CN102223388A - Cluster memory - Google Patents
Cluster memory Download PDFInfo
- Publication number
- CN102223388A CN102223388A CN2010101724549A CN201010172454A CN102223388A CN 102223388 A CN102223388 A CN 102223388A CN 2010101724549 A CN2010101724549 A CN 2010101724549A CN 201010172454 A CN201010172454 A CN 201010172454A CN 102223388 A CN102223388 A CN 102223388A
- Authority
- CN
- China
- Prior art keywords
- memory
- hard disk
- node
- speed cache
- connecting port
- 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.)
- Pending
Links
Images
Landscapes
- Memory System Of A Hierarchy Structure (AREA)
Abstract
The invention provides a cluster memory, which comprises a plurality of foundation memory nodes, wherein the foundation memory nodes comprise memory equipment and processing equipment, the memory equipment comprises a hot plug hard disk and a hard disk driver, the processing equipment comprises a processor, a high-speed cache, a front-end Ethernet interface, a rear-end interlinked interface, an error check and ECC (Error Correction Code) internal memory and an NVRAM (Nonvolatile Random Access Memory), and the units in the memory equipment and the processing equipment are inserted or arranged on a mainboard, and are connected through a bus. The invention can avoid the bottle neck of memory throughput and memory expansion capability caused by the bottle neck of the property of a nose, and improves the memory property of the whole body on the basis of furthest reducing the resource configuration and cost.
Description
Technical field
The present invention relates to the technical field of memory, particularly relate to a kind of large scale network data cluster memory of centralized stores fast and safely that is used for.
Background technology
The industry user of, high concurrent visit data-intensive for the Internet, Streaming Media, high-performance calculation, the energy etc., the linear expansion of capacity and performance is necessary, and for traditional storage mode, this but is an insurmountable barrier.
Comparatively popular storage mode comprises NAS (Network Attached Storage at present, network attached storage) data storage method and SAN (Storage Area Network) storage mode, wherein, the NAS storage mode adopts and is independent of server, a kind of file server of developing for the network data storage connects institute's memory device, network of self-forming separately.Storage is the attached of server no longer just like this, but is present among the network as the separate network node, can be shared by all network users.The SAN storage mode has been created the networking of storage.Storage networkingization has been complied with the trend of computer server architecture networking.The support technology of SAN is optical-fibre channel (FC, a Fibre Channel) technology.It is that ANSI is that a standard setting up of network and passage I/O interface is integrated.FC technical support HIPPI, IPI, SCSI, multiple high-level protocols such as IP, ATM, its maximum characteristic is that the communication protocol of network and equipment and transmitting physical dielectric isolation are opened, various protocols can transmit on same physical connection simultaneously like this.
The storage architecture that NAS and SAN adopt is generally a plurality of dish battle array of a head server node band server node, usually be equipped with two head server nodes, in this case, storage system disposal ability, performance is limited to head, because there is a limiting value in the character of storage after the head, if the value of reaching capacity then need to change head, this storage architecture can be understood as the train that does not have speed-raising preceding with forming, the passenger's all adorned in each compartment, head provides power, and when head has broken down, whole storage just can not provide and serve, the storage that is to say the rear end is much regardless of having, and relies on head fully power is provided.
In a word, need the urgent technical problem that solves of those skilled in the art to be exactly at present: how can propose a kind of cluster memory, with storage throughput of avoiding bringing and the bottleneck of storing extended capability owing to the head performance bottleneck, and on the basis that reduces resource distribution as far as possible and expend, improve whole memory property.
Summary of the invention
Technical problem to be solved by this invention provides a kind of cluster memory, in order to storage throughput of avoiding bringing and the bottleneck of storing extended capability owing to the head performance bottleneck, and on the basis that reduces resource distribution as far as possible and expend, improve whole memory property.
