WO2018032368A1 - Procédé de traitement de données de système de chaîne de blocs sur la base d'une acquisition comprimée et d'un algorithme de reconstruction creuse - Google Patents

Procédé de traitement de données de système de chaîne de blocs sur la base d'une acquisition comprimée et d'un algorithme de reconstruction creuse Download PDF

Info

Publication number
WO2018032368A1
WO2018032368A1 PCT/CN2016/095572 CN2016095572W WO2018032368A1 WO 2018032368 A1 WO2018032368 A1 WO 2018032368A1 CN 2016095572 W CN2016095572 W CN 2016095572W WO 2018032368 A1 WO2018032368 A1 WO 2018032368A1
Authority
WO
WIPO (PCT)
Prior art keywords
sparse
algorithm
data
processing method
compressed sensing
Prior art date
Application number
PCT/CN2016/095572
Other languages
English (en)
Chinese (zh)
Inventor
张丛
Original Assignee
深圳市樊溪电子有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市樊溪电子有限公司 filed Critical 深圳市樊溪电子有限公司
Publication of WO2018032368A1 publication Critical patent/WO2018032368A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
    • G16Z99/00—Subject matter not provided for in other main groups of this subclass

Definitions

  • the invention relates to a method for processing blockchain system data, in particular to a blockchain system data processing method based on compressed sensing and sparse reconstruction algorithm.
  • blockchain is a secure account book database, composed of data blocks, users can constantly update and upgrade here.
  • the platform looks for data.
  • the blockchain can speed up transaction processing, reduce costs, reduce middlemen, improve market insight, and increase business transparency.
  • Blockchain the underlying technology for cryptocurrency bitcoin, is a great innovation.
  • Blockchain technology can be used to combat fraud and illegal transactions.
  • Many industries are now using blockchain technology to get faster transaction processing speed. , is the main factor affecting the wide-ranging application of the private blockchain.
  • NASDAQ handles transactions at a rate of 1 million times per second, and VISA transaction processing peaks at 45,000 times per second. It can be seen that it is very important to improve the transaction processing capability of the blockchain system.
  • the transaction speed of the private blockchain that is currently on the market is relatively low.
  • the birth of compression sensing theory is related to the structure and content of the signal, and is sampled, encoded and reconstructed at a lower frequency than the Nyquist sampling theorem.
  • the theory was originally used in traditional industries such as machinery and electronics. , aviation, optics, but for the current blockchain data processing speed is very slow, resulting in data redundancy and resource waste bottlenecks serious technical problems, the theory provides a new opportunity for its development, significantly affecting the blockchain
  • the sparse reconstruction algorithm is based on the traditional signal decomposition method which decomposes the signal into a complete set of orthogonal bases. It has great limitations. When expressing arbitrary signals, a more efficient non-orthogonal algorithm is needed. put forward.
  • the signal is sparsely decomposed under the redundant dictionary.
  • the representation of the signal is sparse enough to be sparse, thus ensuring that the number of compression measurements directly related to the non-zero coefficient is sufficiently small and the reconstruction data is high probability.
  • the object of the present invention is to provide a blockchain system intersection data processing method based on compression sensing and sparse reconstruction algorithm, comprising the following steps: (1) converting blockchain system transaction data into multi-time-space multi-dimensional data, and then converting The latter data is compressed and sparsely expressed, that is, linear decomposition of multi-time-space multi-dimensional data; (2) constructing an M*N sparse pairwise matrix that is not related to the sparse transformation matrix, linearly projecting multi-time-space multidimensional data, and acquiring perceptual data Calculate the value so that the transaction data object will be dimensioned, where M is much smaller than N, and the perceived data pair is calculated as M*1 order matrix; (3) transaction information reconstruction, using low-dimensional transaction data acquired by compressed sensing The sparse algorithm accurately reconstructs multi-time-space multi-dimensional original transaction data, that is, reconstructs N-dimensional transaction data by using M-dimensional sensing measurements.
  • the sparse representation of step (1) has one of the following methods: sparse transform, Fourier transform, wavelet transform, Gabor transform, Curvelet transform, Bandelet transform, Contourlet transform.
  • the sparse representation algorithm of step (1) comprises a tracking algorithm, a greedy matching tracking algorithm, and an orthogonal matching tracking algorithm.
  • the sparse pairwise matrix in step (2) needs to satisfy the incoherence and the restricted isotropic condition, that is, the sparse transform is used to extract the signal based on the uncorrelated sparse pairwise matrix, and the original acquired transaction data signal can pass through Sparse representation after some transformation.
  • the sparse pairwise matrix in step (2) can be selected from a random Gaussian sparse measurement matrix, a random shell effort measurement matrix, a partial orthogonal matrix, and a sparse random matrix.
  • the step (3) reconstruction algorithm has three types: a convex optimization algorithm based on the L1 norm, a greedy algorithm based on the L0 norm, and a combination algorithm.
  • the convex optimization algorithm based on the L1 norm includes a base tracking algorithm, a gradient projection method, a convex set alternating projection algorithm and an interior point iterative method.
  • the L0 norm-based greedy algorithm includes a matching tracking algorithm, an orthogonal matching tracking algorithm, and a segmentation orthogonal matching tracking algorithm.
  • the combination algorithm includes a chain tracking algorithm, an HHS tracking algorithm and an I-wen algorithm.
  • the data processing method of the invention effectively solves the technical problem that the blockchain data processing speed is slow, and data redundancy and resource waste are generated.
  • FIG. 2 is a flowchart of a blockchain system data processing method based on a compressed sensing and sparse reconstruction algorithm according to an embodiment of the present invention.
  • the essence of a transaction is a relational data structure that contains information about the value transfer of the trading participants. These transaction information is called the accounting ledger.
