A multicast file distribution method, device and medium based on intelligent compression

Abstract

The application discloses a multicast file distribution method and device based on intelligent compression and a medium, relates to the technical field of edge multicast file distribution, and comprises the following steps: multicast pressure layer triggering is performed on a folder block expansion chain, common continuously missed small files in the expansion chain position are merged into a candidate small file cluster, the candidate small file cluster is screened and folded according to the yield, an aggregated compressed segment is generated, and a large file chain position with an original block reservation mark is connected to generate an intelligent compression folding chain; based on the intelligent compression folding chain, the aggregated compressed segment is subjected to directory anchoring type bearing reorganization to generate a compressed sending chain position, the large file chain position with the original block reservation mark is subjected to original block path binding to obtain a reservation sending chain position, and the compressed sending chain position and the reservation sending chain position are mixed and arranged to generate a mixed multicast object sequence. The application reduces the multicast scheduling burden and the missing supplement range, and improves the client file distribution continuity and the transmission resource utilization rate.

Description

Technical Field

[0001] This invention relates to the field of edge multicast file distribution technology, and in particular to a multicast file distribution method, device and medium based on intelligent compression. Background Technology

[0002] With the development of edge computing, industrial internet and local area network collaborative applications, edge data transmission is gradually shifting from single-point file copying to multicast file distribution to multiple edge clients. Existing technologies typically involve the server establishing a multicast group, dividing the files to be distributed into fixed block lengths, compressing them uniformly or sending them directly, and combining this with the client's cache status to reduce duplicate transmissions. At the same time, missing block feedback is used to complete subsequent retransmissions, thereby meeting the need for rapid synchronization of software packages, configuration files, model files and batch business files between edge nodes.

[0003] Existing multicast file distribution typically employs a uniform compression strategy at the file or task level, without simultaneously considering the number of multicast receivers, differences in cache hit rates among clients, and the granularity of resending small and large files. In scenarios where multiple clients only share a missing set of consecutive small files while large files only lack partial file blocks, uniform file-by-file transmission increases the number of multicast scheduling entries, and uniform aggregation compression expands the scope of large file resending. This is because the existing process lacks a mechanism to orchestrate the aggregation compression of small files and the preservation of original blocks of large files in the same chain according to the receiving needs of multiple receivers. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a multicast file distribution method based on intelligent compression to solve the problem of difficulty in coordinating compression triggering and file granular distribution in edge multi-client multicast.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a multicast file distribution method based on intelligent compression, comprising: The server receives file distribution access requests from clients within the edge LAN, establishes signaling connections, and assigns multicast group addresses to verified clients. Under these multicast group addresses, it performs receive bearer replies, forming an edge multicast bearer chain. The server retrieves the folder to be distributed through this chain, performs multi-receiver block pre-occupancy orchestration on the folder, generates data block placeholders, and performs hit-based distribution and filtering of these placeholders according to cache hit status, forming a folder block expansion chain. Multicast pressure is layered and triggered on this expansion chain, grouping consecutive small files that commonly miss in the expansion chain into candidate small file clusters. These candidate small file clusters are then subjected to benefit filtering and folding compression to generate clusters. The system combines compressed fragments and connects them with large file links marked as retaining original blocks to generate a smart compressed folded chain. Based on this chain, the system performs directory-anchored reassembly of the compressed fragments to generate compressed sending links. It then binds the large file links marked as retaining original blocks to their original block paths to obtain retained sending links. Finally, it mixes and orchestrates the compressed and retained sending links to generate a hybrid multicast object sequence. The system performs multicast backfeedback on the hybrid multicast object sequence, while simultaneously obtaining the locations of missing links reported by clients. Links missing by multiple clients are prioritized for backfeedback, while missing links from individual clients are subsequently resent, resulting in a multicast file distribution result.

[0007] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the establishment of the signaling connection is specifically as follows: The server receives file distribution access requests sent by clients at the file distribution access entry point on the edge local area network, and performs access request purification and queuing according to the uniqueness of the source and the integrity of the receiving entry point to form an access request queue; According to the order of the access request queue, the signaling handshake content is sent to the client in the edge LAN. The connection response and cache hit status returned by the client in the edge LAN are received. The connection response is closed and attached according to the client identifier consistency and connection response validity to form a signaling connection chain.

[0008] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the formation of the edge multicast bearer chain is specifically as follows: The receiving bearer is verified using the client in the edge LAN in the signaling connection chain as the verification object, and the multicast group address is sent to the client in the edge LAN that is retained in the multicast access range. The multicast group address is written into the corresponding signaling connection location to form a multicast group address allocation table. According to the multicast group address and signaling connection position in the multicast group address allocation table, the receiving bearer aggregation and back-attachment are performed on the clients in the edge LAN. The number of clients in the edge LAN after multicast reception preparation, the number of multicast receivers and the cache hit status are written into the same bearer chain position to form the edge multicast bearer chain.

[0009] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the placeholder for the generated data block is specifically as follows: Extract the bearer link bit, the number of multicast receivers and the cache hit status from the edge multicast bearer chain, perform bearer chain traction retrieval and folder entry attachment on the folder to be distributed, and connect the number of multicast receivers to the transmission bearer position of the folder to be distributed to form the folder bearer bit; The folders to be distributed in the folder carrier position are sorted according to their original directory order to form a file path order chain. The file path order chain is then pre-occupied and arranged, and the file path, file size and number of file blocks are attached to the same directory chain position to form a directory chain position pre-occupancy table. Based on the directory chain bit pre-occupancy table, the receiver block volume is amplified and arranged and multicast block bits are taken over. The receiver demand bits are formed according to the number of file blocks and the number of missing clients. Then, the small file directory chain bits that are adjacent in the original folder directory order are connected to the same continuous bit interval to generate data block bits.

[0010] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the formation of the folder block expansion chain is specifically as follows: The file paths in the data block placeholders are matched with the cache hit status one by one, and the data block placeholders are split into hit difference streams to form the no-transmission chain, the chain to be sent, and the difference chain, respectively. The hit chain is then collected to form the hit split placeholders. Based on the hit-based traffic splitting placeholders, small files are continuously expanded and clustered into the to-send chain position and the difference chain position according to the original folder directory order to form an expanded chain position. For file data block placeholders that need to maintain independent resending granularity, large file boundary preservation is performed to form a reserved chain position. These are then merged to form a folder block expansion chain.

[0011] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the generation of the intelligent compression fold chain is specifically as follows: Based on the expansion chain bit and the reserved chain bit in the folder block expansion chain, multicast pressure mapping is performed on the number of multicast receivers and the cache hit status carried by the expansion chain bit and the reserved chain bit. The expansion chain bit is mapped to the compression pressure chain bit, and the reserved chain bit is mapped to the reserved pressure chain bit. Then, the multicast pressure layered chain is formed by hierarchical triggering based on the compression pressure chain bit and the reserved pressure chain bit. Based on the cache hit status and original folder directory order in the multicast pressure hierarchical chain, consecutive small files are matched together for common misses. Small files that have not been cached by clients in the edge LAN within the same consecutive directory range are continuated to the same cluster position, and small files with cache hit divergences are retained in the range of different chain positions to form candidate small file clusters. The candidate small file clusters are aligned with the entry occupancy position to generate a multicast entry saving chain. Based on the multicast entry saving chain, the candidate small file clusters with an aggregated sending number of entries less than the number of per-file sending entries are continued to the folded candidate position. The candidate small file clusters with an aggregated sending number of entries not less than the number of per-file sending entries are retained as per-file reserved chain positions according to the original folder directory order, forming folded candidate small file clusters. Based on the number of small files in the folded candidate small file cluster and the original folder directory order, the file-by-file sending entries in the folded candidate small file cluster are continuously connected to the aggregation sending position according to the directory adjacency relationship to obtain the aggregation sending entries. The contents of the small files carried by the aggregation sending entries are merged and compressed into an aggregation compressed fragment according to the original folder directory order. Based on the aggregated compressed fragments, the file-by-file reserved chain bits, and the reserved pressure chain bits in the multicast pressure layered chain, the file-by-file reserved chain bits and the reserved chain bits inherited by the reserved pressure chain bits are marked and pasted back to the original block to generate the original block reserved chain bits. Then, the aggregated compressed fragments and the original block reserved chain bits are folded and connected according to the original folder directory order to generate a smart compressed folded chain.

