CN104050015B - A kind of system of virtual machine image storage distribution - Google Patents
A kind of system of virtual machine image storage distribution Download PDFInfo
- Publication number
- CN104050015B CN104050015B CN201410301202.XA CN201410301202A CN104050015B CN 104050015 B CN104050015 B CN 104050015B CN 201410301202 A CN201410301202 A CN 201410301202A CN 104050015 B CN104050015 B CN 104050015B
- Authority
- CN
- China
- Prior art keywords
- virtual machine
- management server
- storage
- dht
- cluster management
- 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.)
- Expired - Fee Related
Links
- 238000013507 mapping Methods 0.000 claims abstract description 36
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 230000005540 biological transmission Effects 0.000 claims description 11
- 230000001360 synchronised effect Effects 0.000 claims description 5
- 239000000203 mixture Substances 0.000 claims description 4
- 230000000737 periodic effect Effects 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 2
- 238000006116 polymerization reaction Methods 0.000 claims description 2
- 238000000034 method Methods 0.000 abstract description 26
- 230000007812 deficiency Effects 0.000 abstract description 2
- 238000007726 management method Methods 0.000 description 40
- 230000007246 mechanism Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- BYACHAOCSIPLCM-UHFFFAOYSA-N 2-[2-[bis(2-hydroxyethyl)amino]ethyl-(2-hydroxyethyl)amino]ethanol Chemical group OCCN(CCO)CCN(CCO)CCO BYACHAOCSIPLCM-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of virtual machine image storage distribution system, and the system includes cluster management server and native object storage server, DHT MAP mapping servers and the driving of QEMU blocks on the test machine for operating in distributed cloud computing platform.The system can realize the virtual machine batch of distributed cross-region, rapid deployment, solve legacy system and the method deployment virtual machine cycle it is long, can not the low deficiency of dynamic expansion, availability.
Description
Technical field
The present invention relates to a kind of system of field of cloud computer technology, and in particular to a kind of virtual machine image storage distribution is
System.
Background technology
The fast development of cloud computing, the continuous expansion of cloud infrastructure scale, cloud platform bearer service species is increasingly various,
Higher requirement is proposed to the batch under distributed environment, rapid deployment virtual machine.On the one hand, a large amount of enterprises are constantly by tradition
Business migrates to cloud platform, and existing infrastructure dilatation influenceed by various factors such as energy consumption costs, private clound deployment by
Geographic distributed nature is gradually presented；On the other hand, increasing enterprise and user's selection will apply is placed into data
In public cloud and mixed cloud, cloud provider also provides public cloud or mixing cloud service energetically so that distributed cloud computing gradually into
For the cloud computing deployment pattern of main flow.Batch deployment of the distributed cloud computing development trend to virtual machine is proposed than collection middle part
More technological challenges are affixed one's name to, it is necessary under the premise of cross-region, distributed deployment is met, there is provided compared to concentrate deployment mode equity or
Preferably storage and distribution mechanisms.
When facing the service request of a large number of users, under traditional deploying virtual machine strategy, certain wait is generally required
Time, this greatly reduces expectation of the user to " cloud ".For the problem, it has been proposed that much solve method, but most of sides
The improvement for the links that method is concentrated in deployment virtual machine process, the especially optimization to mirror image template File conveyer.
Virtual machine image template includes the software of the operating system and user's needs on virtual machine, a complete mirror image text
Part may have more than ten G, or even tens G sizes, and the transmission of virtual machine image template file occupies the big of whole deploying virtual machine
The time is measured, especially under distributed deployment pattern, transmits the time used in mirror image template file often due to wide area network band
The factors such as width, the time of deploying virtual machine is caused to be difficult to receive.
