CN106603665B - Cloud platform continuous data synchronous method and its device - Google Patents
Cloud platform continuous data synchronous method and its device Download PDFInfo
- Publication number
- CN106603665B CN106603665B CN201611167491.4A CN201611167491A CN106603665B CN 106603665 B CN106603665 B CN 106603665B CN 201611167491 A CN201611167491 A CN 201611167491A CN 106603665 B CN106603665 B CN 106603665B
- Authority
- CN
- China
- Prior art keywords
- data
- storage device
- strange land
- client
- cloud platform
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1095—Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of cloud platform continuous data synchronous method and its device, this method to include：The request that client is sent received from data generating device, and transmitted in parallel will be asked to local storage and strange land storage device；Client is not to wait for the data associated by the request and data persistence is completed in local storage and/or strange land storage device, and waits the local storage not responded or strange land storage device to respond, and task terminates.By the present invention, on the premise of ensureing that data are not lost, and local storage can be taken into account with strange land storage device to the handling capacity of data, avoid the decline of the calculated performance of cloud platform, improve user experience and reduce the computing cost on backstage.
Description
Technical field
The present invention relates to field of cloud computer technology, more particularly to a kind of cloud platform continuous data synchronous method and its device.
Background technology
In field of cloud calculation, snapping technique is generally used when performing data protection to magnetic disk of virtual machine.User can select
Aperiodically do periodically or to system disk and data disks snapshot (snapshot).Snapping technique accident can be occurred after number
It is reduced to according to loss in a certain range, such as, it is allowed to the snapshot number that user creates.But snapping technique can only be extensive by data
The multiple state to a nearest snapshot point, the data that this snapshot point produces afterwards just can not be recovered.Although snapping technique can
To return to some past snapshot point, but when local data center breaks down, it is stored in the fast of local data center
According to can not also use, so as to cause magnetic disk of virtual machine data can not ensure uniformity.
In addition, in field of cloud calculation magnetic disk of virtual machine data are protected with the technical side that can also use remote backup
Case.The implementation of remote backup has two kinds, and one kind is strange land snapshot, i.e., periodically or aperiodically by local data
The snapshot of the heart, when local data center breaks down, can be incited somebody to action to another data center in strange land data center
Data recovery is to a nearest snapshot point；Another is by way of mirror image, and phase is stored in another data center
Same data.The former the shortcomings that is that data can only be recovered to a nearest strange land snapshot point；And the shortcomings that the latter, is once source
Data center is written with the data of mistake, and strange land data center can also store the data of mistake.
As enterprise is incremented by data processing degree of dependence, the loss of low volume data can cause substantial amounts of economic loss.
Traditional data backup and resume is realized using periodic backup and snapshot always.Traditional backup mode and snapshot can only
The data recovery of set time point is accomplished to, the data update between backing up twice can not be protected.With the increase of data volume, pass
The backup mode of system and the RPO (Recovery Point Objective, data recovery point target) and RTO of snapshot
(Recovery Time Objective, recovery time target) just seems long, this is very high to data dependency for some
Information system is unacceptable.This kind of system usually require that one day 24 it is small when run incessantly, the interruption service of short time
Or a small amount of loss of data may all bring huge traffic lost.
In the practical application based on cloud storage technology, Clustering is generally used, by two or multiple stage computers section
Point forms a kind of calculate node set of loose couplings, has provided more efficient fault-tolerant and failover capability to the user, increases
To the processing capability in real time of magnanimity task, so as to provide more preferable dependable with function.
Realized for continuous data protection generally use asynchronous system in cloud platform or the method for synchronization.