[0012] As a preferred embodiment of the intelligent compression-based multicast file distribution method of the present invention, the generation of the hybrid multicast object sequence is as follows: Using the intelligent compression and folding chain as the disassembly object, the chain position content is located item by item according to the folding chain position sequence. The position of the aggregated compression fragment and the position of the original block retained chain position are distinguished according to the content carried by the chain position. The aggregated compression fragment is extracted from the position of the aggregated compression fragment and connected to the bearing and reorganization range. The original block retained chain position is extracted from the position of the original block retained chain position and connected to the original block binding range, thus forming the folding and reorganization entry point. Based on the reorganization range in the folding and reorganization entry, the directory anchor point is extracted from the aggregated and compressed fragment to obtain the corresponding fragment starting directory position and the file order within the fragment, forming a fragment anchor point. The fragment anchor point, the file order within the fragment, and the restored directory position are then bound to form a directory anchored fragment. The directory anchoring fragment restores the directory position according to the fragment anchor point alignment, pastes the files in the fragment back to the compressed bearer position corresponding to the original folder directory order, and connects the aggregated compressed fragment to the same compressed bearer position according to the fragment anchor point and the original folder directory order to obtain the compressed bearer fragment. The fragment anchor point and the compressed bearer fragment are bound to the same sending chain position to generate a compressed sending chain position. Based on the original block binding range in the folding and reorganization entry, the file path and data block placeholders of the original block reserved chain are anchored to the original block path, the file path is mapped to the reserved sending position in the original folder directory order, and the original block path is bound around the reserved sending position. The data block placeholder and the original block reserved mark are connected to the same reserved sending position to obtain the reserved sending chain. According to the order of the folded chain positions in the intelligent compression folded chain, the compressed sending chain position and the reserved sending chain position are folded in a folding order and then pasted back to the corresponding folded chain position. The chain position succession relationship after pasting is then arranged in a hybrid multicast to generate a hybrid multicast object sequence.

[0013] As a preferred embodiment of the multicast file distribution method based on intelligent compression described in this invention, the generation of multicast file distribution results is as follows: The server sends compressed send chain bits and reserved send chain bits according to the multicast send sequence position in the hybrid multicast object sequence. The client receives the compressed send chain bits and reserved send chain bits, restores the aggregated compressed fragment according to the fragment anchor point, and concatenates the original block content according to the original block reserved mark to form a multicast backfeed restoration chain. Receive chain bit reception feedback returned by clients within the edge LAN, re-attach missing chain bits to the chain bit reception feedback according to the chain bit order in the multicast re-feedback and restoration chain, and mark the missing chain bit positions according to the number of clients from which the feedback comes, forming a missing chain bit feedback table. Based on the common missing label in the missing link feedback table, the common missing compressed transmission link and reserved transmission link are pasted back to the multicast transmission sequence position, and common missing priority is injected according to the multicast transmission sequence position and multicast group address after the injection, forming a common missing injection chain; Based on the single-end missing label in the common missing feedback chain and missing chain bit feedback table, the missing chain bit content of the client in a single edge LAN is resent and connected in a single-end missing direction. The resent compressed sending chain bit and reserved sending chain bit are then closed and solidified in the directory reception, generating multicast file distribution results.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the multicast file distribution method based on intelligent compression as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the multicast file distribution method based on intelligent compression as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by triggering multicast pressure layering on the folder block expansion chain, consecutive small files that have missed a common hit are merged and their benefits are screened. Small file clusters with entry savings are folded and compressed, while retaining the original block boundaries and independent resending granularity of large files. This achieves co-chain orchestration of small file aggregation compression and large file retention distribution, reduces the multicast scheduling burden and missing resending range in edge data transmission, and improves the continuity of client file distribution and the utilization rate of transmission resources. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a multicast file distribution method based on intelligent compression.

[0019] Figure 2 A flowchart for generating data block placeholders and folder block expansion chains.

[0020] Figure 3 A flowchart for the formation of an intelligent compression folding chain.

[0021] Figure 4 A flowchart for forming a sequence of hybrid multicast objects. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a multicast file distribution method based on intelligent compression, including the following steps: S1: The server receives file distribution access requests sent by clients within the edge LAN, establishes a signaling connection, and assigns multicast group addresses to verified clients. Under the multicast group address, it performs receive bearer replies to form an edge multicast bearer chain.

[0026] S1.1: The server receives file distribution access requests sent by clients at the file distribution access entry point on the edge LAN, and performs access request purification and grouping according to the uniqueness of the source and the integrity of the receiving entry point to form an access request queue.

[0027] A file distribution access request is an access request from a client within an edge LAN to the server indicating that it wants to join the current file distribution process, and carrying the client's identifier and access information.

[0028] Furthermore, the server receives file distribution access requests at the file distribution access entry point of the edge LAN, performs source uniqueness comparison on the file distribution access requests according to the client identifier and the receiving entry information, classifies the file distribution access requests with the same client identifier and the same receiving entry information into the same source position, retains the file distribution access requests with the earlier receiving time, removes the remaining duplicate file distribution access requests from the access range, and writes the retained file distribution access requests according to the receiving time into the source order.

[0029] The server will match the receiving entry information in the file distribution access request with the client identifier one by one after completing the source uniqueness comparison, and confirm that the receiving entry information can point to the valid receiving location of the client in the edge LAN. File distribution access requests with missing receiving entry information or whose receiving entry information cannot match the client identifier will be removed from the access range. Then, the remaining file distribution access requests will be sequentially connected to the same queue position according to the receiving time and source order to form an access request queue.

[0030] S1.2: Send signaling handshake content to the client in the edge LAN according to the order in the access request queue, receive the connection response and cache hit status returned by the client in the edge LAN, and perform signaling response closure hooking on the connection response according to the client identifier consistency and connection response validity to form a signaling connection chain.

[0031] The signaling handshake content is the connection establishment content sent by the server to the client in the edge LAN according to the access request queue. It is used to indicate the correspondence between the signaling connection entry and the client identifier, and to require the client in the edge LAN to return the connection response and cache hit status.

[0032] Furthermore, the server extracts the client identifier and receiving entry information according to the order in the access request queue, takes over the file path to be verified within the current file distribution range and assigns it to the corresponding signaling handshake content, and sends the signaling handshake content to the client in the edge LAN pointed to by the receiving entry information in sequence, maintaining the correspondence between the signaling handshake content, client identifier, file path to be verified and access request queue position.

[0033] Within the edge LAN, the client receives the signaling handshake content and locates the local cached content item by item according to the file path to be verified. If the local cached content matching the file path can be located, the corresponding file path to be verified is written to the cache hit status. If the local cached content cannot be located at the file path to be verified, or if the located local cached content is inconsistent with the content to be distributed, the corresponding file path to be verified is written to the cache miss status. Then, the client returns a connection response and cache hit status through the connection entry corresponding to the receiving entry information. The server receives the connection response and the cache hit status arranged according to the file path to be verified and places them in the corresponding access request queue.

[0034] The server compares the client identifier in the connection response with the client identifier in the corresponding access request queue position item by item, and confirms that the connection response can completely point to the connection entry corresponding to the signaling handshake content. It binds the client in the edge LAN with the same client identifier and a valid connection response, the connection response, the file path to be verified, and the cache hit status to the same signaling connection position, so that each file path to be verified and the corresponding cache hit status are matched item by item. The client in the edge LAN with inconsistent client identifiers and invalid connection responses is removed from the signaling connection range, and then the signaling connection positions are continued according to the order in the access request queue to form a signaling connection chain.

