current round, and the receiving end decrypts them using the corresponding historical key parameters from the local buffer; after the retransmission data is confirmed by the receiving end, the confirmation is marked by the receiving status bits of all data segments of the corresponding round in the received synchronization confirmation frame being in the first position. The sending end resumes the key update process and continues the normal encrypted transmission of subsequent rounds. During the retransmission, the sending end maintains the exact same block granularity and heterogeneous encrypted block sequence as the original transmission. That is, the retransmission data does not re-execute the block division and reordering operations of S2, but directly reads a copy of the heterogeneous encrypted block sequence from the buffer and retransmits it; the block granularity and heterogeneous encrypted block sequence structure must not be changed during the retransmission operation, and the sending end must not re-execute the data type parsing, block granularity determination, and data transmission in S2 during the retransmission. The process involves segmentation, reordering, and redundancy padding. During retransmission, the sender reads a copy of the original heterogeneous encrypted block sequence from the buffer. The ciphertext and initialization vector in the data payload remain unchanged. In other words, the retransmitted data packet is a complete copy of the original heterogeneous encrypted block sequence. The sender does not recalculate or replace the ciphertext or initialization vector. The retransmission number is carried in the retransmission identifier field in the packet header, which is used by the receiver to identify which retransmission the packet has undergone in order to perform deduplication. This number does not participate in the initialization vector generation process. After receiving a data packet with the retransmission identifier, the receiver reads the original initialization vector from the packet header, uses the historical decryption key corresponding to this round in the local cache, and the original initialization vector to decrypt the data payload. The receiver then reassembles the data using the original block mapping table without re-parsing the mapping table structure.
[0063] If the requested missing round exceeds the range of historical versions retained in the cache, i.e., the missing round number is less than the current round number minus the cache retention quantity of five, then the key parameters for that round have been destroyed. Furthermore, since the temporary private key used during key negotiation follows the forward security principle and is destroyed after use, the key parameters for that historical round cannot be recovered through renegotiation, and the data for that round is unrecoverable. The sender returns an unrecoverable error to the receiver and terminates the current meeting.
[0064] During the meeting, if a new participant terminal requests access, the meeting initiator first executes the identity verification and hardware signature verification process in S1. After successful verification, the following synchronization operations are performed: The conference initiator performs asymmetric encryption on all key components in the current dynamic session key matrix, including the encryption key, obfuscation key, and update key corresponding to the current round, using the public key of the new participant terminal, and independently sends them to the terminal through the second-layer tunnel; at the same time, it sends the current round sequence number, the current block mapping table version number, and the historical key version in the current buffer that has not yet been confirmed by all receivers. After the new participating terminal decrypts the key, it stores the received key component in the local session context, marks its own state as synchronized to the current round, and then returns a synchronization completion confirmation frame to the initiator. After receiving the acknowledgment frame, the initiating end adds the new participating terminal to the receiving end's acknowledgment status table. During the synchronization process, in addition to obtaining the current round key component and round number, the initiating end also needs to send the currently active speaking role identifier and its corresponding key index to the new participating terminal through the second-layer tunnel to determine the encryption subkey combination to be used in the current time slice, as well as the SHA-256 hash value of the complete content of the latest verification frame in the current verification frame chain. The new terminal uses this hash value as the initial anchor point of its local verification frame chain. The trustworthiness of this anchor point is determined by the non-parallelism of the second-layer tunnel. The encrypted transmission guarantee means that the new terminal does not need to verify the continuity of the historical verification frame chain. It only needs to verify whether the hash value of the previous round of verification frame is consistent with the anchor point sent by the initiator when it receives the first subsequent verification frame, and whether the round number of the verification frame is the current round plus one. If both verifications pass, the verification frame is considered valid and the anchor point is marked as trustworthy. The continuity verification of the subsequent verification frame chain is performed according to the normal process in S4. If either verification fails, a synchronization error report is returned to the initiator, and the initiator resends all synchronization parameters for the current round. If a new participant encounters an error or timeout during the synchronization process, the initiating party can reissue the current key, retrying up to three times. If the attempt still fails, the new participant's access will be rejected.
[0065] Example 2 Please see Figure 2 This embodiment provides a paperless meeting data encryption transmission system, including the following modules: Session key initialization module The session key initialization module is deployed at the conference initiator. Its functional components include an identity acquisition unit, a random key distribution unit, a hardware verification unit, a key negotiation unit, and a key derivation and segmentation unit.
