An offshore data distributed storage system based on blockchain fast addressing technology

The offshore data distributed storage system based on blockchain fast addressing technology solves the problems of data loss and privacy leakage in cloud storage, achieves high reliability and security, and meets the storage needs of offshore data centers.

CN115348275BActive Publication Date: 2024-11-05YINUO (SHENZHEN) OFFSHORE CLOUD DATA TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210809515.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-07-11
Publication Date
2024-11-05
Estimated Expiration
2042-07-11

AI Technical Summary

Technical Problem

In existing technologies, cloud storage suffers from problems such as data loss, privacy leaks, and excessive centralization, making it difficult to effectively preserve the ever-increasing amount of data and ensure its reliable and secure storage.

Method used

An offshore data distributed storage system based on blockchain fast addressing technology is adopted, including a distributed storage layer, a distributed network layer, a blockchain service layer and an application layer. It achieves high availability and security through polymorphic data storage and P2P network, and provides object storage, block storage and file system storage services by adopting multi-replica storage and consortium blockchain network technology.

Benefits of technology

It achieves high data reliability and security, avoids data loss, optimizes access speed and efficiency, meets the storage needs of offshore data centers, reduces costs, and improves data privacy protection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115348275B_ABST
    Figure CN115348275B_ABST
Patent Text Reader

Abstract

The present invention discloses an offshore data distributed storage system based on blockchain fast addressing technology, which consists of a distributed storage layer, a distributed network layer, a blockchain service layer and an application layer; the present invention utilizes distributed blockchain technology to realize a low-cost and high-availability distributed storage system, balancing cost and high availability. The system of the present invention has high reliability. Each copy of data will exist in the system in the form of three copies, and each copy of data is divided into multiple blocks and stored in the distributed network of the system. The damage of at most 2 copies of data will not cause the loss of data files; the present invention utilizes the distributed storage, non-tamperability, traceability, multi-party maintenance, cross-validation and other characteristics of blockchain technology, which can effectively define data ownership, track and supervise data circulation, realize safe co-governance, optimize the blockchain addressing algorithm, and meet the actual business needs of offshore data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of big data and distributed storage technology, and relates to a distributed storage system, and specifically to an offshore data distributed storage system based on blockchain fast addressing technology. Background Art

[0002] With the development of new technologies and new business scenarios such as computing technology, virtual reality, artificial intelligence, and digital transformation of government and enterprises, human society has gradually entered the digital age, and along with it, the amount of data in cyberspace has exploded. There is no doubt that relying on technologies such as big data and machine learning, the value of data can be mined from these massive data, but how to preserve the growing data is a rather tricky problem.

[0003] In order to overcome the shortcomings of traditional independent data storage media in terms of reliability and easy loss, cloud storage technology has been widely accepted and applied. However, over-centralized cloud storage also faces the problems of data loss and privacy leakage. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an offshore data distributed storage system based on blockchain fast addressing technology, which has the storage efficiency and usage experience of cloud storage, and at the same time effectively solves the problem of excessive centralization of current data storage, ensuring the trusted and secure storage of privacy and important data.

[0005] The technical solution adopted by the present invention is: an offshore data distributed storage system based on blockchain fast addressing technology, which consists of a distributed storage layer, a distributed network layer, a blockchain service layer and an application layer;

[0006] The distributed storage layer adopts a polymorphic data storage access interface to provide three types of storage services to upper-layer applications in the form of APIs: object storage, block storage, and file system storage; each piece of data will exist in the system in the form of N copies, and each piece of data will be divided into multiple blocks and stored in the distributed network of the system; where N is a preset value;

[0007] The object storage, the upper layer application calls the object storage service access interface, the object storage module automatically completes the mapping of the user object storage space to the underlying unified distributed object storage space, and the user data is stored in the distributed object storage engine in the form of object data;

[0008] The block storage, the upper layer application calls the block storage service access interface, the block storage module automatically completes the user's block storage operation, data read and write operations to the underlying unified distributed object storage space mapping, the user's data on the block storage is ultimately stored in the distributed object storage engine in the form of object data;