[0035] It should be noted that client identifier consistency means that the client identifier in the connection response is the same as the client identifier in the corresponding position of the access request queue, and both point to the same client within the same edge LAN.

[0036] Connection response validity refers to the connection response being able to correspond to the sent signaling handshake content and receiving entry information, and fully carry the response status required to establish a signaling connection.

[0037] S1.3: Use the client in the edge LAN in the signaling connection chain as the verification object to perform receiving bearer verification, and send the multicast group address to the client in the edge LAN that is retained in the multicast access range. Write the multicast group address into the corresponding signaling connection position to form a multicast group address allocation table.

[0038] A multicast group address is a network multicast identifier assigned by the server for the same file distribution range. It is used to simultaneously deliver the file content sent by the server to multiple edge LAN clients that have joined the corresponding multicast group.

[0039] The server extracts the client identifier from the connection response item by item along the signaling connection chain, searches for the signaling connection location with the same client identifier in the signaling connection chain, and writes the connection response to the found signaling connection location. If the client identifier in the connection response cannot be found in the signaling connection chain, the client in the corresponding edge LAN is removed from the multicast access range. If the client identifier in the connection response can point to a unique signaling connection location and the connection response can accept multicast reception, the client in the corresponding edge LAN is kept in the multicast access range.

[0040] The server locates the corresponding signaling connection position according to the client identifier retained within the multicast access range, writes the multicast group address allocated to the same file distribution range into the corresponding signaling connection position, and saves the multicast group address and the client identifier already recorded in the signaling connection position in parallel, so that the client identifier, signaling connection position and multicast group address point to the client in the same edge LAN. Then, it connects each signaling connection position according to the order in the signaling connection chain to form a multicast group address allocation table.

[0041] It should be noted that the signaling connection position is the chain position in the signaling connection chain used to receive the client identifier, connection response and cache hit status of the client within the same edge LAN, and is also the write position used to establish an association between the multicast group address and the corresponding client within the edge LAN.

[0042] S1.4: According to the multicast group address and signaling connection position in the multicast group address allocation table, perform receive bearer aggregation and back-attachment for the clients in the edge LAN, and write the number of clients in the edge LAN after multicast reception preparation, the number of multicast receivers and the cache hit status into the same bearer chain position to form an edge multicast bearer chain.

[0043] Furthermore, following the order in the multicast group address allocation table, the multicast group address and signaling connection location are extracted item by item. The return content after the client in the edge LAN completes the multicast group address access is received through the signaling connection location. The client identifier in the return content is compared item by item with the client identifier recorded in the signaling connection location. If the client identifier matches and the client in the edge LAN can receive the file content through the multicast group address, the client in the edge LAN is retained under the corresponding multicast group address.

[0044] Using the multicast group address as the basis for bearer affiliation and the signaling connection position as the reattachment position, clients within the edge LAN that have completed multicast reception preparation under the same multicast group address are assigned to the same bearer chain position. The cache hit status returned by each client within the edge LAN for each file path to be verified is then attached to the corresponding bearer chain position, ensuring that the client, client identifier, file path to be verified, signaling connection position, and cache hit status within the edge LAN correspond item by item in the bearer chain position. The server counts the number of clients within the edge LAN assigned to the bearer chain position under the same multicast group address, forming the number of multicast receivers. The number of multicast receivers is then attached to the corresponding bearer chain position, and each bearer chain position is then connected according to the multicast group address order in the multicast group address allocation table to complete the receiver bearer aggregation and reattachment, forming the edge multicast bearer chain.

[0045] S2: Retrieve the folder to be distributed through the edge multicast bearer chain, pre-arrange the block quantity of the folder to be distributed for multiple receiving ends, generate data block placeholders, and perform hit distribution and screening of the data block placeholders according to the cache hit status to form a folder block quantity expansion chain.

[0046] S2.1: Extract the bearer link bit, the number of multicast receivers and the cache hit status from the edge multicast bearer chain, perform bearer chain traction retrieval and folder entry attachment on the folder to be distributed, and connect the number of multicast receivers to the transmission bearer position of the folder to be distributed to form the folder bearer bit.

[0047] The folder to be distributed is a folder object selected by the server and prepared to be distributed to clients in the edge LAN via multicast group address. The internal files are arranged in the original folder directory order, and the correspondence between file path, file size and number of file blocks is preserved.

[0048] Furthermore, the corresponding file distribution range is located item by item along the bearer chain position in the edge multicast bearer chain. The number of multicast receivers and the cache hit status of the bearer chain position are extracted. The bearer chain position is used as the basis for retrieving the folder to be distributed. The folders to be distributed corresponding to the same file distribution range are connected to the corresponding bearer chain position, so that the folders to be distributed, the number of multicast receivers and the cache hit status are kept in the same position in the bearer chain position.

[0049] Locate the starting directory of the folder to be distributed, use the starting directory as the folder entry point to connect to the transmission bearer position corresponding to the bearer chain, attach the number of multicast receivers to the transmission bearer position, and attach the cache hit status to the corresponding position of the folder to be distributed, so that the bearer chain, the folder to be distributed, the number of multicast receivers, and the cache hit status all point to the same transmission bearer position, forming the folder bearer position.

[0050] S2.2: Collect the folders to be distributed in the folder carrier position according to their original directory order to form a file path order chain, and pre-arrange the file path order chain to pre-occupy the chain position, and attach the file path, file size and file block number to the same directory chain position to form a directory chain position pre-occupancy table.

[0051] Furthermore, the starting directory of the folder to be distributed is located from the folder carrier position, and the file path is expanded sequentially along the directory hierarchy of the folder to be distributed. The file paths within the same directory level are connected in the order of the folder to be distributed, and the hierarchical paths between different directory levels are kept in the corresponding carrier positions. The continuous relationship between adjacent file paths in the original folder directory order is also preserved, forming a file path sequence chain.

[0052] Directory chain positions are set sequentially along the file path sequence chain, with each file path assigned to a unique corresponding directory chain position. Then, the file content is divided into consecutive file blocks according to the file size and the multicast block division rules already in place on the server. The number of file blocks corresponding to each file path is counted, and the file size and the number of file blocks are assigned to the directory chain position that records the same file path. This ensures that the file path, file size, and number of file blocks maintain a one-to-one correspondence in the directory chain positions. The directory chain positions are then connected according to the order in the file path sequence chain to form a directory chain position pre-occupancy table.

[0053] It should be noted that the multicast chunking rule is that before sending the file multicast, the server divides the same file content into consecutive file blocks according to a uniform block length, configures a corresponding block number for each file block, and maintains the correspondence between the file path, file block position and block number.

[0054] S2.3: Based on the directory chain bit pre-occupancy table, the receiver block volume is amplified and arranged and multicast block bits are taken over. The receiving demand bits are formed according to the number of file blocks and the number of missing clients. Then, the small file directory chain bits that are adjacent in the original folder directory order are connected to the same continuous bit interval to generate data block bits.

[0055] Furthermore, following the file path sequence in the data block placeholders, the file paths attached to the data block placeholders are extracted item by item, and the same file paths to be verified are searched in the edge multicast bearer chain according to the file paths. Then, the cache hit status returned by each client in the edge LAN for the same file path to be verified is extracted according to the client identifier. The file path, client identifier, cache hit status and corresponding receiving requirement placeholders are attached to the same splitting position, so that each file path can correspond to the cache hit status of each client in the edge LAN.

[0056] When all clients within the same edge LAN corresponding to the same file path have cached the corresponding file content, the corresponding data block placeholder is assigned to the no-transmission chain position; when none of the clients within the same edge LAN corresponding to the same file path have cached the corresponding file content, the corresponding data block placeholder is assigned to the pending transmission chain position; when the same file path corresponds to both cache hit and cache miss states, the clients within the edge LAN corresponding to the cache miss state and their corresponding receiving request placeholders are assigned to the difference chain position, and the clients within the edge LAN corresponding to the cache hit state are excluded from the continued transmission range of the corresponding file path.