In asynchronous system, request of data first writes data into local cluster and interior after client sends over
Among depositing, one response of client is then returned to, represents that the request of data has been completed, client can send follow-up reading
Write request, when the data of region of memory reach certain threshold value, such as internal storage data reaches size set in advance or reaches
The threshold value of certain time, system can start the asynchronous data cluster that internal storage data is copied to distal end of an extra thread and work as
In, ensure the uniformity of this part local cluster and distant clusters data, the program also has some simply to change, but essential
Or asynchronous replication of data achievees the purpose that data protection to distant clusters.There is local collect in the continuous data protection of asynchronous scheme
Group and distant clusters data out of step conditions, local cluster because certain reason need by distant clusters recover data when
Wait, it is understood that there may be newest data cannot be returned to, had been written in data in local cluster and memory, but data do not have also
In asynchronous replication distant clusters, if this when, client place physical node was because certain reason causes internal storage data to be lost
Lose, this when, data may just be recovered less than newest version.
There are local cluster and distant clusters, there are the nonsynchronous problem of data, synchronization side for asynchronous scheme for synchronization scenario
Case takes request of data to issue local cluster and distant clusters at the same time, and only local cluster and distant clusters ensures request of data
Data have been persisted to local cluster and distant clusters, and client only receives the response just meeting of local cluster and distant clusters
Think that this request of data has been completed, client can send subsequent request.
The content of the invention
It is an object of the invention to disclose a kind of cloud platform continuous data synchronous method and its a kind of cloud based on this method
Platform continuous data sychronisation, overcomes asynchronous continuous data protection mode between local cluster and distant clusters due to difference
The defects of walking caused loss of data；Meanwhile overcome needed in synchronous continuous data protection mode wait pending data in local collection
The problem of group and distant clusters persistence, avoid the decline of the calculated performance of cloud platform in the method for synchronization.
For achieving the above object, the present invention provides a kind of cloud platform continuous data synchronous method, client to receive
The request sent from data generating device, by request transmitted in parallel to local storage and strange land storage device；Client is not
Wait the data associated by the request to complete data persistence in local storage and/or strange land storage device, and wait
The local storage or strange land storage device not responded respond, and task terminates.
As a further improvement on the present invention, client is not to wait for data associated by the request in local storage
And/or data persistence is completed in the storage device of strange land, and wait the local storage not responded or strange land storage
Device responds, and allows to differ at least one described from client reception between local storage and strange land storage device
Request corresponding to the delay operation cycle.
As a further improvement on the present invention, the local storage and/or strange land storage device are deposited including volatibility
Reservoir, nonvolatile memory, database, calculate node, cluster server or software definition storage；
The client is configured as Qemu or KVM；
The data generating device is configured as virtual machine, application program, the computer installation for running UI.
As a further improvement on the present invention, the Qemu or KVM are driven by Librbd and send out data generating device
The request transmitted in parallel sent is to local storage and strange land storage device.
As a further improvement on the present invention, the request includes read operation, write operation, delete operation, Data Migration behaviour
Make or replicate operation.
As a further improvement on the present invention, further include：Performance is carried out to local storage or strange land storage device
Judge, to determine the delay operation cycle.
Meanwhile the invention also discloses a kind of cloud platform continuous data sychronisation, including：
Client, local storage, strange land storage device；
The request that client is sent received from data generating device, by request transmitted in parallel to local storage and strange land
Storage device；Client is not to wait for the data associated by the request and is completed in local storage and/or strange land storage device
Data persistence, and wait the local storage not responded or strange land storage device to respond, task terminates.
As a further improvement on the present invention, the cache module communicated with client is further included, the cache module is used
In the data corresponding to the request that interim storage local storage and strange land storage device send data generating device；And
Performance judgment module, the performance judgment module are used to carry out performance judgment to local storage or strange land storage device,
The data difference tolerated with definite local storage with strange land storage device a delay operation cycle.
As a further improvement on the present invention, further include homogeneous with client, local storage and strange land storage device
The update module and acquisition module of mutual communication, time threshold T of the acquisition module according to set by setting update module, to this
The performance data of body storage device and strange land storage device is acquired, and is sent to update module；The update module includes
A kind of logic, the logic can determine the delay operation cycle according to the performance data that acquisition module is sent.
As a further improvement on the present invention, the local storage and/or strange land storage device are deposited including volatibility
Reservoir, nonvolatile memory, database, calculate node, cluster server or software definition storage；
The client is configured as Qemu or KVM；
The data generating device is configured as virtual machine, application program, the computer installation for running UI.
