Cloud storage service implementation method and system for unified management of multiple types of storage

Abstract

The present invention discloses a cloud storage service implementation method and system for unified management of multiple types of storage, belonging to the field of cloud storage technology. The method specifically includes: establishing a unified abstract model for heterogeneous storage media, mapping distributed files, object storage, and block storage devices into standardized storage units, deploying a cloud blockchain, and uploading it to the chain, constructing an off-chain storage index table, and dynamically updating the index status through a loosely coupled consensus mechanism. The loosely coupled consensus includes a weighted voting algorithm based on the reputation value of the storage node and an asynchronous verification mechanism based on time and space proof, dividing data into data slices and verification slices, writing the slice information into a storage metadata registration chain and a service quality certification chain at the same time, storing data across chains, and dynamically allocating cloud storage resources; decoupling storage management functions through the blockchain, improving retrieval accuracy, effectively reducing retrieval response time, and reducing data transmission costs.

Description

Technical Field

[0001] The present invention belongs to the field of cloud storage technology, and more specifically relates to a cloud storage service implementation method and system for unified management of multiple types of storage. Background Art

[0002] With the rapid development of cloud computing, big data, and artificial intelligence technologies, the demand for storage services from businesses and individuals is becoming increasingly diverse. Traditional cloud storage systems are typically based on a single storage architecture, such as object storage, block storage, or file storage, each suited to specific application scenarios. However, in real-world business needs, different types of data, such as structured data, unstructured data, and archived data, have varying access patterns, storage cycles, and security requirements.

[0003] Traditional cloud storage architectures are typically optimized for a single storage type, making it difficult to manage different storage systems in a compatible manner. Furthermore, current cloud storage faces the following challenges: low storage resource utilization and complex data consistency and redundancy management. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention proposes a cloud storage service implementation method and system for unified management of multiple types of storage. By utilizing blockchain, it can efficiently integrate different types of storage resources such as block storage, object storage, and file storage, realize intelligent scheduling, automatic migration, and cross-storage consistency management, and improve the overall performance and availability of the storage system.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A cloud storage service implementation method for unified management of multiple types of storage includes:

[0007] Establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud-based blockchain, and upload data to the chain;

[0008] Construct an off-chain storage index table and dynamically update the index status through a loosely coupled consensus mechanism. The loosely coupled consensus mechanism includes a weighted voting algorithm based on the reputation of storage nodes and an asynchronous verification mechanism based on proof of space and time.

[0009] The data to be stored is divided into data slices and verification slices, and the slice information is written into the storage metadata registration chain and the service quality certification chain at the same time. The data to be stored is stored across chains and cloud storage resources are dynamically allocated.

[0010] Specifically, the unified abstract model for heterogeneous storage media includes:

[0011] Storage medium type identification code, access protocol conversion interface, performance characteristic vector, geographic topology coordinates and storage cost dynamic parameters.

[0012] Specifically, the deployment of cloud blockchain and on-chaining includes:

[0013] Deploy three functionally separated blockchains, including: storage metadata registration chain, access control chain, and service quality certification chain;

[0014] Putting multiple types of storage on-chain includes: writing metadata of object storage into the blockchain and automating the allocation and management of block storage resources through smart contracts.

[0015] Specifically, the off-chain storage index table is constructed and the index status is dynamically updated through a loosely coupled consensus mechanism, including:

[0016] Generate an initial index based on the deployed blockchain and reconstruct the initial index using a multi-level index structure;

[0017] Evaluate the reputation of storage nodes and perform weighted voting based on the reputation value of storage nodes, and verify asynchronously through time-space proof;

[0018] Detect and update abnormal index status.

[0019] Specifically, generating an initial index based on the deployed blockchain and reconstructing the initial index using a multi-level index structure include:

[0020] The storage node generates an initial index based on its own storage content and submits the initial index data to the off-chain storage index table;

[0021] Establish a mapping relationship between logical paths and globally unique identifiers, build a global namespace layer, and support regular expression retrieval and fuzzy matching;

[0022] Build a storage type feature matrix to define the characteristics of multiple storage policies, including object storage characteristics, file storage characteristics, and block storage characteristics. Use machine learning models to automatically match data characteristics with storage types and build a storage type mapping layer.