  • the transaction needs to go through three creation, verification, and writing blockchains. The transaction must be digitally signed to ensure the legality of the transaction.
  • Block All transaction information is stored in the block, and a transaction information is a record, which is stored as a separate record in the blockchain.
  • the block consists of a block header and a data part.
  • the block header field contains various characteristics of the block itself, such as the previous block information, the merkle value, and the timestamp.
  • the block header hash value and block height are the two most important indicators for identifying the block.
  • the block primary identifier is its cryptographic hash value, a digital fingerprint obtained by performing a second hash calculation on the block header by the SHA algorithm.
  • the resulting 32-byte hash value is called the block hash value, or the block header hash value, and only the block header is used for calculation.
  • the block hash value can uniquely and unambiguously identify a block, and any node can independently obtain the block hash value by simply hashing the block header.
  • Blockchain A data structure in which blocks are chained in an orderly fashion.
  • a blockchain is like a vertical stack, with the first block being the first block at the bottom of the stack, and each block is then placed on top of the other blocks.
  • a block When a block is written to a blockchain, it will never change and is backed up to another blockchain server.
  • Bitcoin uses the public blockchain technology, which runs on the p2p network, and each connected PC can participate. This brings many drawbacks in the actual application process, for example, low efficiency, slow block generation, and transaction. Processing is not timely. It is not allowed in some specific applications. Private blockchain nodes run at high In the fast network, the transmission rate is fast, the accounting is immediate, and the security is higher. These features ensure that private blockchain technology is used in large-scale transaction processing.
  • Figure 1 shows the network model of a private blockchain, which is essentially a multi-client-multi-server model.
  • multi-server mode there are 3x+1 nodes with one or more leader nodes.
  • the communication mechanism between BC nodes is very cumbersome. Since multiple nodes work together to perform a task, a message mechanism is needed to communicate, and messages may be lost, out of order, and duplicated during the sending process.
  • This structure is called Client/Multi-Server mode, referred to as C/. MS mode.
  • the multi-server model guarantees system security, scalability and fault tolerance.
  • the network model includes: a client node: an initiator of a transaction; a multi-server cluster: a distributed task system, which is composed of a plurality of nodes therein, and tasks in the nodes cooperate through message communication with each other to complete tasks consistently; Leader node: The master node is responsible for receiving tasks from client nodes and distributing them to other nodes. The blockchain node is responsible for receiving and completing the tasks assigned by the leader. At the same time, if the master node fails, multiple BC nodes can be selected spontaneously. A new master node; transactions: transactions that need to be verified and counted into blocks; messages used for communication between nodes.
  • the processing speed and efficiency of blockchain transaction data are evaluated by system node utilization, the average number of transactions being processed, the average number of transactions in the system, the average number of transactions waiting in the system, the average length of stay of the transaction, and the average waiting time of the transaction.
  • the algorithm starts from the following three aspects: (1) sparse representation; (2) sparse measurement; (3) signal reconstruction.
  • the blockchain system transaction data processing method based on the compressed sensing and sparse reconstruction algorithm includes the following steps: (1) converting the blockchain system transaction data into multi-time-space multi-dimensional data, and then performing the converted data.
  • Compression and sparse representation which is to linearly decompose multi-time-space multidimensional data; (2) construct an M*N sparse pairwise matrix that is not related to sparse transformation matrix, and linearly cast multi-time-space multidimensional data Shadow, obtain the perceptual data pair calculation value, so that the transaction data object will be dimensioned, where M is far less than N, and the perceptual data pair is calculated as M*1 order matrix; (3) transaction information reconstruction, using compressed sensing acquisition Low-dimensional transaction data uses sparse algorithm to accurately reconstruct multi-temporal multi-dimensional original transaction data, that is, reconstruct M-dimensional transaction data by using M-dimensional perceptual measurement values.
  • the blockchain application environment is grid power trading.
  • the transaction is generated by the client's application.
  • the transaction is processed by the BC server inside the virtual local area network.
  • the bottleneck affecting transaction processing speed is the network status and computing power of the BC server.
  • the number of BC servers in the cluster is 3x+1. We respectively test the number of BC nodes under 10, 20, 30, 40, 50, 60, etc., and perform Fourier and wavelet transform on the transaction data collected every 10 nodes, construct a sparse dictionary as a redundant dictionary, for the transaction.
  • the matching pursuit algorithm performs sparse representation.
  • the method of QR decomposition and orthogonal transform improved random sparse pair matrices a is constructed, so as to construct a sparse reconciliation matrix which is consistent with the blockchain transaction data eigen, which lays a foundation for transaction information reconstruction.
  • the sparse check matrix requires that the K transaction information values obtained from the conversion of f from trading to f can retain all the information of the original transaction, and ensure the accurate reconstruction of the transaction information, wherein the acquisition of the K quantity satisfies the finite isometric nature. That is, the choice of the constant depends on the size of the blockchain transaction data, and the convex optimization method is used to reconstruct the transaction information. After the big data transaction information is acquired, the alternating direction algorithm and the greedy tracking algorithm are used to synchronize the iteration, and the appropriate iteration is selected. Initial values and thresholds take into account the separation structure.