As a further improvement on the present invention, the nonvolatile memory is filled by the solid-state storage of support random access
Put, flash memory, phase transition storage, strong dielectric body memory, magnetic recording formula write and read memory, two-way uniformly deposit at any time
One of reservoir, resistance RAM, NVDIMM or RAMDISK are formed.
As a further improvement on the present invention, the time threshold T set by the update module is less than 10 seconds.
Compared with prior art, the beneficial effects of the invention are as follows：By the present invention, before ensureing that data are not lost
Put, and local storage can be taken into account with strange land storage device to the handling capacity of data, avoid the calculating of cloud platform
The decline of performance, improves user experience, reduces the computing cost on backstage.
Brief description of the drawings
Fig. 1 is based on a kind of cloud platform continuous data synchronous method of the present invention and its cloud platform continuous data sychronisation institute
The topological diagram being related to；
Fig. 2 is that local cluster promises the sequence chart prior to distant clusters response in embodiment one；
Fig. 3 is that distant clusters promise the sequence chart prior to local cluster response in embodiment two；
Fig. 4 is a kind of topological diagram of the cloud platform continuous data sychronisation in variation；
Fig. 5 is topological diagram of the cloud platform continuous data sychronisation in another variation；
Fig. 6 is the topological diagram of local storage.
Embodiment
The present invention is described in detail for shown each embodiment below in conjunction with the accompanying drawings, but it should explanation, these
Embodiment is not limitation of the present invention, those of ordinary skill in the art according to these embodiment institute work energy, method,
Or equivalent transformation or replacement in structure, belong within protection scope of the present invention.
KVM (Kernel-based Virtual Machine) is built in linux kernel, is a kind of hardware based complete
Virtualization system.KVM basic structures are made of KVM Driver and Qemu.Wherein, KVM Driver have become in Linux
One module of core, and be responsible for creating virtual machine, the distribution of virtual memory, the read/write of virtual cpu register and operation etc.；
The user's space that Qemu is used to simulate PC hardware is set up, there is provided the device model such as hard disk, network and access approach.
Qemu is the application program operated in HOST OS (host operating system), and with the shape of host file
Formula storage virtual machine hard disc data, and provide raw and qcow2 two kinds of common file formats.Qcow2 provides cow and (is answered when writing
System) technology, data block is only allocated in renewal, and provides snapshot (snapshot functions)
Overview
Fig. 1 is involved by a kind of cloud platform continuous data synchronous method of the present invention and its cloud platform continuous data sychronisation
Topological diagram.In various application scenarios shown by the present invention, client 20 receives the request of the transmission of data generating device 10,
The request can be read operation request, or the request of write operation requests, delete operation, data migration operation are asked or duplication
Operation requests, and other based on cloud platform, computer cluster, cluster server or unlatching and run on the void in physical machine
Plan machine be created, run or closing process in involved all data actions.
Client 20 is by above-mentioned request transmitted in parallel to local storage 30 and strange land storage device 40, client 20
It is not to wait for the data associated by the request and completes data persistence in local storage 30 and/or strange land storage device 40,
And waiting the local storage 30 not responded or strange land storage device 40 to respond, task terminates；Then start
Perform next operation.Client 20 is run in HOST OS.
It by the data model translation in memory is storage model that data persistence, which is exactly, and in storage model is converted to
The general designation data models of data model in depositing can be any data structure or object model, and storage model can be relation
Model, XML, binary stream etc..
