WO2016180055A1 - 数据存储、读取的方法、装置及系统 - Google Patents

数据存储、读取的方法、装置及系统 Download PDF

Info

Publication number
WO2016180055A1
WO2016180055A1 PCT/CN2016/072035 CN2016072035W WO2016180055A1 WO 2016180055 A1 WO2016180055 A1 WO 2016180055A1 CN 2016072035 W CN2016072035 W CN 2016072035W WO 2016180055 A1 WO2016180055 A1 WO 2016180055A1
Authority
WO
WIPO (PCT)
Prior art keywords
storage
data
interface
read
storage resource
Prior art date
Application number
PCT/CN2016/072035
Other languages
English (en)
French (fr)
Inventor
李丽彬
韩银俊
高洪
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2016180055A1 publication Critical patent/WO2016180055A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers

Definitions

  • the present invention relates to the field of communications, and in particular, to a data storage and reading method, apparatus, and system.
  • the storage architecture can be divided into two types: one is a traditional dedicated storage architecture, such as a network attached storage (SAN) and a network access server (Storage Area Network and SAN Protocols, referred to as NAS). ), etc., consisting of ordinary servers, which are expensive in unit cost and poor in scalability, and are not suitable for large-scale deployment, but have certain advantages in performance and reliability.
  • SAN network attached storage
  • NAS Storage Area Network and SAN Protocols
  • the other is a traditional cloud storage architecture, such as a distributed file system.
  • This distributed file system is low in cost and linearly expandable, so its scale and system-wide throughput can be made very large.
  • the downside is that the reliability and performance of a single point is not as good as SAN and NAS.
  • the above two storage architectures are incompatible with each other. Users cannot flexibly choose the system they want according to the actual situation. At the same time, the storage systems of these two architectures also use their fixed interfaces and cannot cope with various Ethernet networks. ) Customers and their respective storage protocols.
  • the storage architectures are incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared.
  • no effective solution has been proposed.
  • the invention provides a data storage and reading method, device and system, so as to at least solve the problem that the storage architectures in the related art are mutually incompatible, the interfaces are not unified, and the storage resources cannot be shared.
  • a data storage method including:
  • the data distribution processing system acquires the feature and/or metadata in the preset storage resource corresponding to the storage policy according to the feature and/or metadata of the data to be stored and the storage policy of the data distribution processing system.
  • the data distribution processing system further includes:
  • Key-value KV application programming interface API client and/or key-value KV interface are examples of keys-value KV application programming interface API client and/or key-value KV interface.
  • the feature and/or metadata of the data to be stored includes:
  • the key value KV of the data to be stored is the key value KV of the data to be stored.
  • storing the to-be-stored data in a storage node corresponding to the storage resource includes:
  • the storage node corresponding to the to-be-stored data, where the storage node is configured to receive and write the to-be-stored data.
  • the storage resource fragment is a fragment on a ring of a consistent hash hash algorithm, where the ring corresponds to the Preset storage resources.
  • storing the to-be-stored data in a storage node corresponding to the storage resource includes:
  • the method comprises:
  • the data to be stored is obtained through a key value KV interface, an object access client OAC interface, a block access client BAC interface, or a file access client FAC interface.
  • the method when the data to be stored is obtained through the OAC interface, the method further includes:
  • the data distribution processing system receives the REST protocol for the REST protocol, and the object information is submitted to the OAC, and the OAC obtains the data to be stored corresponding to the object information by using the object data server OMDS.
  • the method when the data to be stored is obtained by using the BAC interface, the method further includes:
  • the data distribution processing system receiving application receives the ISCSI command through the small computer system interface ISCSI object, and then converts to the internal interface of the BAC, and the BAC converts the block data into the corresponding data to be stored.
  • the method when the data to be stored is obtained by using the FAC interface, the method further includes:
  • the data distribution processing system receives an application through a network file system NFS interface, a portable operating system interface POSIX interface, or a general Ethernet file system CIFS interface into an internal file access client FAC interface, the FAC puts the file
  • the metadata information is organized and managed by the file data server FMDS, and the data to be stored corresponding to the metadata information is obtained.
  • a data reading method including:
  • the data distribution processing system is based on characteristics and/or metadata of the data to be read, and the storage of the data distribution processing system a storage policy, where a storage resource corresponding to the feature and/or metadata is obtained, where the preset storage resource is a storage resource of at least one or more storage nodes composition;
  • the data distribution processing system further includes:
  • Key-value KV application programming interface API client and/or key-value KV interface are examples of keys-value KV application programming interface API client and/or key-value KV interface.
  • the feature and/or metadata of the data to be read includes:
  • the key value KV of the data to be read is the key value KV of the data to be read.
  • the storage node corresponding to the storage resource that reads the data to be read includes:
  • the storage resource fragment is a fragment on a ring of a consistent hash hash algorithm, where the ring corresponds to the Preset storage space.
  • the storage node corresponding to the storage resource that reads the data to be read includes:
  • the method comprises:
  • the data to be read is read by a key value KV interface, an object access client OAC interface, a block access client BAC interface, or a file access client FAC interface.
  • the method when the data to be read is read by using the OAC interface, the method further includes:
  • the data distribution processing system reads the data to be read and returns it to the object data server OMDS, and the OMDS returns the data to be read to the OAC interface, and transmits the REST gateway through the representation state to perform REST protocol parsing. , returned to the application.
  • the method when the data to be read is read by the BAC interface, the method further includes:
  • the data distribution processing system reads the data to be read back to the small computer system interface ISCSI object through the interface of the BAC, and the ISCSI object returns the data to be read to the application.
  • the method when the data to be read is read through the FAC interface, the method further includes:
  • the data distribution processing system reads the data to be read and returns it to the file data server FMDS, and passes the The FAC interface reads the data to be read, and then returns the POSIX interface of the operating system interface or the CIFS interface of the universal Ethernet file system to the application through the network file system NFS interface.
  • a data storage device including:
  • the first obtaining module is configured to acquire, according to the feature and/or metadata of the data to be stored, and the storage policy of the data distribution processing system, in the preset storage resource corresponding to the storage policy, a storage resource corresponding to the feature and/or the metadata, wherein the preset storage resource is composed of storage resources of at least one or more storage nodes;
  • a storage module configured to store the to-be-stored data into a storage node corresponding to the storage resource.
  • a data reading apparatus including:
  • a second obtaining module configured to acquire, according to the feature and/or metadata of the data to be read, and the storage policy of the data distribution processing system, in the preset storage resource corresponding to the storage policy, a storage resource corresponding to the feature and/or the metadata, where the preset storage resource is composed of storage resources of at least one or more storage nodes;
  • a reading module configured to read a storage node corresponding to the storage resource of the data to be read.
  • a data processing system including: a key value KV storage database, a key value KV application programming interface API interface;
  • the key value KV stores a database, and is configured to integrate storage resources on each storage node to form a unified preset storage resource;
  • the key value KV application programming interface API interface is configured to apply different storage nodes of the preset storage space according to the storage or read request of the data, and the key value Key of the data, wherein the preset storage resource It consists of storage resources of multiple storage nodes.
  • the function of the key value KV application programming interface API interface comprises at least one of the following:
  • system further includes: an operation and maintenance module OMM, a monitoring module,
  • the OMM is configured to set at least one of the following functions: management of the storage resource, division of the storage resource, configuration of the storage resource, and alarm display of the system;
  • the monitoring module is configured to monitor a status of the storage node and the disk.
  • system further includes: an object access client OAC interface, a block access client BAC interface, and a file access client FAC interface,
  • the object access client OAC interface, the block access client BAC interface, or the file access client FAC interface is configured to store or read data of the key value KV storage database.
  • the method further includes: configuring a module
  • the configuration module is configured to configure a storage policy of the storage resource, where the storage policy includes: a copy mode, a snapshot support mode, a compression mode, and a cache mode.
  • the data distribution processing system acquires the feature and/or the preset storage resource corresponding to the storage policy according to the feature and/or metadata of the data to be stored and the storage policy of the data distribution processing system.
  • a storage resource corresponding to the metadata wherein the preset storage resource is composed of storage resources of at least one or more storage nodes; storing the data to be stored in a storage node corresponding to the storage resource, and solving the storage architecture It is compatible, the interface is not uniform, and the storage resources cannot be shared. The unified sharing of storage resources is realized, which saves resources and maintenance overhead.
  • FIG. 1 is a flow chart of a data storage method according to an embodiment of the present invention.
  • FIG. 2 is a flow chart of a data reading method according to an embodiment of the present invention.
  • FIG. 3 is a block diagram showing the structure of a data storage device according to an embodiment of the present invention.
  • FIG. 4 is a block diagram showing the structure of a data reading apparatus according to an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of an overall architecture of a storage system in accordance with a preferred embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a KV database storage process in accordance with a preferred embodiment of the present invention.
  • FIG. 7 is a schematic diagram of the internal composition of a KV API in accordance with a preferred embodiment of the present invention.
  • FIG. 8 is a schematic diagram of distributed object storage in accordance with a preferred embodiment of the present invention.
  • FIG. 9 is a schematic diagram of distributed block storage in accordance with a preferred embodiment of the present invention.
  • FIG. 10 is a schematic diagram of distributed file storage in accordance with a preferred embodiment of the present invention.