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

L'invention concerne un procédé de traitement de données de système de chaîne de blocs basé sur une acquisition comprimée et un algorithme de reconstruction creuse, qui comprend : (1) la conversion de données de transaction de système de chaîne de blocs en données multi spatio-temporelles et multidimensionnelles, et la réalisation d'une compression et d'une représentation creuse sur les données converties; (2) la construction d'une matrice de calcul creuse M*N non liée à une matrice de transformation creuse et la réalisation d'une projection linéaire sur les données multi spatio-temporelles et multidimensionnelles pour obtenir une valeur calculée de données d'acquisition; et (3) la reconstruction des informations de transaction : reconstruction avec précision des données de transaction multi spatio-temporelles et multidimensionnelles d'origine à l'aide de données de transaction de dimension réduite obtenues par acquisition comprimée et au moyen d'un algorithme creux. Le procédé de traitement de données résout les problèmes techniques de redondance de données et de gaspillage de ressources provoqués par une vitesse de traitement de données de chaîne de blocs peu élevée.
PCT/CN2016/095572 2016-08-13 2016-08-16 Procédé de traitement de données de système de chaîne de blocs sur la base d'une acquisition comprimée et d'un algorithme de reconstruction creuse WO2018032368A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610670838.0 2016-08-13
CN201610670838.0A CN106326641A (zh) 2016-08-13 2016-08-13 基于压缩感知和稀疏重构算法的区块链系统数据处理方法