[0009] The file system storage provides a POSIX file system interface compatible with the POSIX standard, and supports both kernel file system and user space file system modes. The upper layer application calls the POSIX file system interface, and the POSIX file system module and the POSIX file system metadata manager jointly complete the mapping of the user's POSIX file operation to the underlying unified distributed object storage space. The user's data in the POSIX file system is ultimately stored in the distributed object storage engine in the form of object data; the POSIX file system metadata manager is used for mapping and conversion from the POSIX file system space to the object storage space;

[0010] The distributed network layer adopts the alliance chain network technology based on the P2P network as the underlying layer to realize a highly available distributed network;

[0011] The blockchain service layer uses a consortium chain to provide blockchain basic services for the entire network;

[0012] The consortium chain includes a member organization MemberOrg and an ordering organization OrdererOrg. The member organization MemberOrg is a participant in the consortium chain and an entity that needs to upload data to the chain. It will maintain one or more Peer nodes in the network. The ordering organization OrdererOrg is responsible for transaction ordering and block creation in the consortium chain. Each organization contains a member identity MSP, which fully represents an organization in the network to participate in network consensus.

[0013] The application layer is used to provide data sharing and data transaction services for offshore data;

[0014] The nodes of the system are divided into three roles: the key holding node is responsible for keeping the private key of the chameleon hash, the computing node is responsible for calculating new blocks, and the consensus node is responsible for voting on editing requests and editing results; when erroneous, harmful, or expired information is detected, a node issues a blockchain editing request for the first time and broadcasts it to the entire network; all nodes in the system verify the request and check the rationality of the request; when the number of nodes that agree to the request reaches a predetermined number, the request is approved, otherwise the request is rejected; the system selects a node to edit the blockchain according to a predetermined strategy; the node with the right to modify edits the block information according to the request content and broadcasts it to the entire network; the remaining nodes verify the operation results, and update their own ledgers after the verification is passed.

[0015] The beneficial effects of the present invention include:

[0016] (1) The system of the present invention has high reliability. Each piece of data will exist in three copies on the system, and each piece of data will be divided into multiple blocks and stored in the distributed network of the system. The damage of at most two copies of data will not cause the loss of data files.

[0017] (2) The present invention adopts an original technical solution to realize distributed storage, which can replace some commercial storage scenarios under the premise of high availability and performance;

[0018] (3) The present invention optimizes the traditional blockchain network addressing algorithm and improves access speed and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the system framework of an embodiment of the present invention;

[0020] Figure 2 This is a schematic diagram of a distributed storage layer framework according to an embodiment of the present invention;

[0021] Figure 3 A schematic diagram of subtree partitioning of node 0011 in a binary tree according to an embodiment of the present invention;

[0022] Figure 4 This is a schematic diagram of locating a target node by ID in a binary tree according to an embodiment of the present invention;

[0023] Figure 5 A schematic diagram of the relationship between online duration and probability of continuing online in a Gnutella network according to an embodiment of the present invention;

[0024] Figure 6 A schematic diagram of a query process when a=1 in a binary tree according to an embodiment of the present invention;

[0025] Figure 7 A schematic diagram of adding nodes to a binary tree according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to facilitate ordinary technicians in the field to understand and implement the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the implementation examples described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0027] Please see Figure 1 and Figure 2 , the present invention provides an offshore data distributed storage system based on blockchain fast addressing technology, which consists of a distributed storage layer, a distributed network layer, a blockchain service layer and an application layer;

[0028] The distributed storage layer of this embodiment adopts a polymorphic data storage access interface to provide three types of storage services, namely object storage, block storage, and file system storage, to upper-layer applications in the form of APIs; each piece of data will exist in the form of N copies on the system, and each piece of data will be divided into multiple blocks and stored in a distributed network of the system; where N is a preset value;

[0029] In the object storage of this embodiment, the upper layer application calls the object storage service access interface, and the object storage module automatically completes the mapping of the user object storage space to the underlying unified distributed object storage space. The user data is stored in the distributed object storage engine in the form of object data.

[0030] In the block storage of this embodiment, the upper-layer application calls the block storage service access interface, and the block storage module automatically completes the mapping of the user's block storage operations and data read and write operations to the underlying unified distributed object storage space. The user's data on the block storage is ultimately stored in the distributed object storage engine in the form of object data; it supports functions such as snapshots and cloning.