  • FIG. 1 is a flowchart of a data storage method according to an embodiment of the present invention. As shown in FIG. 1, the process includes the following steps:
  • Step S102 The data distribution processing system acquires the feature and/or metadata in the preset storage resource corresponding to the storage policy according to the feature and/or metadata of the data to be stored and the storage policy of the data distribution processing system.
  • Corresponding storage resource where the preset storage resource is composed of storage resources of at least one or more storage nodes;
  • Step S104 storing the to-be-stored data in a storage node corresponding to the storage resource.
  • the data distribution processing system acquires the storage resource corresponding to the feature and/or the metadata according to the feature and/or metadata of the data to be stored and the algorithm of the storage policy of the data distribution processing system, and stores the to-be-stored
  • the data is stored in the storage node corresponding to the storage resource, which solves the problem that the storage architecture is incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared, and the storage resources are uniformly shared, thereby saving resources and maintenance overhead.
  • the data distribution processing system may include: a key value (KV) application programming interface client (Application Programming Interface, API for short) client and/or a key value (KV) interface.
  • KV key value
  • the feature and/or metadata of the data to be stored includes: a key value (KV) of the data to be stored.
  • storing the to-be-stored data in the storage node corresponding to the storage resource includes:
  • the storage resource fragment is a fragment on the ring of the consistent hash hash algorithm, where the ring pair presets the storage resource.
  • storing the to-be-stored data in the storage node corresponding to the storage resource includes:
  • the key access KV interface the Object Access Client (OAC) interface, the Block Access Client (BAC) interface, or the file access client (File Access) may be used.
  • the client referred to as the FAC file interface, obtains the data to be stored.
  • the data distribution processing system receives the application to perform REST protocol analysis through a Representational State Transfer (REST) gateway, and submits the object information to the OAC, the OAC.
  • REST Representational State Transfer
  • the data to be stored corresponding to the object information is obtained by an object data server (Object Meta Data Server, OMDS for short).
  • the data distribution processing system receiving application receives the object through an Internet Small Computer System Interface (ISCSI) interface. After the ISCSI command, it is converted to the internal interface of the BAC, and the BAC converts the block data into the corresponding data to be stored.
  • ISCSI Internet Small Computer System Interface
  • the data distribution processing system receives the application through a Network File System (NFS) interface, and the Portable Operating System Interface (Portable Operating System Interface, POSIX for short)
  • NFS Network File System
  • POSIX Portable Operating System Interface
  • CIFS Common Internet File System
  • FMDS File Meta Data Server
  • FIG. 2 is a flowchart of a data reading method according to an embodiment of the present invention. As shown in FIG. 2, the process includes the following steps:
  • step S202 the data distribution processing system acquires the feature and/or element in the preset storage resource corresponding to the storage policy according to the feature and/or metadata of the data to be read and the storage policy of the data distribution processing system.
  • a storage resource corresponding to the data wherein the preset storage resource is composed of storage resources of at least one or more storage nodes;
  • Step S204 reading a storage node corresponding to the storage resource of the data to be read.
  • the data distribution processing system acquires the storage resource corresponding to the feature and/or the metadata according to the feature and/or metadata of the data to be read and the algorithm of the storage policy of the data distribution processing system, and reads the
  • the storage node corresponding to the storage resource to be read data solves the problem that the storage architecture is incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared, and the storage resources are uniformly shared, thereby saving resources and maintenance overhead.
  • the data distribution processing system further includes:
  • Key-value KV application programming interface API client and/or key-value KV interface are examples of keys-value KV application programming interface API client and/or key-value KV interface.
  • the feature and/or metadata of the data to be read includes:
  • the key value KV of the data to be read is the key value KV of the data to be read.
  • the storage node corresponding to the storage resource that reads the data to be read includes:
  • the storage resource fragment is a fragment on the ring of the consistent hash hash algorithm, where the ring corresponds to a preset storage space.
  • the storage node corresponding to the storage resource that reads the data to be read includes:
  • the client BAC can be accessed through the key value KV interface, the object access client OAC interface, and the block access client BAC.
  • the interface or file access client FAC interface reads the data to be read.
  • the data distribution processing system When the data to be read is read by the OAC interface, the data distribution processing system reads the data to be read and returns it to the object data server OMDS, and the OMDS returns the data to be read to the OAC interface, and
  • the REST protocol is passed through the expressive state to perform REST protocol parsing and returned to the application.
  • the data distribution processing system reads the data to be read through the interface of the BAC and returns the data to the small computer system interface ISCSI object, the ISCSI object to be read The data is returned to the application.
  • the data distribution processing system When the data to be read is read by the FAC file, the data distribution processing system reads the data to be read and returns it to the file data server FMDS, and reads the data to be read through the FAC interface, and then passes the data.
  • the network file system NFS interface, the portable operating system interface POSIX interface, or the general Ethernet file system CIFS interface is returned to the application.
  • a data storage device is also provided in the embodiment, and the device is used to implement the foregoing embodiments and preferred embodiments, and details are not described herein.
  • the term "module” may implement a combination of software and/or hardware of a predetermined function.
  • the apparatus described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware, is also possible and contemplated.
  • FIG. 3 is a structural block diagram of a data storage device according to an embodiment of the present invention. As shown in FIG. 3, the device includes
  • the first obtaining module 32 is configured to acquire, according to the feature and/or metadata of the data to be stored, the storage policy of the data distribution processing system, the preset storage resource corresponding to the storage policy, and the feature And a storage resource corresponding to the metadata, where the preset storage resource is composed of storage resources of at least one or more storage nodes;
  • the storage module 34 is configured to store the to-be-stored data into a storage node corresponding to the storage resource.
  • the data distribution processing system acquires a storage resource corresponding to the feature and/or metadata according to the feature and/or metadata of the data to be stored and the algorithm of the storage policy of the data distribution processing system, and stores the to-be-stored
  • the data is stored in the storage node corresponding to the storage resource, which solves the problem that the storage architecture is incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared, and the storage resources are uniformly shared, thereby saving resources and maintenance overhead.
  • FIG. 4 is a block diagram showing the structure of a data reading apparatus according to an embodiment of the present invention. As shown in FIG. 4, the apparatus includes
  • the second obtaining module 42 is configured to acquire, according to the feature and/or metadata of the data to be read, and the storage policy of the data distribution processing system, in the preset storage resource corresponding to the storage policy, a storage resource corresponding to the feature and/or the metadata, wherein the preset storage resource is composed of storage resources of the at least one or more storage nodes;
  • the reading module 44 is configured to read a storage node corresponding to the storage resource of the data to be read.
  • the data distribution processing system acquires the storage resource corresponding to the feature and/or the metadata according to the feature and/or metadata of the data to be read and the algorithm of the storage policy of the data distribution processing system, and reads the
  • the storage node corresponding to the storage resource to be read data solves the problem that the storage architecture is incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared, and the storage resources are uniformly shared, thereby saving resources and maintenance overhead.
  • a data processing system including: a key value KV storage database, a key value KV application programming interface API interface,
  • the key value KV stores a database, and is configured to integrate storage resources on each storage node to form a unified preset storage resource;
  • the key value KV application programming interface API interface is configured to apply different storage nodes of the preset storage space according to the storage or read request of the data, and the key value Key of the data, wherein the preset storage resource is composed of at least one Or a storage resource consisting of multiple storage nodes.
  • the function of the key value KV application programming interface API interface includes at least one of the following:
  • the preset storage space is fragmented by a consistent hash hash algorithm to generate a storage resource fragment
  • system further includes: an operation and maintenance module OMM, a monitoring module,
  • the OMM is configured to set at least one of the following functions: management of the storage resource, division of the storage resource, configuration of the storage resource, and display of an alarm of the system;
  • the monitoring module is configured to monitor the status of the storage node and the disk.
  • system further includes: an object access client OAC interface, a block access client BAC interface, and a file access client FAC interface.
  • the object accesses the client OAC interface, the block access client BAC interface, or the file access client FAC interface is set to store or read the data of the key value KV storage database.
  • the method further includes: configuring a module
  • the configuration module is configured to configure a storage policy of the storage resource, and the storage policy includes: a copy mode, a snapshot support mode, a compression mode, and a cache mode.
  • the preferred embodiment provides a very large-scale, flexible, unified storage system system and method, which can fully utilize and share the original storage resources and storage architecture of the customer, and transform into an advanced, secure, unified mass storage architecture. Meet the ability to be compatible with distributed cache, block storage, traditional database storage, and distributed structured data storage.
  • the storage system of the preferred embodiment is capable of intelligently coping with the storage requirements of various different Internet devices, including conventional devices and new types of smart devices.
  • the invention adopts a novel algorithm to provide a unified client access layer, and supports access of multiple devices through a unified Key Value API (KV API) interface.
  • KV API Key Value API
  • the invention is compatible with the KV key value database interface, the object storage interface, the block interface, and the file interface, and can be applied to various application scenarios such as a public cloud, a private cloud, and a hybrid cloud.