In the prior art, different compressing file mechanism is proposed for deploying virtual machine transmission optimization, to the full extent
The volume of mirror image template file is optimized, reduces transmission volume.Some documents propose application template pre-copy strategy
To reduce template file transmission time.This method is to be stored in advance in one by template interested in user, that deployment frequency is high
So that copy next time uses in individual pre-copy ATL.Here, pre-copy ATL uses the storage different from conventional template storehouse
Method.Give in the prior art and a kind of the rapid deployment side of virtual machine is realized by virtual machine image incremental backup mechanism
Method；Or the thought with reference to fork functions, substantial amounts of sub- virtual machine is cloned rapidly using father's virtual machine, it is fictitious by creating parallel
The method that the mode of plan machine improves deploying virtual machine efficiency；Again or create virtual disk and internal memory shape using Copy on write technology
The snapshot of state, and further ask and transmit these status informations using distribution according to need memory techniques and multicast technology
Virtual machine rapid clon method.
It also proposed and divided by changing in conventional method from a centrally stored center to multiple hosts in the prior art
The mode of a super large file is sent out, introduces virtual machine image template file distributed storage in each template file storage center
Mechanism so that host can download virtual machine image template file from different storage centers, avoid centrally stored center
The problem of turning into transmission bottleneck when a large amount of requests reach.Another similar BT storage mode is disclosed, by building
Coordinator servers are exchanged each other's needs to the data block of virtual machine image template file between more hosts, with P2P
Load mode realizes that mirror image template file accelerates transmission, while by distinguishing common template file data and user data, using pre-
Copy and incremental transmission mechanism realize the virtual machine rapid deployment concentrated under deployment mode.It is limited in that existing for this method,
In order to keep the uniformity of data, it is necessary to by the way of similar flood, this is infeasible in the wide area network.It is also a kind of poly-
The memory space of each host is closed, forms a general pool, mirror image template file is respectively stored in storage pool, then adopted
The method that deployment efficiency is improved with parallel transmission mode.It is disadvantageous in that, because every kind of template lacks redundant storage, when
Batch deployment same type virtual machine when, frequently data transfer will greatly increase the load of the memory node, especially across
It is more prominent when geographic region is transmitted.
It is, therefore, desirable to provide the virtual machine image storage distribution system and method under a kind of distributed cloud computing environment,
To realize the virtual machine batch of distributed cross-region, rapid deployment, solve legacy system and method deployment virtual machine cycle length, nothing
The deficiencies such as method dynamic expansion, availability be low.
The content of the invention
The defects of in order to overcome above-mentioned prior art, the invention provides a kind of virtual machine image storage distribution system and side
Method.
In order to realize foregoing invention purpose, the present invention adopts the following technical scheme that：
A kind of virtual machine image storage distribution system, it is theed improvement is that：The system includes cluster management server
And operate on the test machine of distributed cloud computing platform native object storage server, DHT-MAP mapping servers and
QEMU blocks drive.
Further, the DHT-MAP mapping servers carry out registration, the route of the data block to obtain data block
With the module of inquiry；
The native object storage server communicates with the DHT-MAP mapping servers, the object storage server
For data, each host memory space of polymerization LAN, composition template volume and the storage pool of snapped volume to be locally stored；
The QEMU blocks driving is communicated with the native object storage server, and the QEMU blocks, which drive, to be used for virtual machine
Mirror image is divided into the data block object of fixed size；
The cluster management server is used to realize cluster management and message transmission, and the mirror image data block between cluster
Obtain.
Further, the cluster management server is according to the super ring topology of chord Algorithm constitutions, the object storage
Server is according to chord Algorithm constitutions local ring topology.
Further, the storage work of the virtual machine image storage distribution system comprises the following steps：
I, virtual machine image is divided into the data block object of fixed size by the QEMU blocks driving, with uniformity Hash
Algorithm generates object identity, and the data block object and its redundant copy storage are arrived into the virtual machine mirror by iSCSI protocol
As storage distribution system；If local have object management server, local cache data block copy；
II, the DHT-MAP mapping servers receive mirrored storage I/O requests, differentiate purpose according to the object identity
Memory node, carry out route locating and be forwarded to the cluster management server, the cluster management server turns I/O requests
Purpose memory node is dealt into, template volume and snapped volume are carried out classification storage by the purpose memory node；
The DHT-MAP mapping servers form local mirrored storage information updating to the cluster management server
Super ring；
III, the cluster management server receive the information of the DHT-MAP servers, by chord algorithms by information
It is synchronized to purpose cluster management server, the local cluster server buffer backup information.