In order to address the above problem, the invention discloses a kind of cluster memory, described cluster memory comprises several basic memory nodes, described basic memory node comprises memory device and treatment facility; Wherein, described memory device comprises:
Hot plug hard disk and hard disk drive;
Described treatment facility comprises:
Connecting port, error checking and correction ECC internal memory and nonvolatile random access memory NVRAM in processor, high-speed cache, front end Ethernet interface, the rear end;
Each element is pegged graft or is arranged on the mainboard in described memory device and the treatment facility, links to each other by bus.
Preferably, the capacity of described memory device is from 1.92TB to 12TB, and described hot plug hard disk is 12 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 12 7200RPM;
Described processor is 1 four a nuclear 2.0GHz processor;
Described high-speed cache is the 4GB high-speed cache;
Described front end Ethernet interface is 2 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
Preferably, the capacity of described memory device is 36TB, and described hot plug hard disk is 36 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
Described processor is 2 four nuclear 2.33GHz processors;
Described high-speed cache is the 8GB high-speed cache;
Described front end Ethernet interface is 4 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
Preferably, the capacity of described memory device is 36TB, and described hot plug hard disk is 36 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
Described processor is 1 four a nuclear 2.33GHz processor;
Described high-speed cache is the 4GB high-speed cache;
Described front end Ethernet interface is 2 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
Preferably, the capacity of described memory device is 5.4TB, and described hot plug hard disk is 12 serial SCSI technology SAS dishes, and described hard disk drive is that the serial SCSI technology SAS of 12 450GB 15K RPM drives the strive forward device;
Described processor is 2 four nuclear 2.33GHz processors;
Described high-speed cache is the 16GB high-speed cache;
Described front end Ethernet interface is 4 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
Preferably, described basic memory node is at least three.
Preferably, described cluster memory also comprises:
With the extension storage node that described basic memory node is connected, described extension storage node comprises memory device, and described memory device comprises hot plug hard disk and hard disk drive.
Preferably, described extension storage node is connected with basic memory node by the SAS line, the capacity of described extension storage node memory device is from 6TB to 12TB, described hot plug hard disk is 12 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 12 7200RPM.
Preferably, described cluster memory also comprises:
With the accelerator node that described basic memory node is connected, described accelerator node comprises treatment facility, and described treatment facility comprises connecting port and hot plug disk in processor, high-speed cache, front network interface, the rear end; Each element is pegged graft or is arranged on the mainboard in the described treatment facility, links to each other by bus.
Preferably, the processor of described accelerator node is 2 four nuclear 2.33GHz processors;
Described high-speed cache is 8GB or 32GB high-speed cache;
Described front network interface is 2 10Gbps front network interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
And, 2 hot plug hard disks;
Perhaps, the processor of described accelerator node is 1 four a nuclear 2.0GHz processor;
Described high-speed cache is the 8GB high-speed cache;
Described front network interface is 2 1Gbps front network interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
And, 2 hot plug hard disks.
Compared with prior art, the present invention has the following advantages:
Cluster memory of the present invention is made of some nodes, and each node had both had disposal ability, has memory space again; Node of every increase, disposal ability and memory space are all followed increase, so disposal ability and space all are linear increasing.The benefit that adopts this framework to bring is exactly, and do not exist because there is not the bottleneck of storing extended capability in the bottleneck of the storage throughput that the controller performance bottleneck brings yet, promptly can continue stably to provide handling capacity by adding new node.
Moreover the memory node of cluster memory of the present invention can be divided into three classes: basic memory node, accelerator node and disk extension storage node.The characteristics of basis memory node are the expansion along with node, and memory property and memory space increase simultaneously.After using base node to carry out configuration, in use, can store the demand of expansion, as insufficient memory, but the throughput performance of disk still can satisfy use; Disk performance can not satisfy use, but memory space still can satisfy use; Memory space and disk performance all can't satisfy user demand.
Work as insufficient memory, but the throughput performance of storage can satisfy still when using, and can use disk extension storage node, the external one or more disk extension storage nodes of each basic memory node, thus reach the purpose of quick expansion.