  • the storage system of the preferred embodiment perfectly supports the existing multiple traditional storage models and storage architectures, and integrates and utilizes existing resources well.
  • Traditional storage methods are various, but there are common drawbacks such as single interface and simple function, which cannot meet the requirements of increasingly developed Internet storage.
  • Currently existed storage devices some use structured algorithms, some use unstructured algorithms to organize; some provide file-based access, some provide object-based access, and some provide access based on block storage, not Do the unification, sharing and elastic scaling of storage resources.
  • the invention provides a scalable, highly reliable, flexible and scalable unified storage resource through the unified distributed KV Store under the bottom layer, and provides the user with a KV interface, an object interface, a block interface, and a file interface, which can well satisfy various applications. The need for storage.
  • the storage system of the preferred embodiment has the characteristics of large capacity, low cost, high performance, and simple expansion. Through the integration and unified management of traditional storage devices, it can provide hundreds of petabytes of storage capacity, efficiently manage hundreds of billions of files, and a single directory can effectively support tens of millions of files.
  • the client application supports tera-scale record storage, efficient and flexible distributed record data storage architecture to meet the explosive growth of data storage and query in the information age. At the same time, the system is built with common hardware and common traditional systems, and supports online capacity expansion. The overall cost is much lower than that of a single traditional storage architecture, and it has good scalability and flexibility.
  • the preferred embodiment of the storage system supports large-scale concurrent reading and writing to meet the needs of large-scale cloud storage.
  • the smartness of the invention slices the stored data and distributes it to each storage node in the form of KV, so that the read and write operations can be concurrently executed at each storage node, thereby greatly improving performance.
  • a three-tier storage system that supports memory/SSD/hard disk, accelerates application access to files, and provides large concurrent read and write capabilities.
  • the KV Store provides high reliability of data through multiple copies or erasure codes, and realizes dynamic expansion of the storage system through redistribution of data fragments and data migration.
  • the unified storage mode of the preferred embodiment is particularly applicable to a virtualized storage portion in a cloud computing environment.
  • the block storage interface can be mounted as a local disk of the cloud computing cloud host.
  • the file interface is suitable for file sharing services in a cloud computing environment, or for storing virtual machine images, and object interfaces and KV interfaces. Applicable to provide universal storage platform services for third-party applications in the cloud environment.
  • the advantage of the present invention is that the underlying storage resources are uniformly shared, and no separate storage resources need to be established for each storage interface, which can greatly save resources and reduce maintenance overhead.
  • the invention may, of course, be embodied in various other embodiments and various modifications and changes can be made in accordance with the present invention without departing from the spirit and scope of the invention. And modifications are intended to fall within the scope of the appended claims.
  • Each of the preferred embodiments of the preferred embodiment satisfies a carrier-grade mass storage solution that is operational, manageable, maintainable, highly reliable, low-cost, and open, standardized, and supports KV storage, file storage, block storage, and object storage. It provides a rich external API interface and standard file interface and block interface access to meet the different needs of various applications.
  • FIG. 5 is a schematic diagram of an overall architecture of a storage system according to a preferred embodiment of the present invention. As shown in FIG. 5, the overall architecture of the storage system is provided in two layers.
  • the next layer is the storage core of the system.
  • the SSD disks distributed on each host node and the SAS/SATA hard disks are integrated to provide unified storage resources.
  • each storage space consists of some or all of the disks of different server nodes. Based on the storage space, you can configure the storage space policy, including copy mode, snapshot support, compression support, cache mode, and so on.
  • the Operation Maintenance Module (OMM) is set to manage system storage resources, partition storage space, storage space configuration, and system alarm display.
  • the monitoring module (Monitor) is configured to monitor the status of the storage node and the disk. When the storage node or disk is abnormal, an alarm is generated and the disk or service node fault handling process is entered.
  • the upper layer is the access layer of the data distribution processing system.
  • the application can access the underlying storage resources directly through the KV API, or access through the OAC object interface, or through the BAC block interface, or access the storage space through the FAC file interface.
  • FIG. 6 is a schematic diagram of a KV database storage process according to a preferred embodiment of the present invention. As shown in FIG. 6, the implementation method proposed by the present invention when a client initiates a data write request is described in detail. Including the following steps:
  • Step 601 When the KV API client (Client) writes data into the KV database, the hash value of the key value Key is obtained and mapped to a position on the ring;
  • Step 602 According to the location of the hash value of the key, obtain the fragment of the corresponding data, find the master storage node corresponding to all the replica sets of the storage domain, and write the data to the master node by searching the routing table of the fragment. on;
  • Step 603 The master node writes multiple copies of the data to other storage nodes of the KV Stroe according to the copy policy.
  • Step 604 after all the nodes are successfully stored, return a read/write request success message to the client, and complete the current reading and writing process.
  • the implementation method of the corresponding KV API Client read request includes the following steps:
  • Step 605 when the KV API Client reads data from the KV Store, find the hash value of the read key, and map it to a position on the ring;
  • Step 606 According to the location of the hash value of the key, obtain the fragment where the corresponding data is located, and find the master storage node corresponding to all the replica sets of the storage domain by searching the routing table of the fragment.
  • Step 607 if the state of the master node is normal, the stored data is read from the master of the replica set, otherwise the data is read from the other replica node;
  • Step 608 after reading the data, the data distribution processing system returns the response data to the application.
  • FIG. 7 is a schematic diagram of the internal composition of a KV API according to a preferred embodiment of the present invention.
  • the internal composition of the KV API is described in detail, including a routing table of spatial fragmentation, and the key is calculated by using a consistent hash algorithm. Fragmentation, find the storage server node corresponding to the fragment by looking up the routing table.
  • the data distribution processing system includes the underlying link management; the provision of the API interface mainly includes a Set write data interface, a Get read data interface, a Delete delete data interface, a consistency Hash algorithm and slice management, and a routing table management.
  • the KV API is a very versatile and simple interface, such as should be set to store user data.
  • FIG. 8 is a schematic diagram of distributed object storage according to a preferred embodiment of the present invention.
  • the external application accesses the unified storage system through the Rest interface.
  • REST Gateway performs REST protocol parsing, submits the data information and object information to be stored to the OAC (Object Access Client) object access client, and OAC stores the metadata information of the object into the KV Store through OMDS (Object Meta Data Server).
  • OAC Object Access Client
  • OMDS Object Meta Data Server
  • FIG. 9 is a schematic diagram of distributed block storage in accordance with a preferred embodiment of the present invention.
  • the external application accesses the block storage system through the iSCSI interface.
  • the ISCSI Target is converted to the internal interface of the BAC (Block Access Client).
  • the BAC converts the block data into the corresponding KV data and stores it in the KV Store through the KV API Client.
  • the ISCSI Target module converts the read request into the internal BAC interface.
  • the BAC reads the data from the KV Store through the KV API and returns it to the ISCSI Target.
  • the further ISCSI Target returns the data to the client application.
  • FIG. 10 is a schematic diagram of distributed file storage according to a preferred embodiment of the present invention.
  • the external application accesses the file storage system through an NFS interface, or a Posix, or CIFS interface.
  • NFS protocol, the Posix interface, or the CIFS interface is converted into an internal FAC (File Access Client) interface.
  • the FAC processes the metadata information of the file through FMDS (File Meta Data Server) for metadata organization and management, and the FMDS puts the file metadata information.
  • FMDS Fe Meta Data Server
  • the KV API is stored in the KV Store, and the FMDS returns the content location key information of the file.
  • the FAC then stores the corresponding file data information into the KV Store through the KV API Client interface.
  • the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware, but in many cases, the former is A better implementation.
  • the technical solution of the present invention which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a cell phone, a computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present invention.
  • Embodiments of the present invention also provide a storage medium.
  • the foregoing storage medium may be configured to store program code for performing the method steps of the above embodiment:
  • the foregoing storage medium may include, but not limited to, a USB flash drive, a Read-Only Memory (ROM), a Random Access Memory (RAM), a mobile hard disk, and a magnetic memory.
  • ROM Read-Only Memory
  • RAM Random Access Memory
  • a mobile hard disk e.g., a hard disk
  • magnetic memory e.g., a hard disk
  • the processor executes the method of the above embodiment according to the stored program code in the storage medium.
  • modules or steps of the present invention described above can be implemented by a general-purpose computing device that can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein.
  • the steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps thereof are fabricated as a single integrated circuit module.
  • the invention is not limited to any specific combination of hardware and software.
  • the data distribution processing system acquires the preset storage resource corresponding to the storage policy according to the feature and/or metadata of the data to be stored and the storage policy of the data distribution processing system.
  • a storage resource corresponding to the feature and/or the metadata wherein the preset storage resource is composed of storage resources of at least one or more storage nodes; and the data to be stored is stored in a storage node corresponding to the storage resource,
  • the storage architecture is incompatible with each other, the interfaces are not unified, and the storage resources cannot be shared.
  • the unified sharing of storage resources is realized, which saves resources and maintenance overhead.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种数据存储、读取的方法、装置及系统,其中,该方法通过数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成(S102);将该待存储数据存储至与该存储资源对应的存储节点中(S104),解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。