[0023] Address chain nodes through GeoHash geocoding and network topology maps to build a physical addressing layer;

[0024] The initial index is reconstructed using the global namespace layer, storage type mapping layer, and physical addressing layer.

[0025] Specifically, the evaluation of storage node reputation and weighted voting based on the storage node reputation value are performed through asynchronous verification through time-space proof, including:

[0026] Build a dynamic node reputation evaluation model. Based on the historical performance and real-time status of storage nodes, dynamically calculate the weight of storage nodes participating in consensus, and calculate the reputation value of storage nodes based on the weight of storage nodes participating in consensus.

[0027] Set a time interval △T, and recalculate the reputation ranking of storage nodes every △T time. The top 20% storage nodes will receive double voting weight, and the reputation value of storage nodes will be dynamically adjusted;

[0028] Generate a time-space proof verification task and split it into independent subtasks. Idle nodes obtain tasks from the independent subtasks and perform verification. The time-space proof verification task includes: a time proof task and a space proof task.

[0029] The verification results of the nodes are merged into a single proof through an aggregation algorithm, and the aggregated proof is written into the blockchain in batches.

[0030] Specifically, the detecting and updating of abnormal index status includes:

[0031] Generate a four-dimensional version vector for each changed index. The four-dimensional version vector includes: timestamp, geographic region, node ID, and operation type. Compare the version vectors of different nodes to identify abnormal index status.

[0032] Analyze whether the abnormal index status is reconcilable. If so, use CRDT semantics to automatically merge and update the abnormal index status. Otherwise, retain the high-reputation node version and store the remaining versions in the arbitration queue. Confirm the final version through the multi-signature mechanism and update the abnormal index status.

[0033] Specifically, the method of dividing the data to be stored into data slices and verification slices, writing the slice information into the storage metadata registration chain and the service quality certification chain at the same time, performing cross-chain storage of the data to be stored, and dynamically allocating cloud storage resources includes:

[0034] Divide the data to be stored into K data slices of equal or unequal length according to a preset rule. Based on the erasure code, calculate the K data slices to generate M parity slices, and satisfy K+M≤N, where N is the preset upper limit.

[0035] The basic information of each data slice and verification slice is written into the storage metadata registration chain and the service quality certification chain, and a dual-chain information synchronization mechanism is established to update the association information between the storage metadata registration chain and the service quality certification chain in real time;

[0036] The generated K data slices and M verification slices are distributed and uploaded to the cloud storage nodes. Each cloud storage node stores the data slices after receiving them and performs cross-chain verification regularly.

[0037] A cloud storage service implementation system for unified management of multiple types of storage, used to implement the cloud storage service implementation method for unified management of multiple types of storage, including: an on-chain module, an index update module, and a cross-chain storage module;

[0038] The on-chain module is used to establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud blockchain, and perform on-chain.

[0039] The index update module is used to build an off-chain storage index table and dynamically update the index status through a loosely coupled consensus mechanism. The loosely coupled consensus mechanism includes a weighted voting algorithm based on the reputation of storage nodes and an asynchronous verification mechanism based on time-space proof.

[0040] The cross-chain storage module is used to divide the data to be stored into data slices and verification slices, write the slice information into the storage metadata registration chain and the service quality certification chain at the same time, perform cross-chain storage on the data to be stored, and dynamically allocate cloud storage resources.

