Data recovery method and system based on multi-cloud platform

Through the data recovery method of multi-cloud platform, the problem of low migration efficiency in multi-cloud scenarios is solved, efficient and secure data migration and system recovery is achieved, data reading efficiency and system stability are improved, and data adapted to various network environments.

CN120429169AActive Publication Date: 2025-08-05上海安云无界软件有限公司

Patent Information

Application Number
CN202510923810.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-04
Publication Date
2025-08-05
Estimated Expiration
2045-07-04

AI Technical Summary

Technical Problem

The prior art cannot efficiently and securely migrate systems and data in multi-cloud scenarios, especially in the case of large number of files and complex file system levels, the transmission efficiency is significantly reduced, which greatly limits the practicality and convenience of migration.

Method used

The multi-cloud platform-based data recovery method is adopted, including intelligent data acquisition layer, adaptive transmission protocol stack, heterogeneous system recovery architecture and spatiotemporal dimensional data management model. Cross-platform data capture, dynamic transmission and automated recovery are achieved through a multi-modal data identification engine, virtual environment abstraction layer, cloud native storage architecture and workflow engine.

Benefits of technology

It improves data reading efficiency, reduces resource consumption at the source side, saves labor costs, improves system reliability and stability, supports high-frequency replication and synchronization, adapts to multiple network scenarios, and ensures data consistency and rapid recovery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120429169A_ABST
    Figure CN120429169A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data recovery, and discloses a data recovery method and system based on a multi-cloud platform, and the method comprises the following steps: S1, constructing an intelligent data acquisition layer: carrying out the dynamic adaption of four data collection channels based on a multi-modal data recognition engine; s2, constructing an adaptive transmission protocol stack: dynamically selecting based on a network quality real-time evaluation model; s3, heterogeneous system architecture recovery: realizing architecture recovery through a virtual environment abstraction layer; s4, establishing a space-time dimension data management model, which is realized based on a cloud native storage architecture; and S5, automatically recovering an arrangement engine through a workflow engine. According to the method, the reliability and the stability of the system are improved while the labor cost is saved, high-frequency copy synchronization is supported, the resource utilization rate is improved, and the data reading efficiency can be improved through block-level data synchronization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data recovery technology, and in particular to a data recovery method and system based on a multi-cloud platform. Background Art

[0002] With the rapid development of cloud computing technology, more and more businesses and individuals are migrating their applications and data to the cloud to leverage its powerful services and flexibility. However, for businesses that already have on-premises systems, or for users with business needs to migrate systems and data between different cloud platforms, effectively migrating on-premises or existing public cloud resources to a new cloud platform presents a new challenge.

[0003] Existing solutions primarily rely on host-to-host file transfer technology and host-to-cloud image importing. However, this technology has significant limitations, such as limited to bidirectional peer-to-peer networks, inability to address one-way transfers from local office networks to the public cloud, low transfer efficiency, and complex operations requiring manual configuration. Consequently, existing technologies cannot meet the requirements for efficient and secure system and data migration in multi-cloud scenarios. This is especially true when faced with large file volumes and complex file system hierarchies, significantly reducing transfer efficiency and significantly limiting the practicality and convenience of migration. Summary of the Invention

[0004] Technical issues solved: In response to the shortcomings of the existing technology, the present invention provides a solution to the problem that the existing technology cannot meet the needs of efficiently and securely migrating systems and data in multi-cloud scenarios, especially when the number of files is large and the file system hierarchy is complex, the transmission efficiency is significantly reduced, which greatly limits the practicality and convenience of migration.

[0005] (2) Technical solution: To achieve the above object, the present invention provides the following technical solutions: A data recovery method based on a multi-cloud platform includes the following steps: Step S1: Intelligent data acquisition layer construction: Based on the multimodal data recognition engine, dynamically adapt to four data acquisition channels: Lightweight client agent module, supporting cross-platform data capture in ARM / x86 heterogeneous environments; VMware agentless channel, using VADP / CBT dual interfaces to achieve accurate identification of changed data; Cloud-native agentless mode, using the cloud platform API gateway to achieve instance-level data extraction; Offline cold data synchronization module, which realizes physical machine data extraction based on PE environment booted by UEFI; Step S2: Adaptive transmission protocol stack construction: Based on the real-time network quality evaluation model, dynamically select: Bidirectional symmetric transmission mode: uses zero-copy technology to achieve efficient transmission of block-level data; One-way secure tunnel mode: builds an encrypted reverse channel based on the TLS1.3 protocol; Hybrid transmission optimization engine: supports multi-channel aggregation transmission with bandwidth adaptation; Step S3: Heterogeneous system recovery architecture: implemented through the virtual environment abstraction layer: Hardware abstraction layer construction based on ACPI table parsing; Driver automatic identification and injection engine; Network configuration intelligent migration module; System startup path reconstruction algorithm; Step S4: Spatiotemporal data management model: implemented based on cloud-native storage architecture: Timeline snapshot indexing system; Incremental change capture engine; Multi-level cache optimization strategy; Intelligent stratification of hot and cold data; Step S5: Automated recovery orchestration engine: implemented through the workflow engine: Intelligent recognition of recovery scenes; Hardware compatibility pre-detection; The recovery path is automatically generated; Post script automatic execution.

