An information encoding, segmenting and recombining method and system based on application layer multi-channel

Abstract

The application provides an information encoding, splitting and recombining method and system based on application layer multi-channel, and belongs to the technical field of communication.The application aims to solve the technical problem that the existing communication scheme depends on the trust of a centralized service provider and has the risk of single-point privacy leakage.The core of the scheme is to construct a multi-factor security model: after the sending end completes information encoding and splitting locally, the sending end transmits the fragments through at least two independent user-level application layer channels, so that the fragments appear as ordinary text data that cannot be automatically identified and classified by the service provider in network transmission; the receiving end receives the fragments from the channels, recombines and decodes the fragments according to the hidden synchronization information, and restores the information.The application does not depend on the trust of any service provider platform, effectively reduces the risk of information theft, and can be conveniently used by ordinary individual users and institutions with commercial privacy protection needs.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and more specifically, to a method and system for information encoding, segmentation and reassembly performed locally on a user terminal. In particular, it relates to an information security scheme that provides enhanced privacy protection for end users based on a multi-factor security model and through private coding transformation and channel separation strategies. Background Technology

[0002] With the development of information technology, instant messaging and email have become the main methods of daily communication. However, whether it's email, WeChat, Lark, or iMessage, all user information typically needs to be stored and relayed through a single service provider's servers. This centralized architecture gives service providers the technical ability to obtain and analyze users' complete communication content, and user privacy essentially relies on one-way trust in the service provider. In recent years, frequent incidents of service provider data breaches and unauthorized access by internal personnel have highlighted the fragility of this trust model. Improving communication privacy and security has become an important issue worthy of attention. Furthermore, in enterprise business communication scenarios, a large amount of confidential business information is transmitted through public instant messaging tools, and enterprises have legitimate concerns about the integrity and privacy of this confidential information. Existing solutions either rely on the credibility of centralized platforms or require enterprises to deploy their own proprietary communication systems, both of which are costly and complex.

[0003] There are currently two main technical approaches to improving communication security: First, there are end-to-end encryption schemes, such as those based on the Signal protocol. The security of these schemes heavily relies on key distribution and management. Keys are typically generated or exchanged with the assistance of a single service provider, and the provider still has the potential to obtain the key and decrypt information. Furthermore, mainstream public-key encryption algorithms rely on mathematical problems for security. With the development of technologies such as quantum computing, these algorithms may face the long-term risk of being computationally broken in the foreseeable future, and the algorithms themselves may contain undiscovered security vulnerabilities.

[0004] The second approach is based on multipath transmission, which fragments data and transmits it through multiple communication paths. However, these approaches typically operate at the physical or network layer, requiring modifications to the underlying network protocol stack of the operating system or relying on specific network devices. Their deployment is complex, technically demanding, and unsuitable for ordinary users without specialized knowledge, making them unsuitable as a convenient consumer-grade privacy protection tool.

[0005] Therefore, there is an urgent need for a communication privacy protection method that does not rely on trust in centralized service providers, can address potential future security threats, and is easy for ordinary users to use. Summary of the Invention

[0006] This invention aims to address the following technical problems in existing communication privacy protection schemes: reliance on one-way trust in centralized service providers poses a risk of single-point data leakage; schemes based on public-key encryption algorithms face long-term potential cracking threats from technologies such as quantum computing, and the algorithms themselves may have unknown vulnerabilities; existing multi-path transmission-based schemes are complex to deploy and difficult for ordinary users to use.

[0007] To address the aforementioned issues and further enhance privacy protection, this invention proposes a method and system for information encoding, segmentation, and reconstruction based on application-layer multi-channels. The core idea is to construct a multi-factor security model that distributes the elements required to reconstruct information across multiple independent dimensions, ensuring communication privacy through channel separation and transformation strategy concealment. In the preferred embodiment, a private dictionary dimension is further introduced, ensuring that information reconstruction requires simultaneous knowledge of multiple independent channels, transformation strategies, and the acquisition of a private dictionary—all are indispensable.

[0008] The method provided by this invention is executed entirely locally on the user's terminal, without relying on an additional centralized server for information storage or relay. Users only need to utilize their existing legitimate communication accounts to securely distribute information. The actual distribution of information depends on a third-party communication service selected by the user; the method and system of this invention do not constitute an independent communication service.