[0041] Specifically, the index update module includes: an index reconstruction unit, a voting verification unit, and an anomaly detection and update unit;

[0042] The index reconstruction unit is used to reconstruct the initial index using a multi-level index structure;

[0043] The voting verification unit is used to evaluate the reputation of the storage node and perform weighted voting based on the reputation value of the storage node, and asynchronously verify through time-space proof;

[0044] The anomaly detection and updating unit is used to detect an anomaly index, and process and update the detected anomaly index.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. This invention proposes a cloud storage service implementation method for unified management of multiple types of storage. Based on a unified abstract model, this method standardizes the mapping of heterogeneous media such as distributed files, object storage, and block storage, constructs a scalable unified storage management architecture, deploys multiple functionally separated blockchains, and combines them with an off-chain indexing mechanism. This method dynamically updates index status through loosely coupled consensus, evaluates the reputation of storage nodes, and performs weighted voting based on the reputation values ​​of storage nodes. This method uses asynchronous verification through proof-of-spacetime verification. When an index anomaly occurs, it analyzes whether the abnormal index status is reconcilable and updates the abnormal index status. This method improves the efficiency and reliability of cloud storage and is suitable for multi-tenant, multi-scenario cloud environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1Flowchart of the method for implementing the cloud storage service with unified management of multiple types of storage provided by the present invention;

[0048] Figure 2 The loosely coupled consensus flow chart provided by the present invention;

[0049] Figure 3 This is a diagram of the system architecture for implementing the cloud storage service with unified management of multiple storage types provided by the present invention. DETAILED DESCRIPTION

[0050] The present application is described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but are not intended to limit the present application in any form. It should be noted that those skilled in the art may make several variations and improvements without departing from the scope of the present application. These all fall within the scope of protection of the present application.

[0051] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0052] It should be noted that, if there is no conflict, the various features in the embodiments of the present application can be combined with each other and are all within the scope of protection of the present application. In addition, although the functional modules are divided in the device schematic and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a different order than the module division in the device or the order in the flow chart. In addition, the "

[0053] The words "first", "second", "third", etc. do not limit the data and execution order, but only distinguish the same or similar items with basically the same functions and effects.

[0054] Unless otherwise defined, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which this application belongs. The terms used in this specification and in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" as used in this specification includes any and all combinations of one or more of the relevant listed items.

[0055] Example 1

[0056] See also Figure 1-Figure 2 The present invention provides an embodiment of a method for implementing a cloud storage service with unified management of multiple types of storage, comprising the following specific steps:

[0057] Step S1: Establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud blockchain, and upload them to the chain.

[0058] In this implementation step, in order to address the storage issues of multiple heterogeneous storage media in a cloud environment, a unified abstract model is constructed to achieve unified identification and management of storage resources. Specifically, the basic characteristics of storage types such as distributed file systems, object storage, and block storage are first identified, including their access protocols, storage structures, capacity granularity, and performance attributes. Then, through abstract modeling, heterogeneous resources are uniformly mapped into standardized storage units, and each storage unit is given a unified type identifier, metadata structure, and interface, so that different storage types have consistent logical representations.

[0059] Deploying the blockchain and uploading it to the chain in step S1 includes:

[0060] Deploy three functionally separated blockchains, including: storage metadata registration chain, access control chain, and service quality certification chain.

[0061] Among them, the storage metadata registration chain is responsible for the basic information management and ownership confirmation of multiple types of storage resources, the access control chain is used to implement fine-grained permission management and access behavior auditing, and the service quality certification chain is used to quantify storage service quality and incentivize nodes.

[0062] Putting multiple types of storage on the chain includes: writing object storage metadata (such as file name, hash value, storage location, access rights) into the blockchain, automatically allocating and managing block storage resources through smart contracts, building a distributed file system based on the blockchain, and storing file shards in the node network.

[0063] In this embodiment, different types of storage are uploaded to the chain in different ways, and then the cross-chain technology of blockchain is used to achieve storage resource sharing and collaboration between different chains. In addition, the blockchain has the characteristic that data cannot be tampered with, and the authenticity of the data is verified through smart contracts to prevent the data from being maliciously tampered with or forged.

[0064] Step S2: Construct an off-chain storage index table and dynamically update the index status through a loosely coupled consensus mechanism. The loosely coupled consensus includes a weighted voting algorithm based on the reputation value of the storage node and an asynchronous verification mechanism based on time-space proof.