[0006] Preferably, the adaptive transmission protocol stack constructed in step S2 dynamically selects a bidirectional symmetric transmission mode or a unidirectional secure tunnel mode through a real-time network quality evaluation model, and implements adaptive tuning of transmission parameters based on a bandwidth prediction algorithm, specifically including: Network quality probe module, real-time collection of packet loss rate, delay, and bandwidth parameters; Transmission mode decision engine, which selects the optimal transmission mode based on reinforcement learning algorithm; Dynamic congestion control module, supporting multiple congestion control algorithms such as BBR and CUBIC; Transmission encryption module supports AES-256 and ChaCha20 encryption algorithms.

[0007] Preferably, the heterogeneous system recovery architecture in step S3 implements compatibility adaptation between different operating systems and hardware platforms through a virtual environment abstraction layer, specifically including: ACPI table parsing module to extract hardware configuration information; Driver warehouse management module, maintaining multi-version driver library; Driver injection engine, automatically matching and injecting appropriate drivers based on hardware information; The system starts the configuration reconstruction module and regenerates the boot configuration file; The network configuration migration module realizes automatic configuration of IP addresses and DNS network parameters.

[0008] Preferably, the spatiotemporal dimension data management model in step S4 implements efficient management and rapid recovery of data copies based on a cloud-native storage architecture, specifically including: The timeline index module creates a unique index for each snapshot at a time point; Incremental change capture module, which identifies data changes based on block-level comparison technology; Multi-level cache module, using LRU and LFU hybrid cache strategy to improve access efficiency; The hot and cold data tiering module automatically migrates data storage tiers based on access frequency.

[0009] Preferably, the automated recovery orchestration engine in step S5 implements automated execution and monitoring of the recovery process through a workflow engine, specifically including: The recovery scene recognition module automatically matches the recovery template based on the source system characteristics; Hardware compatibility detection module, pre-detecting the compatibility between the target environment and the source system; Recovery path generation module, which generates the optimal recovery path based on graph algorithms; Execution monitoring module to track the recovery process in real time and provide visual feedback; Post-processing module supports automated execution of custom scripts.

[0010] Preferably, the multimodal data recognition engine constructed by the intelligent data acquisition layer in step S1 can automatically select the optimal data acquisition channel according to the source environment type, specifically including: The environment detection module identifies the source system's operating system, hardware architecture, and virtualization platform; Channel evaluation module, which evaluates the applicability of each acquisition channel based on performance indicators and environmental characteristics; Intelligent switching module supports dynamic switching of acquisition channels during data acquisition; Data quality verification module ensures the integrity and consistency of collected data.

[0011] A data recovery system based on a multi-cloud platform, comprising: The intelligent data acquisition layer includes four data collection modules: client agent software, VMware agentless module, cloud instance agentless module, and source machine offline data synchronization; Adaptive transport layer, realizing two transmission modes: bidirectional symmetric transmission and unidirectional secure tunnel; Heterogeneous recovery engine supports cross-platform and cross-architecture system recovery; Spatiotemporal data management module, which enables the creation, management and recovery of data snapshots; The automated orchestration center provides visual configuration and execution of the recovery process.

[0012] Preferably, the client agent software adopts a microkernel architecture design, specifically including: Data collection engine, supporting file system filter driver and volume shadow copy service; Data processing engine, which implements data compression, encryption and segmentation; Communication engine, supporting breakpoint resumption and transmission optimization; Resource management module realizes intelligent scheduling of system resources.

[0013] Preferably, the cloud instance agentless module acquires data through the cloud platform API, specifically including: The cloud platform adapter layer provides a unified interface to adapt to different cloud platform APIs; Snapshot management module, which enables the creation and management of cloud disk snapshots; Difference analysis engine, which identifies incremental data based on block-level comparison technology; Data transmission optimization module supports multi-threaded parallel transmission and bandwidth control.

[0014] (3) Beneficial effects: Compared with the existing technology, the present invention provides a data recovery method and system based on a multi-cloud platform, which has the following beneficial effects: 1. The present invention uses block-level technology for source-side data acquisition, which improves data reading efficiency and reduces the consumption of source machine resources. The faster incremental synchronization interval improves the RPO indicator from one day to minutes. The source-side agent software is compatible with mainstream Windows versions and numerous Linux distributions. The software-customized PE system is developed to support startup on multiple cloud platforms and integrate data processing programs to realize information collection, data writing, and system conversion functions. After the target cloud host switches between different states, it can always ensure the retention of task information and will not cause task information loss due to PE system restart.

[0015] 2. The present invention deeply integrates software workflow with cloud platform. Automated cloud resource calls can automatically match the specifications of source production and create disaster recovery resources on the target platform according to template planning. This not only saves tedious supporting operations, but also eliminates the possibility of human error. While saving labor costs, it improves the reliability and stability of the system. The automated agentless function eliminates the process of installing the client one by one on the source side, which not only shortens the deployment cycle, but also avoids compatibility issues of client adaptation. There is no need to trouble the end customer to provide account passwords. Automated drive conversion makes disaster recovery of the entire machine on heterogeneous platforms possible. Disaster recovery of heterogeneous platforms can also avoid security risks caused by security vulnerabilities on the same platform, significantly reducing the RTO of disaster recovery.