Description

数据存储、读取的方法、装置及系统 技术领域
本发明涉及通信领域,具体而言,涉及一种数据存储、读取的方法、装置及系统。
背景技术
随着互联网二代(Web2.0)时代的到来,面对互联网数据中心(Internet Data Center,简称为IDC)爆炸式大数据需求,以及多种多样的智能终端,物联网的普及。传统存储服务的弊端越来越明显,不能满足日益发展的互联网的需求。建立面向服务的“云存储”环境已经从遥不可及的奢望变成一种迫切的现实需求。
在相关技术中,存储架构可以划分为两种:一是传统专用的存储架构,如存储区域网络(Network Attached Storage,简称为SAN)和网络接入服务器(Storage Area Network and SAN Protocols,简称为NAS)等,由普通的服务器组成,其单位成本昂贵,可扩展性差,不适合大范围部署,但是其性能和可靠性有一定的优势。
另外一种是传统云存储架构,如分布式文件系统,这种分布式文件系统成本低、可线性扩展,所以其规模和全系统的吞吐量可以做得很大。但是其缺点在于单个点的可靠性和性能不如SAN和NAS。
上述两种存储架构互不兼容,用户不能灵活的根据实际情况动态选择自己想要的系统,同时这两种架构的存储系统,也使用各自固定的接口,不能应对多种多样的以太网(Internet)客户和他们各自的存储协议。
针对相关技术中,存储架构互不兼容,接口不统一,不能共享存储资源的问题,目前尚未提出有效的解决方案。
发明内容
本发明提供了一种数据存储、读取的方法、装置及系统,以至少解决相关技术中存储架构互不兼容,接口不统一,不能共享存储资源的问题。
根据本发明的一个实施例,提供了一种数据存储方法,包括:
数据分发处理系统根据待存储数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
将所述待存储数据存储至与所述存储资源对应的存储节点中。
在本发明的实施例中,所述数据分发处理系统还包括:
键值KV应用程序编程接口API客户端和/或键值KV接口。
在本发明的实施例中,所述待存储数据的特征和/或元数据包括:
所述待存储数据的键值KV。
在本发明的实施例中,将所述待存储数据存储至与所述存储资源对应的存储节点中,包括:
获取与所述存储资源对应的存储资源分片,其中,所述预设存储资源由所述存储资源分片组成;
依据所述存储资源分片在所述预设存储资源的空间路由表,查找与所述待存储数据对应的存储节点,所述存储节点设置为接收写入所述待存储数据。
在本发明的实施例中,在所述存储策略为一致性哈希hash算法的情况下,所述存储资源分片为一致性哈希hash算法环上的分片,其中,所述环对应所述预设存储资源。
在本发明的实施例中,将所述待存储数据存储至与所述存储资源对应的存储节点中包括:
将所述待存储数据的一个或多个副本到写入到其它存储节点中。
在本发明的实施例中,所述方法包括:
通过键值KV接口、对象访问客户端OAC接口、块访问客户端BAC接口或文件访问客户端FAC接口获取所述待存储数据。
在本发明的实施例中,在通过所述OAC接口获取所述待存储数据时,所述方法还包括:
所述数据分发处理系统接收应用程序通过表述性状态传递REST网关进行REST协议解析,将对象信息提交给OAC,所述OAC通过对象数据服务器OMDS获取所述对象信息对应的所述待存储数据。
在本发明的实施例中,在通过所述BAC接口获取所述待存储数据时,所述方法还包括:
所述数据分发处理系统接收应用程序通过小型计算机系统接口ISCSI对象接收到ISCSI指令后,转换为BAC的内部接口,所述BAC把块数据转换为对应的所述待存储数据。
在本发明的实施例中,通过所述FAC接口获取所述待存储数据时,所述方法还包括:
所述数据分发处理系统接收应用程序通过网络文件系统NFS接口,可移植操作系统接口POSIX接口,或通用以太网文件系统CIFS接口转换成内部的文件接入客户端FAC接口,所述FAC把文件的元数据信息通过文件数据服务器FMDS进行元数据的组织管理,获取所述元数据信息对应的待存储数据。
根据本发明的另一个实施例,还提供了一种数据读取方法,包括:
数据分发处理系统根据待读取数据的特征和/或元数据,以及所述数据分发处理系统的存 储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
读取所述待读取数据的所述存储资源对应的存储节点。
在本发明的实施例中,所述数据分发处理系统还包括:
键值KV应用程序编程接口API客户端和/或键值KV接口。
在本发明的实施例中,所述待读取数据的特征和/或元数据包括:
所述待读取数据的键值KV。
在本发明的实施例中,所述读取所述待读取数据的所述存储资源对应的存储节点包括:
获取与所述存储资源对应的存储资源分片,其中,所述预设存储资源由所述存储资源分片组成;
依据所述存储资源分片在所述预设存储资源的空间路由表,查找与所述待读取数据对应的存储节点,读取所述对应的存储节点。
在本发明的实施例中,在所述存储策略为一致性哈希hash算法的情况下,所述存储资源分片为一致性哈希hash算法环上的分片,其中,所述环对应所述预设存储空间。
在本发明的实施例中,读取所述待读取数据的所述存储资源对应的存储节点包括:
查找所述待读取数据对应的副本集的存储节点,从所述副本集的存储节点上读取所述待读取数据。
在本发明的实施例中,所述方法包括:
通过键值KV接口、对象访问客户端OAC接口、块访问客户端BAC接口或文件访问客户端FAC接口读取所述待读取数据。
在本发明的实施例中,在通过所述OAC接口读取所述待读取数据时,所述方法还包括:
所述数据分发处理系统读取所述待读取数据返回给对象数据服务器OMDS,所述OMDS将所述待读取数据返回给所述OAC接口,并通过表述性状态传递REST网关进行REST协议解析,返回给应用程序。
在本发明的实施例中,在通过所述BAC接口读取所述待读取数据时,所述方法还包括:
所述数据分发处理系统通过所述BAC的接口读取所述待读取数据返回给小型计算机系统接口ISCSI对象,所述ISCSI对象将所述待读取数据返回给应用程序。
在本发明的实施例中,在通过所述FAC接口读取所述待读取数据时,所述方法还包括:
所述数据分发处理系统读取所述待读取数据返回给文件数据服务器FMDS,并通过所述 FAC接口读取所述待读取数据,再通过网络文件系统NFS接口,可移植操作系统接口POSIX接口,或通用以太网文件系统CIFS接口返回给应用程序。
根据本发明的另一个实施例,还提供了一种数据存储装置,包括:
第一获取模块,设置为数据分发处理系统根据待存储数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
存储模块,设置为将所述待存储数据存储至与所述存储资源对应的存储节点中。
根据本发明的另一个实施例,还提供了一种数据读取装置,包括:
第二获取模块,设置为数据分发处理系统根据待读取数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
读取模块,设置为读取所述待读取数据的所述存储资源对应的存储节点。
根据本发明的另一个实施例,还提供了一种数据处理系统,包括:键值KV存储数据库,键值KV应用程序编程接口API接口;
所述键值KV存储数据库,设置为整合各个存储节点上的存储资源形成统一预设存储资源;
所述键值KV应用程序编程接口API接口,设置为根据数据的存储或者读取请求,以及所述数据的关键值Key申请预设存储空间的不同的存储节点,其中,所述预设存储资源由多个存储节点的存储资源组成。
在本发明的实施例中,所述键值KV应用程序编程接口API接口的功能包括以下至少之一:
写数据接口,读数据接口,以及删除数据接口处理所述数据;
对所述预设存储空间通过一致性哈希hash算法进行分片,生成存储资源分片;
管理所述存储资源分片;
管理所述存储资源分片对应的空间路由表。
在本发明的实施例中,所述系统还包括:操作维护模块OMM,监控模块,
所述OMM,设置为设置以下至少之一功能:所述存储资源的管理,所述存储资源的划分,所述存储资源的配置,以及系统的告警显示;
所述监控模块,设置为监控所述存储节点和所述磁盘的状态。
在本发明的实施例中,所述系统还包括:对象访问客户端OAC接口,块访问客户端BAC接口,以及文件访问客户端FAC接口,
对象访问客户端OAC接口,块访问客户端BAC接口,或者文件访问客户端FAC接口设置为对所述键值KV存储数据库的数据进行存储或者读取。
在本发明的实施例中,还包括:配置模块
所述配置模块设置为配置存储资源的存储策略,所述存储策略包括:副本模式,快照支持模式,压缩模式,缓存模式。
通过本发明,通过数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;将该待存储数据存储至与该存储资源对应的存储节点中,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。
附图说明
此处所说明的附图用来提供对本发明的进一步理解,构成本申请的一部分,本发明的示意性实施例及其说明用于解释本发明,并不构成对本发明的不当限定。在附图中:
图1是根据本发明实施例的一种数据存储方法的流程图;
图2是根据本发明实施例的一种数据读取方法的流程图;
图3是根据本发明实施例的一种数据存储装置的结构框图;
图4是根据本发明实施例的一种数据读取装置的结构框图;
图5是根据本发明优选实施例的存储系统整体架构的示意图;
图6是根据本发明优选实施例的KV数据库存储过程的示意图;
图7是根据本发明优选实施例的KV API内部组成的示意图;
图8是根据本发明优选实施例的分布式对象存储的示意图;
图9是根据本发明优选实施例的分布式块存储的示意图;
图10是根据本发明优选实施例的分布式文件存储的示意图。
具体实施方式
下文中将参考附图并结合实施例来详细说明本发明。需要说明的是,在不冲突的情况下,本申请中的实施例及实施例中的特征可以相互组合。
需要说明的是,本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象,而不必用于描述特定的顺序或先后次序。