[0057] Following the original folder directory order, locate the directory location corresponding to each file path item by item. Then, use the file paths attached to the no-transmission link, the to-be-sent link, and the difference link to find the same directory location. Connect the no-transmission link to the directory location of the file path that does not need to be sent. Connect the to-be-sent link to the directory location of the file path that all clients in the edge LAN need to receive. Connect the difference link to the directory location of the file path that some clients in the edge LAN need to receive. Then, attach the corresponding data block placeholder and the receiving requirement placeholder to the same directory location to complete the collection of hit link positions and form hit distribution placeholders.

[0058] It should be noted that the number of clients that missed a cache hit refers to the total number of edge LAN clients that did not cache the corresponding file content and need to continue receiving the corresponding file blocks in the cache hit state for the same file path.

[0059] S2.4: Match the file paths in the data block placeholders with the cache hit status item by item, and perform hit difference splitting on the data block placeholders to form the no-transmission chain, the chain to be sent, and the difference chain, respectively. Then, collect the hit chain bits to form the hit split placeholders.

[0060] Furthermore, following the file path order in the data block placeholders, the cache hit status corresponding to each file path is located, and the data block placeholders are taken over item by item according to the caching status of the same file path by the clients in the edge LAN. When all the clients in the edge LAN corresponding to the same file path have cached the file content, the corresponding data block placeholder is assigned to the no-transmission chain. When none of the clients in the edge LAN corresponding to the same file path have cached the file content, the corresponding data block placeholder is assigned to the pending transmission chain. When the same file path corresponds to both cache hit and cache miss status, the receiving demand placeholders corresponding to the clients in the edge LAN that need to continue receiving file content are assigned to the difference chain.

[0061] Following the original folder directory order, locate the directory location corresponding to each file path. Then, use the file paths attached to the no-transmission link, the to-be-sent link, and the difference link to find the same directory location. Connect the no-transmission link to the directory location of the file path that does not need to be sent, connect the to-be-sent link to the directory location of the file path that all clients in the edge LAN need to receive, and connect the difference link to the directory location of the file path that some clients in the edge LAN need to receive. Finally, attach the corresponding data block placeholder to the same directory location to complete the collection of hit link positions and form hit distribution placeholders.

[0062] S2.5: Based on the hit-based traffic splitting placeholder, small files are continuously expanded and aggregated into the to-send chain position and the difference chain position according to the original folder directory order to form an expanded chain position. Large file boundary retention is performed on the placeholder of file data blocks that need to maintain independent resending granularity to form a retained chain position. The two are then merged to form a folder block expansion chain.

[0063] Furthermore, the sendable chain and difference chain are extracted from the hit splitting placeholders. The data block placeholders of adjacent small files are located along the original folder directory order. The data block placeholder at the end of the previous small file is sequentially connected with the data block placeholder at the beginning of the next small file, and the file path and file block order corresponding to each small file are preserved. The server takes over the receiving request placeholders corresponding to each small file in sequence according to the number of clients that missed the hit. The receiving request placeholders of adjacent small files are arranged continuously into the same continuous placeholder interval. Then, the continuous placeholder interval is connected to the corresponding sendable chain and difference chain to form an expanded chain.

[0064] From the hit distribution placeholders, locate the large file data block placeholders that need to maintain independent resending granularity. Mark the start block position and end block position of each large file data block placeholder. Continuously connect the file data block placeholders between the start block position and the end block position according to the file block order, and retain the block position boundary between the large file data block placeholder and the adjacent expansion chain position. The server connects the large file data block placeholders with the boundary preservation to the corresponding original folder directory position to form the reserved chain position. Then, it connects the expansion chain position and the reserved chain position according to the original folder directory order to form the folder block expansion chain.

[0065] It should be noted that independent resending granularity means that the missing range and resending range of a large file are determined separately according to its own file block boundaries. The server can resend only the missing file blocks in the large file without merging and resending the corresponding data blocks of adjacent small files.

[0066] S3: Multicast pressure layering is triggered on the folder block expansion chain. Consecutive small files that are not hit in the expansion chain are merged into candidate small file clusters. The candidate small file clusters are subjected to revenue screening and folding compression to generate aggregated compressed fragments. Then, the large file chain with the original block retention mark is continued to generate a smart compressed folding chain.

[0067] S3.1: Based on the expansion chain bit and the reserved chain bit in the folder block expansion chain, perform multicast pressure mapping on the number of multicast receivers and the cache hit status carried by the expansion chain bit and the reserved chain bit, map the expansion chain bit to the compression pressure chain bit, map the reserved chain bit to the reserved pressure chain bit, and perform hierarchical triggering and traction based on the compression pressure chain bit and the reserved pressure chain bit to form a multicast pressure hierarchical chain.

[0068] Furthermore, the expansion chain position and the reserved chain position are located item by item along the original folder directory order in the folder block expansion chain, and the number of multicast receivers and the cache hit status of each expansion chain position and the reserved chain position are extracted. For the expansion chain position, the server distinguishes between cache hit status and cache miss status according to the client identifier, and takes over the receiving demand placeholder corresponding to the miss client and the continuous small file data block placeholder in the expansion chain position to the same chain position, and arranges the continuous small file data block placeholder according to the coverage of the miss client, and connects the expansion chain position to the compression pressure chain position; for the reserved chain position, the server extracts the independent retransmission range along the start block position and end block position of the large file data block placeholder, takes over the receiving demand placeholder corresponding to the cache miss status to the large file data block placeholder, and preserves the block position boundary of the large file data block placeholder, and connects the reserved chain position to the reserved pressure chain position.

[0069] Following the original folder directory order, the compressed pressure chain and the reserved pressure chain are connected. The compressed pressure chain is connected to the common miss alignment position, and the reserved pressure chain is connected to the original block mark re-paste position. The directory switching positions between adjacent chains are also retained to form a multicast pressure layered chain.

[0070] It should be noted that multicast pressure layering is based on the formation method of the receiving demand occupancy corresponding to the chain position, the file organization method, and the retransmission granularity. It divides the folder block expansion chain into compressed pressure chain positions corresponding to consecutive small files and reserved pressure chain positions corresponding to independent retransmission files. It does not mean that the compressed pressure chain positions and reserved pressure chain positions are sorted by numerical size.

[0071] The total number of receive requests is the logical number of receive requests formed by expanding the number of file blocks in the corresponding chain according to the number of missed clients. It is used to represent the multicast carrying requirements generated when the same file content is missed by multiple receivers. It does not mean that the server needs to generate the same file block content repeatedly for each missed client.

[0072] S3.2: According to the cache hit status and original folder directory order in the multicast pressure hierarchical chain, perform common miss pairing for consecutive small files, connect small files that have not been cached by clients in the edge LAN within the same consecutive directory range to the same cluster position, and retain small files with cache hit divergences within the difference chain position range to form candidate small file clusters.

[0073] Furthermore, along the compression pressure chain position in the multicast pressure hierarchy chain, the cache hit status corresponding to consecutive small files is extracted item by item according to the original folder directory order, and the edge LAN clients that have not cached the corresponding file content in each small file and need to continue receiving the corresponding file block are listed; when the edge LAN clients corresponding to adjacent small files are completely identical and the file paths are continuous, the adjacent small files are continued; when the edge LAN clients are added or removed or the file path is interrupted, the continuation is stopped at the corresponding position.

[0074] Starting from each continuation start position, consecutive small files that have both missed cache hits are moved to the same cluster position according to the file path and file block order; small files with cache hit discrepancies are kept in the difference chain position and are no longer moved to the current cluster position. Then, the cluster positions are arranged according to the original folder directory order to form candidate small file clusters.

[0075] It should be noted that the difference chain range is the consecutive chain range occupied by the data blocks that are hit in the hit splitting placeholder and that are not hit in the client cache of some edge LANs under the same file path.