[0009] The multi-factor security model distributes the elements required to restore information across multiple independent dimensions, achieving an exponential increase in protection capabilities. In the basic scheme, the model includes two dimensions: channel separation and policy concealment. The channel separation refers to splitting the original information and distributing it through at least two independent user-level application layer channels. These independent user-level application layer channels refer to two or more communication accounts, each with independent login credentials and independent communication address spaces. These accounts can belong to the same communication service type (e.g., two different email accounts) or different communication service types (e.g., a combination of an email account and an instant messaging account). For email channels, the account combinations of the communicating parties can be different; for instant messaging channels, both parties must use the group chat function provided by the same service provider. Because the accounts are logically isolated on the server side, any single account's permissions can only access the information fragments carried by that account, and cannot reconstruct the true information; any unauthorized third party must simultaneously obtain access permissions to all independent accounts to obtain all information fragments, which greatly increases the difficulty of obtaining complete data.

[0010] The aforementioned strategy concealment refers to the fact that the rules for splitting and transforming information are agreed upon in advance by both communicating parties through an out-of-band channel, and will not be transmitted in the application layer channel. The out-of-band channel refers to other information transmission methods that do not pass through the application layer channel, such as face-to-face verbal agreements, telephone calls, SMS messages, QR code scanning, NFC tap-to-pay, etc. Without knowledge of these strategies, unauthorized parties are like facing a jigsaw puzzle without knowing the assembly rules, unable to reconstruct it into meaningful original information.

[0011] In the preferred embodiment, the multi-factor security model further introduces a private dictionary dimension: both communicating parties use a completely private encoding dictionary to encode information. This dictionary is generated and stored entirely locally on the user terminal and can be specified by the user, without any network transmission. Even if an unauthorized party obtains all fragments and knows the transformation strategy, it cannot decode and reconstruct the original information without possessing this dictionary. This combination of three independent dimensions—channel separation, policy concealment, and dictionary privacy—creates a three-dimensional protection system.

[0012] The transformation strategy of this invention is a core technical means to achieve information camouflage and privacy protection. The transformation strategy can employ publicly available encoding methods such as Base64, or it can further employ encoding methods based on private large character sets.

[0013] The encoding method based on a private large character set is a preferred embodiment of the present invention. Specifically, the communicating parties agree on a specific electronic document or a specific URL and extraction rules through an out-of-band channel, for example, agreeing to "take the first 1024 non-repeating characters of the second part of 'The Three-Body Problem'". Then, each party parses the electronic document locally on its terminal and extracts the exact same character set as an encoding dictionary. The character set may include various text characters such as Chinese characters, punctuation marks, English letters, numbers, etc., or any combination thereof. After the sending end converts the original data into a byte stream, it encodes it using the encoding dictionary to generate a character sequence composed of characters from the dictionary. This character sequence appears as ordinary text data without fixed encoding features and cannot be classified by automated content recognition devices during transmission. Since Base64 encoding has a fixed character set and padding rules, its encoding results are easily recognized and classified by feature-based automated devices; while the present invention preferably uses a private large character set, which has no fixed encoding features that can be recognized by automated devices. At the same time, since the cardinality of the encoding dictionary is large enough, the length of the encoded character sequence will not be excessively expanded compared to the original data, and the transmission efficiency is controllable.

[0014] For publicly available encoding methods like Base64, since the encoded result is a sequence of alphanumeric symbols, it can be further encapsulated into a pre-defined message shell, such as disguising it as a system log or shared link, to weaken the identifiability of its fixed encoding characteristics. For private large character set encoding schemes, no additional shell is needed; they can also be further disguised as common cultural content such as article excerpts, poetry quotations, scripture copying, or lists of names. Furthermore, generative artificial intelligence models running locally on the terminal can be used to semantically reconstruct character sequences, rewriting them into grammatically correct and logically consistent natural language text, achieving a higher level of semantic camouflage.