在本实施例中提供了一种数据存储方法,图1是根据本发明实施例的一种数据存储方法的流程图,如图1所示,该流程包括如下步骤:
步骤S102,数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;
步骤S104,将该待存储数据存储至与该存储资源对应的存储节点中。
通过上述步骤,数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略的算法,获取与该特征和/或元数据对应的存储资源,存储该待存储数据存储至与该存储资源对应的存储节点中,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。
在本实施例中该数据分发处理系统可以包括:键值(Key Value,简称为KV)应用程序编程接口客户端(Application Programming Interface,简称为API)客户端和/或键值(KV)接口。该待存储数据的特征和/或元数据包括:该待存储数据的键值(KV)。
在本实施例中,将该待存储数据存储至与该存储资源对应的存储节点中,包括:
获取与该存储资源对应的存储资源分片,其中,该预设存储资源由该存储资源分片组成;
依据该存储资源分片在该预设存储资源的空间路由表,查找与该待存储数据对应的存储节点,该存储节点用于接收写入该待存储数据。
在本实施例中,在该存储策略为一致性哈希hash算法的情况下,该存储资源分片为一致性哈希hash算法环上的分片,其中,该环对应该预设存储资源。
在本实施例中,将该待存储数据存储至与该存储资源对应的存储节点中包括:
将该待存储数据的一个或多个副本到写入到其它存储节点中。
在本实施例中,可以通过键值KV接口、对象访问客户端(Object Access Client,简称为OAC)接口、块访问客户端(Block Access Client,简称为BAC)接口或文件访问客户端(File Access Client,简称为FAC)文件接口获取该待存储数据。
其中,在通过该OAC接口获取该待存储数据时,该数据分发处理系统接收应用程序通过表述性状态传递(Representational State Transfer,简称REST)网关进行REST协议解析,将对象信息提交给OAC,该OAC通过对象数据服务器(Object Meta Data Server,简称为OMDS)获取该对象信息对应的该待存储数据。
其中,在通过该BAC接口获取该待存储数据时,该数据分发处理系统接收应用程序通过小型计算机系统接口(Internet Small Computer System Interface,简称为ISCSI)对象接收到 ISCSI指令后,转换为BAC的内部接口,该BAC把块数据转换为对应的该待存储数据。
其中,在通过该FAC文件获取该待存储数据时,该数据分发处理系统接收应用程序通过网络文件系统(Network File System,简称为NFS)接口,可移植操作系统接口(Portable Operating SystemInterface,简称为POSIX)接口,或通用以太网文件系统(Common Internet File System,简称为CIFS)接口转换成内部的文件接入客户端FAC接口,该FAC把文件的元数据信息通过文件数据服务器(File Meta Data Server,简称为FMDS)进行元数据的组织管理,获取该元数据信息对应的待存储数据。
在本实施例中提供了一种数据读取方法,图2是根据本发明实施例的一种数据读取方法的流程图,如图2所示,该流程包括如下步骤:
步骤S202,数据分发处理系统根据待读取数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;
步骤S204,读取该待读取数据的该存储资源对应的存储节点。
通过上述步骤,数据分发处理系统根据待读取数据的特征和/或元数据,以及该数据分发处理系统的存储策略的算法,获取与该特征和/或元数据对应的存储资源,读取该待读取数据的该存储资源对应的存储节点,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。
在本实施例中,该数据分发处理系统还包括:
键值KV应用程序编程接口API客户端和/或键值KV接口。
在本实施例中,该待读取数据的特征和/或元数据包括:
该待读取数据的键值KV。
在本实施例中,该读取该待读取数据的该存储资源对应的存储节点包括:
获取与该存储资源对应的存储资源分片,其中,该预设存储资源由该存储资源分片组成;
依据该存储资源分片在该预设存储资源的空间路由表,查找与该待读取数据对应的存储节点,读取该对应的存储节点。
在本实施例中,在该存储策略为一致性哈希hash算法的情况下,该存储资源分片为一致性哈希hash算法环上的分片,其中,该环对应该预设存储空间。
在本实施例中,读取该待读取数据的该存储资源对应的存储节点包括:
查找该待读取数据对应的副本集的存储节点,从该副本集的存储节点上读取该待读取数据。
在本实施例中,可以通过键值KV接口、对象访问客户端OAC接口、块访问客户端BAC 接口或文件访问客户端FAC接口读取该待读取数据。
其中,在通过该OAC接口读取该待读取数据时,该数据分发处理系统读取该待读取数据返回给对象数据服务器OMDS,该OMDS将该待读取数据返回给该OAC接口,并通过表述性状态传递REST网关进行REST协议解析,返回给应用程序。
其中,在通过该BAC接口读取该待读取数据时,该数据分发处理系统通过该BAC的接口读取该待读取数据返回给小型计算机系统接口ISCSI对象,该ISCSI对象将该待读取数据返回给应用程序。
其中,在通过该FAC文件读取该待读取数据时,该数据分发处理系统读取该待读取数据返回给文件数据服务器FMDS,并通过该FAC接口读取该待读取数据,再通过网络文件系统NFS接口,可移植操作系统接口POSIX接口,或通用以太网文件系统CIFS接口返回给应用程序。
在本实施例中还提供了一种数据存储装置,该装置用于实现上述实施例及优选实施方式,已经进行过说明的不再赘述。如以下所使用的,术语“模块”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置较佳地以软件来实现,但是硬件,或者软件和硬件的组合的实现也是可能并被构想的。
图3是根据本发明实施例的一种数据存储装置的结构框图,如图3所示,该装置包括
第一获取模块32,设置为数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;
存储模块34,设置为将该待存储数据存储至与该存储资源对应的存储节点中。
通过上述装置,数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略的算法,获取与该特征和/或元数据对应的存储资源,存储该待存储数据存储至与该存储资源对应的存储节点中,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。
图4是根据本发明实施例的一种数据读取装置的结构框图,如图4所示,该装置包括
第二获取模块42,设置为数据分发处理系统根据待读取数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;
读取模块44,设置为读取该待读取数据的该存储资源对应的存储节点。
通过上述步骤,数据分发处理系统根据待读取数据的特征和/或元数据,以及该数据分发处理系统的存储策略的算法,获取与该特征和/或元数据对应的存储资源,读取该待读取数据的该存储资源对应的存储节点,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。
在本发明的实施例中,还提供了一种数据处理系统,包括:键值KV存储数据库,键值KV应用程序编程接口API接口,
该键值KV存储数据库,设置为整合各个存储节点上的存储资源形成统一预设存储资源;
该键值KV应用程序编程接口API接口,设置为根据数据的存储或者读取请求,以及该数据的关键值Key申请预设存储空间的不同的存储节点,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成。
在本实施例中,该键值KV应用程序编程接口API接口的功能包括以下至少之一:
写数据接口,读数据接口,以及删除数据接口处理该数据;
对该预设存储空间通过一致性哈希hash算法进行分片,生成存储资源分片;
管理该存储资源分片;
管理该存储资源分片对应的空间路由表。
在本实施例中,该系统还包括:操作维护模块OMM,监控模块,
该OMM,设置为设置以下至少之一功能:该存储资源的管理,该存储资源的划分,该存储资源的配置,以及系统的告警显示;
该监控模块,设置为监控该存储节点和该磁盘的状态。
在本实施例中,该系统还包括:对象访问客户端OAC接口,块访问客户端BAC接口,以及文件访问客户端FAC接口,
对象访问客户端OAC接口,块访问客户端BAC接口,或者文件访问客户端FAC接口设置为对该键值KV存储数据库的数据进行存储或者读取。
在本实施例中,还包括:配置模块
该配置模块设置为配置存储资源的存储策略,该存储策略包括:副本模式,快照支持模式,压缩模式,缓存模式。
下面结合优选实施例和实施方式对本发明进行详细说明。
本优选实施例提供了一种超大规模的,弹性的,统一存储系统系统和方法,能够充分利用和共享客户原有的存储资源和存储架构,转化为先进的,安全的,统一海量存储架构,满足同时兼容分布式高速缓冲、块存储、传统数据库存储、分布式结构化数据存储的能力。
本优选实施例的存储系统能够智能应对各种不同的互联网设备的存储需求,包含传统设备和新型的智能设备。本发明采用一种新型的算法,提供统一的客户端接入层,通过统一的Key Value API(KV API)接口,支持多种设备的接入。本发明兼容KV键值数据库接口,对象存储接口,块接口,以及文件接口,可以应用于公有云,私有云,混合云等多种应用场景, 能够满足多种新型互联网设备,例如传统PC,新型平板电脑,智能移动终端等,智能物联网设备,Andriod设备,苹果iOS设备等对存储的访问需求。