[0066] The identity collection unit is triggered at the scheduled time of the meeting. It reads the set of identity identifiers of all participants in the current meeting from the database of the meeting management system. Each participant's identity identifier is a digital certificate serial number or a unified enterprise identity identifier. At the same time, the identity collection unit reads the system clock to generate a meeting timestamp. The identity identifier set and the timestamp are concatenated according to the lexicographical order of the participant identifiers and then input into the SHA-256 hash function. The generated hash value is stored as a unique meeting session ID in a global session context structure in the initiator's memory. This structure persists throughout the meeting lifecycle.
[0067] When a session starts, the random key distribution unit calls the cryptographically secure pseudo-random number generator interface of the initiating operating system to generate a 256-bit one-time random key, which is temporarily stored in memory as a byte array. The random key distribution unit obtains the session ID from the identity acquisition unit, concatenates the session ID with the one-time random key, iterates through the participant identity set, reads the public key certificate corresponding to each participant terminal from the local key store, and encrypts the concatenated data one packet at a time using the SM2 asymmetric encryption algorithm. The encrypted ciphertext is then encapsulated as a session establishment signaling message and synchronously distributed to each participant terminal through a transport layer socket. Each participant terminal uses its own private key to decrypt and obtain the session ID and one-time random key, which are then stored in the local session context in the terminal's memory.
[0068] The hardware verification unit is triggered after each participating terminal obtains a one-time random key. Each terminal's hardware verification unit reads the hardware signature provided by the operating system firmware layer. This signature is a hash of the terminal's factory serial number, network interface MAC address, and the trusted platform module's PCR metric value. The terminal-side hardware verification unit concatenates the hardware signature, one-time random key, and session ID into an authentication data block, calls the terminal's private key to digitally sign the data block, generates an authentication credential, and returns it to the initiating hardware verification unit via a socket. Upon receiving the authentication credentials from each terminal, the initiating hardware verification unit performs dual verification: the first verification extracts the one-time random key from the authentication credential and compares it with the original value stored in memory over a constant time; after comparison, the key is marked as used. The second verification uses the terminal's public key to verify the validity of the digital signature and compares the hardware signature with the database registration information. The initiating hardware verification unit maintains an admission flag for each participating terminal, setting it to 1 only when both dual verifications pass; otherwise, it sets it to 0 and rejects the terminal's subsequent key negotiation requests.
[0069] The key negotiation unit starts after hardware verification is completed, and only terminals with the access flag set to 1 can participate. The key negotiation units of the initiating end and each participating terminal each call the elliptic curve cryptography library to generate temporary SM2 public-private key pairs. These pairs are stored in memory as a structure, with the private key marked as sensitive data and protected by memory encryption. Each terminal's key negotiation unit concatenates the temporary public key with the hardware signature and encrypts it using the initiating end's public key before sending it to the initiating end. The initiating end's key negotiation unit collects all temporary public keys and verifies the signatures. After successful verification, it constructs a public key using its own temporary public key and all terminal temporary public keys. The key negotiation unit of each terminal calculates the ECDH shared secret value with the initiating end based on the public key set and its own temporary private key. The initiating end synchronously calculates the shared secret value with each terminal. After all shared secret values are collected, the key negotiation unit concatenates the shared secret values in lexicographical order of identity identifiers and inputs them into the key derivation function KDF to derive a 256-bit dynamic root key. This root key is stored as a byte array in the global session context of the initiating end's memory, and each terminal stores it in its local session context. Any participant can destroy the key by calling the system memory clearing function at the end of the meeting.
[0070] The key derivation and segmentation unit takes the dynamic root key as a seed and inputs it into the KDF. It performs three derivations using the ASCII codes of three fixed derivation strings ENCRYPT, OBFUSCATE, and UPDATE as parameters, and outputs three independent 2048-bit key components, namely the encryption key, the confusion key, and the update key. The three components are stored in memory as a dynamic session key matrix in the form of a structure array.