[0065] The specific steps of step S2 include:

[0066] Step S201: Generate an initial index based on the deployed blockchain and reconstruct the initial index using a multi-level index structure.

[0067] The specific steps of step S201 are:

[0068] Step S2011: The storage node generates an initial index based on its own storage content and submits the initial index data to the off-chain storage index table.

[0069] In this embodiment, after completing data storage, the storage node generates a unique fingerprint for each file / block through a hash algorithm and extracts key metadata (such as file name, size, storage time, and node ID). The key metadata information constitutes the core of the initial index. The node generates a localized index structure based on its own storage structure, such as a B+ tree and an inverted index, which only records the local physical location and logical relationship of the data, avoiding complex global calculations and reducing node load.

[0070] The off-chain storage index table serves as a cache layer for the index, receiving the initial index submitted by each node. The table structure is data hash + node ID → metadata + storage location. It supports fast queries and maintains index freshness through a regular synchronization mechanism, which can improve retrieval efficiency and reduce storage costs.

[0071] Step S2012: Establish a mapping relationship between the logical path and the globally unique identifier, build a global namespace layer, and support regular expression retrieval and fuzzy matching.

[0072] In this embodiment, for the unified naming and positioning of multi-source heterogeneous data, a mapping relationship between logical paths and globally unique identifiers is constructed, and a global namespace layer is established based on this. Specifically, each data path is first assigned a set of logical identifiers, which consist of the user directory structure, tenant ID, storage type and file metadata; then, through hash mapping or encoding rules, the logical path is mapped to a non-repeatable globally unique identifier to ensure uniqueness and locatability across storage types, nodes or chains. Based on this mapping result, a global namespace layer is constructed.

[0073] Step S2013: Construct a storage type feature matrix to define the features of multiple storage types, including object storage features, file storage features, and block storage features. Automatically match data features with storage types through a machine learning model to construct a storage type mapping layer.

[0074] In this embodiment, a storage type feature matrix is ​​constructed to address the performance differences and business adaptation requirements of multiple types of storage media. Automatic matching of data features with storage types is achieved based on a machine learning model, thereby forming a storage type mapping layer. Specifically, structured modeling is performed on storage types such as object storage, file storage, and block storage, and their characteristic parameters, such as access latency, throughput, data consistency level, scalability, adaptation interface, and cost coefficient, are extracted to form a structured storage type feature matrix. During the data access process, real-time analysis is performed on the stored data to extract content and behavioral features, including data size, access frequency, lifecycle, modification frequency, and concurrency requirements, as input feature vectors. Using a trained classification model (such as a decision tree, random forest, support vector machine, etc.), the data features are matched with the storage feature matrix to predict the optimal storage type. The matching results are recorded in the storage type mapping layer for storage scheduling and policy selection.

[0075] Step S2014: Address the chain nodes through GeoHash geocoding and network topology maps to build a physical addressing layer;

[0076] In this embodiment, a combination of GeoHash geographic coding and network topology maps is used to address chain nodes, and a physical addressing layer is constructed on this basis. GeoHash is a spatial hash algorithm that encodes geographic coordinates into sortable strings. By configuring physical location information, such as longitude and latitude, for each chain node, the node's physical address is converted into the corresponding GeoHash code, thereby quickly clustering and locating the nodes, and improving the spatial correlation between node selection and data scheduling.

[0077] On the other hand, the network connection relationship between chain nodes is modeled and a network topology map is constructed. This map generates a weighted connection graph by collecting link bandwidth, latency, packet loss rate and other indicators between chain nodes in real time, and jointly indexes it with the GeoHash encoding results to form a physical addressing layer.

[0078] Step S2015: reconstruct the initial index using the global namespace layer, storage type mapping layer, and physical addressing layer.