Local storage 30 and/or strange land storage device 40 include volatile memory (for example, DRAM), non-volatile
Memory (for example, support the solid-state storage device of random access, flash memory, phase transition storage, strong dielectric body memory,
Magnetic recording formula write at any time read memory, two-way Unified Memory, resistance RAM, NVDIMM or RAMDISK), database,
Calculate node, cluster server or software definition storage.Client 20 be configured as Qemu either KVM or other be based on void
Planization technology configures the client driving 20a of formation.Data generating device 10 is configured as virtual machine, application program, operation
The computer installation of UI.In the present specification, data generating device 10 selects application program 101 to elaborate.Application program
101 can independent operating in the PC of user (guest), also can independent operating in the void positioned at strange land based on user's Telnet
In plan machine, and with client 20a can be driven to carry out data transmit-receive by api interface or system bus.In the present specification, should
Client driving 20a is configured as Qemu or KVM, and drives the request for sending data generating device 10 simultaneously by Librbd
Row is sent to local storage 30 and strange land storage device 40.In the present specification, local cluster 30a is local storage
A kind of 30 concrete application example, distant clusters 40a are a kind of instantiations of strange land storage device 40.
Embodiment one：
Present embodiment illustrates local cluster 30a response prior to distant clusters 40a reach client driving 20a when reality
Example scene.
Coordinate with reference to shown in Fig. 2, in the present embodiment, which comprises the following steps：Should
Write request is sent by the use of program 101 as data generating device to client driving 20a；Client driving 20a is parallel by write request
(i.e. in the same time) is respectively sent to local cluster 30a and distant clusters 40a.Due to the meter of local cluster 30a and distant clusters 40a
The difference of calculation machine performance and network environment, so as to cause both disposal ability to the data corresponding to same write request to exist
Certain difference.Therefore, local cluster 30a and distant clusters 40a can cause write request institute right after above-mentioned write request is received
There are the regular hour is poor when writing local cluster 30a and distant clusters 40a for the data answered.
In this implement scene, response of the response prior to distant clusters 40a of local cluster 30a reaches client driving
20a.At this time, the response of the slower distant clusters 40a of client driving 20a meeting wait-for-responses, until it makes to client 20a
Response.Finally, 20a is driven to 101 response of application program from client, and completes this write operation requests.
In the present embodiment, client driving 20a is not to wait for data associated by the request in local cluster 30a and distal end
Data persistence is completed in cluster 40a, and waits the distant clusters 40a not responded to respond, and allows local cluster
The delay operation corresponding at least one request received from client driving 20a is differed between 30a and distant clusters 40a
Cycle.The delay operation cycle depends on every parameter of local cluster 30a and distant clusters 40a, such as CPU floating-points fortune
Calculate speed, network environment etc..
Specifically, when data generating device 10 is configured as VM virtual machines, client 20 is configured as Qemu or KVM,
Local cluster 30a is configured as local Ceph, when distant clusters 40a is configured as distal end Ceph, the execution of above-mentioned write request
Journey can be described in detail below.
Application program in VM virtual machines sends write request to QEMU/KVM virtualization layers.QEMU/KVM passes through Librbd's
Driving sends write request to the Ceph clusters being located locally.QEMU/KVM sends parallel hair also by the driving of Librbd at the same time
Write request is sent to the Ceph clusters positioned at distal end.Local Ceph clusters send out the write request response after the write request has been handled
Give QEMU/KVM.QEMU/KVM judges response of the response for local Ceph clusters, sends responses to QEMU/KVM's
The driving of librbd.The driving of QEMU/KVM is sent to VM virtual machines, represents this write operation requests writing in Ceph clusters
The corresponding data write-in of operation is completed.Then, the response of distal end Ceph clusters is waited to reach QEMU/KVM.Finally, this is completed
Data corresponding to secondary write operation requests are in local Ceph clusters (a kind of instantiation of local storage 30) and distal end
Write while Ceph clusters (a kind of instantiation of strange land storage device 40), so as to ensure that local Ceph clusters and distal end
Data in Ceph clusters keep synchronous.
Embodiment two：
Present embodiment illustrates distant clusters 40a response prior to local cluster 30a reach client driving 20a when reality
Example scene.The present embodiment and the application scenarios of embodiment one are as follows in difference.