Further, the distribution work of the virtual machine image storage distribution system comprises the following steps：
I, periodic synchronous global information between the cluster management server, line number is entered according to hot spot data frequency height
According to prefetching, shorten virtual machine batch deployment cycle；
When II, plan machine deployment, template volume caching and the snapped volume caching inquired about on host, if hit, is locally delayed
Deposit；If do not hit, it is necessary to which the test machine for disposing virtual machine sends the necessary template of image starting to super node and rolls up data block
Request, by iSCSI protocol to the object management server on demand read start needed for data block, data acquisition complete
Cached afterwards locally.
Further, in order to accelerate the efficiency of the system read block, the system is opened up using two-stage chord rings
Flutter, all DHT-MAP mapping servers form local ring or two-stage ring, the DHT-MAP mapping servers fortune in cluster
Row is on all hosts.
Further, each cluster sets the cluster management server, forms super node, is distributed in the cluster of various regions
Management server forms super ring or one-level ring by consistent hash algorithm；
The super node is responsible for part ID spaces, by ground regional partition ID identifier spaces, reduces DHT algorithm complications road
By caused route or inquiry time delay.
Compared with prior art, the beneficial effects of the present invention are：
1st, system of the invention realizes the virtual machine batch of distributed cross-region, rapid deployment, solve legacy system and
The method deployment virtual machine cycle is long, can not the problem such as dynamic expansion, availability be low.
2nd, system of the invention uses non-stop layer full symmetrical configuration, no Single Point of Faliure, and memory capacity and performance can be linear
Extension.
3rd, system of the invention uses distributed structure/architecture, storage and dissemination system to be accelerated using two-stage chord rings topology
The positioning of data block and router efficiency.Wherein, all DHT-MAP mapping servers form local ring or two-stage ring in cluster,
DHT-MAP mapping servers are operated on all hosts；Each cluster sets cluster management server between cluster, forms super
Node, the cluster management server for being distributed in various regions form super ring or one-level ring by consistent hash algorithm.In order to reduce DHT
Route or inquiry time delay caused by the tortuous route of algorithm, each super node is responsible for a part of ID spaces, generally according to ground
Domain carries out the segmentation of ID identifier spaces, and territorial scope and ID compasss of competency overlap as far as possible.
4th, system of the invention completes mirror image data storage using object storage server, by the way that all native object are deposited
Server node is stored up according to chord Algorithm constitutions local ring topology, polymerize each host memory space of LAN, forms template
Roll up the storage pool of (read-only) and snapped volume (read-write).After local first preserves mirror image copies, passed through using uniformity hash algorithm
Mirror image copies are saved on purpose memory node by cluster management server.
5th, system of the invention realizes data pre-fetching, on demand classification caching, the mechanism such as dynamic load, can be according to focus number
Data pre-fetching is carried out according to frequency height, is cached locally after the completion of data acquisition, according to template volume caching and fast during caching
Carried out respectively according to volume caching, by iSCSI protocol to object management server needed for dynamic load starts on demand during data acquisition
Data block, shorten virtual machine batch deployment cycle.
Brief description of the drawings
Fig. 1 is the workflow diagram of virtual machine image storage distribution system in the present embodiment；
Fig. 2 is the storage course of work flow chart of virtual machine image storage distribution system in the present embodiment；
Fig. 3 is the distribution course of work flow chart of virtual machine image storage distribution system in the present embodiment.
2、4、8、13、17、18、20：DHT-MAP mapping servers.
1、8、11、20：Cluster management server.
Embodiment
The invention will be further described below in conjunction with the accompanying drawings.
The invention provides a kind of virtual machine image storage based on DHT (distributed hash tables) algorithm
The system of distribution, the system will use non-stop layer full symmetrical configuration, no Single Point of Faliure, and memory capacity and performance can expand in advance
Exhibition.