[0031] The file system storage of this embodiment provides a POSIX file system interface compatible with the POSIX standard, and supports both kernel file system and user space file system (FUSE) modes. The upper layer application calls the POSIX file system interface, and the POSIX file system module and the POSIX file system metadata manager (responsible for mapping and converting the POSIX file system space to the object storage space) jointly complete the mapping of the user's POSIX file operation to the underlying unified distributed object storage space. The user's data in the POSIX file system is ultimately stored in the distributed object storage engine in the form of object data; the POSIX file system metadata manager is used for mapping and converting the POSIX file system space to the object storage space;

[0032] The distributed network layer of this embodiment adopts the alliance chain network technology based on the P2P network as the underlying layer to realize a highly available distributed network;

[0033] The blockchain service layer of this embodiment uses a consortium chain to provide blockchain basic services for the entire network;

[0034] The consortium chain of this embodiment includes a member organization MemberOrg and an ordering organization OrdererOrg; the member organization MemberOrg is a participant in the consortium chain, an entity that needs to upload data to the chain, and will maintain one or more Peer nodes in the network; the ordering organization OrdererOrg is responsible for transaction ordering and block creation in the consortium chain; it is often operated and maintained by the initiator or regulator of the network (created by an offshore data center and supervised by a superior authority); each organization contains a member identity MSP, which can be understood as the identity of the organization, and fully represents an organization in the network to participate in the network consensus;

[0035] The application layer is used to provide data sharing and data transaction services for offshore data, such as offshore IT service outsourcing management applications, offshore data processing applications, offshore regulatory platform applications, etc.

[0036] The nodes of the system are divided into three roles: the key holding node is responsible for keeping the private key of Chameleon Hash, the computing node is responsible for calculating new blocks, and the consensus node is responsible for voting on editing requests and editing results. This distributed management method enhances the security of editing behavior; when erroneous, harmful, or expired information is detected, a node issues a blockchain editing request for the first time and broadcasts it to the entire network; all nodes in the system verify the request and check the rationality of the request; when the number of nodes that agree to the request reaches a predetermined number, the request is considered to be passed, otherwise the request is rejected; the system selects a node to edit the blockchain according to a predetermined strategy; nodes with modification rights edit block information according to the request content and broadcast it to the entire network; the remaining nodes verify the operation results, and update their own ledgers after verification; this not only retains the distributed characteristics of the platform, but also ensures the effective governance of on-chain data.

[0037] This embodiment is based on the alliance chain network technology with P2P network as the underlying layer, and is based on Kademlia network and uses Kademlia protocol (referred to as Kad) to implement the distributed hash table DHT; all nodes in the Kademlia network are regarded as leaves of a split binary tree, and the position of each node is uniquely determined by the shortest prefix of its ID value, and each node has a 160-bit ID value as an identifier; the entire network node state is represented by a binary tree, and each node ID value is represented in binary form, and the nth bit of the binary corresponds to the nth layer of the binary tree; if the bit is 1, enter the left subtree, and if it is 0, enter the right subtree; finally, the leaf node on the binary tree corresponds to a certain node;

[0038] The Kademlia protocol includes four remote RPC operations: PING, STORE, FIND_NODE, and FIND_VALUE. The PING operation is to detect a node to determine whether it is still online. The STORE operation is to notify a node to store a value.<key,value> Yes, for future query needs, where Key is the unique identifier of the node and value is the payload (the business carried); the FIND_NODE operation uses a 160-bit ID as a parameter. The recipient of this operation returns the (IP address, UDP port, Node ID) information of the K nodes it knows that are closer to the target ID, where IP address represents the IP address of the node Node, UDP represents the port number of the Node node, and Node ID represents the 160-bit ID number of the unique identifier of the Node node; the IND_VALUE operation is similar to the FIND_NODE operation, except that it only needs to return the (IP address, UDP port, Node ID) information of a node. If the recipient of this operation receives a STORE operation with the same key, it will directly return the stored value.

[0039] In the Kad network, all nodes are treated as leaves of a binary tree, and the position of each node is uniquely determined by the shortest prefix of its ID value. Each node has a 160-bit ID value as an identifier. The entire network node status can be represented by a binary tree.