[0076] S3.3: Perform entry occupancy alignment on candidate small file clusters, generate multicast entry saving chains, and based on the multicast entry saving chains, continue candidate small file clusters with an aggregated sending number of entries less than the number of per-file sending entries to the folded candidate position, and retain candidate small file clusters with an aggregated sending number of entries not less than the number of per-file sending entries as per-file reserved chain positions according to the original folder directory order, forming folded candidate small file clusters.

[0077] Furthermore, the server sets up a sending entry point and multicast scheduling order for each small file according to the file path and original folder directory order in the candidate small file cluster, and counts the number of file-by-file sending entries corresponding to each candidate small file cluster; the server then pre-schedules consecutive small files in the same candidate small file cluster to the same aggregate sending entry point, counts the number of aggregate sending entries, and counts the number of missed clients corresponding to all small files in the candidate small file cluster from the cache hit status received by the candidate small file cluster, while extracting the number of multicast receivers from the corresponding bearer chain position of the edge multicast bearer chain.

[0078] The expression for calculating the multicast folding revenue exponent is as follows: ; In the formula, Indicates the first The multicast folding benefit index corresponds to each candidate small file cluster. The multicast folding benefit index is used to characterize the degree of reduction in multicast scheduling entries and the degree of coverage of common missed clients. Indicates the first The number of file-by-file sending entries formed when each small file in a candidate small file cluster occupies a sending entry point is obtained by counting the file-by-file sending entries that have been set in the candidate small file cluster. Indicates the first The number of aggregated sending entries formed after consecutive small files within a candidate small file cluster are pre-arranged to the same aggregated sending entry is obtained by counting the multicast scheduling entries received by the aggregated sending entry. Indicates the first The number of missing clients corresponding to all small files in a candidate small file cluster is obtained by comparing the edge LAN clients corresponding to each small file in the candidate small file cluster. Indicates the first The number of multicast receivers under the multicast group address to which each candidate small file cluster belongs is extracted from the corresponding bearer chain bit of the edge multicast bearer chain.

[0079] The server assigns the number of file-by-file sending entries, the number of aggregated sending entries, and the multicast folding benefit index to the corresponding candidate small file cluster positions, and continues the candidate small file cluster positions according to the original folder directory order to generate a multicast entry saving chain. When the multicast folding benefit index is greater than zero, the sending entry occupied by each small file in the candidate small file cluster is canceled, and the file path, file block order, and clients in the edge LAN that are jointly missed are continued to the same folding candidate position. When the multicast folding benefit index is equal to zero, the corresponding aggregated sending entry is canceled, the small files in the candidate small file cluster are returned to the difference chain position range according to the original folder directory order, and the receiving demand occupancy is restored to the corresponding file path to form a folded candidate small file cluster.

[0080] It should be noted that the above expression incorporates the reduction between the number of file-by-file transmission entries and the number of aggregate transmission entries, as well as the coverage of the number of jointly missed clients in the number of multicast receivers, into the calculation. A harmonic coupling approach is used to limit false triggering caused by any single factor being excessively high. When aggregate transmission significantly reduces multicast scheduling entries but there are few jointly missed clients, the multicast folding benefit index will not be excessively amplified by the saving effect of a single entry. When there are many jointly missed clients but aggregate transmission does not reduce multicast scheduling entries, the multicast folding benefit index will also not generate effective benefits. When the number of aggregate transmission entries is not less than the number of file-by-file transmission entries, the multicast folding benefit index is directly recorded as zero. The multicast folding benefit index is used to filter out small file clusters from candidate small file clusters that simultaneously meet the needs of saving multicast scheduling entries and multiple clients receiving the same data. This provides a basis for the continuation of folding candidate positions and avoids triggering unprofitable compression solely based on the number of small files or the cache miss status.

[0081] The number of aggregated sending entries is the total number of multicast scheduling entries required for the server to complete the multicast sending of the corresponding candidate small file cluster after merging consecutive small files in the same candidate small file cluster into the aggregated sending entry.

[0082] S3.4: Based on the number of small files in the folded candidate small file cluster and the original folder directory order, the file-by-file sending entries in the folded candidate small file cluster are continuously connected to the aggregation sending position according to the directory adjacency relationship to obtain the aggregation sending entry. The contents of the small files carried by the aggregation sending entry are merged and compressed into an aggregation compressed fragment according to the original folder directory order.

[0083] Furthermore, file-by-file sending entries are extracted sequentially along the original folder directory order in the folded candidate small file cluster. The number of small files, file path, and file block corresponding to each file-by-file sending entry are sequentially connected to the same folded candidate position. The sending entry points occupied by each small file are cancelled. Then, the starting file block position of the next file-by-file sending entry is connected to the end file block position of the previous file-by-file sending entry. The file separator position is retained at the connection position of adjacent small files. The consecutive file-by-file sending entries are connected to the same sending entry point to obtain the aggregated sending entry.

[0084] The content of small files is extracted sequentially along the file path in the aggregated sending entry. The content of each small file is arranged continuously according to the file block order, and lossless compression is performed from the starting file position to the ending file position of the aggregated sending entry. During the compression process, the server retains the file separator position, attaches the file path and file block order to the corresponding compression position, and then attaches the compressed small file content, file separator position, file path and file block order together to the same fragment position to generate an aggregated compressed fragment.

[0085] S3.5: Based on the aggregated compressed fragments, file-by-file reserved chain bits, and reserved pressure chain bits in the multicast pressure layered chain, the file-by-file reserved chain bits and the reserved chain bits inherited by the reserved pressure chain bits are marked and pasted back to the original block to generate the original block reserved chain bits. The aggregated compressed fragments and original block reserved chain bits are folded and connected according to the original folder directory order to generate a smart compressed folded chain.

[0086] Furthermore, the reserved chain position is located item by item along the reserved pressure chain position in the multicast pressure layering chain. The file path, start block position, end block position and original folder directory position corresponding to the large file undertaken by the reserved chain position are extracted. The original block reserved mark is attached to the chain position boundary corresponding to the start block position and the end block position respectively, and the original block order is maintained along the file block order between the start block position and the end block position. The server attaches the original block reserved mark, file path and original folder directory position to the same reserved chain position, blocks the block position merging between the reserved chain position and the adjacent expansion chain position, and generates the original block reserved chain position.

[0087] Following the original folder directory order, locate the directory positions corresponding to the aggregated compressed fragment and the original block retained chain position item by item. Connect the aggregated compressed fragment to the original directory position of the consecutive small files, and connect the original block retained chain position to the original directory position of the corresponding large file. When the current directory position corresponds to the aggregated compressed fragment and the next directory position corresponds to the original block retained chain position, connect the end of the aggregated compressed fragment to the starting block position of the original block retained chain position. When the current directory position corresponds to the original block retained chain position and the next directory position corresponds to the aggregated compressed fragment, connect the ending block position of the original block retained chain position to the starting position of the aggregated compressed fragment. The boundaries of the aggregated compressed fragment and the original block retained chain position are preserved respectively to generate a smart compressed fold chain.

[0088] S4: Based on the intelligent compression folding chain, the aggregated compressed fragments are reorganized by directory anchoring to generate compressed sending chain bits. The original block path is bound to the chain bits of large files with original block retention marks to obtain the retained sending chain bits. The compressed sending chain bits and the retained sending chain bits are mixed and arranged to generate a mixed multicast object sequence.

[0089] S4.1: Using the intelligent compression folding chain as the disassembly object, locate the chain position content item by item according to the folding chain position sequence, distinguish the position of the aggregated compression fragment and the position of the original block retained chain position according to the content carried by the chain position, extract the aggregated compression fragment from the position of the aggregated compression fragment and connect it to the bearing recombination range, extract the original block retained chain position from the position of the original block and connect it to the original block binding range, thus forming the folding recombination entry.