[0079] In this embodiment, the initial index is reconstructed by integrating the global namespace layer, the storage type mapping layer, and the physical addressing layer. The initial index is generated by the storage node based on its local storage content, recording basic path information and data element attributes to support unified scheduling of heterogeneous storage in the cloud environment. Specifically, the correspondence between data characteristics and storage strategies is analyzed based on the storage type mapping layer, and the most suitable storage type is matched for each index item. Combined with the GeoHash location coding and topological connection information provided by the physical addressing layer, the optimal physical distribution and routing information is added to the data index. Finally, the information of these three levels is integrated to generate a reconstructed index table with logical naming, type adaptation, and physical positioning capabilities.

[0080] The benefits of index reconstruction include: improving data retrieval efficiency and obtaining required data more quickly, especially for application scenarios with high real-time requirements, such as online video playback and real-time data analysis; optimizing storage resource utilization and evenly distributing data across different nodes, thus avoiding performance degradation caused by excessive load on certain nodes.

[0081] Step S202: Evaluate the reputation of the storage node and perform weighted voting based on the reputation value of the storage node, and verify asynchronously through time-space proof;

[0082] The specific steps of step S202 are:

[0083] Step S2021: Build a dynamic node reputation evaluation model, dynamically calculate the weight of the storage node participating in the consensus based on the historical performance and real-time status of the storage node, and calculate the reputation value of the storage node based on the weight of the storage node participating in the consensus.

[0084] Specifically, we first collect 12 indicators such as node availability, response delay, and data integrity in real time. By setting multi-dimensional scoring rules, we standardize the above indicators and aggregate them into the node's current reputation scoring factor according to the weight. We calculate its participation weight in the consensus algorithm, and use the calculated consensus weight as input to further map it to a specific reputation value.

[0085] Step S2022: Set a time interval ΔT, and recalculate the reputation ranking of the storage nodes every ΔT. The top 20% of the storage nodes will receive double voting weight, and the reputation value of the storage nodes will be dynamically adjusted.

[0086] The benefits of step S2022 are: defense against malicious nodes, the influence of low-reputation nodes is automatically suppressed; resource optimization: high-quality nodes obtain more tasks, improving overall efficiency.

[0087] Step S2023: Generate a space-time proof verification task and break it down into independent subtasks. Idle nodes obtain tasks from the independent subtasks and perform verification.

[0088] The time-space proof verification task includes: a time proof task, in which a timestamp certificate with a digital signature is generated by a trusted time source every 15 minutes. The certificate includes: the current timestamp, a random number, a hash value of the previous certificate, and a signature of the time source. The storage node must submit a new certificate within two time windows after the end of the current time window. The new certificate includes: the current timestamp certificate hash, the previous certificate hash, and the storage node signature.

[0089] The time-space proof verification task also includes: a space proof task, in which part of the total amount of data shards is randomly sampled for verification in each round.

[0090] When idle nodes obtain tasks from independent subtasks, they give priority to stealing tasks from nodes with heavier loads to achieve load balancing.

[0091] Step S2024: The verification results of the nodes are merged into a single certificate through an aggregation algorithm, and the aggregated certificates are written into the blockchain in batches.

[0092] In this embodiment, efficient asynchronous verification of storage proof is achieved through time window constraints, random sharding verification, distributed task scheduling and multi-node consensus, which reduces verification overhead while ensuring data integrity.

[0093] Step S203: Detect and update the abnormal index status.

[0094] The specific steps of step S203 are:

[0095] Step S2031: Generate a four-dimensional version vector for each changed index. The four-dimensional version vector includes: timestamp, geographic region, node ID, and operation type. Compare the version vectors of different nodes to identify abnormal index status.

[0096] In this embodiment, the timestamp in the four-dimensional version vector records the absolute time of the index change and is used to construct an operation sequence chain. The geographic area in the four-dimensional version vector uses GeoHash encoding at the physical addressing layer to mark the geographic area where the operation is initiated. The node ID in the four-dimensional version vector is associated with the node identity in the dynamic reputation evaluation model and records the specific node initiating the operation. The operation types in the four-dimensional version vector include addition, deletion, modification, and query.