When disk performance can not satisfy use, but memory space still can satisfy when using, and can use the accelerator node.Modular equipment of accelerator node can add within less than 30 seconds in any cluster memory, and need not to shut down or interrupt data access.And, because accelerator node set handling equipment device storage equipment not also only can effectively be saved cost.
Description of drawings
Fig. 1 is the structure chart of a kind of cluster memory embodiment 1 of the present invention;
Fig. 2 is the structure chart of a kind of cluster memory embodiment 2 of the present invention;
Fig. 3 is the structure chart of a kind of cluster memory embodiment 3 of the present invention.
Embodiment
For above-mentioned purpose of the present invention, feature and advantage can be become apparent more, the present invention is further detailed explanation below in conjunction with the drawings and specific embodiments.
At present, storage is just carrying out developing rapidly, from being that foundation develops into so that more image, digital content and the unstructured data of high value are foundation with literal and text message.Build the new application that type and aspects such as emulation, amusement, space flight, national defence, life science, Internet service and business diagnosis are rich in creativity at computer, combine, created the digital content of brand-new type with the network of high bandwidth.Along with more enterprises and individuals creates and use these novel data, the demand of the mode of storage, management and the visit data of innovation is how much levels continues to increase.
For satisfying the requirement of modern data intensive enterprise, quicken visit to critical data, significantly reduce the cost and the complexity of storage and management data simultaneously, the present invention is creatively proposed.
One of core idea of the embodiment of the invention is, a kind of cluster storage architecture of innovation is proposed, promptly form cluster by several memory nodes, attributes such as its performance and capacity can be superposeed by the mode of cluster and be expanded, with the high performance file-sharing problem of effective solution, take into account the high scalability of the high availability and the file system of data simultaneously.For the basic memory node in the cluster storage of the present invention, each all is the same, it had both comprised the memory device that is used to provide capacity, comprise the treatment facility that is used to provide power again, that is to say that each basic memory node not only provides power but also provide capacity, avoid because the bottleneck of storage throughput that the head performance bottleneck brings and storage extended capability by the storage architecture of this disperse power, and can effectively reduce resource distribution and expend, improve whole memory property.
With reference to figure 1, show the structure chart of a kind of cluster memory embodiment 1 of the present invention, described cluster memory comprises several basic memory nodes 10, described basic memory node 10 comprises memory device 11 and treatment facility 12; Wherein, described memory device 11 comprises:
Hot plug hard disk 111 and hard disk drive 112;
Described treatment facility 12 comprises:
Each element is pegged graft or is arranged on the mainboard in described memory device 11 and the treatment facility 12, links to each other by bus.
Use the present invention, memory space in the memory device of many stylobates plinth memory node can aggregate into a storage pool that unified access interface and administration interface can be provided to application server, disk on all memory devices be visited and be utilized to application server can pellucidly by this access interface, can give full play to the performance and the disk utilance of memory device.Data will store and read from many memory devices according to certain rule, to obtain higher concurrent access performance.And the treatment facility of each stylobate plinth memory node is used for data processing separately, provides power in order to be respectively the cluster storage, thereby satisfies every demands such as the high bandwidth in the storage application, concurrent visit, enhanced scalability, manageability.
In a preferred embodiment of the present invention, connecting port can adopt the 10000000000 interior connecting ports of InfiniBand in the described rear end.Because the quality of memory property sees mainly read operation and write operation are how to finish.The process of a write operation comprises: client is initiated the file write operation requests, pass through core switch, write then on the basic memory node, suppose that current cluster memory is made of eight basic memory nodes, use the present invention, made full use of the bandwidth of infinibaed, can from eight nodes, pick out the notst busy current node and receive the file that client sends write request, after present node is received file, can automatically this document be divided into eight fritters, write respectively on eight successively the hard disks by the switch of rear end.Correspondingly, read operation is exactly that eight basic memory nodes from the cluster memory are aggregated into after the node, turns back on the client by switch.