[0015] Furthermore, the technical solution of this invention is compatible with and can be superimposed on end-to-end encryption schemes. The sending end can first perform end-to-end encryption on the original information, and then send the encrypted ciphertext to the encoding and segmentation engine for subsequent encoding and splitting processing; the receiving end, after completing reassembly and decoding, then performs end-to-end decryption. The two security mechanisms operate independently and complement each other: end-to-end encryption ensures the confidentiality of the content, while the channel separation and strategy concealment of this solution ensure the concealment and resistance to deconstruction during the distribution process. Even if end-to-end encryption is breached due to quantum computing or other reasons, this solution can still provide an independent security barrier.

[0016] The splitting method of this invention is a specific means of dividing the encoded data sequence into multiple information segments. Any splitting method that makes it insufficient to reconstruct the complete original information when acquiring any segment from any single channel is applicable. This invention provides the following preferred methods or combinations thereof: Fixed-length block splitting: The encoded data sequence is divided into several segments of fixed length, with the number of segments matching the number of channels N. For example, data of total length L is divided into segments of L / N and distributed through N channels. If L is not divisible by N, trailing zeros or random data can be added.

[0017] Modulo-N block splitting: The encoded data sequence is divided into modulo-N blocks based on position. For the i-th data unit in the data sequence (i starts counting from 0), it is assigned to the (i mod N)-th channel, with channel numbers ranging from 0 to N-1. The advantage of this method is that acquiring data from any single channel alone results in discontinuous data units with an interval of N, leading to a high degree of information corruption. When N=2, modulo-2 block splitting is equivalent to parity-even bit splitting.

[0018] Random packet splitting: The communicating parties pre-agree on a random seed via an out-of-band channel. The sending end generates a random allocation sequence based on this seed, assigning a channel number to each data unit. Even if an unauthorized party obtains all channel segments, without knowing the seed, it needs to exhaustively enumerate all possible allocation combinations. The computational load increases exponentially with data length, making it infeasible to reconstruct information with limited computing resources.

[0019] In a preferred embodiment of the present invention, modulo-2 packet splitting is used as the basic scheme for dual-channel scenarios, and modulo-N packet splitting is used as the basic scheme for multi-channel scenarios. When a higher level of security is required, random packet splitting is used as an enhancement scheme.

[0020] Furthermore, a random delay mechanism can be introduced when distributing information segments through various channels. After the sending end completes the segmentation, it does not send all segments simultaneously, but instead introduces a randomly generated time delay for each channel. This delay can be randomly selected within a preset reasonable range (e.g., between 0.5 seconds and 30 seconds), so that the time points at which each segment is sent through different channels are staggered, simulating the behavioral pattern of a natural person operating different communication tools at different times, thereby eliminating the behavioral characteristics caused by high-frequency, high-temporal-similarity transmission behavior.

[0021] When the length of any encoded information fragment exceeds the single message length limit of its corresponding channel, the sender automatically triggers an adaptation mechanism: further dividing the information fragment into multiple sub-fragments, each sub-fragment's length not exceeding the single message length limit of the channel, and embedding a sub-sequence number into each sub-fragment. The sub-sequence number is embedded covertly along with the original session identifier and sequence number. Upon receiving multiple sub-fragments from the same channel, the receiver first reassembles them sequentially according to the sub-sequence numbers to restore the complete information fragment for that channel, and then sends the complete information fragments from each channel to the reassembly engine for reconstruction and decoding. This mechanism typically does not trigger in everyday text communication, but for large amounts of data such as files and images, the encoded length may exceed the channel limit. In this case, the mechanism ensures the integrity of the distribution, and the entire process is completely transparent to the user.

[0022] This invention employs a covert embedding method for synchronization information. The synchronization information includes session identifiers and sequence numbers, used to indicate that multiple fragments belong to the same session and their respective order. Specifically, the synchronization information is processed together with the content data of the information fragments using a transformation strategy and then embedded into the content data. For large character set encoding schemes, several marker characters representing the synchronization information can be directly inserted at the beginning or end of the character sequence or at agreed-upon specific positions. Since the entire character sequence appears as a random collection of characters to a third party without access to the codebook, the marker characters and content characters are visually identical and cannot be effectively distinguished or extracted by unauthorized parties. Only a legitimate recipient who possesses the codebook and the embedding rules can accurately identify and separate the synchronization information.

[0023] Furthermore, the concealed embedded information may also include the sender's identity information. This identity information, along with the synchronization information, is processed using a transformation strategy, and the receiving end can extract this information during the reassembly and decoding stage for auditing and tracing purposes.