[0040] The ID value is represented in binary form; the nth bit of the binary corresponds to the nth layer of the binary tree; if the bit is 1, it enters the left subtree, and if it is 0, it enters the right subtree; finally, the leaf node on the binary tree corresponds to a certain node. Each node is represented by 160 bits of binary, so there are at most 160 layers, but in reality there are not so many layers, because the shortest unique prefix is ​​used, so the leaf node can appear in any number of layers.

[0041] Please see Figure 3 , the split binary tree of this embodiment can decompose any node into a series of continuous subtrees that do not contain itself; the highest level subtree is composed of the other half of the whole tree that does not contain itself; the next level subtree is composed of the remaining half that does not contain itself; and so on, until the whole tree is split; for each node, when it completes the subtree splitting from its own perspective, it will get n subtrees; for each subtree, if one of the nodes inside is known, the node can use these n nodes for recursive routing to reach any node in the entire binary tree.

[0042] Please see Figure 3, the dotted line contains the subtrees of node 0011, and the entire subtree that does not belong to it is split out from each fork point. For each node, when it completes the subtree splitting from its own perspective, it will get n subtrees; for each subtree, if it knows a node in it, then it can use these n nodes for recursive routing to reach any node in the entire binary tree.

[0043] Please see Figure 4 In this embodiment, node 0011 finds node 1110 through continuous query. Only node 101 in the first query step is known, and other nodes are obtained through step-by-step feedback.

[0044] In the split binary tree of this embodiment, each node records K nodes in each subtree after completing the subtree splitting, wherein the K value is a system-level even constant, which is a preset value; for a certain node, if n subtrees are split from its own perspective, it needs to maintain n routing tables, and the upper limit of the nodes in each routing table is K;

[0045] After each node completes the subtree split, it must record the K nodes in each subtree. The K value mentioned here is a system-level constant (must be an even number). It is set by the software system using Kad (for example, the Kad network used for BT downloads, K is set to 8). K-buckets are actually routing tables. For a certain node, if it splits n subtrees from its own perspective, it needs to maintain n routing tables, and the upper limit of nodes in each routing table is K.

[0046] For each node i, 0≦i≦160, each node stores the distance range between itself and the node in the interval [2 i ~2 (i +1) ), which consists of a list of (IP address, UDP port, Node ID) data. Each such list is called a K bucket, and each bucket has no more than K data items. See the following table for details:

[0047]

[0048]

[0049] The K bucket update method of this embodiment includes an active collection node method, a passive collection node method, and a failed node detection method;

[0050] The active node collection method is that any node can initiate a request to query a node (FIND_NODE), thereby refreshing the node information in the K-bucket;

[0051] The passive node collection method is to add the node ID of the other party to a K-bucket when receiving a request from other nodes (such as FIND_NODE, FIND_VALUE);

[0052] The way to detect failed nodes is to initiate a PING request to determine whether a node in the K-bucket is online, and then clean up the offline nodes in the K-bucket.

[0053] This embodiment takes the passive collection node mode as an example. When node x receives a PRC message, the IP address of the sender node y is used to update the corresponding K bucket. The specific implementation includes the following steps:

[0054] Step 1: Calculate the distance between yourself and the sender d(x,y)=x⊕y, where x and y are node ID values;

[0055] Step 2: Select the corresponding K buckets by distance d for update operation;

[0056] Step 3: Judgement;

[0057] If y's IP address already exists in this K bucket, move the corresponding item to the end of the K bucket;

[0058] If y's IP address is not recorded in the K bucket, go to step 4;

[0059] Step 4: Judgement;

[0060] If the entry in the K bucket is less than the preset value k, then the (IP address, UDP port, Node ID) information of y is directly inserted into the tail of the queue; (where 0≦k≦160;)

[0061] If the record items in the K bucket are greater than the preset value k, select the head record item node z (if it is node z) to perform the RPC_PING operation: if node z does not respond, remove z's information from the K bucket and insert y's information at the end of the queue; if node z responds, move z's information to the end of the queue and ignore y's information.

[0062] From the above analysis, we can see that the update principle is to insert the latest node to the tail of the queue and will not remove the online nodes in the queue.