Coordinate with reference to shown in Fig. 3, in the present embodiment, which comprises the following steps：Should
Write request is sent by the use of program 101 as data generating device to client driving 20a；Client driving 20a is parallel by write request
(i.e. in the same time) is respectively sent to local cluster 30a and distant clusters 40a.Due to the meter of local cluster 30a and distant clusters 40a
The difference of calculation machine performance and network environment, so as to cause both disposal ability to the data corresponding to same write request to exist
Certain difference.Therefore, local cluster 30a and distant clusters 40a can cause write request institute right after above-mentioned write request is received
There are the regular hour is poor when writing local cluster 30a and distant clusters 40a for the data answered.
In this implement scene, response of the response prior to local cluster 30a of distant clusters 40a reaches client driving
20a.At this time, the response of the slower local cluster 30a of client driving 20a meeting wait-for-responses, until it makes to client 20a
Response.Finally, 20a is driven to 101 response of application program from client, and completes this write operation requests.
Embodiment three：
In order to optimize local storage 30 (that is, embodiment one and local cluster 30a in embodiment two), strange land is deposited
In storage device 40 (that is, strange land cluster 40a in embodiment one and embodiment two) to client 20 (that is, embodiment one with it is real
Apply the client driving 20a in example two) accuracy and data reliability that respond.Join shown in Fig. 4 and Fig. 5, in this implementation
In example, further included in the cloud platform continuous data sychronisation：The cache module 21 communicated with client 20.Cache module 21
Corresponding to the request sent for interim storage local storage 30 and strange land storage device 40 to data generating device 10
Data；And performance judgment module 22.Performance judgment module 22 by api interface and client 20, local storage 30 and
Strange land storage device 40 communicates.Performance judgment module 22 is used to carry out local storage 30 or strange land storage device 40
Performance judgment, the data difference tolerated with definite local storage 30 with strange land storage device 40 a delay operation cycle
It is different.
After some in local storage 30 and strange land storage device 40 comparatively fast responds, it can be connect by client 20
Next request is received, and next ask is issued to the very fast local storage 30 for completing response or strange land storage device 40
Ask, and start to perform the request；When the local storage 30 compared with the previous request of slow-response or strange land storage device 40 are complete
Into after response, the response of next request can be got started.
Since local storage 30 and the performance of strange land storage device 40 can convert at any time with network environment,
Both sequencing is simultaneously not fixed, so as to the calculated performance according to local storage 30 and strange land storage device 40 and/
Or network environment and flexibly switch with conversion sequencing so that whole technical solution can either reach current asynchronous data
The efficient advantage of backup, and it is avoided that the computing resource of undue consumption backstage or cloud platform, therefore this programme can be shown
The raising user experience of work, avoids the request that user sends from producing hysteresis and the delay of response.
Join shown in Fig. 5, as a further improvement on the present invention, which further includes and client
End 20, the update module 24 and acquisition module 23 of 40 homogeneous mutual communication of local storage 30 and strange land storage device.The collection
Time threshold T of the module 23 according to set by setting update module 24, to body storage device 30 and strange land storage device 40
Performance data is acquired, and is sent to update module 24.The update module 24 includes a kind of logic, and the logic being capable of basis
The performance data that acquisition module 23 is sent determines the delay operation cycle.By this Technical Architecture, can determine to delay with easier
The capacity set required for storing module 21, so that the capacity setting to cache module 21 provides accurate foundation.Dress is locally stored in this
Put 30 and/or strange land storage device 40 include volatile memory, nonvolatile memory, database, calculate node, cluster clothes
Business device or software definition storage.
The problem of since traditional DRAM can not keep self-refresh there are power down and losing data.In the present embodiment, should
Local storage 30 and strange land storage device 40 are more preferably nonvolatile memory, and more preferably
NVDIMM.NVDIMM is a kind of memory bar specification for being integrated with DRAM and Nonvolatile memory chip.
Join shown in Fig. 6, the present embodiment further discloses a kind of NVDIMM.The local storage 30 and strange land storage device 40
In include at least a memory element 100.
The memory element 100, including：
The memory 20b and Memory Controller 10a that matching is set some side by side, specifically, memory 20b and storage
The quantity of device controller 10a is equal.The memory element 100 further includes control circuit 30.Control circuit 30 includes：It is arranged on described
Cut between memory 20b and Memory Controller 10a in the first selector 301 of cascade structure, second selector 302, bypass
Change the mold 303 and one moderator 304 of block.