[0024] Furthermore, the sequence number can be replaced by a random number encoding based on a random seed instead of a simple sequential number. This random seed can be the same as the random seed used in the random grouping and splitting method, without any additional agreement between the communicating parties. The sending end generates a set of pseudo-random numbers based on this seed as the sequence number of each segment. Since the random number sequence appears irregular to the outside, even if an unauthorized party extracts the synchronization information, it cannot deduce the correct arrangement order of the segments. The receiving end generates the same random number sequence based on the same random seed, and uses this sequence to sort and reassemble the segments.

[0025] Furthermore, when monitoring various channels, the receiving end can quickly distinguish between encrypted fragments and ordinary communication content by detecting whether the message contains a synchronization information marker that conforms to the agreed format. Ordinary messages without synchronization information markers are not processed by the system; encrypted fragments containing synchronization information markers are automatically extracted and sent to the reassembly engine. This identification process is completed locally and is transparent to the user.

[0026] 6. Methods and Steps Based on the above model and strategy, this invention provides an information coding, segmentation and reassembly method based on application layer multi-channel, the method including a transmitting end method and a receiving end method. Figure 1 This is a schematic diagram of the overall process of the method of the present invention. During the information distribution process, the complete information remains only at the sending end and the receiving end; no other link possesses the complete information.

[0027] The sending method includes the following steps: First, it acquires the raw information to be processed input by the user; then, it calls the local encoding and segmentation engine to encode and split the raw information into two or more information segments according to the transformation strategy. If the length of any segment exceeds the single message limit of the corresponding channel, it is automatically further segmented into sub-segments and embeds sub-sequence numbers, while concealing the embedded synchronization information; after the information segmentation is completed, a random delay is introduced for the time of each segment to be sent through different channels to eliminate the time correlation characteristics of the sending behavior; finally, it calls the background communication module to directly distribute the two or more information segments to the receiving end through at least two independent user-level application layer channels associated with the sending end. The user-level application layer channels include, but are not limited to, communication channels that users can directly use, such as email accounts, instant messaging tool accounts, and mobile SMS accounts; different accounts under the same service type are considered as independent channels. The information segmentation and transformation process is completed locally on the user terminal, without relying on additionally deployed centralized servers for storage or relay.

[0028] The receiving end method includes the following steps: First, it calls the background communication module to listen for new messages through at least two independent user-level application layer channels associated with the receiving end. When an encrypted fragment containing a synchronization information marker is detected, it is automatically extracted; ordinary messages without synchronization information markers are not processed. Then, it calls the local reconstruction engine to sort and pair the fragments according to the hidden embedded synchronization information; finally, it reconstructs and decodes the fragments according to the reconstruction strategy corresponding to the transformation strategy, restores the original information, and presents it to the UI module.

[0029] Furthermore, each client can pre-configure a channel pool, which contains configuration information for two or more available channels. Before communication, both parties must agree on at least two common channels to be used for the communication via out-of-band channels, and each party selects the configuration corresponding to the agreed channels from its own channel pool. Each time information is distributed, the sending client dynamically selects an currently available channel from the agreed common channels to distribute the information fragment.

[0030] Figure 2 and Figure 3 The principles of the transmitter and receiver using dual mailboxes as channels are demonstrated respectively. Figure 4 This demonstrates the interface for users to configure channel pools. Figure 5 A structural block diagram of the system of the present invention is shown. Figure 6 The diagram illustrates the configuration and communication process using dual instant messaging as the channel.

[0031] This invention also provides an application-layer multi-channel information coding, segmentation, and reassembly system. For example... Figure 5 As shown, the system runs locally on the user terminal and includes four modules: The UI module is used to acquire the original information input by the user and present the restored original information to the user. It also provides a configuration entry point for the channel pool, allowing users to bind multiple application-layer channels; an entry point for communication channel agreement, allowing both communicating parties to agree on the common channel used for this communication via out-of-band; and a configuration entry point for transformation strategies, allowing users to agree on the rules for generating the encoding dictionary.

[0032] Encoding and Segmentation Engine: This is a locally running software module and the core of the sending end. It is used to execute preset transformation strategies, including the core encoding method based on a private large character set, which encodes the original information and splits it into multiple information fragments with ordinary text characteristics, while concealing and embedding synchronization information.