[0016] 3. The present invention supports both unidirectional and bidirectional data transmission technologies. It can not only meet the data transmission needs of VPN, dedicated line and intranet networks, but also cope with unidirectional transmission scenarios such as office network to Internet and firewall rules restricting unidirectional access. It also has a flexible network transmission architecture. In the scenario of a bidirectional transmission network, the source machine can directly transmit to the target cloud host. In the scenario where the source machine and the target cloud host cannot communicate, data can be transferred through a data gateway.

[0017] 4. The present invention can not only realize the copying and recovery of file data, but also realize the copying and recovery of the entire operating system. The restored target system and the source machine are consistent, and the data are consistent. After the system is restored, the user does not need to perform too much operation to quickly start the application and resume business operation.

[0018] 5. The present invention improves data reading efficiency and realizes efficient input or output operations through block-level data synchronization. It does not need to maintain redundant information such as file attributes and directory structures, can optimize random access, and avoid frequent head seeks caused by file fragmentation by block addressing. It has batch processing capabilities, and a single input or output can process continuous block data, reducing the number of operations. By optimizing memory usage and reducing the number of data copies between kernel state and user state, the CPU load is reduced. Changes are detected in units of storage blocks, and only modified blocks are transmitted instead of the entire file to achieve lower data transmission volume. It supports high-frequency copy synchronization, thereby improving resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a data recovery method based on a multi-cloud platform proposed by the present invention.

[0020] Figure 2 This is a schematic diagram of the process structure of the method for realizing automated target resource configuration in data replication in combination with a cloud platform proposed by the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Example 1: like Figure 1-Figure 2 As shown, a data recovery method based on a multi-cloud platform includes the following steps: Step S1: Intelligent data acquisition layer construction: Based on the multimodal data recognition engine, dynamically adapt to four data acquisition channels; Lightweight client agent module, supporting cross-platform data capture in ARM / x86 heterogeneous environments; VMware agentless channel, using VADP / CBT dual interfaces to achieve accurate identification of changed data; Cloud-native agentless mode, using the cloud platform API gateway to achieve instance-level data extraction; Offline cold data synchronization module, which realizes physical machine data extraction based on PE environment booted by UEFI; The multimodal data recognition engine constructed by the intelligent data acquisition layer in step S1 can automatically select the optimal data acquisition channel according to the source environment type, specifically including: The environment detection module identifies the source system's operating system, hardware architecture, and virtualization platform; Let the source environment feature vector .

[0023] Channel evaluation module, which evaluates the suitability of each acquisition channel based on performance indicators and environmental characteristics; Set the collection channel set , the evaluation function for: ; in, is the weight coefficient, Environmental characteristics and acquisition channels The adaptability.

[0024] Intelligent switching module supports dynamic switching of acquisition channels during data acquisition; Set the switching condition function , when this condition is met, from the channel Switch to channel .

[0025] Data quality verification module to ensure the integrity and consistency of collected data; Set data integrity index , and its calculation formula is: Among them, the data integrity is calculated by comparing the size of the original data with the size of the collected data. If the difference between the two is within the allowable error range, the data is considered complete. At the same time, to ensure the consistency of the collected data, a hash check algorithm is used to check the data. The data consistency indicator is set as the hash check pass rate, and its calculation formula is: The amount of data that passed verification / the total amount of data × 100%.

[0026] : Represents the data integrity index, and the value range is usually between 0 and 1. The closer it is to 1, the better the data integrity. It reflects the proportion of valid data after collection / transmission.

[0027] : Indicates the size of valid data (for example, the amount of complete and undamaged data after collection), which is a data volume statistic that meets quality requirements.

[0028] : Refers to the total data size that should theoretically be collected / transmitted, and is the full target scale for data operations.

[0029] Step S2: Adaptive transmission protocol stack construction: Based on the real-time network quality evaluation model, dynamically select: Bidirectional symmetric transmission mode: uses zero-copy technology to achieve efficient transmission of block-level data; One-way secure tunnel mode: builds an encrypted reverse channel based on the TLS1.3 protocol; Hybrid transmission optimization engine: supports multi-channel aggregation transmission with bandwidth adaptation; The adaptive transmission protocol stack constructed in step S2 dynamically selects a bidirectional symmetric transmission mode or a unidirectional secure tunnel mode through a real-time network quality evaluation model, and implements adaptive tuning of transmission parameters based on a bandwidth prediction algorithm, specifically including: The network quality probe module collects parameters such as packet loss rate, latency, and bandwidth in real time.

[0030] Define network quality assessment function , and its calculation formula is: ; in; is the packet loss rate at time t; is the delay at time t; is the amount of money at time t; 、 、 is the weight coefficient, and + + =1.

[0031] The transmission mode decision engine selects the optimal transmission mode based on the reinforcement learning algorithm.

[0032] Dynamic congestion control module, supporting multiple congestion control algorithms such as BBR and CUBIC; Transmission encryption module supports encryption algorithms such as AES-256 and ChaCha20.

[0033] Step S3: Heterogeneous system recovery architecture: implemented through the virtual environment abstraction layer: Hardware abstraction layer construction based on ACPI table parsing; Driver automatic identification and injection engine; Network configuration intelligent migration module; System startup path reconstruction algorithm; The heterogeneous system recovery architecture in step S3 implements compatibility adaptation between different operating systems and hardware platforms through the virtual environment abstraction layer, specifically including: ACPI table parsing module to extract hardware configuration information; Set hardware configuration information collection ,in express hardware attributes.