[0071] The key derivation and segmentation unit further reads the list of speaking roles in the current meeting and constructs a role weight factor data structure for each role. This structure includes a speaking duration percentage field, a priority coefficient field, and a sensitivity level coefficient field. The unit calls the floating-point arithmetic unit to calculate the normalized weight factor and, based on the weight factor, proportionally extracts the total length of the key fragment for each role from the 2048-bit dynamic root key. If the total length is less than three times the preset security length threshold of 384 bits, a hash expansion function is called to pad it to 384 bits. The padded key fragment is then divided into three equal parts, which are used as the encryption subkey, obfuscation subkey, and update subkey for the corresponding role. The subkey structures for all roles are stored in the key allocation table in the initiator's memory, with the role identifier as the key and the key structure as the value.
[0072] Data block encryption module The data chunking and encryption module is deployed at the sending end, which is the terminal of the participant corresponding to the current speaking role or the conference initiator. The functional components of the module include a data type identification unit, a bandwidth monitoring unit, a chunking unit, an encryption unit, and an obfuscation processing unit. The module receives raw data streams from the conference application as input, and the data is passed in the form of file handles or stream buffer pointers.
[0073] The data type identification unit parses the metadata of the input data and determines the data type by checking the MIME type, file extension, and data stream characteristics. For screen projection data, the unit identifies its video encoding format and frame rate information. For document data, the unit reads the file header identifier. The unit outputs a data type enumeration value, which includes two values: continuous data type and discrete data type, and writes it into the type status register inside the module.
[0074] The bandwidth monitoring unit runs independently in a background thread, periodically sending UDP probe packets to the receiving end and measuring round-trip latency and throughput. The unit maintains a sliding window array of length 10 to store the most recent ten bandwidth measurements. Each time a new measurement arrives, the window average is calculated and compared with the current value to obtain the bandwidth fluctuation value, which is stored in the module's bandwidth status register in floating-point form for the block unit to read in real time.
[0075] At the beginning of each data fragmentation cycle, the chunking unit reads the type status register and bandwidth status register, and calculates a dynamic scaling factor based on bandwidth fluctuations. The unit has a base chunk size configuration item in memory; for continuous data types, this configuration item is stored as an integer value of 100 in milliseconds, and for discrete data types, it is stored as an integer value of 512 in kilobytes. The chunking unit multiplies the base chunk size by the scaling factor to obtain the actual chunk size, and selects different cutting strategies based on the data type. For continuous data, the unit uses the system clock as a reference and cuts the data read from the stream buffer according to the actual time window. For discrete data, the unit reads data sequentially from the file handle in bytes. The chunking unit assigns a zero-based incrementing sequence number to each generated data fragment, encapsulates the fragment data and the sequence number into a chunk data structure, and places it in the encryption queue.
[0076] The encryption unit retrieves data fragments one by one from the queue to be encrypted and reads the encryption subkey from the key structure of the current role. The unit generates an initial vector for each fragment, which is generated by inputting the fragment number and the current round number into the SHA-256 hash function and taking the first 128 bits. The unit calls the CBC mode encryption function of the SM4 block cipher algorithm, passing in the encryption subkey, the initial vector and the fragment data, and outputs a ciphertext byte array. The encryption unit encapsulates the ciphertext and the corresponding fragment number and initial vector into an encryption block structure and puts it into the queue to be obfuscated.
[0077] The obfuscation processing unit reads the obfuscation subkey from the key structure of the current role, and uses this key as a seed to input into the initialization function of the Mason rotation algorithm MT19937 to generate a random permutation sequence array. The array length is the total number of encrypted blocks in the obfuscation queue, and the array elements are integers from 0 to the total number minus one. The unit rearranges the encrypted blocks according to the permutation sequence so that the original adjacent blocks are not adjacent in the transmission sequence. At the same time, the unit reads the current packet loss rate from the feedback channel of the receiving end, calculates the padding length according to the packet loss rate, generates random padding bytes for each rearranged encrypted block, and embeds the CRC32 check value and the original sequence number in them. The unit distributes the padding blocks in reverse order according to the permutation sequence among the encrypted blocks, and finally outputs a heterogeneous encrypted block sequence, writes it into the output buffer of the module, and waits for the composite encrypted stream generation module to read it.
[0078] Composite encrypted stream generation module The composite encrypted stream generation module is deployed at the sending end, and its functional components include a tunnel management unit, a control signaling encryption unit, and a timing interleaving unit.