[0033] Reassembly Engine: A locally running software module and the core of the receiver. It executes the reassembly strategy corresponding to the transformation strategy, receives information segments from multiple channels, sorts, deduplicates, and merges them according to the hidden embedded synchronization information, and finally decodes and restores the original information.

[0034] Background communication module: Used for low-level communication with user-level application layer channels. This module supports multiple communication protocols, such as SMTP, IMAP, and HTTP API, and can simultaneously connect to at least two independent user-level application layer channels registered by the user to complete the silent distribution and reception of information fragments.

[0035] The four modules work together to achieve the multi-channel information encoding, segmentation, and reconstruction functions described in this invention. The information segmentation and transformation process is completed locally at the user terminal, and the actual distribution of information depends on a third-party communication service selected by the user. Each terminal integrates the complete functions of the four modules, enabling it to initiate information segmentation and distribution, as well as receive and reconstruct information.

[0036] The technical solution of this invention is applicable to both personal privacy protection and enterprise business communication scenarios. It can be deployed as a security enhancement component for existing enterprise communication tools without replacing the existing communication platform. For example, a lawyer in a law firm who needs to communicate case details with clients via email can deploy this solution on their terminal to encode and split the communication content and send it through two different email addresses, ensuring that no single platform can obtain the complete communication content, thereby enhancing the privacy protection of client information.

[0037] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention constructs a multi-factor security model, distributing the elements required to restore information to independent dimensions such as channel separation and policy concealment. In the preferred scheme, dictionary privacy is also introduced, so that the security of information does not depend on trust in any single centralized platform, nor on the unsolvability of a single mathematical problem, which greatly increases the security of privacy information transmission.

[0038] 2. This invention employs text data encoding, splitting, and transformation strategies, enabling information fragments to be transmitted over the network as ordinary data traffic, effectively reducing the risk of privacy information being automatically identified and categorized. In particular, the encoding method based on a private large character set, lacking fixed encoding characteristics, has a stronger ability to prevent unauthorized parties from automatically classifying and identifying it.

[0039] 3. The various splitting methods provided by this invention (including fixed-length group splitting, modulo N group splitting, and random group splitting based on an agreed seed) further enhance information security. In particular, with random group splitting, the computational cost of exhaustive splitting increases exponentially with the data length when an unauthorized party does not know the seed. Even if all segments are captured, it is almost impossible to complete the reconstruction computationally.

[0040] 4. This invention introduces a channel pool mechanism, allowing users to pre-configure multiple available channels, from which the system dynamically selects and distributes channels, improving ease of use and communication reliability.

[0041] 5. All operations in this invention are completed locally on the user's terminal, without relying on an additional centralized server. Users only need to configure their existing communication accounts to achieve secure information distribution. The actual distribution of information depends on a third-party communication service selected by the user. The method and system of this invention do not constitute an independent communication service, are easy to deploy, and can be used by ordinary users.

[0042] 6. This invention can provide a technical interface for accountability and compliance auditing by concealing embedded identity information and communication records stored in the channel itself, while protecting the privacy of communication content.

[0043] 7. This invention is applicable to enterprise business communication scenarios. Without changing the existing usage habits of communication tools, it provides enhanced privacy protection for confidential business information and reduces the risk of business information being leaked through a single channel. Attached Figure Description

[0044] Figure 1 A flowchart illustrating the information encoding, segmentation, and reassembly method based on application layer multi-channel provided in an embodiment of the present invention; Figure 2 A schematic diagram illustrating the principle of an email-based sender provided in an embodiment of the present invention; Figure 3 A schematic diagram illustrating the principle of an email-based receiver provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the user interface configuration provided in an embodiment of the present invention; Figure 5 A structural block diagram of the system provided in the embodiments of the present invention; Figure 6 This is a schematic diagram of the channel configuration and communication process based on dual instant messaging provided in an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Example

[0046] This embodiment uses dual mailboxes as dual channels for asynchronous communication, employing a modulo-2 packet splitting method. The App in this embodiment runs as a local information processing tool; all encoding, segmentation, and reassembly operations are completed locally on the user's terminal. The App does not build or operate any message transmission server; the actual distribution of information relies on the basic communication services provided by the user's self-registered email service provider.