In specific implementation, the basic memory node of described cluster memory disposes three at least.
Certainly, those skilled in the art can also dispose other hardware according to actual needs, and as power supply, fan or the like, the present invention does not limit this.
For making those skilled in the art understand the present invention better, below provide the concrete example of several cluster memories.
Example 1:
The cluster memory that comprises several basic memory nodes, the capacity of each basic memory node are that the memory device of each basic memory node comprises from 1.92TB to 12TB (1T=60G):
12 serial line interface SATA hard disks, and, the serial line interface SATA hard disk drive of 12 7200RPM;
The treatment facility of each basic memory node comprises:
1 four nuclear 2.0GHz processor;
The 4GB high-speed cache;
2 1Gbps front end gigabit ethernet interfaces;
Connecting port in 2 20Gbps (DDR) InfiniBand rear end 10,000,000,000;
4GB ECC internal memory;
The NVRAM of 512MB.
Example 2:
The cluster memory that comprises several basic memory nodes, the capacity of each basic memory node are 36TB, and the memory device of each basic memory node comprises:
36 serial line interface SATA hard disks, and, the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
The treatment facility of each basic memory node comprises:
2 four nuclear 2.33GHz processors;
The 8GB high-speed cache;
4 1Gbps front end gigabit ethernet interfaces;
Connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000;
4GB ECC internal memory;
The NVRAM of 512MB.
Example 3:
The cluster memory that comprises several basic memory nodes, the capacity of each basic memory node are 36TB, and the memory device of each basic memory node comprises:
36 serial line interface SATA hard disks, and, the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
The treatment facility of each basic memory node comprises:
1 four nuclear 2.33GHz processor;
The 4GB high-speed cache;
2 1Gbps front end gigabit ethernet interfaces;
Connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000;
4GB ECC internal memory;
The NVRAM of 512MB.
Example 4:
The cluster memory that comprises several basic memory nodes, the capacity of each basic memory node are 5.4TB, and the memory device of each basic memory node comprises:
12 serial SCSI technology SAS dishes, and the serial SCSI technology SAS of 12 450GB 15K RPM drives the strive forward device;
The treatment facility of each basic memory node comprises:
2 four nuclear 2.33GHz processors;
The 16GB high-speed cache;
4 1Gbps front end gigabit ethernet interfaces;
Connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000;
4GB ECC internal memory;
The NVRAM of 512MB.
Certainly, the configuration of above-mentioned basic memory node is only as example, and in practice, the user can use above-mentioned example or makes up in conjunction with actual demand, or adopts other configuration, the present invention these to be need not limited.
In practice, insufficient memory may occur uses, but the situation that the overall performance that the basic memory node of several of current configuration (as 10) provides still can meet the demands, in this case, can on the basis of original base memory node, further dispose the extension storage node, specifically can be with reference to the structure chart of of the present invention a kind of cluster memory embodiment 2 shown in Figure 2, described cluster memory 20 comprises several basic memory node 21 and extension storage nodes 22, and described basic memory node comprises memory device 211 and treatment facility 212; Wherein, described memory device 211 comprises:
Hot plug hard disk and hard disk drive;
Described treatment facility 212 comprises:
Connecting port, error checking and correction ECC internal memory and nonvolatile random access memory NVRAM in processor, high-speed cache, front end Ethernet interface, the rear end;
Each element is pegged graft or is arranged on the mainboard in described memory device 211 and the treatment facility 212, links to each other by bus; And,
With the extension storage node 22 that described basic memory node 21 is connected, described extension storage node 22 comprises memory device 221, and described memory device 221 can comprise hot plug hard disk and hard disk drive.