[0034] Driver warehouse management module, maintaining multi-version driver library; Driver Library ,in Indicates the hardware type, Indicates the driver version.

[0035] Driver injection engine, automatically matching and injecting appropriate drivers based on hardware information; Matching function Defined as: in, For hardware and drive Compatibility score.

[0036] The system starts the configuration reconstruction module and regenerates the boot configuration file; , by configuring the generated function Generate a new boot configuration file, where This is the operating system information.

[0037] Network configuration migration module, which realizes automatic configuration of network parameters such as IP address and DNS; ,in, is the subnet information, and M is the MAC address.

[0038] Step S4: Spatiotemporal data management model: implemented based on cloud-native storage architecture: Timeline snapshot indexing system; Incremental change capture engine; Multi-level cache optimization strategy; Intelligent stratification of hot and cold data; The spatiotemporal data management model in step S4 implements efficient management and rapid recovery of data copies based on the cloud-native storage architecture. Specifically, it includes: The timeline index module creates a unique index for each snapshot at a time point; Incremental change capture module, which identifies data changes based on block-level comparison technology; Multi-level cache module, using LRU and LFU hybrid cache strategy to improve access efficiency; The hot and cold data tiering module automatically migrates data storage tiers based on access frequency.

[0039] Step S5: Automated recovery orchestration engine: implemented through the workflow engine: Intelligent recognition of recovery scenes; Hardware compatibility pre-detection; The recovery path is automatically generated; Post-script automatic execution; In step S5, the automated recovery orchestration engine uses the workflow engine to automate and monitor the recovery process, specifically including: The recovery scene recognition module automatically matches the recovery template based on the source system characteristics; Let the source system eigenvector , restore template collection , matching function ; in, It is the similarity score between the source system characteristics and the recovery template.

[0040] Hardware compatibility detection module, pre-detecting the compatibility between the target environment and the source system; Compatibility Score The calculation formula is: ; in, is the weight coefficient, and The source and target systems are hardware attributes.

[0041] Recovery path generation module, which generates the optimal recovery path based on graph algorithms; Set up recovery task graph , where V is the node set, E is the edge set, and the optimal path P is obtained by Dijkstra algorithm Generate, where s is the starting node and t is the target node.

[0042] Execution monitoring module to track the recovery process in real time and provide visual feedback; Set recovery progress , and its calculation formula is: ; Post-processing module supports automated execution of custom scripts; Set up custom script collection , execute the function through the script Execute the script.

[0043] A data recovery system based on a multi-cloud platform, comprising: The intelligent data acquisition layer includes four data collection modules: client agent software, VMware agentless module, cloud instance agentless module, and source machine offline data synchronization; The cloud instance agentless module acquires data through the cloud platform API, specifically including: The cloud platform adapter layer provides a unified interface to adapt to different cloud platform APIs; Snapshot management module, which enables the creation and management of cloud disk snapshots; Difference analysis engine, which identifies incremental data based on block-level comparison technology; Data transmission optimization module, supporting multi-threaded parallel transmission and bandwidth control; Adaptive transport layer, realizing two transmission modes: bidirectional symmetric transmission and unidirectional secure tunnel; Heterogeneous recovery engine supports cross-platform and cross-architecture system recovery; Spatiotemporal data management module, which enables the creation, management and recovery of data snapshots; The automated orchestration center provides visual configuration and execution of the recovery process.

[0044] Preferably, the client agent software adopts a microkernel architecture design, specifically including: Data collection engine, supporting file system filter driver and volume shadow copy service; Data processing engine, which implements data compression, encryption and segmentation; Communication engine, supporting breakpoint resumption and transmission optimization; Resource management module realizes intelligent scheduling of system resources.

[0045] Example 2: A data recovery method based on a multi-cloud platform includes the following steps: Step 1: Data acquisition. Using block-level data reading and writing technology, we developed four data acquisition methods: client agent software, VMware agentless, cloud instance agentless module, and source machine offline data synchronization. This technology directly operates on the physical blocks or logical blocks of the storage device, bypassing the file system level, reducing the following overheads: metadata operations: no need to maintain redundant information such as file attributes and directory structures; random access optimization: block addressing avoids frequent head seeks caused by file fragmentation; batch processing capabilities: a single input or output can process continuous block data, reducing the number of operations. For example, when backing up a 1GB file, the traditional method requires thousands of small file operations, while the block level only requires dozens of batch transfers; improving resource utilization and optimizing memory usage: block-level reading and writing can accurately control the buffer size, such as aligning 4 1K / 8K blocks, avoiding file caching and uncontrollable memory usage; reduced CPU load: reducing the number of data copies between kernel state and user state; lower data transmission volume; only synchronizing changed blocks: detecting changes in storage blocks, such as 4KB-64MB, and only transmitting modified blocks instead of the entire file; millisecond-level incremental detection, fast change location: quickly comparing the status of source and target blocks through hashes and bitmaps; write-time copy, recording data block modification logs; support high-frequency replication and synchronization; no fear of frequent updates, even if the file is modified hundreds of times per hour, block-level synchronization can still handle it efficiently.