[0047] Users Alice and Bob wished to conduct private and secure text communication. They agreed in person to use the first 4096 distinct characters (including Chinese characters and punctuation marks) of the second part of the novel *The Three-Body Problem* as their encoding dictionary, and to employ a modulo-2 grouping method. They each had two separate email addresses: a 163 email address and a QQ email address.

[0048] First, the communication channels were configured. Alice and Bob agreed in person to use a 163 email address and a QQ email address as the two shared channels for this communication. Figure 4 As shown, Alice configures a "channel pool" on her terminal app, binding her 163 email's SMTP / IMAP authorization code to channel A and her QQ email's corresponding information to channel B. Bob also independently configures his terminal, binding his corresponding email account to the same two shared channels. Both parties can also configure other backup channels in their respective channel pools for failover.

[0049] The sending process is as follows Figure 1 and Figure 2As shown, when Alice sends raw information to Bob through the App, the App first obtains the raw information input by the user. Then, it calls the local encoding and segmentation engine to convert the raw information into a byte stream, encodes it using the agreed-upon dictionary from the "Three-Body Problem" series, and generates a character sequence. The engine then processes this character sequence using a modulo-2 grouping method: characters with even-numbered sequences are extracted to form segment A; characters with odd-numbered sequences are extracted to form segment B. After segmentation, the engine appends a session identifier and sequence number to each segment, and may also append sender identification information (such as the sender's email address prefix or a hashed user identifier). This information is also encoded and inserted into the first M positions of the character sequence of each segment (for example, M=8, meaning the first 8 character positions are used to carry synchronization information, and M must be large enough to accommodate the complete synchronization information encoding). If the length of any segment exceeds the single email body length limit of the corresponding email address, the engine automatically further divides the segment into multiple sub-segments, embeds a sub-sequence number in each sub-segment, and sends them separately. Finally, in the background, the App introduces a random delay (e.g., a random value between 0.5 and 30 seconds) for the sending time of the two segments before sending segment A and segment B respectively via the SMTP protocol, and then sends them to Bob's corresponding email addresses via 163 email and QQ email respectively.

[0050] The receiving end process is as follows Figure 1 and Figure 3 As shown, Bob's app continuously monitors his 163 and QQ email accounts via the IMAP protocol. Upon receiving two emails from Alice, the app automatically extracts the email body and checks if it conforms to the agreed-upon synchronization information format to determine if it's an encrypted fragment. If confirmed as encrypted, and multiple fragments are received on the same channel, they are first concatenated in order according to their sequence numbers to reconstruct the complete fragment for that channel. Then, the local reconstruction engine is invoked. It first extracts and decodes the synchronization information and sender identification information from the agreed-upon positions (e.g., the first 8 characters) of each fragment's character sequence. Based on the session identifier and sequence number, it confirms that the two fragments belong to the same session and determines that fragment A is an even-numbered sequence and fragment B is an odd-numbered sequence. The reconstruction engine alternates between the two sequences according to their odd and even positions—that is, taking the first character from fragment A, the first character from fragment B, the second character from fragment A, the second character from fragment B, and so on—to recover the complete character sequence. The engine then performs reverse decoding according to the agreed-upon character dictionary from *The Three-Body Problem* to reconstruct the original information. Bob's app interface eventually displays the restored information.

[0051] It should be noted that the above dual-mail solution also applies to two different accounts from the same email service provider (e.g., two 163 email accounts or two QQ email accounts). Since each account has its own independent login credentials and independent inbox, which are logically isolated storage spaces on the server side, they also constitute independent user-level application layer channels, thus achieving the channel separation effect described in this invention.

[0052] Because this embodiment uses a standard email protocol, both the sender's and receiver's email service providers retain records of the sending and receiving of each segment on their servers, including the sender's email address, sending time, and email body. Simultaneously, the receiving end can extract the sender's embedded identity information during the reconstruction and decoding stage. This mechanism protects the privacy of the communication content while preserving the traceability of the communication participants.