本优选实施例存储系统完美支持现有的多种传统存储模型和存储架构,很好的整合和利用已有的资源。传统的存储方式多种多样,但是普遍存在接口单一,功能简单的弊病,不能满足日益发展互联网存储要求。目前存在的存储设备,有的采用结构化算法,有的采用非结构化算法组织;有的提供基于文件的方式访问,有的提供基于对象的方式访问,有的提供基于块存储方式访问,不能做到存储资源的统一,共享和弹性伸缩。本发明通过底层统一的分布式KV Store存储数据,提供可扩展,高可靠,弹性伸缩的统一存储资源,为用户提供KV接口,对象接口,块接口,文件接口,能够很好的满足各种应用对存储的需求。
本优选实施例存储系统具有容量大,成本低,性能高,扩容简单的特点。通过对传统存储设备的整合和统一管理,可以提供上百PB级的存储容量,高效的管理千亿个文件,单目录可有效支持千万级的文件数量。客户应用支持万亿级记录存储,高效灵活的分布式记录数据存储架构,可满足信息时代爆炸式增长的数据存储和查询。同时系统利用通用硬件和通用传统系统构建,支持在线扩容功能,整体成本远低于单一的传统存储架构,并且具有良好的可扩充性和灵活性。
本优选实施例存储系统支持大规模并发读写,满足大规模云存储需要。本发明智能将存储数据进行切片,以KV的形式分布到各个存储节点中去,使读写操作可在各存储节点并发执行,大大提升性能。支持内存/SSD/硬盘的三级存储系统,加速应用对文件的访问,提供大并发的读写能力。KV Store通过多副本或纠删码提供数据的高可靠性,通过数据分片的重分布和数据迁移实现存储系统的动态伸缩。
本优选实施例的统一存储方式尤其适用于云计算环境下的虚拟化存储部分。云计算对外提供云服务时,块存储接口可以作为云计算云主机的本地磁盘挂载,文件接口适用于云计算环境下的文件共享服务,或用于存储虚拟机镜像,而对象接口和KV接口适用于应用于为云环境下的第三方应用提供通用的存储平台服务。本发明的优势在于底层的存储资源是统一共享的,不需要为每一种存储接口建立单独的存储资源,这样可以大大的节约资源和减少维护的开销。当然,本发明还可有其他多种实施例,在不背离本发明精神及其实质的情况下,熟悉本领域的技术人员可根据本发明作出各种相应的改变和变形,但这些相应的改变和变形都应属于本发明所附的权利要求的保护范围。
本优选实施例的各个实例满足可运营、可管理、可维护、高可靠性、低成本和具备开放性、标准化的电信级海量存储解决方案,支持KV存储,文件存储、块存储,对象存储,对外提供丰富的对外API接口和标准的文件接口和块接口访问,满足各种应用的不同需求。
图5是根据本发明优选实施例的存储系统整体架构的示意图,如图5所示,提供的存储系统整体架构图,分为二层。
下面一层是系统的存储核心。通过软件的方法,把分布在各个主机节点上的SSD盘和SAS/SATA硬盘整合起来,对外提供统一的存储资源。在存储资源的基础上,申请不同的存储 空间,每个存储空间由不同服务器节点的部分或全部磁盘构成。在存储空间的基础上,可以配置存储空间的策略,包括副本模式,快照支持,压缩支持,缓存模式等。操作维护模块(Operation Maintenance Module,简称为OMM)设置为对系统存储资源的管理,存储空间的划分,存储空间的配置,已经系统的告警显示等。监控模块(Monitor),设置为监控存储节点和磁盘的状态,在存储节点或磁盘异常时,产生告警,并进入磁盘或服务节点故障处理流程。
上面一层是数据分发处理系统访问层,应用可以直接通过KV API访问底层的存储资源,或通过OAC对象接口访问,或通过BAC块接口访问,或通过FAC文件接口访问存储空间。
图6是根据本发明优选实施例的KV数据库存储过程的示意图,如图6所示,详细描述客户端发起数据写请求时本发明提出的实现方法。包括如下步骤:
步骤601,KV API客户端(Client)向KV数据库(Store)中写入数据时,求出关键值Key的Hash值,映射到环上一个位置;
步骤602,根据Key的Hash值的位置,得到对应的数据所在分片,通过查找分片的路由表,找出对应存储域所有副本集的主(Master)存储节点,将数据写入到Master节点上;
步骤603,由该Master节点,根据副本策略,把数据写多个副本到KV Stroe的其它存储节点中;
步骤604,在所有节点存储成功后,给客户端返回读写请求成功消息,完成本次读写流程。
相应的KV API Client读请求的实现方法包括如下步骤:
步骤605,KV API Client从KV Store中读取数据时,求出读取Key的Hash值,映射到环上一个位置;
步骤606,根据Key的Hash值的位置,得到对应的数据所在分片,通过查找分片的路由表,找出对应存储域所有副本集的Master存储节点
步骤607,如果由该Master节点状态正常,则从副本集的Master读取存储的数据,否则从其它副本节点读取数据;
步骤608,读取到数据后,有数据分发处理系统给应用返回响应的数据。
图7是根据本发明优选实施例的KV API内部组成的示意图,如图7所示,详细描述KV API内部组成,包括空间分片的路由表,key使用一致性hash算法计算得出对应的所在分片,通过查找路由表,找到分片对应的存储服务器节点。数据分发处理系统包含了底层的链路管理;API接口的提供,主要包括Set写数据接口,Get读数据接口,Delete删除数据接口;一致性Hash算法和分片的管理;路由表的管理等。KV API是非常通用的简单的接口,比如应设置为存储用户数据等。
图8是根据本发明优选实施例的分布式对象存储的示意图,如图8所示,详细描述对象存储接口场景下的数据存储过程。外部应用程序通过Rest接口访问统一存储系统。REST  Gateway进行REST的协议解析,把需要存储的数据信息和对象信息提交给OAC(Object Access Client)对象访问客户端,OAC把对象的元数据信息通过OMDS(Object Meta Data Server)存储到KV Store中,同时生成对应对象数据的key信息,把对象的数据信息直接通过KV API存储到KV Store中。
图9是根据本发明优选实施例的分布式块存储的示意图,如图9所示,详细描述块存储接口场景下的数据存储过程。外部应用程序通过iSCSI接口访问块存储系统。ISCSI Target接收到iSCSI指令后,转换为BAC(Block Access Client)的内部接口,BAC把块数据转换为对应的KV数据,通过KV API Client存储到KV Store中。对应的读取数据时,有ISCSI Target模块把读请求转换为内部的BAC的接口,BAC通过KV API从KV Store中读取数据返回给ISCSI Target,进一步的ISCSI Target把数据返回给客户端应用。
图10是根据本发明优选实施例的分布式文件存储的示意图,如图10所示,详细描述文件存储接口场景下的数据存储过程。外部应用程序通过NFS接口,或Posix,或CIFS接口访问文件存储系统。结果NFS协议,Posix接口,或CIFS接口转换成内部的FAC(File Access Client)接口,FAC把文件的元数据信息通过FMDS(File Meta Data Server)进行元数据的组织管理,FMDS把文件元数据信息通过KV API存储到KV Store中,同时FMDS返回文件的内容位置key信息,FAC再把相应的文件数据信息通过KV API Client接口存储到KV Store中。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到根据上述实施例的方法可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件,但很多情况下前者是更佳的实施方式。基于这样的理解,本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中,包括若干指令用以使得一台终端设备(可以是手机,计算机,服务器,或者网络设备等)执行本发明各个实施例所述的方法。
本发明的实施例还提供了一种存储介质。可选地,在本实施例中,上述存储介质可以被设置为存储用于执行上述实施例方法步骤的程序代码:
可选地,在本实施例中,上述存储介质可以包括但不限于:U盘、只读存储器(ROM,Read-Only Memory)、随机存取存储器(RAM,Random Access Memory)、移动硬盘、磁碟或者光盘等各种可以存储程序代码的介质。
可选地,在本实施例中,处理器根据存储介质中已存储的程序代码执行上述实施例的方法。
显然,本领域的技术人员应该明白,上述的本发明的各模块或各步骤可以用通用的计算装置来实现,它们可以集中在单个的计算装置上,或者分布在多个计算装置所组成的网络上,可选地,它们可以用计算装置可执行的程序代码来实现,从而,可以将它们存储在存储装置中由计算装置来执行,并且在某些情况下,可以以不同于此处的顺序执行所示出或描述的步骤,或者将它们分别制作成各个集成电路模块,或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样,本发明不限制于任何特定的硬件和软件结合。
以上所述仅为本发明的优选实施例而已,并不用于限制本发明,对于本领域的技术人员来说,本发明可以有各种更改和变化。凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。
工业实用性
基于本发明实施例提供的上述技术方案,数据分发处理系统根据待存储数据的特征和/或元数据,以及该数据分发处理系统的存储策略,在该存储策略对应的预设存储资源中,获取与该特征和/或元数据对应的存储资源,其中,该预设存储资源由至少一个或多个存储节点的存储资源组成;将该待存储数据存储至与该存储资源对应的存储节点中,解决了存储架构互不兼容,接口不统一,不能共享存储资源的问题,实现了存储资源的统一共享,节约了资源和维护开销。