The system includes cluster management server, the native object operated on the test machine of distributed cloud computing platform is deposited
Store up server, the DHT-MAP mapping servers operated on the test machine of distributed cloud computing platform and operate in distributed cloud
QEMU blocks driving on the test machine of calculating platform.The system can carry out the storage distribution of the data block of mirror image, register update, road
By the operation such as inquiry, by data pre-fetching, increment mirror image, the measure such as reading, query optimization is shortened virtual machine batch and disposed on demand
Cycle.
Test machine：Test machine is the Cloud Server node for covering national each province, including virtual on host and host
Machine.
DHT-MAP mapping servers：Registration, route and inquiry that data block obtains are realized, manages DHT-MAP mapping services
Device node, detect local failure/newly-increased node and report node member change.
Native object storage server：Communicated with the DHT-MAP mapping servers, complete being locally stored for data, gathered
Close each host of LAN (i.e. physical server) memory space, the storage of composition template volume (read-only) and snapped volume (read-write)
Pond；All native object storage server nodes are according to chord Algorithm constitutions local ring topology.
QEMU blocks drive：Communicated with the native object storage server, the QEMU blocks, which drive, to be used for virtual machine mirror
Data block object as being divided into fixed size；For example, fixed size is 4M.
Cluster management server：Manage local DHT-MAP mapping servers node member and other cluster management servers
Node member, detection failure/newly-increased node and report node member change, complete cluster management and message transmission, and across cluster
Between mirror image data block obtain, all cluster management servers are according to the super ring topology of chord Algorithm constitutions.
Present invention storage use distributed storage, according to distributed memory system CAP principles, uniformity, availability and divides
Fault-tolerance three demands in area's are at best able to meet two.To realize in the case of there is network interruption, the system of separation also can be just
Often operation, and require that system can use at any time, mirrored storage and dissemination system meet subregion fault-tolerance and availability.
The nerve of a covering that present system uses is chord, and the one kind belonged in DHT algorithms, master data is object, each
One 128bits of object Global ID's unique mark.Here, object size immobilizes, read by object identity/
Write/establishment/and the operation such as delete.
The storage work of virtual machine image storage distribution system comprises the following steps：
I, virtual machine image is divided into the data block object of fixed size by the driving of QEMU blocks, with hash algorithm generation pair
As mark, it is to virtual machine image storage distribution by the data block object and its redundant copy storage by iSCSI protocol
System；If local have object management server, local cache mirror image data block copy；
II, DHT-MAP mapping server receive mirrored storage I/O requests, differentiate purpose storage according to the object identity
Node, carry out route locating and be forwarded to the cluster management server, I/O requests are forwarded to purpose by cluster management server
Memory node, on purpose memory node, template volume and snapped volume are subjected to classification storage；Meanwhile DHT-MAP is by local mirror image
Super ring of the storage information renewal to cluster management server composition；
III, cluster management server receive the information of the DHT-MAP servers, by chord algorithms by synchronizing information
To purpose cluster management server, the local cluster server buffer backup information.
The distribution work of the system of virtual machine image storage distribution comprises the following steps：
I, periodic synchronous global information between cluster management server, it is pre- that data are carried out according to hot spot data frequency height
Take, shorten virtual machine batch deployment cycle；
The template volume caching and snapped volume caching inquired about first when II, deploying virtual machine on local host, if hit, enters
Row local cache；If do not hit, it is necessary to which the test machine for disposing virtual machine sends the necessary template of image starting to super node
The request of data block is rolled up, the data block needed for starting is read to object management server by iSCSI protocol on demand；Data acquisition
After the completion of locally cached；
In the system, in order to accelerate the positioning of data block and router efficiency in reading process on demand, storage and dissemination system
Using two-stage chord rings topology, all DHT-MAP mapping servers form local ring or two-stage ring, DHT-MAP mappings in cluster
Server is operated on all hosts；Each cluster sets cluster management server, forms super node, is distributed in various regions
Cluster management server forms super ring or one-level ring by consistent hash algorithm.Each super node is responsible for a part of ID
Space, the segmentation of ID identifier spaces is generally carried out according to region, territorial scope and ID compasss of competency overlap as far as possible, reduce DHT and calculate
Route or inquiry time delay caused by the tortuous route of method.In identifier space, node ID and object ID property are just the same.