[0063] Please see Figure 5, which is a diagram of the relationship between the user's online time and the probability of continuing to be online. According to the K-bucket update principle, nodes with a long online time have a higher probability of continuing to remain in the K-bucket list. Therefore: by keeping nodes with a long online time in the K-bucket, Kad significantly increases the probability that the nodes in the K-bucket will still be online in the next time period, which greatly improves the stability of the Kad network and reduces network maintenance costs (no need to frequently build node routing tables). At the same time, it can defend against DOS attacks to a certain extent, because Kad will only update the K-bucket information when the old node fails, which avoids flooding routing information through the addition of new nodes.

[0064] Kad can quickly perform node search, but when searching for a node, it is not known whether the node to be queried exists. For example, node x needs to search for node t.

[0065] Please see Figure 6 , Kad performs route lookup according to the following recursive operation steps:

[0066] (1) Calculate the distance from node x to node: d(x,t) = x⊕t;

[0067] (2) Take out the information of α nodes from the [㏒d]th K bucket of node x and perform FIND_NODE operation at the same time. If the information in this K bucket is less than α, select a total of α nodes closest to d from several nearby buckets; where [] is a rounding symbol and α is a preset value; in the BitTorrent implementation, the value is α=3.

[0068] (3) For each node that receives the query operation, if it finds itself to be node t, it replies that it is the closest to node t; otherwise, it measures the distance between itself and node t and selects information of α nodes from its corresponding K bucket to give to node x;

[0069] (4) Node x performs the FIND_NODE operation again for each newly received node. This process is repeated until a node in each branch responds that it is the closest to node t. Finally, node x obtains the node information of the k nodes closest to node t.

[0070] Since each query can obtain information from the K buckets closer to t, this mechanism ensures that each recursive operation can at least halve the distance, so the time complexity is: O(logN).

[0071] Please see Figure 7In the split binary tree of this embodiment, if node u wants to join the Kademlia network, it first needs to contact a node w that is already in the Kademlia network. Node u first inserts node w into its appropriate K bucket, then performs a FIND_NODE operation on its own node ID, and then updates its own K bucket content according to the received information; by gradually querying its neighboring nodes from near to far, the node completes the construction of the empty K bucket information, and also publishes its own information to the K buckets of other nodes;

[0072] Among them, when node u learns new node information, it tries to insert the new node information into the corresponding K-bucket according to its prefix value: if the K-bucket is not full, the new node is directly inserted into this K-bucket; if the K-bucket is full, continue to judge: if the coverage range of the K-bucket includes the ID of node u, then split the K-bucket into two new K-buckets of the same size, and redistribute the node information in the original K-bucket according to the new K-bucket prefix value; otherwise, directly discard the new node information.

[0073] Nodes do not need to publish any information when leaving the Kad network. One of the goals of the Kademlia protocol is to be able to work flexibly even when any node fails at any time. To this end, Kad requires each node to periodically publish all the stored<key,value> The data is cached in its k nearest neighbors, so that the data stored in the failed node can be quickly updated to other new nodes.

[0074] In the business scenario, the present invention is based on the characteristics of actual offshore data center business. Because offshore data centers require huge storage capacity, commercial storage is expensive, and the core technology is in the hands of foreign giants. In order to realize independent and controllable storage technology, the present invention uses distributed alliance blockchain technology to realize highly available distributed storage, and relies on multi-copy disaster recovery technology to achieve high availability. At the same time, traditional blockchain networks have the disadvantage of long addressing time. The present invention optimizes the relevant implementation technologies and methods, improves the addressing performance, and meets the actual storage needs of offshore data centers.

[0075] It should be understood that the above description of the preferred embodiment is relatively detailed and cannot be regarded as limiting the scope of patent protection of the present invention. Under the enlightenment of the present invention, ordinary technicians in this field can also make substitutions or modifications without departing from the scope of protection of the claims of the present invention, which all fall within the scope of protection of the present invention. The scope of protection requested for the present invention shall be based on the attached claims.