[0079] During module initialization, the tunnel management unit calls the operating system's socket application programming interface to create two socket descriptors, which are bound to different ports and configured as a first-layer tunnel and a second-layer tunnel, respectively. The socket for the first-layer tunnel is configured for the UDP protocol, with the socket options SO_SNDBUF and SO_RCVBUF set to larger buffers to support high throughput, and the IP_TOS field set to a low-latency service type. The socket for the second-layer tunnel is configured for the TCP protocol, with the SO_KEEPALIVE keep-alive option and TCP_NODELAY option set to disable the Nag algorithm to reduce signaling latency. The tunnel management unit maintains configuration structures for the two transport protocol stacks. The first-layer structure contains a verification policy field, a retransmission policy field, and a window size field, while the second-layer structure contains a forward error correction coding field and an ARQ parameter field.
[0080] The control signaling encryption unit reads the current dynamic session key matrix from system memory and extracts the update subkey. The unit constructs a control signaling data structure, which includes a key index field and a block mapping table field. The key index includes four subfields: encryption key identifier, obfuscation key identifier, update key identifier, and current version number. The block mapping table includes an original sequence array and a padding position array. The unit uses the update subkey to call the SM4 algorithm to serialize and encrypt the control signaling data structure, and outputs the encrypted control signaling byte stream.
[0081] The timing interleaving unit uses a fixed number of heterogeneous encrypted blocks as an interleaving cycle. During the transmission cycle, it reads the sequence of heterogeneous encrypted blocks from the module input buffer. At the beginning of each cycle, the unit queries the urgency level of the control signaling queue to be transmitted. The urgency level is calculated by a weighted sum of the key index update frequency and the mapping table change magnitude, with weighting coefficients of 0.6 and 0.4 preset in the unit configuration. The unit inserts the control signaling with the highest urgency level into the fixed time slot reserved between the third and fourth data blocks in the cycle. The time slot timer starts at the beginning of the insertion and triggers an interrupt when the time value reaches the average transmission time of a data block, forcibly terminating the current insertion operation and postponing the remaining signaling to the next cycle. The unit writes the composite encrypted stream containing data blocks and control signaling sequentially into the transmission buffer of the first-layer tunnel and sends it to the receiving end via a UDP socket. The temporarily borrowed control signaling data packets are appended with a CRC check value of the second verification strategy, which the receiving end uses for independent verification.
[0082] Synchronous update module The synchronization update module is deployed on both the sending and receiving ends. The sending end module of the synchronization update module includes a verification frame generation unit, a key transformation unit, and a retransmission processing unit. The receiving end module of the synchronization update module includes a verification frame parsing unit, a synchronization confirmation unit, and a buffer management unit.
[0083] The verification frame generation unit is triggered at the sending end by a timer at a preset interval. The timer period is dynamically calculated based on the throughput and packet loss rate. The base interval of 1000 milliseconds, the throughput impact coefficient of 50 milliseconds per megabit per second, and the packet loss rate impact coefficient of 20 milliseconds per 1 / 100th are preset as constants in the unit configuration. The lower limit of 200 milliseconds is enforced by the comparator inside the unit. At the trigger time, the unit constructs a verification frame structure, fills in three fields: the current encryption round number, the HMAC-SM3 authentication code generated based on the update key, and the SHA-256 hash value of the complete content of the previous round verification frame, and inserts the verification frame into the sending queue of the composite encrypted stream.
[0084] The receiving end's verification frame parsing unit extracts the verification frame from the composite encrypted stream, extracts the authentication code field, and calls the local HMAC-SM3 to recalculate it, confirming the sender's identity by comparison; the parsing unit also extracts the hash value field of the previous round and compares it with the hash value of the previous frame in the local cache to verify the continuity of the verification frame chain; after the continuity verification is passed, the parsing unit triggers the synchronization confirmation unit.
[0085] The synchronization confirmation unit constructs a synchronization confirmation frame structure at the receiving end, filling in three fields: the current confirmation round number, the current timestamp of the receiving end, and the reception status bitmap. The status bitmap is generated by the data reception bitmap array in the transmission control block of the receiving end. The array length is equal to the total number of segments in the current round, and each bit indicates the reception status of a segment. The unit returns the synchronization confirmation frame to the sending end through the second-layer tunnel.