[0046] The present invention improves data reading efficiency and realizes efficient input or output operations through block-level data synchronization. There is no need to maintain redundant information such as file attributes and directory structure, and it can optimize random access and avoid frequent head seeks caused by file fragmentation by block addressing. It has batch processing capabilities, and a single input or output can process continuous block data, reducing the number of operations. By optimizing memory usage, the number of data copies between kernel state and user state is reduced to reduce CPU load. Changes are detected in units of storage blocks, and only modified blocks are transmitted instead of entire files to achieve lower data transmission volume. It supports high-frequency copy synchronization, thereby improving resource utilization. The block-level data processing technology of this software achieves comprehensive breakthroughs in performance, flexibility and resource management through underlying architecture innovation and dynamic block adjustment strategies. Whether it is large-scale data migration, real-time backup, or high-performance computing scenarios, it can provide efficiency and stability far exceeding traditional file synchronization solutions.

[0047] Step 2. Data transmission supports two transmission connection modes, namely peer-to-peer connection and unidirectional connection, supports various network scenarios such as VPN, dedicated line and Internet, and technically supports two-way transmission and unidirectional transmission. In two-way transmission, two-way traffic is required between the two nodes on the network. Through two-way data flow, after the TCP connection is established, the communicating parties have independent sending buffers and receiving buffers, and can send and receive data at the same time. When data is separated, the port number, IP address and sequence number are used to distinguish data streams in different directions. In unidirectional transmission, between the two nodes on the network, only one-way release is required, and the receiver does not need a reverse link. The reverse transmission tunnel is established using https to realize the one-way transmission function between the source and target ends. In terms of architecture, it supports data replication in various scenarios.

[0048] Source → Network 1 → Data Gateway Node → Network 2 → Data Writing Node; Source → Network 1 → Data is written to the node.

[0049] Step 3, data replication task is established, the target host uses the customized image to start in PE mode, and data processing related programs are run in PE. The target host maintains PE mode during data replication, and after the system conversion process during recovery, the virtual machine can enter the target system after restarting. During the data replication process, it is necessary to combine the cloud platform to realize automatic target resource configuration. The specific process is as follows: synchronization task configuration, initiating synchronization, creating a cloud host using the customized image on the target cloud platform, judging whether the creation is successful, reporting an error when it fails, stopping the task, reading the cloud host information when the creation is successful, registering to the management console, and then creating a data disk and mounting it to the cloud host. The program in the cloud host recognizes the disk and corresponds to the cloud platform record, and then matches it with the source machine, establishes a complex data task, and finally completes it. The present invention can not only realize the replication and recovery of file data, but also realize the replication and recovery of the entire operating system, and the restored target system is consistent with the source machine, and the data is consistent. After the system is restored, the user can quickly start the application and resume business operation without performing too much operation.

[0050] Step 4: Retain and manage historical copies of data. Use the cloud platform's own disk snapshot function to create a snapshot of the corresponding time point after each synchronization is completed, so as to achieve the ability to retain historical copies. The main tasks of data copy management include: according to the user-set policy, automatically create a disk snapshot of the target cloud host after each synchronization is completed; according to the user-set configuration, automatically maintain the number of retained copies of the target cloud host snapshot; cooperate with the recovery function to roll back the cloud host to the snapshot point selected by the user; cooperate with the recovery function to create a cloud host using the snapshot selected by the user to complete the recovery function.

[0051] Step 5, operating system and data recovery, addresses the issue of heterogeneous system startup. It supports automated network card configuration after system recovery and user-defined execution scripts. Through the integration of processes and architecture, diverse recovery capabilities are achieved, including the following: target cloud host state switching; system conversion; recovery mode 1: using the target host for recovery; recovery mode 2: using historical point recovery; and recovery mode 3: using snapshots for test recovery. During different recovery processes, the target cloud host needs to switch states to match the current task. This switching process requires the integration of software processes and cloud platform functions. In the replication synchronization state, the system within the target cloud host is a customized mirror system for receiving and processing data. In the recovery state, the system within the target cloud host undergoes system conversion, completes relevant configurations such as driver replacement and network card settings, and boots into the user's operating system. This operating system and internal data remain consistent with the user's source machine. By invoking the cloud platform's snapshot rollback capability, the cloud host is restored to the customized mirror system, allowing the data replication task to be continued or other recovery tasks to be re-executed.

[0052] It should be noted that after the user confirms the configuration of the replication task, the entire process will be automatically completed through the management console to schedule the source machine, cloud platform, and cloud host related resources. The specific process includes: a. The user determines the replication task configuration and initiates the task. The user needs to determine the target cloud platform, image, cloud platform network, security group, and storage-related resource configuration. The software automatically creates a cloud host on the cloud platform based on these configurations. The parameters required may vary depending on the cloud platform. b. Based on the configuration selected by the user, the software automatically creates a cloud host by calling the cloud platform API; c. Obtain cloud host status information through the cloud platform API, including start and stop status, network information, and disk information; d. Register the target cloud host to the management console and obtain internal information of the cloud host, including network information and disk information; e. Match the information in the virtual machine with the information on the cloud platform; f. Based on the matching information, a replication task is created between the source machine and the target cloud host, and a full synchronization is initiated. A full synchronization requires copying all data from the source machine to the target host. G. After full synchronization is completed, incremental synchronization tasks are initiated periodically according to the user-configured policy. The present invention supports both one-way and two-way data transmission technologies. It can not only be applied to data transmission requirements under networks such as VPN, dedicated lines, and intranets, but can also cope with one-way transmission scenarios such as office networks to the Internet and firewall rules restricting one-way access. It also has a flexible network transmission architecture. In the scenario of a two-way transmission network, the source machine can directly transmit to the target cloud host. In the scenario where the source machine and the target cloud host cannot communicate, data can be transferred through a data gateway.