Claims (27)

  1. 一种数据存储方法,包括:
    数据分发处理系统根据待存储数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
    将所述待存储数据存储至与所述存储资源对应的存储节点中。
  2. 根据权利要求1所述的方法,其中,所述数据分发处理系统还包括:
    键值KV应用程序编程接口API客户端和/或键值KV接口。
  3. 根据权利要求1所述的方法,其中,所述待存储数据的特征和/或元数据包括:
    所述待存储数据的键值KV。
  4. 根据权利要求1所述的方法,其中,将所述待存储数据存储至与所述存储资源对应的存储节点中,包括:
    获取与所述存储资源对应的存储资源分片,其中,所述预设存储资源由所述存储资源分片组成;
    依据所述存储资源分片在所述预设存储资源的空间路由表,查找与所述待存储数据对应的存储节点,所述存储节点用于接收写入所述待存储数据。
  5. 根据权利要求4所述的方法,其中,在所述存储策略为一致性哈希hash算法的情况下,所述存储资源分片为一致性哈希hash算法环上的分片,其中,所述环对应所述预设存储资源。
  6. 根据权利要求1所述的方法,其中,将所述待存储数据存储至与所述存储资源对应的存储节点中包括:
    将所述待存储数据的一个或多个副本到写入到其它存储节点中。
  7. 根据权利要求1所述的方法,其中,所述方法包括:
    通过键值KV接口、对象访问客户端OAC接口、块访问客户端BAC接口或文件访问客户端FAC接口获取所述待存储数据。
  8. 根据权利要求7所述的方法,其中,在通过所述OAC接口获取所述待存储数据时,所述方法还包括:
    所述数据分发处理系统接收应用程序通过表述性状态传递REST网关进行REST协议解析,将对象信息提交给OAC,所述OAC通过对象数据服务器OMDS获取所述对象信息对应的所述待存储数据。
  9. 根据权利要求7所述的方法,其中,在通过所述BAC接口获取所述待存储数据时,所述 方法还包括:
    所述数据分发处理系统接收应用程序通过小型计算机系统接口ISCSI对象接收到ISCSI指令后,转换为BAC的内部接口,所述BAC把块数据转换为对应的所述待存储数据。
  10. 根据权利要求7所述的方法,其中,在通过所述FAC接口获取所述待存储数据时,所述方法还包括:
    所述数据分发处理系统接收应用程序通过网络文件系统NFS接口,可移植操作系统接口POSIX接口,或通用以太网文件系统CIFS接口转换成内部的文件接入客户端FAC接口,所述FAC把文件的元数据信息通过文件数据服务器FMDS进行元数据的组织管理,获取所述元数据信息对应的待存储数据。
  11. 一种数据读取方法,包括:
    数据分发处理系统根据待读取数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由由至少一个或多个存储节点的存储资源组成;
    读取所述待读取数据的所述存储资源对应的存储节点。
  12. 根据权利要求11所述的方法,其中,所述数据分发处理系统还包括:
    键值KV应用程序编程接口API客户端和/或键值KV接口。
  13. 根据权利要求11所述的方法,其中,所述待读取数据的特征和/或元数据包括:
    所述待读取数据的键值KV。
  14. 根据权利要求11所述的方法,其中,所述读取所述待读取数据的所述存储资源对应的存储节点包括:
    获取与所述存储资源对应的存储资源分片,其中,所述预设存储资源由所述存储资源分片组成;
    依据所述存储资源分片在所述预设存储资源的空间路由表,查找与所述待读取数据对应的存储节点,读取所述对应的存储节点。
  15. 根据权利要求11所述的方法,其中,在所述存储策略为一致性哈希hash算法的情况下,所述存储资源分片为一致性哈希hash算法环上的分片,其中,所述环对应所述预设存储空间。
  16. 根据权利要求11所述的方法,其中,读取所述待读取数据的所述存储资源对应的存储节点包括:
    查找所述待读取数据对应的副本集的存储节点,从所述副本集的存储节点上读取所 述待读取数据。
  17. 根据权利要求11所述的方法,其中,所述方法包括:
    通过键值KV接口、对象访问客户端OAC接口、块访问客户端BAC接口或文件访问客户端FAC接口读取所述待读取数据。
  18. 根据权利要求17所述的方法,其中,在通过所述OAC接口读取所述待读取数据时,所述方法还包括:
    所述数据分发处理系统读取所述待读取数据返回给对象数据服务器OMDS,所述OMDS将所述待读取数据返回给所述OAC接口,并通过表述性状态传递REST网关进行REST协议解析,返回给应用程序。
  19. 根据权利要求17所述的方法,其中,在通过所述BAC接口读取所述待读取数据时,所述方法还包括:
    所述数据分发处理系统通过所述BAC的接口读取所述待读取数据返回给小型计算机系统接口ISCSI对象,所述ISCSI对象将所述待读取数据返回给应用程序。
  20. 根据权利要求17所述的方法,其中,在通过所述FAC接口读取所述待读取数据时,所述方法还包括:
    所述数据分发处理系统读取所述待读取数据返回给文件数据服务器FMDS,并通过所述FAC接口读取所述待读取数据,再通过网络文件系统NFS接口,可移植操作系统接口POSIX接口,或通用以太网文件系统CIFS接口返回给应用程序。
  21. 一种数据存储装置,包括:
    第一获取模块,设置为数据分发处理系统根据待存储数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
    存储模块,设置为将所述待存储数据存储至与所述存储资源对应的存储节点中。
  22. 一种数据读取装置,包括:
    第二获取模块,设置为数据分发处理系统根据待读取数据的特征和/或元数据,以及所述数据分发处理系统的存储策略,在所述存储策略对应的预设存储资源中,获取与所述特征和/或元数据对应的存储资源,其中,所述预设存储资源由至少一个或多个存储节点的存储资源组成;
    读取模块,设置为读取所述待读取数据的所述存储资源对应的存储节点。
  23. 一种数据处理系统,包括:键值KV存储数据库,键值KV应用程序编程接口API接口,
    所述键值KV存储数据库,设置为整合各个存储节点上的存储资源形成统一预设存储资源;
    所述键值KV应用程序编程接口API接口,设置为根据数据的存储或者读取请求,以及所述数据的关键值Key申请预设存储空间的不同的存储节点,其中,所述预设存储资源由多个存储节点的存储资源组成。
  24. 根据权利要求23所述的系统,其中,所述键值KV应用程序编程接口API接口的功能包括以下至少之一:
    写数据接口,读数据接口,以及删除数据接口处理所述数据;
    对所述预设存储空间通过一致性哈希hash算法进行分片,生成存储资源分片;
    管理所述存储资源分片;
    管理所述存储资源分片对应的空间路由表。
  25. 根据权利要求23所述的系统,其中,所述系统还包括:操作维护模块OMM,监控模块,
    所述OMM,设置为设置以下至少之一功能:所述存储资源的管理,所述存储资源的划分,所述存储资源的配置,以及系统的告警显示;
    所述监控模块,设置为监控所述存储节点和所述磁盘的状态。
  26. 根据权利要求23所述的系统,其中,所述系统还包括:对象访问客户端OAC接口,块访问客户端BAC接口,以及文件访问客户端FAC接口,
    对象访问客户端OAC接口,块访问客户端BAC接口,或者文件访问客户端FAC接口设置为对所述键值KV存储数据库的数据进行存储或者读取。
  27. 根据权利要求23所述的系统,其中,还包括:配置模块
    所述配置模块设置为配置存储资源的存储策略,所述存储策略包括:副本模式,快照支持模式,压缩模式,缓存模式。
PCT/CN2016/072035 2015-05-12 2016-01-25 数据存储、读取的方法、装置及系统 WO2016180055A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510240369.4 2015-05-12
CN201510240369.4A CN106293490A (zh) 2015-05-12 2015-05-12 数据存储、读取的方法、装置及系统