[0086] After receiving the synchronization confirmation frame, the key transformation unit at the sending end extracts the round number and timestamp from the synchronization confirmation frame to construct a transformation seed. The unit converts the round number into a 32-bit unsigned integer and the timestamp into a 64-bit unsigned integer, concatenates them into a 96-bit sequence, and inputs it into SHA-256 to obtain a 32-byte transformation seed. The unit uses the current update key as the initial value, concatenates the initial value with the transformation seed, and inputs it into SHA-256 for hash iteration. The sum of the basic iteration count of 256 and the current round number is used as the actual iteration count, with an upper limit of 1000 iterations enforced by the unit's internal saturation counter. After the iteration is completed, a new 32-byte update key is output, and the same transformation is applied synchronously to the encryption key and the obfuscation key. After the transformation is completed, the unit encapsulates the three old key parameters before replacement into a historical version structure and writes it to the cache management unit.
[0087] The cache management unit maintains a circular queue with a capacity of five. Each unit stores five fields: round number, encryption key, obfuscation key, update key, and acknowledgment flag. When a unit receives a new historical version, it calls the enqueue function and checks if the queue is full. If it is full, it retrieves the oldest version from the head pointer position, checks its acknowledgment flag, and if it is 0, it triggers a retransmission request and overwrites it. If all five units have 0 flags, they are all marked as overwriteable and the new version is written. The acknowledgment flag is updated by the unit when it receives a synchronization acknowledgment frame and the status bitmap is all 1. Versions with the flag set to 1 are detected and released immediately during the polling task.
[0088] The retransmission processing unit is triggered upon receiving a retransmission request from the receiver, which includes the missing start and end rounds. The unit first pauses the trigger enable signal of the key transformation unit, causing the transformation timer to stop counting. The unit queries the circular queue of the buffer management unit. If the missing round is within the queue range, it retrieves the encryption key and update key from the corresponding historical version and controls the sender to retransmit from the beginning of that round. After retransmission, the unit receives a synchronization confirmation frame. After confirming that all bits in the status bitmap are 1, it restores the trigger enable signal of the key transformation unit. If the missing round exceeds the queue range, the data for that round is unrecoverable, and an unrecoverable error is returned to the receiver.
[0089] Example 3 This example illustrates a scenario where a company's R&D center holds a project review meeting with six participants: a moderator (who initiated the meeting), two main speakers, two discussion speakers, and one observer. The meeting is scheduled to last 90 minutes, with the agenda consisting of project background introduction, technical solution explanation, expert discussion, and summary resolution.
[0090] At the scheduled start time, the initiator collects the identity identifiers and system timestamps of the six participating parties to generate a unique session identifier. Simultaneously, the initiator generates a one-time random key, encrypts it using each party's public key, and distributes it to each terminal. Upon receiving the key, each terminal constructs an authentication credential using its own hardware signature and the random key and returns it to the initiator. After the initiator verifies all credentials, the six parties jointly execute an elliptic curve-based key negotiation protocol to derive a 256-bit dynamic root key. The initiator sets the weights for each role according to the agenda: host 0.25, two main speakers 0.20 each, two discussion speakers 0.15 each, and observers 0.05. After allocation, the host receives a total key segment length of 512 bits, each speaker receives 410 bits, each discussion speaker receives 307 bits (padded to 384 bits), and each attendee receives 102 bits (padded to 384 bits). The total key segment length for each role is divided into three equal parts, which serve as the corresponding encryption subkey, obfuscation key, and update subkey for that role. The lengths of each subkey are as follows: approximately 170.7 bits for the host, approximately 136.7 bits for the speaker, 128 bits for the discussion speaker, and 128 bits for each attendee, all of which are not less than the preset security length threshold. The three sets of encryption, obfuscation, and update subkeys for each role are stored in the dynamic session key matrix.
[0091] The meeting moved into the technical solution explanation phase, and speaker A began speaking. The data block encryption module on speaker A's terminal detected that the current network bandwidth was 8 megabits per second, a decrease of about 15% compared to the previous period, with a scaling factor of 0.85. At this time, the data to be transmitted was a screen-projected video stream, the data type was marked as continuous data, the basic block size was 100 milliseconds, and the actual block size was adjusted to 85 milliseconds. The module cut the video stream according to the 85-millisecond time window, and each segment of data was encrypted using the encryption subkey corresponding to speaker A's SM4 encryption, and then a random permutation sequence was generated using the obfuscation subkey to globally scramble the encrypted segments. The current network packet loss rate was 3%, the basic padding length was 32 bytes, the scaling factor was 256, and the calculated padding length was about 40 bytes. The padding content embedded the CRC32 check value and the original segment sequence number, and was distributed in reverse order between each segment to form a heterogeneous encryption block sequence.