As shown in figure 1, Fig. 1 is the workflow diagram of virtual machine image storage distribution system in the present embodiment；Virtual machine mirror
Picture storage distribution system is topological using two-stage chord rings, and all DHT-MAP mapping servers form local ring or two level in cluster
Ring, DHT-MAP mapping servers are operated on all test machines.
Each cluster sets cluster management server, forms super node, and the cluster management server for being distributed in various regions is led to
Cross uniformity hash algorithm and form super ring or one-level ring.
Each super node is responsible for a part of ID spaces, the segmentation of ID identifier spaces, ground generally being carried out according to region
Domain scope and ID compasss of competency overlap as far as possible, route or inquiry time delay caused by reducing the tortuous route of DHT algorithms.Identifier space
In, node ID and object ID property are just the same.
In the present embodiment, virtual machine image storage distribution system has linear scalability, new node adition process such as Fig. 1
It is shown, comprise the following steps：
I, node 15 should be inserted between node 13 and node 17, therefore, after node adds, it is necessary to change node 17 before
Drive, 15 are changed to by original 13, while be also required to the descendant node of more new node 13, renewal process is similar；
II, node 17 report change to the super node 20 that cluster management server is formed；
III, super node 20 update finger table route entrys, and the predecessor information of node 17 is updated, and by node 11
Finger tables Section 2 and the 1st, the finger tables of node 13 be changed to 15 by original 17；
IV, super node notice figner routing table update information to all cluster servers.
As shown in Fig. 2 Fig. 2 is the storage course of work flow chart of virtual machine image storage distribution system in the present embodiment；
The storage work of virtual machine image storage distribution system comprises the following steps in the present embodiment：
I, the template volume and user volume of virtual machine (can be distinguished mould by QEMU blocks driver with blktrace monitoring tools
Coiled sheet and user volume) it is divided into the data block object of fixed size (in this implementation, default size 4M), send to native object pipe
Manage server；
II, Object Manager are that each data block generates ID by uniformity hash algorithm, and by relevant information with ID
Information (including：ID, offset, length, action type etc.) and data send together to DHT-MAP mapping servers；
III, DHT-MAP mapping server receive the storage I/O requests of native object manager, as shown in Figure 2, DHT-
MAP mapping servers 2 receive the storage I/O requests for the data block that an ID is 17, it is first determined whether in same cluster pipe
In the ID spaces that reason server super node 3 " can be seen ", if directly being redirected by super node 3 and issuing local
On the DHT-MAP mapping server nodes of ring, i.e. data block 4 and data block 6；If it was not then pass through chord in super ring
Method for routing is forwarded to the next-hop that the finger tables of super node 3 point to --- super node 14, pass through super node
In the range of the ID that finger table search " can be seen " to data block 17 for super node 20, data block is transmitted to super node
20, and redirected by the latter one and issue DHT-MAP mapping servers 17；
IV, DHT-MAP mapping server 17 is transmitted to native object manager by request is stored, by native object manager
The write operation in local is performed, completes storing process.
V, DHT-MAP mapping servers form local mirrored storage information updating to the cluster management server super
Level ring；
Cluster management server receives the information of the DHT-MAP servers, by chord algorithms by synchronizing information to mesh
Cluster management server, the local cluster server buffer backup information.
As described in Figure 3, Fig. 3 is the distribution course of work flow chart of virtual machine image storage distribution system in the present embodiment；
The distribution work of virtual machine image storage distribution system comprises the following steps in the present embodiment：
I, need the test machine for disposing virtual machine to initiate the acquisition of template volume data block to super node 14 to ask, the data
Block is the necessary data block of image starting, and following information is included in request<Object identity ID, offset, length, action type>,
In the present embodiment, action type is reads, length 4M；
After II, super node 14 receive the solicited message from node 10, calculated in super ring according to uniformity Hash
Method will ask to carry out routing forwarding in identifier space according to object identity ID, finds to store from data block according to finger tables and saves
Next nearest node 20 of point, and application is forwarded to super node 20；
III, super node 20 inquire about the finger tables of oneself, find the ID scopes that node 17 " can be seen " at oneself
Interior, then a direct jump forwards the request to node 17；
Data are directly transmitted by iSCSI protocol between IV, node 10 and node 17, complete the template of node 10 volume data
Loading procedure；
V, node 10 obtains user data loading according to above-mentioned identical mode from node 6, and is finally completed virtual machine mirror
The complete distribution of picture, deploying virtual machine process are completed；
The caching of template volume data and user volume data is realized on VI, node using flash cache group modes.This
In used flash cache technologies, using device mapper frameworks realize by multiple block devices invent one band caching
Block device, for local data block cache.