[0090] Furthermore, the aggregated compressed fragment and the original block retention chain position are located sequentially along the folded chain position in the intelligent compression folded chain. The boundaries of the aggregated compressed fragment and the original block retention chain position are used as the disassembly positions, and the aggregated compressed fragment and the original block retention chain position are separated from the adjacent chain positions respectively. The server connects the aggregated compressed fragment to the corresponding carrier reorganization range, and takes over the file path, file block order and file separation position in the aggregated compressed fragment to the same chain position interval. Then, the original block retention chain position is connected to the corresponding original block binding range, and takes over the file path, original folder directory position, start block position, end block position and original block retention mark in the original block retention chain position to the same chain position interval. The carrier reorganization range and the original block binding range are arranged according to the original folder directory position in the intelligent compression folded chain to form the folded reorganization entry.

[0091] It should be noted that the scope of the recombination is the continuous chain interval occupied by the aggregated compressed fragment from the starting file position to the ending file position, which is used to carry the file path, file block order and file delimiter position in the aggregated compressed fragment.

[0092] The original block binding range is the continuous range of chain bits occupied by the original block reserved chain bits from the start block bit to the end block bit bit. It is used to take over the file path, original folder directory location and original block reserved mark in the original block reserved chain bits.

[0093] S4.2: Based on the reorganization range in the folding and reorganization entry, extract the directory anchor point of the aggregated compressed segment, obtain the corresponding segment starting directory position and file order within the segment, form the segment anchor point, and bind the segment anchor point, file order within the segment and the restored directory position to form the directory anchored segment.

[0094] Furthermore, the server locates the file delimiter position item by item along the carrying reorganization range in the folding and reorganization entry point, from the starting chain position of the aggregated compressed fragment to the ending chain position. It determines the starting directory position of the fragment using the file path corresponding to the starting file, and then arranges the file paths and file block order corresponding to each small file according to the order of the file delimiter positions to form the file order within the fragment. The server connects the starting directory position of the fragment and the file order within the fragment to the same chain position to form the fragment anchor point.

[0095] The server locates the corresponding restoration directory for each small file according to the position of each file path in the original folder directory order in the aggregated compressed fragment. Then, it connects each small file with its corresponding restoration directory position according to the file order within the fragment. Finally, it binds the fragment anchor point, the file order within the fragment, and the restoration directory position to the same aggregated compressed fragment within the scope of the reorganization, forming a directory anchored fragment.

[0096] S4.3: The directory anchoring fragment restores the directory position according to the fragment anchor point alignment, pastes the file order in the fragment back to the compressed bearer position corresponding to the original folder directory order, and connects the aggregated compressed fragment to the same compressed bearer position according to the fragment anchor point and the original folder directory order to obtain the compressed bearer fragment. The fragment anchor point and the compressed bearer fragment are bound to the same sending chain position to generate a compressed sending chain position.

[0097] Furthermore, based on the directory anchoring segment, extract the segment anchor point, file order within the segment, restored directory location, and aggregated compressed segment. Using the segment starting directory location corresponding to the segment anchor point as the starting point of the compressed carrier, connect the aggregated compressed segment to the same compressed carrier location, and arrange the file separator positions according to the file order within the segment, so that the compressed content corresponding to each small file corresponds item by item with the restored directory location in the original folder directory order, thus obtaining the compressed carrier segment.

[0098] The compressed carrier segment is connected to the send chain position, the segment anchor point is connected to the beginning of the send chain position, the file order and the restored directory position within the segment are connected to the corresponding file separator positions respectively, and then the send chain position is connected to the fold chain position originally corresponding to the aggregated compressed segment in the intelligent compression fold chain, so that the segment anchor point, the compressed carrier segment and the fold chain position all point to the same send position, thus generating a compressed send chain position.

[0099] S4.4: Based on the original block binding range in the folding and reorganizing entry, anchor the file path and data block placeholder that are attached to the original block reserved chain position, map the file path to the reserved sending position in the original folder directory order, and bind the original block path around the reserved sending position. Connect the data block placeholder and the original block reserved mark to the same reserved sending position to obtain the reserved sending chain position.

[0100] Furthermore, the server locates the original block retention chain position item by item along the original block binding range in the folding and reassembly entry, extracts the file path, original folder directory position, start block position, and end block position of the original block retention chain position, searches for the corresponding directory position in the original folder directory order according to the file path, and sets the retention sending position according to the corresponding directory position; the server takes over each data block placeholder sequentially from the start block position to the end block position, keeps the file path and retention sending position corresponding, and preserves the file block order between adjacent data block placeholders.

[0101] The server sequentially arranges data block placeholders from the start block to the end block along the reserved transmission position, connects the start block to the beginning of the reserved transmission position, connects the end block to the end of the reserved transmission position, and attaches the original block reservation markers to the corresponding chain boundaries. The server then attaches the file path to the reserved transmission position, maintains the original block reservation markers and data block placeholders in the same position, and connects the reserved transmission position to the original fold chain position corresponding to the original block reservation chain position in the smart compression fold chain, thus obtaining the reserved transmission chain position.

[0102] S4.5: According to the order of the folded chain positions in the intelligent compression folded chain, the compressed sending chain position and the reserved sending chain position are folded in a traction order, and the compressed sending chain position and the reserved sending chain position are pasted back to the corresponding folded chain position. The chain position succession relationship after pasting is arranged in a hybrid multicast, and a hybrid multicast object sequence is generated.

[0103] Furthermore, the server sequentially locates the folded chain positions originally occupied by the aggregated compressed fragment and the original block reserved chain position in the folded chain position of the intelligent compression folded chain. It uses the fragment anchor point in the compressed sending chain position to find the folded chain position corresponding to the aggregated compressed fragment, and uses the file path and original folder directory location in the reserved sending chain position to find the folded chain position corresponding to the original block reserved chain position. Then, the compressed sending chain position and the reserved sending chain position are respectively pasted back to the found folded chain positions. When the type of adjacent folded chain positions changes, the boundary of the compressed sending chain position and the boundary of the reserved sending chain position are preserved, and the end of the previous folded chain position is connected to the beginning of the next folded chain position.

[0104] Starting from the first folded chain position in the intelligent compression folded chain, the server arranges the compressed sending chain positions and reserved sending chain positions in sequence according to the order of the folded chain positions. Each compressed sending chain position is used as an aggregate sending position, and each reserved sending chain position is arranged according to the file block order from the start block position to the end block position, while retaining independent resending granularity. The server then alternately connects the compressed sending chain positions and reserved sending chain positions to the same multicast sending sequence according to the original folder directory order, so that adjacent sending chain positions and corresponding folded chain positions maintain the same sequential relationship, generating a hybrid multicast object sequence.

[0105] It should be noted that the folded chain position order is the sequential relationship between the aggregated compressed fragments and the original block retained chain positions in the intelligent compression folded chain, arranged according to the original folder directory positions.

[0106] The corresponding folded chain position is the original chain position in the intelligent compression folded chain that matches the segment anchor point in the compressed sending chain position or the file path and original folder directory position in the preserved sending chain position.

[0107] S5: Perform multicast reflow and restoration on the hybrid multicast object sequence, and at the same time obtain the missing link positions reported by the clients. Prioritize the reflow of the link content that is missing in common with multiple clients, and subsequently resend the link content that is missing from a single client, and generate the multicast file distribution result.

[0108] S5.1: The server sends compressed send chain bits and reserved send chain bits according to the multicast send sequence position in the hybrid multicast object sequence. The client receives the compressed send chain bits and reserved send chain bits, restores the aggregated compressed fragment according to the fragment anchor point, and concatenates the original block content according to the original block reserved mark to form a multicast backfeed restoration chain.

[0109] Furthermore, the server sequentially locates the compressed transmission chain and the reserved transmission chain along the multicast transmission sequence position in the hybrid multicast object sequence. It then connects the aggregated compressed segment corresponding to the compressed transmission chain to the corresponding multicast transmission sequence position. The original block content between the start block and the end block in the reserved transmission chain is connected to the subsequent multicast transmission sequence positions according to the file block order, and sent sequentially through the multicast group address. The server preserves the chain boundary at the switching position between the compressed transmission chain and the reserved transmission chain. Clients within the edge LAN receive the content according to the multicast transmission sequence position and the chain boundary, respectively, and receive the content at the compressed transmission chain receiving position and the reserved transmission chain receiving position.