Claims

1. An offshore data distributed storage system based on blockchain fast addressing technology, characterized by: It consists of distributed storage layer, distributed network layer, blockchain service layer and application layer; The distributed storage layer uses a polymorphic data storage access interface to provide upper-layer applications with three types of storage services: object storage, block storage, and file system storage. Each piece of data will exist in the system in the form of N copies, and each piece of data will be divided into multiple blocks and stored in the system's distributed network. N is a preset value. For the object storage, the upper layer application calls the object storage service access interface, and the object storage module automatically completes the mapping of the user object storage space to the underlying unified distributed object storage space. The user data is stored in the distributed object storage engine in the form of object data; For block storage, upper-layer applications call the block storage service access interface, and the block storage module automatically completes the mapping of user block storage operations and data read and write operations to the underlying unified distributed object storage space. The user's data on the block storage is ultimately stored in the distributed object storage engine in the form of object data; The file system storage provides a POSIX file system interface compatible with the POSIX standard and supports both kernel file system and user space file system modes. Upper-layer applications call the POSIX file system interface, and the POSIX file system module and the POSIX file system metadata manager jointly complete the mapping of the user's POSIX file operations to the underlying unified distributed object storage space. The user's data in the POSIX file system is ultimately stored in the distributed object storage engine in the form of object data. The POSIX file system metadata manager is used for mapping and converting the POSIX file system space to the object storage space. The distributed network layer uses alliance chain network technology based on P2P network as the underlying layer to achieve a highly available distributed network; The blockchain service layer uses a consortium chain to provide blockchain basic services for the entire network; The consortium chain includes a member organization, MemberOrg, and an ordering organization, OrdererOrg. The MemberOrg is a participant in the consortium chain and an entity that needs to upload data to the chain. It maintains one or more peer nodes in the network. The OrdererOrg is responsible for transaction ordering and block creation in the consortium chain. Each organization has a membership identity, MSP, which fully represents an organization in the network and participates in network consensus. The application layer is used to provide data sharing and data transaction services for offshore data; The nodes of the system are divided into three roles: the key holding node is responsible for keeping the private key of the Chameleon Hash, the computing node is responsible for calculating new blocks, and the consensus node is responsible for voting on editing requests and editing results; when erroneous, harmful, or expired information is detected, a node issues a blockchain editing request for the first time and broadcasts it to the entire network; all nodes in the system verify the request and check the rationality of the request; when the number of nodes that agree to the request reaches a predetermined number, the request is approved, otherwise the request is rejected; the system selects a node to edit the blockchain according to a predetermined strategy; the node with modification rights edits the block information according to the request content and broadcasts it to the entire network; the remaining nodes verify the operation results and update their own ledgers after the verification is passed.

2. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 1 is characterized by: The consortium chain network technology based on the P2P network as the underlying layer is based on the Kademlia network and uses the distributed hash table DHT implemented by the Kademlia protocol. All nodes in the Kademlia network are treated as leaves of a split binary tree, and the position of each node is uniquely determined by the shortest prefix of its ID value. Each node has a 160-bit ID value as an identifier. The entire network node state is represented by a binary tree, and each node ID value is represented in binary form. The nth bit of the binary corresponds to the nth layer of the binary tree. If the bit is 1, it enters the left subtree, and if it is 0, it enters the right subtree. Finally, the leaf node on the binary tree corresponds to a certain node. The Kademlia protocol includes four remote RPC operations: PING, STORE, FIND_NODE, and FIND_VALUE. The PING operation is to detect a node to determine whether it is still online. The STORE operation is to notify a node to store a value.<key,value> Yes, for future query needs, where Key is the unique identifier of the node and Value is the payload; the FIND_NODE operation uses a 160-bit ID as a parameter, and the recipient of this operation returns the (IP address, UDP port, Node ID) information of the K nodes it knows that are closer to the target ID, where IP address represents the IP address of the node, UDP represents the port number of the Node, and Node ID represents the 160-bit ID number of the unique identifier of the Node; the IND_VALUE operation is similar to the FIND_NODE operation, except that it only needs to return the (IPaddress, UDP port, Node ID) information of a node. If the recipient of this operation receives a STORE operation with the same key, it will directly return the stored value.

3. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 2 is characterized by: The binary tree splitting method can decompose any node into a series of continuous subtrees that do not contain itself; the highest-level subtree is composed of the other half of the entire tree that does not contain itself; the next-level subtree is composed of the remaining half that does not contain itself; for each node, when it completes the subtree splitting from its own perspective, it will obtain n subtrees; for each subtree, if a node in it is known, the node can use these n nodes for recursive routing to reach any node in the entire binary tree.

4. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 2 is characterized by: In the split binary tree, an exclusive-OR operation is used to calculate the distance between nodes; wherein, the distance from node A to node B is equal to the distance from node B to node A, the distance between a node itself is 0, the distance between any two nodes must be greater than 0, and the distance from node A to node C via node B is always greater than or equal to the distance from node A directly to node C.

5. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 2 is characterized by: In the split binary tree, each node records the K nodes in each subtree after completing the subtree split, where the K value is a system-level even constant and a preset value. For a node, if n subtrees are split from its own perspective, it needs to maintain n routing tables, and the upper limit of the node number in each routing table is K. For each node i, 0≦i≦160, each node stores the distance range between itself and the node in the interval [2 i ~2 (i+1) ), the node information is composed of (IP address, UDP port, Node ID) data lists, each of which is called a K bucket, and each bucket has no more than K data items.

6. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 5 is characterized by: The K bucket update method includes active collection node method, passive collection node method and detection failure node method; The active node collection method is that any node can initiate a request to query the node, thereby refreshing the node information in the K-bucket; The passive node collection method is to add the node ID of the other node to a K-bucket when receiving a request from the other node; The method for detecting failed nodes is to determine whether a node in the K-bucket is online by initiating a PING request, and then clear the offline nodes in the K-bucket.

7. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 6 is characterized by: In the passive collection node method, when node x receives a PRC message, the IP address of the sender node y is used to update the corresponding K bucket; the specific implementation includes the following steps: Step 1: Calculate the distance between yourself and the sender d(x,y) = x⊕y, where x and y are node ID values; Step 2: Select the corresponding K buckets by distance d for update operation; Step 3: Judgement; If y's IP address already exists in this K bucket, move the corresponding item to the end of the K bucket; If y's IP address is not recorded in the K bucket, go to step 4; Step 4: Judgement; If the number of entries in the K bucket is less than the preset value k, then the (IP address, UDP port, Node ID) information of y is directly inserted into the tail of the queue; If the number of entries in the K bucket is greater than the preset value k, the head entry node z is selected for RPC_PING operation: if node z does not respond, z's information is removed from the K bucket and y's information is inserted at the end of the queue; if node z responds, z's information is moved to the end of the queue and y's information is ignored.

8. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 5 is characterized by: In the K bucket, when node x needs to search for node t, the specific implementation includes the following steps: (1) Calculate the distance from node x to node t: d(x,t) = x⊕t; (2) Take out the information of α nodes from the [㏒d]th K bucket of node x and perform FIND_NODE operation at the same time. If the information in this K bucket is less than α, select a total of α nodes with the closest distance d from several nearby buckets; where [] is the rounding symbol and α is the preset value; (3) For each node that receives the query operation, if it finds itself to be node t, it replies that it is the closest to node t; otherwise, it measures the distance between itself and node t and selects information of α nodes from its corresponding K bucket to give to node x; (4) Node x performs the FIND_NODE operation again for each newly received node. This process is repeated until a node in each branch responds that it is the closest to node t. Finally, node x obtains the node information of the k nodes closest to node t.

9. The offshore data distributed storage system based on blockchain fast addressing technology according to claim 5 is characterized by: In the split binary tree, if node u wants to join the Kademlia network, it first needs to contact a node w that is already in the Kademlia network. Node u first inserts node w into its appropriate K-bucket, then performs a FIND_NODE operation on its own node ID, and then updates its own K-bucket content based on the received information. By gradually querying its neighboring nodes from near to far, the node completes the construction of empty K-bucket information and also publishes its own information to the K-buckets of other nodes. Among them, when node u learns new node information, it tries to insert the new node information into the corresponding K-bucket according to its prefix value: if the K-bucket is not full, the new node is directly inserted into this K-bucket; if the K-bucket is full, it continues to judge: if the coverage range of the K-bucket includes the ID of node u, the K-bucket is split into two new K-buckets of the same size, and the node information in the original K-bucket is redistributed according to the new K-bucket prefix value; otherwise, the new node information is directly discarded.