[0097] The comparison mechanism for anomaly detection includes: temporal consistency verification: comparing the timestamps of version vectors of different nodes; geographic-node binding verification: verifying the matching relationship between node ID and geographic location to detect cross-regional operation anomalies; and conflict resolution marking: when multiple version vectors of the same index have similar timestamps, are geographically dispersed, or have conflicting operation types, they are marked as an abnormal state requiring manual intervention.

[0098] Specifically, it accurately locates distributed index anomalies, solves data synchronization conflicts and node malicious operations, and reduces operation and maintenance costs.

[0099] Step S2032: Analyze whether the abnormal index status is reconcilable. If so, use CRDT semantics to automatically merge and update the abnormal index status. Otherwise, retain the high-reputation node version and store the remaining versions in the arbitration queue. Confirm the final version through the multi-signature mechanism and update the abnormal index status.

[0100] In this embodiment, the arbitration queue mechanism: abnormal versions are stored in the queue, accompanied by a four-dimensional vector and the cause of the conflict; the multi-signature process is triggered: the system automatically selects the topN nodes whose total credibility is greater than the preset threshold (such as N=5, based on the dynamic weight of S2021), and at least 3 nodes are required to sign to confirm the final version; multi-signature semantics: the signing node needs to verify the rationality of the four-dimensional vector, perform a secondary check on the deletion operation, and check whether there is backup storage to support the deletion. After the signature is passed, the final version of the vector is added with the arbitration timestamp and the signing node set, and the old version is marked as a historical conflict and archived.

[0101] Step S3: Split the data to be stored into data slices and verification slices, write the slice information into the storage metadata registration chain and the service quality certification chain at the same time, store the data to be stored across chains, and dynamically allocate cloud storage resources.

[0102] The specific steps of step S3 are:

[0103] Step S301: Divide the data to be stored into K data slices of equal or unequal length according to preset rules. Based on the erasure code, calculate the K data slices to generate M check slices, and satisfy K+M≤N, where N represents the preset upper limit, aiming to balance storage redundancy and resource consumption and prevent the additional burden caused by excessive redundancy.

[0104] In this embodiment, the data to be stored includes different formats, and each data piece is accompanied by a unique identifier, such as a piece number, a hash value, etc. The check piece is used to realize data recovery when part of the data piece is lost or damaged, store error correction information, and also record the parameters of the generation process to ensure that the original data can be restored when the data is recovered.

[0105] Step S302: Write the basic information of each data slice and verification slice into the storage metadata registration chain and the service quality certification chain, establish a dual-chain information synchronization mechanism, and update the association information between the storage metadata registration chain and the service quality certification chain in real time.

[0106] In this embodiment, the basic information of the data slice and the verification slice includes: unique identifier, size, generation time, encryption parameters, hash value, etc. The information written into the service quality proof chain includes: node information of the storage shard, storage latency, node response time, health status, historical fault records, etc. Through regular verification and cross-validation, the consistency of records on the two chains is ensured, and data inconsistency problems caused by node failure or information transmission abnormalities are discovered in time, thereby realizing cross-chain storage.

[0107] Step S303: The generated K data slices and M verification slices are distributed and uploaded to the cloud storage nodes. Each cloud storage node stores the data slices after receiving them and performs cross-chain verification regularly.

[0108] In this embodiment, a cross-chain verification process is initiated periodically or in a triggered manner to compare the shard information recorded on the storage metadata registration chain with the storage status of the corresponding node in the service quality certification chain. The verification process includes multi-dimensional indicators such as data integrity verification, storage latency assessment, and data access statistics to ensure that each storage node provides high-quality storage services as agreed. When it is detected that a cloud storage node data slice is lost or damaged, the problem node is locked through the service quality certification chain, and the verification slice information on other nodes in the storage network is called to start data recovery. Based on the verification slice and the remaining data slices, the lost data slices are reconstructed through the erasure code algorithm to complete the data recovery process and ensure the overall data integrity of the system.