[0110] Within the edge LAN, the client extracts the segment anchor point, file order within the segment, and restoration directory location from the compressed sending chain position. It then splits and aggregates the compressed segment along the file delimiter, restores the content of each small file according to the file order within the segment, and concatenates the content of each small file to the corresponding restoration directory location. Within the edge LAN, the client extracts the file path, start block position, and end block position from the reserved sending chain position. It then concatenates the original block content to the corresponding file path according to the file block order, and finally connects the restored compressed sending chain position and reserved sending chain position according to the multicast sending sequence position to form a multicast backfeed restoration chain.

[0111] S5.2: Receive chain position reception feedback returned by clients within the edge LAN, paste missing chain positions back into the chain position reception feedback according to the chain position order in the multicast re-feedback restoration chain, and mark the missing chain position according to the number of clients from which the feedback comes, forming a missing chain position feedback table.

[0112] The link bit reception feedback is the feedback content returned by the client in the edge LAN after completing the reception of compressed and reserved transmission link bits, according to the multicast transmission sequence position, to return the link bit reception status and the position of missing link bits.

[0113] Furthermore, the server receives the chain position reception feedback returned by the client in the edge LAN, distinguishes the feedback source of the chain position reception feedback according to the client identifier, and then searches for the compressed transmission chain position or the reserved transmission chain position corresponding to the missing chain position in the chain position order in the multicast refeedback restoration chain. The client that provided the feedback is connected to the found chain position position, and the chain position reception feedback corresponding to the same missing chain position is grouped into the same feedback position, thus completing the missing chain position reattachment.

[0114] The server counts the number of client feedback sources that receive feedback at the same feedback location. When the same missing chain position corresponds to multiple clients within the edge LAN, a common missing label is written at the corresponding feedback location. When the same missing chain position corresponds to only a single client within the edge LAN, a single-end missing label is written at the corresponding feedback location. Then, the missing chain position, the client feedback source, and the missing label are arranged according to the chain position order in the multicast refeedback restoration chain to form a missing chain position feedback table.

[0115] It should be noted that the number of feedback source clients refers to the total number of clients within the edge LAN that receive feedback for the same missing link position.

[0116] S5.3: Based on the common missing label in the missing link feedback table, the common missing compressed transmission link and reserved transmission link are pasted back to the multicast transmission sequence position, and common missing priority is injected according to the multicast transmission sequence position and multicast group address after the injection, forming a common missing injection chain.

[0117] The common missing label is a missing category marker set for the corresponding missing link position when the same missing link position corresponds to the link position received feedback from multiple clients within the edge LAN in the missing link position feedback table.

[0118] Furthermore, the server extracts common missing identifiers item by item from the missing link feedback table. Based on the missing link positions corresponding to the common missing identifiers, it searches for the same multicast transmission sequence positions in the mixed multicast object sequence to locate the commonly missing compressed transmission links and reserved transmission links. The server then reconnects the compressed transmission links along with the fragment anchor points to the original multicast transmission sequence positions, and reconnects the reserved transmission links along with the file path, start block position, and end block position to the original multicast transmission sequence positions.

[0119] The server arranges the commonly missing compressed transmission chain bits and reserved transmission chain bits according to the position of the multicast transmission sequence after the reconnection, and sends the corresponding chain bit content to the edge LAN client that returns the common missing label using the multicast group address, and preserves the chain bit boundary between the compressed transmission chain bits and the reserved transmission chain bits; the server then completes the transmission of the compressed transmission chain bits and reserved transmission chain bits according to the position of the multicast transmission sequence to form the common missing reconnection chain.

[0120] S5.4: Based on the single-end missing label in the common missing feedback chain and missing chain bit feedback table, perform single-end missing targeted retransmission and continuation of the missing chain bit content of a client in a single edge LAN, and perform directory reception closure and solidification on the retransmitted compressed sending chain bit and reserved sending chain bit to generate multicast file distribution results.

[0121] Single-ended missing labeling is a missing category marker set for a missing link position when the same missing link position corresponds to only one link position received feedback from a client within the edge LAN in the missing link position feedback table.

[0122] Furthermore, the server excludes chain positions that have already completed priority re-feedback of common missing links along the common missing re-feedback chain. Then, it extracts the single-end missing label, the feedback source client, and the missing link position from the missing link feedback table. According to the missing link position, it locates the corresponding compressed transmission link or reserved transmission link in the hybrid multicast object sequence. Based on the client identifier of the feedback source client, it locates the corresponding signaling connection position. Through the corresponding signaling connection position, it re-sends the missing link content to the client in the single edge LAN. Then, it connects the re-sent link to the original multicast transmission sequence position.

[0123] After receiving the resent link content, the client in the edge LAN restores the compressed sending link according to the segment anchor point, the file order within the segment, and the location of the restore directory. It then concatenates the retained sending link according to the file path, start block, end block, and file block order. The client in the edge LAN attaches the resent compressed sending link and the retained sending link to the corresponding missing link positions, and then connects the received link content and the resent link content according to the original folder directory order to generate the multicast file distribution result.

[0124] The result of multicast file distribution is that after the client in the edge LAN completes the process of restoring the aggregated compressed fragments, continuing the original block content, prioritizing the re-uploading of common missing data, and targeted resending of single-end missing data, the complete folder content is restored according to the original folder directory order.

[0125] This embodiment also provides a computer device applicable to the multicast file distribution method based on intelligent compression, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the multicast file distribution method based on intelligent compression as proposed in the above embodiment.

[0126] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0127] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the multicast file distribution method based on intelligent compression as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0128] In summary, this invention achieves multicast pressure layering and triggering on the folder block expansion chain, merging consecutive small files that have common misses and performing benefit screening, folding and compressing small file clusters with entry savings, while preserving the original block boundaries and independent resending granularity of large files. This enables co-chain orchestration of small file aggregation compression and large file retention distribution, reducing the multicast scheduling burden and missing resending range in edge data transmission, and improving the continuity of client file distribution and transmission resource utilization.

[0129] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multicast file distribution method based on intelligent compression, characterized in that, include: The server receives file distribution access requests from clients within the edge LAN, establishes a signaling connection, and assigns multicast group addresses to verified clients. Under the multicast group address, it performs receive bearer replies to form an edge multicast bearer chain. The folder to be distributed is retrieved through the edge multicast bearer chain, and the folder to be distributed is pre-arranged for multiple receiving end blocks to generate data block placeholders. The data block placeholders are then split and screened according to the cache hit status to form a folder block expansion chain. Multicast pressure layering is triggered on the folder block expansion chain. Consecutive small files that are jointly missed in the expansion chain are merged into candidate small file clusters. The candidate small file clusters are subjected to revenue screening and folding compression to generate aggregated compressed fragments. Then, the large file chain with the original block retention mark is continued to generate a smart compressed folding chain. Based on the intelligent compression folding chain, the aggregated compressed fragments are reorganized by directory anchoring to generate compressed sending chain bits. The original block path is bound to the chain bits of large files with original block retention marks to obtain the retained sending chain bits. The compressed sending chain bits and the retained sending chain bits are mixed and arranged to generate a mixed multicast object sequence. Multicast reflow is performed on the sequence of mixed multicast objects. At the same time, the positions of missing links reported by the clients are obtained. The links that are missing in common to multiple clients are reflowed first, and the links that are missing to a single client are resent later, generating the multicast file distribution result.