Publications (1)

Publication Number Publication Date
WO2018032368A1 true WO2018032368A1 (fr) 2018-02-22

Family

ID=57740883

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/095572 WO2018032368A1 (fr) 2016-08-13 2016-08-16 Procédé de traitement de données de système de chaîne de blocs sur la base d'une acquisition comprimée et d'un algorithme de reconstruction creuse

Country Status (2)

Country Link
CN (1) CN106326641A (fr)
WO (1) WO2018032368A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109214975A (zh) * 2018-09-01 2019-01-15 哈尔滨工程大学 一种基于二维稀疏信号恢复的二维逐步正交匹配追踪方法
TWI694700B (zh) * 2018-08-01 2020-05-21 香港商阿里巴巴集團服務有限公司 資料處理方法和裝置、用戶端
CN114115730A (zh) * 2021-11-02 2022-03-01 北京银盾泰安网络科技有限公司 一种应用容器存储引擎平台
CN114463962A (zh) * 2020-10-21 2022-05-10 中国石油化工股份有限公司 智能节点数据采集方法、电子设备及储存介质

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10560270B2 (en) 2017-05-03 2020-02-11 International Business Machines Corporation Optimal data storage configuration in a blockchain
CN107181797B (zh) * 2017-05-11 2020-03-06 中国农业银行股份有限公司 一种区块链的区块压缩方法和系统
CN108959280B (zh) * 2017-05-17 2021-08-06 中国移动通信有限公司研究院 一种存储虚拟资源关联信息的方法及装置
CN107527371B (zh) * 2017-09-07 2020-05-01 中国科学院光电技术研究所 一种在压缩感知中逼近光滑l0范数的图像重建算法的设计构建方法
CN107689797B (zh) * 2017-09-14 2020-04-28 中南大学 基于维数任意的分块哈达玛测量矩阵的压缩感知信号重建方法
CN108233943B (zh) * 2018-01-23 2020-06-19 浙江大学 一种基于最小相关性测量矩阵的压缩感知方法
CN108663606A (zh) * 2018-05-17 2018-10-16 国网辽宁省电力有限公司电力科学研究院 一种以低采样频率对局部放电信号进行采集的方法及系统
CN109542908B (zh) * 2018-11-23 2019-08-13 中科驭数（北京）科技有限公司 key-value数据库中的数据压缩方法、存储方法、访问方法和系统
CN112750037A (zh) * 2019-04-30 2021-05-04 创新先进技术有限公司 基于区块链的数据压缩、查询方法及装置和电子设备
US10795882B2 (en) 2019-04-30 2020-10-06 Alibaba Group Holding Limited Blockchain-based data compression and searching
CN110532329B (zh) * 2019-09-02 2022-06-21 智慧谷(厦门)物联科技有限公司 一种基于区块链技术的智能手环数据处理与共享方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102867413A (zh) * 2012-07-18 2013-01-09 浙江工业大学 车路协作环境下的车辆传感数据压缩感知获取方法
CN102868885A (zh) * 2012-08-27 2013-01-09 中国科学院长春光学精密机械与物理研究所 基于压缩感知的星上实时图像合成压缩系统
CN103944579A (zh) * 2014-04-10 2014-07-23 东华大学 一种压缩感知重构的编码解码系统
WO2015131396A1 (fr) * 2014-03-07 2015-09-11 中国科学院微电子研究所 Procédé d'échantillonnage aléatoire de signal unidimensionnel sur la base d'une détection compressée