Finally it should be noted that：The above embodiments are merely illustrative of the technical scheme of the present invention and are not intended to be limiting thereof, to the greatest extent
The present invention is described in detail with reference to above-described embodiment for pipe, those of ordinary skills in the art should understand that：Still
The embodiment of the present invention can be modified or equivalent substitution, and without departing from any of spirit and scope of the invention
Modification or equivalent substitution, it all should cover among scope of the presently claimed invention.
Claims (4)
- A kind of 1. virtual machine image storage distribution system, it is characterised in that：The system includes cluster management server and operation Native object storage server, DHT-MAP mapping servers and QEMU blocks on the test machine of distributed cloud computing platform are driven It is dynamic；The DHT-MAP mapping servers carry out the module of the registration of the data block, route and inquiry to obtain data block；The native object storage server communicates with the DHT-MAP mapping servers, the native object storage server For data, each host memory space of polymerization LAN, composition template volume and the storage pool of snapped volume to be locally stored；The QEMU blocks driving is communicated with the native object storage server, and the QEMU blocks, which drive, to be used for virtual machine image It is divided into the data block object of fixed size；The cluster management server is used to realize cluster management and message transmission, and the mirror image data block between cluster obtains Take；Each cluster sets the cluster management server, forms super node, and the cluster management server for being distributed in various regions is led to Cross consistent hash algorithm and form super ring, i.e. one-level ring；The super node is responsible for part ID spaces, by ground regional partition ID identifier spaces, reduces the tortuous route of DHT algorithms and leads The route or inquiry time delay of cause；The storage work of the virtual machine image storage distribution system comprises the following steps：I, virtual machine image is divided into the data block object of fixed size by the QEMU blocks driving, with uniformity hash algorithm Object identity is generated, is deposited the data block object and its redundant copy storage to the virtual machine image by iSCSI protocol Store up dissemination system；If local have native object management server, local cache data block copy；II, the DHT-MAP mapping servers receive mirrored storage I/O requests, differentiate purpose storage according to the object identity Node, carry out route locating and be forwarded to the cluster management server, I/O requests are forwarded to by the cluster management server Template volume and snapped volume are carried out classification storage by purpose memory node, the purpose memory node；The DHT-MAP mapping servers form local mirrored storage information updating to the cluster management server super Ring；III, the cluster management server receive the information of the DHT-MAP mapping servers, by chord algorithms by information It is synchronized to purpose cluster management server, local cluster server buffer backup information.
- 2. the system as claimed in claim 1, it is characterised in that：The cluster management server surpasses according to chord Algorithm constitutions Level ring topology, the native object storage server is according to chord Algorithm constitutions local ring topology.
- 3. the system as claimed in claim 1, it is characterised in that：The distribution saddlebag of the virtual machine image storage distribution system Include following steps：I, periodic synchronous global information between the cluster management server, it is pre- that data are carried out according to hot spot data frequency height Take, shorten virtual machine batch deployment cycle；When II, deploying virtual machine, the template volume caching and snapped volume caching inquired about on host, if hit, is locally cached； If do not hit, it is necessary to which template necessary to the test machine for disposing virtual machine sends image starting to super node rolls up data block Request, read on demand to object management server by iSCSI protocol and start required data block, at this after the completion of data acquisition Ground is cached.