[0092] The composite encrypted stream generation module establishes a first-layer UDP tunnel and a second-layer TCP tunnel. The presenter's terminal sends a heterogeneous encrypted block sequence through the first-layer tunnel, with every ten data blocks constituting an interleaving cycle. Before the cycle begins, the module queries the current control signaling queue. Since the key index has not changed, the mapping table changes slightly with the block structure adjustment, indicating a low urgency level. The control signaling is inserted into the reserved time slot between the third and fourth data blocks in the cycle in normal order. The control signaling is encrypted independently using the updated subkey and temporarily borrows the first-layer tunnel for transmission, but a strong checksum is appended to the end. After the receiving end extracts the control signaling from the first-layer tunnel, it uses the second-layer tunnel verification strategy to verify its integrity.
[0093] The synchronization update module calculates a preset interval based on the current throughput of 7.5 megabits per second and a packet loss rate of 3%. The base interval is 1000 milliseconds. Subtracting the throughput impact of 50 multiplied by 7.5 gives 625 milliseconds. Subtracting the packet loss impact of 20 multiplied by 3 gives 60 milliseconds, resulting in a final interval of approximately 565 milliseconds. The sending end embeds a verification frame every 565 milliseconds. The verification frame contains the current round sequence number, the HMAC authentication code generated based on the update key, and the hash value of the previous round verification frame. The receiving end verifies the continuity of the verification chain frame by frame. After three consecutive frames pass verification, a synchronization confirmation frame is returned. The confirmation frame carries the current round sequence number, the local timestamp, and a bitmap of the reception status of all segments. After receiving the confirmation frame, the sending end uses the concatenation of the round sequence number and the timestamp as the transformation seed. The current round is round 1. The number of iterations is 256 plus 1, which equals 257. The update key, encryption key, and obfuscation key are synchronously subjected to hash iteration transformation. After transformation, the old version is stored in the buffer.
[0094] During the expert discussion phase, the speaker role changes to speaker B. The initiating end generates a role switch control signaling message, which includes the target role identifier and the corresponding key index. After the receiving end confirms the switch, the key combination of speaker B is activated in the next time slice. During the discussion, network congestion occurs suddenly, and the packet loss rate rises to 8%. A check frame chain of a certain round is broken, and the receiving end detects that the missing round is round 7. A retransmission request is sent to the sending end, which suspends the current key update process, checks the buffer, and finds that the historical key of round 7 is still retained. It then uses the key of that round to retransmit the composite encrypted stream from the corresponding starting point. After all the retransmitted data is confirmed to be received, the key update process resumes, and the key continues to be changed in subsequent rounds according to the incremental iteration number.
[0095] The meeting entered the summary and resolution phase. The host spoke, and the time slice switched back to the host's key combination. At this point, the meeting had progressed to the 23rd round, with 279 iterations. The key strength gradually increased with each round. At the end of the meeting, all six terminals called the memory clearing function to destroy the dynamic root key and all subkeys. The historical cache was also cleared. All encryption parameters from this meeting were unrecoverable, ensuring that subsequent meetings could not revert to this data. Throughout the meeting, the video stream latency remained stable within 200 milliseconds, discrete data such as document annotations arrived intact, control signaling did not affect the main data throughput, and there were no decryption failures during key version switching.
[0096] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.
Claims
1. A method for encrypted transmission of paperless meeting data, characterized in that, Includes the following steps: S1: Collect the set of participant identities and meeting timestamps for the current meeting to generate a unique session ID; generate a dynamic root key for the session ID through a key negotiation protocol based on the hardware feature code of each participant's terminal and a real-time random number; and perform hierarchical expansion of the dynamic root key to form a dynamic session key matrix containing encryption keys, obfuscation keys, and update keys. S2: Parse the data type of the conference data to be transmitted, determine the block granularity based on the current network bandwidth fluctuation value; divide the conference data to be transmitted into multiple data segments according to the block granularity, extract the encryption key from the dynamic session key matrix and perform symmetric encryption on each data segment; use the confusion key to rearrange the encryption results of each data segment and fill in the redundancy to generate a heterogeneous encryption block sequence. S3: Establish a first-layer tunnel and a second-layer tunnel with the receiving end, wherein the first-layer tunnel is used to transmit the heterogeneous encrypted block sequence, and the second-layer tunnel is a control signaling logic channel containing a key index and a block mapping table; The update key is extracted from the dynamic session key matrix to independently encrypt the control signaling. The encrypted control signaling is then interleaved into the heterogeneous encryption block sequence in time order to form a composite encryption stream. S4: When the sending end transmits the composite encrypted stream to the receiving end, it embeds a verification frame carrying the current encryption round within a preset interval; the receiving end performs integrity verification based on the verification frame and confirms identity based on the authentication code generated based on the updated key in the verification frame. If the verification passes, it returns a synchronization confirmation frame to the sending end; after receiving the synchronization confirmation frame, the sending end triggers an irreversible transformation of the updated key in the dynamic session key matrix and applies the transformed new key parameters to the data encryption of the next round until the meeting ends.