Memory Controller 10a receives data access request 101a, 101b of each memory 20b and sends to moderator
304.Moderator 304 receives each data access request 101a, 101b and judges the order of priority.First selector 301
According to the order of priority, multiple interface channel 102a, 102b are established with Memory Controller 10a successively, and pass through the second selection
Device 302 carries out write-in/reading of data with the normal data passage 312 that multiple memories 20 being set up in parallel are established.
In the present embodiment, priority ranking uses following ordering strategy：
First priority：The management data of memory 20；
Second priority：System state data；
Third priority：Need the general data for writing memory 20b or being read from memory 20b.
Specifically, in the present embodiment, moderator 304 includes data according to data access request 101a, 101b
Importance, judges the order of priority.That is in Fig. 6, which the Memory Controller 10a meeting of data access request 101a, 101b
Data connection passage 102a, 102b are preferentially established with first selector 301.
When Memory Controller 10a monitors the abnormal conditions of data write-in/reading in memory 20b, pass through bypass
Handover module 303 sends interruptive command 3041 to moderator 304, to interrupt first selector 301 and second selector 302
Connection, then connection is established between bypass changeover module 303 and second selector 302, i.e. interface channel 313 in Fig. 6.
Specifically, the Memory Controller 10a is connected by interface bus 305 with system bus 40.The interface is total
Line 305 includes I2C or SMBUS, and preferably SMBUS.SMBUS (System Management Bus) is a kind of two-wire system
Universal serial bus.It has abandoned controlling bus and the form of data/address bus in traditional central processing unit, therefore passes through interface bus
305 can transmit data-signal, clock signal, address signal.
Memory 20b is nonvolatile memory or volatile memory, and preferably nonvolatile memory.It is described
Nonvolatile memory is by flash memory (Flash memory), phase transition storage (FCM), strong dielectric body memory
(FeRAM), magnetic recording formula write at any time read memory (MRAM), two-way Unified Memory (OUM), resistance RAM (RRAM) it
One is formed.Flash memory (Flash memory) is made of one of nand flash memory, NOR flash memory；The volatile memory by
One of static RAM (SRAM), dynamic random access memory (DRAM) are formed, and preferably dynamic random is deposited
Access to memory (DRAM).Memory 20b is nand flash memory, and Memory Controller 10a is NAND controller.Since NAND dodges
It is the highly developed prior art to deposit with NAND controller, is not repeated herein.Due to the inherent characteristic of nand flash memory, each
Nand flash memory all has write-in/read cycle of its setting, therefore can be set up in parallel multiple nand flash memories, and with multiple NAND
Controller matches setting side by side.
Those listed above is a series of to be described in detail only for feasibility embodiment of the invention specifically
Bright, they simultaneously are not used to limit the scope of the invention, all equivalent implementations made without departing from skill spirit of the present invention
Or change should all be included in the protection scope of the present invention.
It is obvious to a person skilled in the art that the invention is not restricted to the details of above-mentioned one exemplary embodiment, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Profit requires rather than described above limits, it is intended that all in the implication and scope of the equivalency of claim by falling
Change is included in the present invention.Any reference numeral in claim should not be considered as to the involved claim of limitation.
Moreover, it will be appreciated that although the present specification is described in terms of embodiments, not each embodiment is only wrapped
Containing an independent technical solution, this narrating mode of specification is only that those skilled in the art should for clarity
Using specification as an entirety, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
It is appreciated that other embodiment.
Claims (10)
- A kind of 1. cloud platform continuous data synchronous method, it is characterised in that：Client is asked received from what data generating device was sent Ask, and transmitted in parallel will be asked to local storage and strange land storage device；Client is not to wait for the number associated by the request According to completing data persistence in local storage and/or strange land storage device, and wait what is do not responded to be locally stored Device or strange land storage device respond, and allow to differ at least one between local storage and strange land storage device Delay operation cycle corresponding to the request received from client, task terminate；WhereinThe delay operation cycle is determined by update module according to the performance data that acquisition module is sent.