- 4. the system as claimed in claim 1, it is characterised in that：It is described in order to accelerate the efficiency of the system read block System is topological using two-stage chord rings, and all DHT-MAP mapping servers form local ring, i.e. two-stage ring, institute in cluster DHT-MAP mapping servers are stated to operate on all hosts.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410301202.XA CN104050015B (en) | 2014-06-27 | 2014-06-27 | A kind of system of virtual machine image storage distribution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410301202.XA CN104050015B (en) | 2014-06-27 | 2014-06-27 | A kind of system of virtual machine image storage distribution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104050015A CN104050015A (en) | 2014-09-17 |
| CN104050015B true CN104050015B (en) | 2018-01-19 |
Family
ID=51502917
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410301202.XA Expired - Fee Related CN104050015B (en) | 2014-06-27 | 2014-06-27 | A kind of system of virtual machine image storage distribution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104050015B (en) |
Families Citing this family (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105556473A (en) * | 2014-12-27 | 2016-05-04 | 华为技术有限公司 | I/O task processing method, device and system |
| CN105117346A (en) * | 2015-09-24 | 2015-12-02 | 上海爱数软件有限公司 | Automatic testing method and system for distributed system of virtualization platform |
| CN105512053B (en) * | 2015-12-07 | 2017-06-16 | 中南大学 | The mirror cache method of mobile transparent computing system server end multi-user access |
| CN105721545B (en) * | 2016-01-20 | 2019-01-22 | 浪潮(北京)电子信息产业有限公司 | A kind of multi-level cluster management implementation method |
| CN107479942A (en) * | 2016-06-08 | 2017-12-15 | 深圳市深信服电子科技有限公司 | The client operating system of virtual machine starts method and device |
| CN106873911A (en) * | 2017-02-10 | 2017-06-20 | 济南浪潮高新科技投资发展有限公司 | The implementation method of Distributed Storage is carried out using Classfication of containers pattern |
| CN107153513B (en) * | 2017-03-22 | 2020-07-24 | 佛山科学技术学院 | Storage control method of distributed system server and server |
| CN107066206B (en) * | 2017-03-22 | 2020-07-24 | 佛山科学技术学院 | Storage control method and system for distributed physical disk |
| CN107168645B (en) * | 2017-03-22 | 2020-07-28 | 佛山科学技术学院 | Storage control method and system of distributed system |
| CN107491343B (en) * | 2017-09-08 | 2020-04-17 | 中电莱斯信息系统有限公司 | Cross-cluster resource scheduling system based on cloud computing |
| CN107728940A (en) * | 2017-09-27 | 2018-02-23 | 柏科数据技术（深圳）股份有限公司 | The method, apparatus and storage medium of a kind of management copy in data-storage system |
| CN107678688A (en) * | 2017-09-27 | 2018-02-09 | 柏域信息科技（上海）有限公司 | A kind of method, apparatus and storage medium of the managing redundant copy based in cloud storage system |
| CN107729020B (en) * | 2017-10-11 | 2020-08-28 | 北京航空航天大学 | Method for realizing rapid deployment of large-scale container |
| CN110019013B (en) * | 2017-12-08 | 2021-11-05 | 中国电信股份有限公司 | Mirror image storage method, device and computer readable storage medium |
| CN108958886B (en) * | 2018-07-02 | 2022-03-04 | 哈尔滨工业大学 | Virtual machine rapid deployment system and method based on network flow |
| CN109918229B (en) * | 2019-02-18 | 2021-03-30 | 国家计算机网络与信息安全管理中心 | Database cluster copy construction method and device in non-log mode |
| CN110442601B (en) * | 2019-06-29 | 2022-04-22 | 苏州浪潮智能科技有限公司 | Openstack mirror image data parallel acceleration method and device |
| CN110677348B (en) * | 2019-09-17 | 2021-07-27 | 创新先进技术有限公司 | Data distribution method, access method and respective devices based on cache cluster routing |
| CN112099803B (en) * | 2020-07-30 | 2024-04-12 | 浪潮通用软件有限公司 | Incremental deployment method and system of ERP system |
| CN112532701B (en) * | 2020-11-19 | 2022-07-15 | 苏州浪潮智能科技有限公司 | Distributed accelerated storage method and system based on QEMU (quantum execution unit) |
| CN112948052B (en) * | 2021-02-23 | 2023-10-13 | 中国联合网络通信集团有限公司 | Virtual machine migration method across data centers, data center and computer medium |