[0053] When transmitting non-text information with large amounts of data, such as files or images, the length of the encoded sequence may significantly exceed the length limit of a single email body. In this case, the aforementioned segmentation and splicing mechanism automatically takes effect to ensure the complete transmission of large files. For everyday short text communication, the encoded sequence is usually much shorter than the channel length limit, and this mechanism does not incur additional overhead. Example

[0054] This embodiment uses "dual instant messaging" as a dual-channel synchronous communication method and employs a random group splitting method.

[0055] Alice and Bob each have two separate instant messaging accounts, one for a WeChat group and the other for a Lark group. They agreed in person to use the first 4096 unique Chinese characters from "Dream of the Red Chamber" as the encoding dictionary and to use "2024Meeting" as the random seed.

[0056] like Figure 4 and Figure 6 As shown, Alice configures a channel pool in her terminal app, binding the Webhook address of the WeChat group robot to channel A and the Webhook address of the Lark group robot to channel B. Bob also performs the same independent configuration.

[0057] When Alice sends a message, the app calls the encoding and segmentation engine to encode the message using a pre-defined private Chinese character dictionary, generating a character sequence. Then, based on the pre-defined random seed "2024Meeting," the engine generates a random allocation sequence, which randomly assigns each Chinese character in the sequence to either channel A or channel B. The engine then splits the character sequence into two segments based on this allocation sequence, and sends both segments as group messages to the shared group chat via HTTP calls to the WeChat group robot API and the Lark group robot API.

[0058] Bob's app continuously listens to two group chats via HTTP long connections or polling. Upon receiving two messages, it extracts character sequence fragments. The reconstruction engine generates identical randomly assigned sequences based on an agreed-upon random seed "2024Meeting," thereby determining the corresponding position of each Chinese character in the two fragments. The two fragments are then reconstructed into a complete character sequence, decoded using a Chinese character dictionary, and finally displayed on the app interface as the restored original information.

[0059] This embodiment leverages the high real-time performance of instant messaging tools, combined with a random grouping and splitting method, to achieve near real-time communication while further enhancing information security. Even if a third party obtains all group messages, without knowing the random seed, they cannot determine the correct attribution and order of each Chinese character, thus making it impossible to reconstruct the original information. Example

[0060] This embodiment provides a hybrid channel solution that combines email and instant messaging.

[0061] The message segments, containing key real-time metadata such as message type and timestamp, are sent via instant messaging tools, while the core content segments are sent via email. Upon receiving the instant messaging message, the receiving end enters a waiting state, reassembling and restoring the message as soon as the email arrives, thus balancing real-time performance with content security. Example

[0062] This embodiment provides a method for dynamic switching and fault tolerance of channel pools.

[0063] When a channel in the channel pool fails to distribute data due to a service anomaly, the App automatically detects the fault and, according to a preset priority strategy, calls a backup channel from the channel pool to complete the transmission of the segment. This process is completely transparent to the user.

[0064] When no available backup channels remain in the channel pool, the App cannot securely transmit the message fragment and will notify the user of the distribution failure. Simultaneously, the App continuously checks the availability of each channel at preset intervals (e.g., every 30 seconds). When at least two channels are detected to be available again, the App automatically retrieves the message to be transmitted from the failure queue and re-executes the encoding, splitting, and distribution process. The entire process requires no manual user intervention, ensuring eventual reachability of communication.

[0065] The aforementioned fault-tolerance mechanism ensures high availability of communication when a channel is available, promptly provides feedback on the status to users in extreme cases, and automatically retransmits after the channel is restored, thus avoiding information loss.

[0066] This specific embodiment is only used to illustrate the present invention and is not intended to limit the present invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also fall within the protection scope of the present invention.

Claims

1. A method for information coding, segmentation, and reassembly based on application-layer multi-channel systems, characterized in that, Includes sending and receiving methods; The sending end method includes: acquiring the original information to be processed; calling a local encoding and segmentation engine to encode and split the original information into two or more information segments according to a preset transformation strategy, wherein the transformation strategy can convert the original information into ordinary text data that cannot be classified by an automated content recognition device based on fixed encoding features, while embedding the synchronization information required for reassembly into the information segments; calling a background communication module to distribute the two or more information segments to the receiving end through at least two independent user-level application layer channels associated with the sending end, this process is completed locally on the user terminal and does not rely on an additionally deployed centralized server for storing or relaying the information segments; The receiving end method includes: invoking a background communication module to receive the two or more information fragments respectively through the at least two independent user-level application layer channels associated with the receiving end; invoking a local reconstruction engine to sort and pair the fragments according to the hidden embedded synchronization information, and reconstructing and decoding the fragments according to the reconstruction strategy corresponding to the transformation strategy to restore the original information.