[0053] The present invention also proposes a data recovery system based on a multi-cloud platform, including four types: client agent software, VMware agentless, cloud instance agentless module and source machine offline data synchronization. The present invention deeply integrates the software workflow with the cloud platform. The automated cloud resource call can automatically match the specifications of the source production and create disaster recovery resources on the target platform according to the template plan, which not only saves tedious supporting operations, but also eliminates the possibility of human error, saves labor costs, and improves the reliability and stability of the system. The automated agentless function eliminates the process of installing the client one by one on the source end, which not only shortens the deployment cycle, but also avoids the compatibility problem of client adaptation. The end customer is not required to provide an account and password. Automated driver conversion makes disaster recovery of the entire heterogeneous platform possible. Disaster recovery of heterogeneous platforms can also avoid security risks caused by security vulnerabilities on the same platform, significantly reducing the RTO of disaster recovery. The data acquisition method of the client agent software is to install the agent software on the source machine in a physical machine environment or a single-machine scenario such as a cloud platform virtual machine. By installing the agent software, data is directly acquired within the operating system, and the data is packaged and transmitted according to the software's unified data structure. The transmitted data can be compressed and encrypted; data transmission is carried out via the TCP / IP protocol; the client software supports running in ARM and x86 computing environments. The client software supports the current mainstream Windows distributions and mainstream Linux distributions; VMware's agentless data acquisition method uses the VADP and CBT interfaces provided by the VMware virtualization platform to achieve virtual machine data acquisition, without the need to install agent software in the virtual machine; the cloud instance agentless module data acquisition method is to use the disk created by the source machine snapshot, mount it to the data reading server of this software, read the source machine data and complete the data transmission synchronization, under the premise of cloud platform support; when incremental difference transmission is required, the read data disk will be fully verified, and after the verification is completed, only the data that differs between the source and target ends will be transmitted ; The data acquisition method for offline data synchronization is to use the source end reading PE provided by this software, start the source machine to the PE state, and the PE runs the data reading program of this software, and completes the synchronization of the source machine data offline through TCP / IP. The automated data and recovery method based on the multi-cloud platform in the present invention combines a multi-level, new system architecture for data acquisition, data transmission and data writing. Different from the traditional file transfer mode, the architecture innovatively combines the operations within the cloud host with the operations of the external cloud platform. When performing tasks such as data replication and recovery, coordinated interaction can be achieved, and the automation of the entire process can be achieved through the unified scheduling of the management console server.

[0054] This software supports any P2V and V2V whole-machine recovery, supports system conversion of heterogeneous platforms, and supports platform adaptation for multi-cloud disaster recovery, which involves three aspects.

[0055] Drive conversion: Physical machines generally use SAS drivers, while virtual machines use MPTSAS, virtio, Hyper-V, Xen and other drivers depending on the type of hypervisor. This software system can implant the corresponding driver of the target platform before the target machine is started, ensuring that its bus, disk, network card and other devices can be loaded and booted normally.

[0056] Boot type conversion, there are two operating system boot modes: Legacy and UEFI; Legacy boot mode, as the name suggests, is the traditional boot mode. It was a common boot mode before WIN8. Legacy boot mode has good system compatibility and can run any operating system, whether 32-bit or 64-bit. Legacy mode uses the MBR disk format, and its system can only be installed on MBR formatted disks; it only supports up to 4 primary partitions; it does not support hard drives larger than 2TB; and a single MBR can only store the boot record of one operating system. UEFI boot mode, UEFI is the successor to Legacy and is also the current mainstream boot mode. Compared with Legacy, UEFI has better programmability, better scalability, higher performance, and higher security. Windows systems have supported UEFI since WIN7, and WIN8 has used UEFI by default. UEFI mode uses the GPT disk format (GUID Partition Table). UEFI has no restrictions on the number of partitions and disk size, and provides a secure boot function to prevent viruses from loading during boot. The software supports the same type of boot and boot conversion of the target machine.

[0057] Network card settings: The target cloud host's network card name and IP address may be different from those of the source machine due to different hardware drivers and network architectures. Therefore, this software system will inject new network card information and IP addresses into the target system, supporting DHCP and Static respectively according to the conditions of different platforms to meet the needs of user recovery scenarios.

[0058] Recovery mode 1: Recover using the target host; In this recovery mode, operations are performed directly on the target cloud host. During this process, heavy disk IO operations such as invalid data rollback and disk creation are performed. Therefore, no matter how large the data volume of the target cloud host is, the operating system and data can be restored within minutes, minimizing RTO and providing users with the highest switching efficiency.

[0059] Key points of the specific process: Users customize the recovery task and set parameters such as the cloud host instance size, network card configuration, post-script, security group, etc. after recovery; if the replication task is interrupted, subsequent replication tasks will not be initiated again to avoid errors caused by re-initiating synchronization after recovery; decide whether to create a rollback snapshot based on the user's task configuration; perform system conversion; shut down the cloud host; apply the relevant parameters set by the user to the cloud host configuration through the cloud platform API, including but not limited to instance specifications, network card configuration, security group, elastic IP, etc.; turn on the cloud host, and the cloud host enters the user's target system; after the cloud host starts to the target system, it performs an operation, sends a successful startup message to the management console, and executes the post-script set by the user; in order to use complex and changeable actual application scenarios and to enhance the integrity of the overall recovery mode, a re-execution process is designed. The user must enable "rollback snapshot" to run the following process: the user decides to re-execute the recovery task; the target cloud host is shut down; execute the cloud host snapshot to roll back to the state before recovery; start the target host; the management console confirms that the target host is connected; re-execute the process to complete.