2. The multicast file distribution method based on intelligent compression as described in claim 1, characterized in that, The establishment of the signaling connection is specifically as follows: The server receives file distribution access requests sent by clients at the file distribution access entry point on the edge local area network, and performs access request purification and queuing according to the uniqueness of the source and the integrity of the receiving entry point to form an access request queue; According to the order of the access request queue, the signaling handshake content is sent to the client in the edge LAN. The connection response and cache hit status returned by the client in the edge LAN are received. The connection response is closed and attached according to the client identifier consistency and connection response validity to form a signaling connection chain.

3. The multicast file distribution method based on intelligent compression as described in claim 2, characterized in that, The formation of the edge multicast bearer chain is as follows: The receiving bearer is verified using the client in the edge LAN in the signaling connection chain as the verification object, and the multicast group address is sent to the client in the edge LAN that is retained in the multicast access range. The multicast group address is written into the corresponding signaling connection location to form a multicast group address allocation table. According to the multicast group address and signaling connection position in the multicast group address allocation table, the receiving bearer aggregation and back-attachment are performed on the clients in the edge LAN. The number of clients in the edge LAN after multicast reception preparation, the number of multicast receivers and the cache hit status are written into the same bearer chain position to form the edge multicast bearer chain.

4. The multicast file distribution method based on intelligent compression as described in claim 1 or 3, characterized in that, The placeholder for the generated data block is as follows: Extract the bearer link bit, the number of multicast receivers and the cache hit status from the edge multicast bearer chain, perform bearer chain traction retrieval and folder entry attachment on the folder to be distributed, and connect the number of multicast receivers to the transmission bearer position of the folder to be distributed to form the folder bearer bit; The folders to be distributed in the folder carrier position are sorted according to their original directory order to form a file path order chain. The file path order chain is then pre-occupied and arranged, and the file path, file size and number of file blocks are attached to the same directory chain position to form a directory chain position pre-occupancy table. Based on the directory chain bit pre-occupancy table, the receiver block volume is amplified and arranged and multicast block bits are taken over. The receiver demand bits are formed according to the number of file blocks and the number of missing clients. Then, the small file directory chain bits that are adjacent in the original folder directory order are connected to the same continuous bit interval to generate data block bits.

5. The multicast file distribution method based on intelligent compression as described in claim 4, characterized in that, The formation of the folder block expansion chain is as follows: The file paths in the data block placeholders are matched with the cache hit status one by one, and the data block placeholders are split into hit difference streams to form the no-transmission chain, the chain to be sent, and the difference chain, respectively. The hit chain is then collected to form the hit split placeholders. Based on the hit-based traffic splitting placeholders, small files are continuously expanded and clustered into the to-send chain position and the difference chain position according to the original folder directory order to form an expanded chain position. For file data block placeholders that need to maintain independent resending granularity, large file boundary preservation is performed to form a reserved chain position. These are then merged to form a folder block expansion chain.

6. The multicast file distribution method based on intelligent compression as described in claim 5, characterized in that, The generation of the intelligent compression folding chain is as follows: Based on the expansion chain bit and the reserved chain bit in the folder block expansion chain, multicast pressure mapping is performed on the number of multicast receivers and the cache hit status carried by the expansion chain bit and the reserved chain bit. The expansion chain bit is mapped to the compression pressure chain bit, and the reserved chain bit is mapped to the reserved pressure chain bit. Then, the multicast pressure layered chain is formed by hierarchical triggering based on the compression pressure chain bit and the reserved pressure chain bit. Based on the cache hit status and original folder directory order in the multicast pressure hierarchical chain, consecutive small files are matched together for common misses. Small files that have not been cached by clients in the edge LAN within the same consecutive directory range are continuated to the same cluster position, and small files with cache hit divergences are retained in the range of different chain positions to form candidate small file clusters. The candidate small file clusters are aligned with the entry occupancy position to generate a multicast entry saving chain. Based on the multicast entry saving chain, the candidate small file clusters with an aggregated sending number of entries less than the number of per-file sending entries are continued to the folded candidate position. The candidate small file clusters with an aggregated sending number of entries not less than the number of per-file sending entries are retained as per-file reserved chain positions according to the original folder directory order, forming folded candidate small file clusters. Based on the number of small files in the folded candidate small file cluster and the original folder directory order, the file-by-file sending entries in the folded candidate small file cluster are continuously connected to the aggregation sending position according to the directory adjacency relationship to obtain the aggregation sending entries. The contents of the small files carried by the aggregation sending entries are merged and compressed into an aggregation compressed fragment according to the original folder directory order. Based on the aggregated compressed fragments, the file-by-file reserved chain bits, and the reserved pressure chain bits in the multicast pressure layered chain, the file-by-file reserved chain bits and the reserved chain bits inherited by the reserved pressure chain bits are marked and pasted back to the original block to generate the original block reserved chain bits. Then, the aggregated compressed fragments and the original block reserved chain bits are folded and connected according to the original folder directory order to generate a smart compressed folded chain.

7. The multicast file distribution method based on intelligent compression as described in claim 6, characterized in that, The generation of the hybrid multicast object sequence is as follows: Using the intelligent compression and folding chain as the disassembly object, the chain position content is located item by item according to the folding chain position sequence. The position of the aggregated compression fragment and the position of the original block retained chain position are distinguished according to the content carried by the chain position. The aggregated compression fragment is extracted from the position of the aggregated compression fragment and connected to the bearing and reorganization range. The original block retained chain position is extracted from the position of the original block retained chain position and connected to the original block binding range, thus forming the folding and reorganization entry point. Based on the reorganization range in the folding and reorganization entry, the directory anchor point is extracted from the aggregated and compressed fragment to obtain the corresponding fragment starting directory position and the file order within the fragment, forming a fragment anchor point. The fragment anchor point, the file order within the fragment, and the restored directory position are then bound to form a directory anchored fragment. The directory anchoring fragment restores the directory position according to the fragment anchor point alignment, pastes the files in the fragment back to the compressed bearer position corresponding to the original folder directory order, and connects the aggregated compressed fragment to the same compressed bearer position according to the fragment anchor point and the original folder directory order to obtain the compressed bearer fragment. The fragment anchor point and the compressed bearer fragment are bound to the same sending chain position to generate a compressed sending chain position. Based on the original block binding range in the folding and reorganization entry, the file path and data block placeholders of the original block reserved chain are anchored to the original block path, the file path is mapped to the reserved sending position in the original folder directory order, and the original block path is bound around the reserved sending position. The data block placeholder and the original block reserved mark are connected to the same reserved sending position to obtain the reserved sending chain. According to the order of the folded chain positions in the intelligent compression folded chain, the compressed sending chain position and the reserved sending chain position are folded in a folding order and then pasted back to the corresponding folded chain position. The chain position succession relationship after pasting is then arranged in a hybrid multicast to generate a hybrid multicast object sequence.

8. The multicast file distribution method based on intelligent compression as described in claim 1, characterized in that, The distribution results of the generated multicast file are as follows: The server sends compressed send chain bits and reserved send chain bits according to the multicast send sequence position in the hybrid multicast object sequence. The client receives the compressed send chain bits and reserved send chain bits, restores the aggregated compressed fragment according to the fragment anchor point, and concatenates the original block content according to the original block reserved mark to form a multicast backfeed restoration chain. Receive chain bit reception feedback returned by clients within the edge LAN, re-attach missing chain bits to the chain bit reception feedback according to the chain bit order in the multicast re-feedback and restoration chain, and mark the missing chain bit positions according to the number of clients from which the feedback comes, forming a missing chain bit feedback table. Based on the common missing label in the missing link feedback table, the common missing compressed transmission link and reserved transmission link are pasted back to the multicast transmission sequence position, and common missing priority is injected according to the multicast transmission sequence position and multicast group address after the injection, forming a common missing injection chain; Based on the single-end missing label in the common missing feedback chain and missing chain bit feedback table, the missing chain bit content of the client in a single edge LAN is resent and connected in a single-end missing direction. The resent compressed sending chain bit and reserved sending chain bit are then closed and solidified in the directory reception, generating multicast file distribution results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the multicast file distribution method based on intelligent compression as described in any one of claims 1 to 8.

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

Images