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100246920A1 (en) * 2009-03-31 2010-09-30 Iowa State University Research Foundation, Inc. Recursive sparse reconstruction
CN105608146A (zh) * 2015-12-17 2016-05-25 布比（北京）网络技术有限公司 一种区块链溯源追踪方法
CN105636094A (zh) * 2016-03-16 2016-06-01 中国地质大学（武汉） 基于分簇压缩感知的无线传感网预警方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102867413A (zh) * 2012-07-18 2013-01-09 浙江工业大学 车路协作环境下的车辆传感数据压缩感知获取方法
CN102868885A (zh) * 2012-08-27 2013-01-09 中国科学院长春光学精密机械与物理研究所 基于压缩感知的星上实时图像合成压缩系统
WO2015131396A1 (fr) * 2014-03-07 2015-09-11 中国科学院微电子研究所 Procédé d'échantillonnage aléatoire de signal unidimensionnel sur la base d'une détection compressée
CN103944579A (zh) * 2014-04-10 2014-07-23 东华大学 一种压缩感知重构的编码解码系统

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI694700B (zh) * 2018-08-01 2020-05-21 香港商阿里巴巴集團服務有限公司 資料處理方法和裝置、用戶端
CN109214975A (zh) * 2018-09-01 2019-01-15 哈尔滨工程大学 一种基于二维稀疏信号恢复的二维逐步正交匹配追踪方法
CN109214975B (zh) * 2018-09-01 2023-03-28 哈尔滨工程大学 一种基于二维稀疏信号恢复的二维逐步正交匹配追踪方法
CN114463962A (zh) * 2020-10-21 2022-05-10 中国石油化工股份有限公司 智能节点数据采集方法、电子设备及储存介质
CN114115730A (zh) * 2021-11-02 2022-03-01 北京银盾泰安网络科技有限公司 一种应用容器存储引擎平台
CN114115730B (zh) * 2021-11-02 2023-06-13 北京银盾泰安网络科技有限公司 一种应用容器存储引擎平台

Also Published As

Publication number Publication date
CN106326641A (zh) 2017-01-11

Similar Documents

Publication Publication Date Title
WO2018032368A1 (fr) Procédé de traitement de données de système de chaîne de blocs sur la base d'une acquisition comprimée et d'un algorithme de reconstruction creuse
Sun et al. Generative tensor network classification model for supervised machine learning
Shen et al. A fast optimization method for general binary code learning
Gong et al. Web scale photo hash clustering on a single machine
CN104008174A (zh) 一种海量图像检索的隐私保护索引生成方法
Yuan et al. An effective image classification method for shallow densely connected convolution networks through squeezing and splitting techniques
Wang et al. Duplicate discovery on 2 billion internet images
CN106599917A (zh) 一种基于稀疏表示的近似图像重复检测方法
CN116361256B (zh) 基于日志解析的数据同步方法及系统
Zhang et al. Effective image retrieval via multilinear multi-index fusion
Guðmundsson et al. Towards engineering a web-scale multimedia service: A case study using Spark
CN114065850A (zh) 基于统一锚点与子空间学习的谱聚类方法及系统
Li et al. Fast principal component analysis for hyperspectral imaging based on cloud computing
CN104408039B (zh) 基于Hilbert曲线与R‑tree的HBase多维查询系统的构建及其查询方法
CN107958259A (zh) 一种基于卷积神经网络的图像分类方法
Chandrasekhar et al. Compact global descriptors for visual search
Yang et al. Large scale video data analysis based on spark
Al-Shammary et al. Dynamic fractal clustering technique for SOAP web messages
Cao et al. A new approach for large‐scale scene image retrieval based on improved parallel K‐means algorithm in MapReduce environment
Zheng et al. Low-rank matrix recovery with discriminant regularization
Cao et al. Subspace clustering based on latent low rank representation with schatten-p norm
CN114565772B (zh) 集合特征提取方法、装置、电子设备及存储介质
Asante-Mensah et al. Image reconstruction using superpixel clustering and tensor completion
CN112861965B (zh) 基于多特征交叉一致性模型的图像匹配方法
Wei et al. Sketch and Scale Geo-distributed tSNE and UMAP

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16913128

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16913128

Country of ref document: EP

Kind code of ref document: A1