[0109] This application effectively solves the problem of multi-type storage by putting storage resources on-chain. It uses dual-chain cross-chain storage collaboration and multi-level index reconstruction including global namespace layer, storage type mapping layer and physical addressing layer. Compared with traditional cloud storage or single-chain storage, it realizes cross-verification of storage information and service quality, breaking through the scalability limitations of blockchain. At the same time, it uses loosely coupled consensus, combined with reputation weighting and time-space proof, to reduce consensus latency while ensuring security. It uses four-dimensional version vectors to incorporate geographic information and operational behavior into version management, solving the spatial conflict problem in distributed index synchronization.

[0110] Example 2

[0111] See also Figure 3 , another embodiment provided by the present invention: a cloud storage service implementation system for unified management of multiple types of storage, including: an on-chain module, an index update module and a cross-chain storage module;

[0112] The on-chain module is used to establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud blockchain, and perform on-chain.

[0113] The index update module is used to build an off-chain storage index table and dynamically update the index status through a loosely coupled consensus mechanism. The loosely coupled consensus mechanism includes a weighted voting algorithm based on the reputation of storage nodes and an asynchronous verification mechanism based on time-space proof.

[0114] The cross-chain storage module is used to divide the data to be stored into data slices and verification slices, write the slice information into the storage metadata registration chain and the service quality certification chain at the same time, perform cross-chain storage on the data to be stored, and dynamically allocate cloud storage resources.

[0115] Index update module, including: index reconstruction unit, voting verification unit and anomaly detection and update unit;

[0116] The index reconstruction unit is used to reconstruct the initial index using a multi-level index structure;

[0117] The voting verification unit is used to evaluate the reputation of the storage node and perform weighted voting based on the reputation value of the storage node, and asynchronously verify through time-space proof;

[0118] The anomaly detection and updating unit is used to detect an anomaly index, and process and update the detected anomaly index.

[0119] In addition, the parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0120] The above-described specific embodiments further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is merely a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for implementing a cloud storage service with unified management of multiple types of storage, characterized in that: include: Establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud-based blockchain, and upload data to the chain; Construct an off-chain storage index table and dynamically update the index status using a loosely coupled consensus mechanism. The loosely coupled consensus mechanism includes a weighted voting algorithm based on the reputation of storage nodes and an asynchronous verification mechanism based on proof of space and time. Split the data to be stored into data slices and verification slices, write the slice information into the storage metadata registration chain and the service quality certification chain at the same time, store the data to be stored across chains, and dynamically allocate cloud storage resources; The construction of the off-chain storage index table and the dynamic update of the index status through the loosely coupled consensus mechanism include: Generate an initial index based on the deployed blockchain and reconstruct the initial index using a multi-level index structure; Evaluate the reputation of storage nodes and perform weighted voting based on the reputation value of storage nodes, and verify asynchronously through time-space proof; Detect and update abnormal index status; The initial index is generated according to the deployed blockchain, and the initial index is reconstructed using a multi-level index structure, including: The storage node generates an initial index based on its own storage content and submits the initial index data to the off-chain storage index table; Establish a mapping relationship between logical paths and globally unique identifiers, build a global namespace layer, and support regular expression retrieval and fuzzy matching; Build a storage type feature matrix to define the characteristics of multiple storage policies, including object storage characteristics, file storage characteristics, and block storage characteristics. Use machine learning models to automatically match data characteristics with storage types and build a storage type mapping layer. Address chain nodes through GeoHash geocoding and network topology maps to build a physical addressing layer; Reconstruct the initial index using the global namespace layer, storage type mapping layer, and physical addressing layer; The evaluation of storage node reputation and weighted voting based on the storage node reputation value are performed through asynchronous verification through time-space proof, including: Build a dynamic node reputation evaluation model. Based on the historical performance and real-time status of storage nodes, dynamically calculate the weight of storage nodes participating in consensus, and calculate the reputation value of storage nodes based on the weight of storage nodes participating in consensus. Set a time interval △T, and recalculate the reputation ranking of storage nodes every △T time. The top 20% storage nodes will receive double voting weight, and the reputation value of storage nodes will be dynamically adjusted; Generate a time-space proof verification task and split it into independent subtasks. Idle nodes obtain tasks from the independent subtasks and perform verification. The time-space proof verification task includes: a time proof task and a space proof task. The verification results of the nodes are combined into a single proof through an aggregation algorithm, and the aggregated proof is written into the blockchain in batches; The detecting and updating of abnormal index status includes: Generate a four-dimensional version vector for each changed index. The four-dimensional version vector includes: timestamp, geographic region, node ID, and operation type. Compare the version vectors of different nodes to identify abnormal index status. Analyze whether the abnormal index status is reconcilable. If so, use CRDT semantics to automatically merge and update the abnormal index status. Otherwise, retain the high-reputation node version and store the remaining versions in the arbitration queue. Confirm the final version through the multi-signature mechanism and update the abnormal index status.