| CN113468045B (en) * | 2021-06-04 | 2023-12-29 | 济南浪潮数据技术有限公司 | Test system, method and component for server batch configuration software |
| CN113791796B (en) * | 2021-11-15 | 2022-03-04 | 北京金山云网络技术有限公司 | Cross-domain mirror image generation method, mirror image installation method, device and electronic equipment |
| CN114780949B (en) * | 2022-05-20 | 2022-09-16 | 北京数安行科技有限公司 | Method and system for lightweight data security protection based on virtual container |
| CN114995960A (en) * | 2022-07-19 | 2022-09-02 | 银河麒麟软件(长沙)有限公司 | Virtual machine resource pool starting optimization method, system and medium |
| CN118093725A (en) * | 2024-04-22 | 2024-05-28 | 极限数据(北京)科技有限公司 | Ultra-large-scale distributed cluster architecture and data processing method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101834914B (en) * | 2010-04-23 | 2012-09-05 | 北京交通大学 | Method for realizing mapping relation between identity and position |
| CN102521014B (en) * | 2011-12-06 | 2015-02-04 | 北京航空航天大学 | Deploying method and deploying device for virtual machine |
| CN104365058B (en) * | 2012-04-13 | 2018-03-20 | 思杰系统有限公司 | For the system and method in multinuclear and group system high speed caching SNMP data |
| CN102780769B (en) * | 2012-07-04 | 2015-01-28 | 南京邮电大学 | Cloud computing platform-based disaster recovery storage method |
| CN102882983B (en) * | 2012-10-22 | 2015-06-10 | 南京云创存储科技有限公司 | Rapid data memory method for improving concurrent visiting performance in cloud memory system |
-
2014
- 2014-06-27 CN CN201410301202.XA patent/CN104050015B/en not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| 层次p2p存储系统关键技术研究;钟海峰;《中国博士学位论文全文数据库信息科技辑》;20120515(第05期);第18-66页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104050015A (en) | 2014-09-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104050015B (en) | A kind of system of virtual machine image storage distribution | |
| CN102831120B (en) | A kind of data processing method and system | |
| CN101645922B (en) | CDN network system based on geographical position information encoding and distribution treatment method | |
| Surana et al. | Load balancing in dynamic structured peer-to-peer systems | |
| CN101187931A (en) | Distribution type file system multi-file copy management method | |
| CN106161610A (en) | A kind of method and system of distributed storage | |
| CN106663022A (en) | Live migration of virtual machines that use externalized memory pages | |
| CN102855284A (en) | Method and system for managing data of cluster storage system | |
| CN107079046A (en) | For providing system and method for the subnet through pipe (SA) query caching for dynamic cloud | |
| CN102546782A (en) | Distribution system and data operation method thereof | |
| JP2008533564A (en) | Method and apparatus for data management | |
| CN103310000A (en) | Metadata management method | |
| CN101697526A (en) | Method and system for load balancing of metadata management in distributed file system | |
| CN104410655B (en) | The storage system and method for distributed network based on naming mechanism | |
| CN106775446A (en) | Based on the distributed file system small documents access method that solid state hard disc accelerates | |
| US11271800B1 (en) | Leaderless, parallel, and topology-aware protocol for achieving consensus with recovery from failure of all nodes in a group | |
| CN103595780A (en) | Cloud computing resource scheduling method based on repeat removing | |
| CN103885811B (en) | Method, system and device that dummy machine system total system is migrated online | |
| CN101753405A (en) | Cluster server memory management method and system | |
| CN110417883A (en) | A kind of design method of the point to point network structure applied to block chain | |
| CN102098228A (en) | Integrated management system for mobility of identification network and method thereof | |
| Zhu et al. | CCDN: Content-centric data center networks | |
| CN101399765B (en) | Method and system for reducing hot node load in P2P network | |
| CN103294799B (en) | A kind of data parallel batch imports the method and system of read-only inquiry system | |
| CN103401951A (en) | Peer-to-peer architecture based elastic cloud distribution method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180119 |