[0060] Recovery mode 2: Use historical point recovery; In a disaster recovery scenario, the target host's data is synchronized with the source host, so file errors on the source host are also synchronized to the target host, such as logical errors in the database or files that have been tampered with after a ransomware attack. In this case, it is necessary to restore the user system and data through historical copies. Since different cloud platforms have different snapshot capabilities, there are two main process designs for recovery mode 2: one is to achieve this by rolling back the target cloud host to a specific point in time, and the other is to create a new cloud host through the target cloud host snapshot.

[0061] Key points of the specific process: Users customize the recovery task and set parameters such as the cloud host instance size, network card configuration, post-script, security group, etc. after recovery; interrupt the replication task, and subsequent replication tasks will not be initiated again, to avoid errors caused by initiating replication synchronization after recovery; create a rollback snapshot; shut down the cloud host; roll back the target cloud host to the selected time point according to the user's task configuration; perform system conversion; apply the relevant parameters set by the user to the cloud host configuration through the cloud platform API, including but not limited to instance specifications, network card configuration, security group, elastic IP, etc.; turn on the cloud host, and the cloud host enters the user's target system; after the cloud host starts to the target system, it performs an operation, sends a successful startup message to the management console, and executes the post-script set by the user; also in order to use complex and changeable actual application scenarios, and to enhance the integrity of the overall recovery model, a re-execution process is designed: the user decides to re-execute the recovery task; the target cloud host is shut down; execute the cloud host rollback snapshot to the pre-recovery state; start the target host; the management console confirms that the target host is connected; re-execute the process to complete.

[0062] Recovery mode three: Use snapshots to implement test recovery; In complex and ever-changing actual application scenarios, users want to test and verify the target cloud host, or perform operations such as disaster recovery drills, but do not want regular replication and synchronization interruptions. To meet this requirement, the software has designed a test mode using snapshot recovery. The difference between this mode and mode 2 is that the replication and synchronization tasks will not be interrupted. The specific process points are as follows: the user customizes the recovery task and sets parameters such as the size of the restored cloud host instance, network card configuration, post-script, security group, etc.; by calling the cloud platform snapshot capability, a new cloud host is created using the selected snapshot point; the user-defined cloud host configuration, including but not limited to instance specifications, network card configuration, security group, elastic IP, etc., is applied; the cloud host is powered on; the system conversion is performed; the cloud host is restarted, and the cloud host enters the user's target system; after the cloud host is booted into the target system, an operation is performed, a startup success message is sent to the management console, and the user-defined post-script is executed. To enhance the integrity of the overall recovery model, the process for deleting a recovery task is as follows: the user decides to delete the recovery task; an instruction to delete the corresponding cloud host is sent to the cloud platform; and the deletion is completed.

[0063] The present invention uses block-level technology to obtain data from the source end, thereby improving data reading efficiency and reducing the consumption of source machine resources; faster incremental synchronization intervals improve the RPO indicator from one day to minutes; the source-end agent software is adapted to mainstream Windows versions and numerous Linux distributions; the software is compatible with ARM and x86 platforms; the developed two-way and unidirectional transmission technologies design and implement flexible transmission architectures to meet the needs of changing network scenarios; the software customizes the PE system to support startup on multiple cloud platforms and integrate data processing programs to realize information collection, data writing, and system conversion functions; after the target cloud host switches between different states, it can always ensure the retention of task information and will not cause task information loss due to PE system restart; the system conversion program is adapted to mainstream Windows versions and numerous Linux distributions, and matches drivers of different cloud platforms; the software process is combined with the cloud platform API, and the operation process is automated by matching the internal resources of the virtual machine with the cloud platform layer; so that the software can achieve diverse and rich data recovery capabilities on this basis.

[0064] In the description herein, it should be noted that relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Furthermore, the terms "include," "comprise," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

Claims

1. A data recovery method based on a multi-cloud platform, characterized in that: The following steps are involved: Step S1: Intelligent data acquisition layer construction: Based on the multimodal data recognition engine, dynamically adapt to four data acquisition channels: Lightweight client agent module, supporting cross-platform data capture in ARM / x86 heterogeneous environments; VMware agentless channel, using VADP / CBT dual interfaces to achieve accurate identification of changed data; Cloud-native agentless mode, using the cloud platform API gateway to achieve instance-level data extraction; Offline cold data synchronization module, which realizes physical machine data extraction based on PE environment booted by UEFI; Step S2: Adaptive transmission protocol stack construction: Based on the real-time network quality evaluation model, dynamically select: Bidirectional symmetric transmission mode: uses zero-copy technology to achieve efficient transmission of block-level data; One-way secure tunnel mode: builds an encrypted reverse channel based on the TLS1.3 protocol; Hybrid transmission optimization engine: supports multi-channel aggregation transmission with bandwidth adaptation; Step S3: Heterogeneous system recovery architecture: implemented through the virtual environment abstraction layer: Hardware abstraction layer construction based on ACPI table parsing; Driver automatic identification and injection engine; Network configuration intelligent migration module; System startup path reconstruction algorithm; Step S4: Spatiotemporal data management model: implemented based on cloud-native storage architecture: Timeline snapshot indexing system; Incremental change capture engine; Multi-level cache optimization strategy; Intelligent stratification of hot and cold data; Step S5: Automated recovery orchestration engine: implemented through the workflow engine: Intelligent recognition of recovery scenes; Hardware compatibility pre-detection; The recovery path is automatically generated; Post script automatic execution.