- 2. cloud platform continuous data synchronous method according to claim 1, it is characterised in thatThe local storage and/or strange land storage device include volatile memory, nonvolatile memory, database, Calculate node, cluster server or software definition storage；The client is configured as Qemu or KVM；The data generating device is configured as virtual machine, application program, the computer installation for running UI.
- 3. cloud platform continuous data synchronous method according to claim 2, it is characterised in that the Qemu or KVM lead to Librbd drivings are crossed by the request transmitted in parallel that data generating device is sent to local storage and strange land storage device.
- 4. cloud platform continuous data synchronous method as claimed in any of claims 1 to 3, it is characterised in that described Request includes read operation, write operation, delete operation, data migration operation or replicates operation.
- 5. cloud platform continuous data synchronous method according to claim 4, it is characterised in that further include：To being locally stored Device or strange land storage device carry out performance judgment, to determine the delay operation cycle.
- A kind of 6. cloud platform continuous data sychronisation, it is characterised in that including：Client, local storage, strange land storage device；The request that client is sent received from data generating device, and request transmitted in parallel to local storage and strange land is deposited Storage device；Client is not to wait for the data associated by the request and completes number in local storage and/or strange land storage device According to persistence, and the local storage not responded or strange land storage device is waited to respond, task terminates；AndThe cache module communicated with client, the cache module are used for interim storage local storage and are filled with strange land storage Put the data corresponding to the request sent to data generating device；And performance judgment module, the performance judgment module are used for Performance judgment is carried out to local storage or strange land storage device, to determine that local storage and strange land storage device exist The data difference that one delay operation cycle is tolerated；WhereinThe delay operation cycle is determined by update module according to the performance data that acquisition module is sent.
- 7. cloud platform continuous data sychronisation according to claim 6, it is characterised in that further include with client, originally The update module and acquisition module of ground storage device and the homogeneous mutual communication of strange land storage device, the acquisition module according to setting more Time threshold T set by new module, is acquired the performance data of body storage device and strange land storage device, and send To update module；The update module includes a kind of logic, and the performance data that the logic can be sent according to acquisition module is true Surely the operation cycle is postponed.
- 8. cloud platform continuous data sychronisation according to claim 7, it is characterised in thatThe local storage and/or strange land storage device include volatile memory, nonvolatile memory, database, Calculate node, cluster server or software definition storage；The client is configured as Qemu or KVM；The data generating device is configured as virtual machine, application program, the computer installation for running UI.
- 9. cloud platform continuous data sychronisation according to claim 8, it is characterised in that the nonvolatile memory By the solid-state storage device of support random access, flash memory, phase transition storage, strong dielectric body memory, magnetic recording formula with When write-in reads memory, two-way Unified Memory, one of resistance RAM, NVDIMM or RAMDISK are formed.