Publications (1)

Publication Number Publication Date
WO2016180055A1 true WO2016180055A1 (zh) 2016-11-17

Family

ID=57248717

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/072035 WO2016180055A1 (zh) 2015-05-12 2016-01-25 数据存储、读取的方法、装置及系统

Country Status (2)

Country Link
CN (1) CN106293490A (zh)
WO (1) WO2016180055A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109358813A (zh) * 2018-10-10 2019-02-19 郑州云海信息技术有限公司 一种分布式存储系统的扩容方法及装置
CN111290700A (zh) * 2018-12-10 2020-06-16 北京京东尚科信息技术有限公司 分布式数据读写方法和系统
CN111796767A (zh) * 2020-06-24 2020-10-20 北京青云科技股份有限公司 一种分布式文件系统及数据管理方法
CN112434015A (zh) * 2020-12-08 2021-03-02 新华三大数据技术有限公司 数据存储的方法、装置、电子设备及介质
WO2023125507A1 (zh) * 2021-12-29 2023-07-06 华为技术有限公司 生成块组的方法、装置和设备

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108415854A (zh) * 2018-02-11 2018-08-17 中国神华能源股份有限公司 基于共享缓存的数据采集系统及方法
CN108920725B (zh) * 2018-08-02 2020-08-04 网宿科技股份有限公司 一种对象存储的方法及对象存储网关
CN109302499B (zh) * 2018-11-30 2021-08-03 成都华栖云科技有限公司 一种可兼容访问nas存储和oss存储的存储网关
CN111857592A (zh) * 2020-07-21 2020-10-30 星辰天合(北京)数据科技有限公司 基于对象存储系统的数据存储方法及装置、电子设备
CN111930316B (zh) * 2020-09-09 2021-04-20 上海七牛信息技术有限公司 一种内容分发网络的缓存读写系统和方法
CN112839076B (zh) * 2020-12-29 2022-07-05 杭州海康威视系统技术有限公司 数据存储、读取方法、网关、电子设备及存储介质
CN113010526A (zh) * 2021-04-19 2021-06-22 星辰天合(北京)数据科技有限公司 基于对象存储服务的存储方法及装置
CN113296714B (zh) * 2021-06-17 2022-03-04 杭州海康威视系统技术有限公司 一种基于nas协议的数据存储系统
CN113806125B (zh) * 2021-09-07 2023-12-22 济南浪潮数据技术有限公司 一种卸载卷异常的处理方法、装置、设备及可读介质
CN114465770A (zh) * 2021-12-29 2022-05-10 天翼云科技有限公司 数据处理方法及相关装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102567495A (zh) * 2011-12-22 2012-07-11 国网信息通信有限公司 一种海量信息存储系统及实现方法
CN102821398A (zh) * 2011-06-09 2012-12-12 中国科学院声学研究所 一种面向用户多需求的传感器网络数据存取方法及系统
CN103019960A (zh) * 2012-12-03 2013-04-03 华为技术有限公司 分布式缓存方法及系统
CN103984607A (zh) * 2013-02-08 2014-08-13 华为技术有限公司 分布式存储的方法、装置和系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008059315A (ja) * 2006-08-31 2008-03-13 Hitachi Ltd 負荷分散方法及び計算機システム
WO2010025746A1 (de) * 2008-09-03 2010-03-11 Dspace Digital Signal Processing And Control Engineering Gmbh Kommunikationssystem und schnittstellengerät fur ein kommunikationssystem
CN101800762B (zh) * 2009-12-30 2014-03-19 中兴通讯股份有限公司 一种对多个业务进行融合的业务云系统及业务实现方法
CN102625307B (zh) * 2011-01-31 2014-07-09 电信科学技术研究院 一种无线网络接入系统
CN102799620B (zh) * 2012-06-19 2014-09-10 中国电力科学研究院 一种iec 61850通用数据库信息模型接口的设计方法
CN103218175B (zh) * 2013-04-01 2015-10-28 无锡成电科大科技发展有限公司 多租户的云存储平台访问控制系统
CN104281414B (zh) * 2013-07-10 2018-07-06 南京中兴软件有限责任公司 分布式文件系统及其小文件访问方法
CN103647712B (zh) * 2013-12-17 2018-02-16 南京联创科技集团股份有限公司 分布式路由处理业务的方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102821398A (zh) * 2011-06-09 2012-12-12 中国科学院声学研究所 一种面向用户多需求的传感器网络数据存取方法及系统
CN102567495A (zh) * 2011-12-22 2012-07-11 国网信息通信有限公司 一种海量信息存储系统及实现方法
CN103019960A (zh) * 2012-12-03 2013-04-03 华为技术有限公司 分布式缓存方法及系统
CN103984607A (zh) * 2013-02-08 2014-08-13 华为技术有限公司 分布式存储的方法、装置和系统

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109358813A (zh) * 2018-10-10 2019-02-19 郑州云海信息技术有限公司 一种分布式存储系统的扩容方法及装置
CN109358813B (zh) * 2018-10-10 2022-03-04 郑州云海信息技术有限公司 一种分布式存储系统的扩容方法及装置
CN111290700A (zh) * 2018-12-10 2020-06-16 北京京东尚科信息技术有限公司 分布式数据读写方法和系统
CN111796767A (zh) * 2020-06-24 2020-10-20 北京青云科技股份有限公司 一种分布式文件系统及数据管理方法
CN111796767B (zh) * 2020-06-24 2023-12-08 北京青云科技股份有限公司 一种分布式文件系统及数据管理方法
CN112434015A (zh) * 2020-12-08 2021-03-02 新华三大数据技术有限公司 数据存储的方法、装置、电子设备及介质
CN112434015B (zh) * 2020-12-08 2022-08-19 新华三大数据技术有限公司 数据存储的方法、装置、电子设备及介质
WO2023125507A1 (zh) * 2021-12-29 2023-07-06 华为技术有限公司 生成块组的方法、装置和设备

Also Published As

Publication number Publication date
CN106293490A (zh) 2017-01-04

Similar Documents

Publication Publication Date Title
WO2016180055A1 (zh) 数据存储、读取的方法、装置及系统
US11422982B2 (en) Scaling stateful clusters while maintaining access
US10223361B2 (en) Methods and systems for restoring a data container archived at an object-based storage
CN107169083B (zh) 公安卡口海量车辆数据存储与检索方法及装置、电子设备
US10013185B2 (en) Mapping systems and methods of an accelerated application-oriented middleware layer
Chandrasekar et al. A novel indexing scheme for efficient handling of small files in hadoop distributed file system
US10659225B2 (en) Encrypting existing live unencrypted data using age-based garbage collection
US9426219B1 (en) Efficient multi-part upload for a data warehouse
US11048591B1 (en) Efficient name space organization in a global name space cluster
US20130218934A1 (en) Method for directory entries split and merge in distributed file system
US20150248443A1 (en) Hierarchical host-based storage
WO2019001017A1 (zh) 集群间数据迁移方法、系统、服务器及计算机存储介质
WO2018119601A1 (zh) 数据转换方法及备份服务器
US10860604B1 (en) Scalable tracking for database udpates according to a secondary index
US11106625B2 (en) Enabling a Hadoop file system with POSIX compliance
US11442927B1 (en) Storage performance-based distribution of deduplicated data to nodes within a clustered storage environment
US9110820B1 (en) Hybrid data storage system in an HPC exascale environment
US10503693B1 (en) Method and system for parallel file operation in distributed data storage system with mixed types of storage media
Dwivedi et al. Analytical review on Hadoop Distributed file system
US11151162B2 (en) Timestamp consistency for synchronous replication
US10831719B2 (en) File consistency in shared storage using partial-edit files
US11625192B2 (en) Peer storage compute sharing using memory buffer
US11520744B1 (en) Utilizing data source identifiers to obtain deduplication efficiency within a clustered storage environment
US11288132B1 (en) Distributing multiple phases of deduplication processing amongst a set of nodes within a clustered storage environment
US20150227537A1 (en) Non-uniform file access in a distributed file system

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: 16791905

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: 16791905

Country of ref document: EP

Kind code of ref document: A1