2. The paperless conference data encryption transmission method according to claim 1, characterized in that, The dynamic root key in S1 also adopts an identity-based authentication mechanism and is bound to a one-time random key. The one-time random key is generated by the conference initiator when the session starts and is synchronously distributed to each participating terminal along with the session ID. Each participating terminal must verify the validity of the one-time random key and the consistency of the hardware signature code before it can participate in the key negotiation protocol.
3. The paperless conference data encryption transmission method according to claim 1, characterized in that, The hierarchical expansion described in S1 further generates role weight factors based on the current meeting's list of speaking roles. The dynamic root key is allocated bit lengths according to these role weight factors, serving as the total key fragment length for each speaking role. The key fragments for each role are expanded using a key derivation function, ensuring the expanded total length is not less than three times a preset security length threshold. These fragments are then divided into encryption subkeys, obfuscation subkeys, and update subkeys corresponding to the speaking role, with each subkey's length not less than the preset security length threshold. The dynamic session key matrix activates key combinations matching the current speaking role for encryption operations in a time-slice rotation manner. The time-slice rotation is performed within the same encryption round, while the irreversible transformation is performed between different encryption rounds. When the current role is switched, the control signal used to indicate the role switch is confirmed by the receiving end before the key combination corresponding to the next time slice is activated.
4. The paperless conference data encryption transmission method according to claim 1, characterized in that, When determining the chunk granularity based on the current network bandwidth fluctuation value as described in S2, the network bandwidth fluctuation value is converted into a dynamic scaling factor, and the preset basic chunk size is adjusted in real time based on the scaling factor. When the bandwidth decreases, the chunk size is reduced to reduce the load of a single transmission, and when the bandwidth increases, the chunk size is increased to reduce protocol overhead. At the same time, the data type is used to identify whether the conference data to be transmitted is continuous data. If it is continuous data, the chunk granularity is measured in time length; if it is discrete data, it is measured in byte length. The chunk size under both measurement units is adjusted by the same proportion of the scaling factor.
5. The paperless conference data encryption transmission method according to claim 1, characterized in that, When performing reordering and redundancy padding on the encryption results of each data segment using the obfuscation key as described in S2, the obfuscation key is used as a seed to generate a random permutation sequence, and the global positions of all encrypted data segments are shuffled according to this sequence, so that segments that are adjacent in the original order are not adjacent in the transmission sequence; the padding length of the redundancy padding is dynamically determined by the current network packet loss rate, and the higher the packet loss rate, the larger the padding length, and the padding content includes the original sequence number and verification information of the segment, and the padding positions are distributed among the segments in reverse order of the random permutation sequence, so that the receiving end can remove the padding and restore the original segment order according to the verification information and the original sequence number.
6. The paperless conference data encryption transmission method according to claim 1, characterized in that, S3 The first layer tunnel and the second layer tunnel each use independent transmission protocol stacks. The transmission protocol stack of the first layer tunnel is configured with a first verification strategy that prioritizes throughput, and the transmission protocol stack of the second layer tunnel is configured with a second verification strategy that prioritizes integrity. When the encrypted control signaling in the second layer tunnel is interspersed in the heterogeneous encrypted block sequence of the first layer tunnel, the control signaling temporarily borrows the transmission channel of the first layer tunnel, but its verification method still follows the second verification strategy of the second layer tunnel. After the control signaling completes transmission, the independent logical channel state of the second layer tunnel is restored.