- 10. cloud platform continuous data sychronisation according to claim 7, it is characterised in that set by the update module Fixed time threshold T is less than 10 seconds.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611167491.4A CN106603665B (en) | 2016-12-16 | 2016-12-16 | Cloud platform continuous data synchronous method and its device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611167491.4A CN106603665B (en) | 2016-12-16 | 2016-12-16 | Cloud platform continuous data synchronous method and its device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106603665A CN106603665A (en) | 2017-04-26 |
| CN106603665B true CN106603665B (en) | 2018-04-13 |
Family
ID=58599456
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611167491.4A Active CN106603665B (en) | 2016-12-16 | 2016-12-16 | Cloud platform continuous data synchronous method and its device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106603665B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107547160B (en) * | 2017-06-28 | 2019-06-04 | 新华三技术有限公司 | A kind of method for synchronizing time and device |
| CN109710699B (en) * | 2019-01-02 | 2021-01-22 | 神州数码融信软件有限公司 | Recording method and playback method of transaction data |
| CN109819048B (en) * | 2019-02-27 | 2022-03-15 | 北京字节跳动网络技术有限公司 | Data synchronization method, device, terminal and storage medium |
| CN110158430B (en) * | 2019-05-08 | 2021-03-02 | 中铁北京工程局集团有限公司 | Automatic napping laminating machine for bridge concrete surface |
| US11126364B2 (en) * | 2019-07-18 | 2021-09-21 | Pure Storage, Inc. | Virtual storage system architecture |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102412985A (en) * | 2011-10-31 | 2012-04-11 | 烽火通信科技股份有限公司 | Method for managing object of network management system based on operation separation |
| CN103078945A (en) * | 2013-01-07 | 2013-05-01 | 北京奇虎科技有限公司 | Method and system for processing browser crash data |
| CN104239476A (en) * | 2014-09-04 | 2014-12-24 | 上海天脉聚源文化传媒有限公司 | Method, device and system for synchronizing databases |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751415B (en) * | 2008-12-09 | 2012-03-28 | 中国移动通信集团公司 | Metadata service system, metadata synchronized method and writing server updating method |
| CN105045531B (en) * | 2015-07-01 | 2018-01-02 | 山东超越数控电子有限公司 | Cache synchronization mechanism between one kind storage dual controller |
-
2016
- 2016-12-16 CN CN201611167491.4A patent/CN106603665B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102412985A (en) * | 2011-10-31 | 2012-04-11 | 烽火通信科技股份有限公司 | Method for managing object of network management system based on operation separation |
| CN103078945A (en) * | 2013-01-07 | 2013-05-01 | 北京奇虎科技有限公司 | Method and system for processing browser crash data |
| CN104239476A (en) * | 2014-09-04 | 2014-12-24 | 上海天脉聚源文化传媒有限公司 | Method, device and system for synchronizing databases |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106603665A (en) | 2017-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106603665B (en) | Cloud platform continuous data synchronous method and its device | |
| CN102203741B (en) | Performing a data write on a storage device | |
| CN109857445A (en) | Storage system and control software layout method | |
| US8010837B2 (en) | Storage sub system and data restoration method thereof | |
| CN101609419B (en) | Continuous on-line transferring data backup method of virtual machine and device thereof | |
| CN103827843B (en) | A kind of data writing method, device and system | |
| CN102467408B (en) | Method and device for accessing data of virtual machine | |
| CN106030500B (en) | The data migration method of storage system | |
| US20220121533A1 (en) | Creating remote backup copies in asynchronous replication arrangements involving relatively long cycle times | |
| CN102123176B (en) | Space distribution and management method and device for network storage system | |
| CN108235751A (en) | Identify the method, apparatus and data-storage system of object storage device inferior health | |
| US20060047926A1 (en) | Managing multiple snapshot copies of data | |
| CN107798130A (en) | A kind of Snapshot Method of distributed storage | |
| CN106126371A (en) | Data return a grade method, Apparatus and system | |
| CN102158540A (en) | System and method for realizing distributed database | |
| CN101763224A (en) | Method of managing data snapshot images in a storage system | |
| CN102073739A (en) | Method for reading and writing data in distributed file system with snapshot function | |
| CN100530069C (en) | Virtualizing system and method for non-homogeny storage device | |
| CN105074675A (en) | Computer system including hierarchical block storage device, storage controller, and program | |
| CN106227621A (en) | The data back up method of logic-based volume management simplification volume and system | |
| CN104272242A (en) | Creating encrypted storage volumes | |
| CN101945131A (en) | Storage virtualization-based data migration method | |
| CN104040525A (en) | Accessing overlay media over a network connection | |
| CN106104515A (en) | Utilize File system design and the fault recovery method of nonvolatile memory | |
| WO2015116078A1 (en) | Memory data versioning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 214125 Wuxi science and Technology Park, Jiangsu Binhu District No. 6 Patentee after: Huayun data holding group Co., Ltd Address before: No.6, science and education software park, Binhu District, Wuxi City, Jiangsu Province Patentee before: WUXI CHINAC DATA TECHNICAL SERVICE Co.,Ltd. |