2. The method according to claim 1, characterized in that, The transformation strategy includes an encoding transformation method based on a private large character set. This method includes: the two communicating parties agree on electronic documents and character extraction rules through an out-of-band channel, and generate identical character sets as encoding dictionaries on their respective terminals; after the sending end converts the original data into a byte stream, it encodes the data using the encoding dictionary to generate a character sequence composed of characters from the dictionary.

3. The method according to claim 1, characterized in that, The method of splitting the original information into two or more information segments is such that acquiring any segment of any channel alone is insufficient to reconstruct the complete original information. It can be one or a combination of fixed-length group splitting, modulo N group splitting, random group splitting, or any other splitting method that can achieve the same technical effect. Modulo N group splitting allocates the i-th data unit in the data sequence to the (i mod N)-th channel, where i starts counting from 0 and N is the number of channels, with channel numbers ranging from 0 to N-1; random group splitting uses a random seed pre-agreed by both communicating parties, and the sending end generates a random allocation sequence based on the seed to assign a channel number to each data unit.

4. The method according to claim 1, characterized in that, The at least two independent user-level application layer channels refer to two or more independent user-level application layer communication accounts, each with independent login credentials and independent communication address space; the communication accounts include two or more email accounts, or two or more instant messaging tool accounts, or two or more mobile SMS accounts, or any combination of the above types; wherein different accounts under the same communication service type are considered as independent user-level application layer channels.

5. The method according to claim 1, characterized in that, Each client is pre-configured with a channel pool, which contains configuration information for two or more available channels. Before communication, the two parties agree on at least two common channels. When distributing information, the currently available channels are dynamically selected from the common channels to distribute information fragments, and a random delay is introduced for the time of each fragment to be sent through different channels, so as to eliminate the time correlation characteristics of the sending behavior.

6. The method according to claim 1, characterized in that, The synchronization information attached to the information fragment is processed together with the information fragment using a transformation strategy and is hidden and embedded within the information fragment, so that a third party without a codebook cannot distinguish between the synchronization information and the content data.

7. The method according to claim 5, characterized in that, When a distribution failure is detected in any of the currently used application layer channels, other currently unused channels from the common channels are automatically called as backup channels to complete the distribution or reception of information segments. If no backup channel is available, a distribution failure prompt is sent to the user, and the availability status of each channel is continuously checked at preset intervals. When at least two channels are detected to be available again, the information to be distributed is automatically retrieved from the failure queue and the distribution process is re-executed.

8. The method according to claim 1, characterized in that, When the length of any encoded information fragment exceeds the single message length limit of its corresponding user-level application layer channel, the sending end automatically further divides the information fragment into multiple sub-fragments, embeds a sub-sequence number into each sub-fragment, and distributes them through the channel respectively; before reassembling and decoding, the receiving end first splices multiple sub-fragments of the same channel in sequence according to the sub-sequence number to restore the complete information fragment of the channel.

9. A system for information coding, segmentation, and reassembly based on application-layer multi-channels, characterized in that, Running locally on the user terminal, including: The UI module is used to obtain the original information input by the user and to present the restored original information, and provides the configuration entry for the channel pool and transformation strategy. The encoding and segmentation engine is used to execute preset transformation strategies, encode the original information and split it into multiple information fragments, and hide and embed synchronization information. The reassembly engine is used to sort and pair received information fragments based on the hidden embedded synchronization information, and decode and restore the original information according to the reassembly strategy. The background communication module is used to distribute and receive information fragments with at least two independent user-level application layer channels registered by the user. The above four modules work together to realize the multi-channel information encoding, segmentation and reconstruction functions described in this invention; the encoding and segmentation engine, reconstruction engine and background communication module all run locally on the user terminal and do not rely on an additional centralized server; the actual distribution of information depends on the third-party communication service selected by the user. Each terminal integrates the complete functions of the above four modules, and can both initiate information segmentation and distribution, and receive information and reconstruct it.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images