2. The data recovery method based on a multi-cloud platform according to claim 1, characterized in that: The adaptive transmission protocol stack constructed in step S2 dynamically selects a bidirectional symmetric transmission mode or a unidirectional secure tunnel mode through a real-time network quality evaluation model, and implements adaptive tuning of transmission parameters based on a bandwidth prediction algorithm, specifically including: Network quality probe module, real-time collection of packet loss rate, delay, and bandwidth parameters; Transmission mode decision engine, which selects the optimal transmission mode based on reinforcement learning algorithm; Dynamic congestion control module, supporting multiple congestion control algorithms such as BBR and CUBIC; Transmission encryption module supports AES-256 and ChaCha20 encryption algorithms.

3. The data recovery method based on a multi-cloud platform according to claim 1, characterized in that: The heterogeneous system recovery architecture in step S3 implements compatibility adaptation between different operating systems and hardware platforms through the virtual environment abstraction layer, specifically including: ACPI table parsing module to extract hardware configuration information; Driver warehouse management module, maintaining multi-version driver library; Driver injection engine, automatically matching and injecting appropriate drivers based on hardware information; The system starts the configuration reconstruction module and regenerates the boot configuration file; The network configuration migration module realizes automatic configuration of IP addresses and DNS network parameters.

4. The data recovery method based on a multi-cloud platform according to claim 1, characterized in that: The spatiotemporal data management model in step S4 implements efficient management and rapid recovery of data copies based on the cloud-native storage architecture, specifically including: The timeline index module creates a unique index for each snapshot at a time point; Incremental change capture module, which identifies data changes based on block-level comparison technology; Multi-level cache module, using LRU and LFU hybrid cache strategy to improve access efficiency; The hot and cold data tiering module automatically migrates data storage tiers based on access frequency.

5. The data recovery method based on a multi-cloud platform according to claim 1, characterized in that: The automated recovery orchestration engine in step S5 implements automated execution and monitoring of the recovery process through the workflow engine, specifically including: The recovery scene recognition module automatically matches the recovery template based on the source system characteristics; Hardware compatibility detection module, pre-detecting the compatibility between the target environment and the source system; Recovery path generation module, which generates the optimal recovery path based on graph algorithms; Execution monitoring module to track the recovery process in real time and provide visual feedback; Post-processing module supports automated execution of custom scripts.

6. The data recovery method based on a multi-cloud platform according to claim 1, characterized in that: The multimodal data recognition engine constructed by the intelligent data acquisition layer in step S1 can automatically select the optimal data acquisition channel according to the source environment type, specifically including: The environment detection module identifies the source system's operating system, hardware architecture, and virtualization platform; Channel evaluation module, which evaluates the suitability of each acquisition channel based on performance indicators and environmental characteristics; Intelligent switching module supports dynamic switching of acquisition channels during data acquisition; Data quality verification module ensures the integrity and consistency of collected data.

7. A data recovery system based on a multi-cloud platform, characterized in that: include: The intelligent data acquisition layer includes four data collection modules: client agent software, VMware agentless module, cloud instance agentless module, and source machine offline data synchronization; Adaptive transport layer, realizing two transmission modes: bidirectional symmetric transmission and unidirectional secure tunnel; Heterogeneous recovery engine supports cross-platform and cross-architecture system recovery; Spatiotemporal data management module, which enables the creation, management and recovery of data snapshots; The automated orchestration center provides visual configuration and execution of the recovery process.

8. The data recovery system based on a multi-cloud platform according to claim 7, characterized in that: The client agent software adopts a microkernel architecture design, specifically including: Data collection engine, supporting file system filter driver and volume shadow copy service; Data processing engine, which implements data compression, encryption and segmentation; Communication engine, supporting breakpoint resumption and transmission optimization; Resource management module realizes intelligent scheduling of system resources.

9. The data recovery system based on a multi-cloud platform according to claim 7, characterized in that: The cloud instance agentless module acquires data through the cloud platform API, specifically including: The cloud platform adapter layer provides a unified interface to adapt to different cloud platform APIs; Snapshot management module, which enables the creation and management of cloud disk snapshots; Difference analysis engine, which identifies incremental data based on block-level comparison technology; Data transmission optimization module supports multi-threaded parallel transmission and bandwidth control.

Citation Information

Patent Citations

  • Data migration system and method

    CN117111836A

  • Data management system of Internet of Things platform based on cloud technology

    CN118631873A

  • Cross-system data intercommunication method, platform, equipment and medium

    CN119892514A

  • AI-based enterprise business data processing system and method

    CN119917495A

  • Synchronous acquisition error detection method and system of multi-channel dynamic data acquisition system

    CN120028043A

Cited By

  • Monitoring method, electronic equipment, chip, storage medium and program product

    CN122204634A

  • A monitoring method, electronic device, chip, storage medium and program product

    CN122204634B