2. The method for implementing a cloud storage service with unified management of multiple types of storage according to claim 1, wherein: The unified abstract model for heterogeneous storage media includes: Storage medium type identification code, access protocol conversion interface, performance characteristic vector, geographic topology coordinates and storage cost dynamic parameters.

3. The method for implementing a cloud storage service with unified management of multiple types of storage according to claim 1, wherein: The deployment of cloud blockchain and on-chaining includes: Deploy three functionally separated blockchains, including: storage metadata registration chain, access control chain, and service quality certification chain; Putting multiple types of storage on-chain includes: writing metadata of object storage into the blockchain and automating the allocation and management of block storage resources through smart contracts.

4. The method for implementing a cloud storage service with unified management of multiple types of storage according to claim 1, wherein: The method of dividing the data to be stored into data slices and verification slices, writing the slice information into the storage metadata registration chain and the service quality certification chain at the same time, performing cross-chain storage of the data to be stored, and dynamically allocating cloud storage resources includes: Divide the data to be stored into K data slices of equal or unequal length according to a preset rule. Based on the erasure code, calculate the K data slices to generate M parity slices, and satisfy K+M≤N, where N is the preset upper limit. The basic information of each data slice and verification slice is written into the storage metadata registration chain and the service quality certification chain, and a dual-chain information synchronization mechanism is established to update the association information between the storage metadata registration chain and the service quality certification chain in real time; The generated K data slices and M verification slices are distributed and uploaded to the cloud storage nodes. Each cloud storage node stores the data slices after receiving them and performs cross-chain verification regularly.

5. A cloud storage service implementation system for unified management of multiple types of storage, used to implement the cloud storage service implementation method for unified management of multiple types of storage according to any one of claims 1 to 4, characterized in that: include: On-chain module, index update module and cross-chain storage module; The on-chain module is used to establish a unified abstract model for heterogeneous storage media, map distributed files, object storage, and block storage devices into standardized storage units, deploy cloud blockchain, and perform on-chain. The index update module is used to build an off-chain storage index table and dynamically update the index status through a loosely coupled consensus mechanism. The loosely coupled consensus mechanism includes a weighted voting algorithm based on the reputation of storage nodes and an asynchronous verification mechanism based on time-space proof. The cross-chain storage module is used to divide the data to be stored into data slices and verification slices, write the slice information into the storage metadata registration chain and the service quality certification chain at the same time, perform cross-chain storage on the data to be stored, and dynamically allocate cloud storage resources.

6. The cloud storage service implementation system for unified management of multiple types of storage according to claim 5, characterized in that: The index update module includes: an index reconstruction unit, a voting verification unit and an anomaly detection and update unit; The index reconstruction unit is used to reconstruct the initial index using a multi-level index structure; The voting verification unit is used to evaluate the reputation of the storage node and perform weighted voting based on the reputation value of the storage node, and asynchronously verify through time-space proof; The anomaly detection and updating unit is used to detect anomaly indexes, and process and update the detected anomaly indexes.

Images