7. The paperless conference data encryption transmission method according to claim 1, characterized in that, When the encrypted control signaling is interleaved into the heterogeneous encrypted block sequence in sequence as described in S3, the interleaving priority is determined according to the urgency level of the control signaling. The urgency level is jointly determined by the update frequency of the key index and the change range of the block mapping table. The interleaving position is based on a fixed number of heterogeneous encrypted blocks as an interleaving cycle. A fixed time slot is reserved for the encrypted control signaling in each interleaving cycle. When the fixed time slot is exhausted, the interleaving of control signaling in the current cycle is forcibly terminated, and the remaining control signaling is postponed to the next interleaving cycle. The key index contains the version number corresponding to the current encryption round, so that the receiving end can extract the matching historical key parameters from the buffer according to the version number to decrypt the control signaling.
8. The paperless conference data encryption transmission method according to claim 1, characterized in that, When embedding a verification frame carrying the current encryption round within a preset interval as described in S4, the length of the preset interval is jointly determined by the data throughput of the current composite encryption stream and the historical packet loss rate. The higher the throughput, the smaller the interval to make the verification frames more densely distributed; the higher the packet loss rate, the smaller the interval to enhance fault tolerance. In addition to carrying the current encryption round, the verification frame also carries the hash value of the verification frame from the previous round to form a verification frame chain. During integrity verification, the receiving end sequentially verifies the continuity of the verification frame chain. If the continuity verification of any verification frame fails, a retransmission request carrying missing round information is returned to the sending end. The sending end suspends the key update process of the current round based on the retransmission request and retransmits the composite encryption stream from the beginning of the corresponding round using the historical key parameters corresponding to the missing round. After the retransmitted data is confirmed to be received by the receiving end, the key update process is resumed and subsequent rounds continue. If the requested missing round exceeds the range of historical versions retained in the buffer, the data of that round is unrecoverable, and the sending end terminates the current meeting.
9. The paperless conference data encryption transmission method according to claim 1, characterized in that, When the update key in the dynamic session key matrix is triggered to undergo an irreversible transformation as described in S4, the transformation seed is formed by the round number corresponding to the current encryption round and the timestamp carried in the synchronization confirmation frame returned by the receiving end. The update key is then subjected to a hash iteration operation with the transformation seed. The number of iterations is dynamically determined by the current encryption round number. The larger the round number, the more iterations there are, but it does not exceed the preset upper limit of the number of iterations. The new key parameters generated after the transformation replace all key components of the previous round. The hierarchical relationship between the replaced key components is maintained. The historical key parameters before the replacement in the previous round are stored in the buffer and a preset number of historical versions are retained until the data of the corresponding round is confirmed to be received by the receiving end and then destroyed, so as to support the retransmission of missing round data within the buffer range.
10. A paperless conference data encryption transmission system, characterized in that, Includes the following modules: The session key initialization module is used to collect the set of participant identity identifiers and the meeting timestamp of the current meeting, and generate a unique session ID for the meeting; based on the hardware feature code of each participant's terminal and a real-time random number, a dynamic root key is generated for the session ID through a key negotiation protocol; the dynamic root key is hierarchically expanded to form a dynamic session key matrix containing an encryption key, an obfuscation key and an update key; The data segmentation and encryption module is used to parse the data type of the conference data to be transmitted and determine the segmentation granularity based on the current network bandwidth fluctuation value; divide the conference data to be transmitted into multiple data segments according to the segmentation granularity, and extract the encryption key from the dynamic session key matrix to perform symmetric encryption on each data segment; The encryption results of each data segment are rearranged and redundantly padded using the obfuscation key to generate a heterogeneous encrypted block sequence. A composite encrypted stream generation module is used to establish a first-layer tunnel and a second-layer tunnel with the receiving end, wherein the first-layer tunnel is used to transmit the heterogeneous encrypted block sequence, and the second-layer tunnel is a control signaling logic channel containing a key index and a block mapping table; The update key is extracted from the dynamic session key matrix to independently encrypt the control signaling, and the encrypted control signaling is interleaved into the heterogeneous encryption block sequence in time sequence to form a composite encryption stream; The synchronization update module is used to embed a verification frame carrying the current encryption round and an authentication code generated based on the update key within a preset interval when the sending end transmits the composite encrypted stream to the receiving end. The receiving end confirms the identity based on the authentication code in the verification frame and performs integrity verification based on the verification frame. If the verification passes, it returns a synchronization confirmation frame to the sending end. After receiving the synchronization confirmation frame, the sending end triggers an irreversible transformation of the update key in the dynamic session key matrix and applies the transformed new key parameters to the data encryption of the next round until the meeting ends.