Example as a kind of concrete application, described extension storage node can be connected with basic memory node by the SAS line, the capacity of described extension storage node memory device can be from 6TB to 12TB, in this case, described hot plug hard disk can be set to 12 serial line interface SATA hard disks, and described hard disk drive can be set to the serial line interface SATA hard disk drive of 12 7200RPM.
In specific implementation, use present embodiment, can below each basic memory node, hang one or more disk expanding nodes, so just can expand whole memory space easily, for the user, can reduce cost as far as possible, also can reach the purpose of storage expansion easily, easily.
In practice, it is enough also might memory space to occur, but the situation that the overall performance that the basic memory node of several of current configuration (as 10) provides can not meet the demands, in this case, can on the basis of original base memory node, further dispose the accelerator node, specifically can be with reference to the structure chart of of the present invention a kind of cluster memory embodiment 3 shown in Figure 3, described cluster memory 30 comprises several basic memory node 31 and accelerator nodes 32, and described basic memory node comprises memory device 311 and treatment facility 312; Wherein, described memory device 311 comprises:
Hot plug hard disk and hard disk drive;
Described treatment facility 312 comprises:
Connecting port, error checking and correction ECC internal memory and nonvolatile random access memory NVRAM in processor, high-speed cache, front end Ethernet interface, the rear end;
Each element is pegged graft or is arranged on the mainboard in described memory device 311 and the treatment facility 312, links to each other by bus; And,
With the accelerator node 32 that described basic memory node 31 is connected, described accelerator node comprises treatment facility 321, and described treatment facility 321 can comprise connecting port and hot plug disk in processor, high-speed cache, front network interface, the rear end; Each element is pegged graft or is arranged on the mainboard in the described treatment facility, links to each other by bus.
For example, a kind of accelerator node treatment facility can comprise:
2 four nuclear 2.33GHz processors;
8GB or 32GB high-speed cache;
2 10Gbps front network interfaces;
Connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000;
And, 2 hot plug hard disks.
Or as, another kind of accelerator node treatment facility can comprise:
1 four nuclear 2.0GHz processor;
The 8GB high-speed cache;
2 1Gbps front network interfaces;
Connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000;
And, 2 hot plug hard disks.
In specific implementation, use present embodiment, can will speed up the device node a modular equipment is set, in order to adding to as far as possible in short time in any cluster memory, and need not to shut down or interrupt data access.And, because the accelerator node that increases only disposes treatment facility, and do not have configure storage devices, so can on the basis of saving cost, reach the effect that improves disposal ability.
For aforesaid device embodiment, what those skilled in the art were easy to expect is: the applied in any combination of the foregoing description 1-embodiment 3 all is feasible, but this specification has not just described in detail at this as space is limited.
The present invention can be used in numerous general or special purpose computingasystem environment or the configuration.For example: personal computer, server computer, handheld device or portable set, plate equipment, multicomputer system, the system based on microprocessor, set top box, programmable consumer-elcetronics devices, network PC, minicom, mainframe computer, comprise distributed computing environment (DCE) of above any system or equipment or the like.
More than the video frequency monitoring method of a kind of wireless video monitoring system provided by the present invention, a kind of wireless video monitoring system, a kind of video monitor and a kind of server that is used for wireless video monitoring are described in detail, used specific case herein principle of the present invention and execution mode are set forth, the explanation of above embodiment just is used for helping to understand method of the present invention and core concept thereof; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.
Claims (10)
1. a cluster memory is characterized in that, described cluster memory comprises several basic memory nodes, and described basic memory node comprises memory device and treatment facility; Wherein, described memory device comprises:
Hot plug hard disk and hard disk drive;
Described treatment facility comprises:
Connecting port, error checking and correction ECC internal memory and nonvolatile random access memory NVRAM in processor, high-speed cache, front end Ethernet interface, the rear end;
Each element is pegged graft or is arranged on the mainboard in described memory device and the treatment facility, links to each other by bus.
2. cluster memory as claimed in claim 1, it is characterized in that, the capacity of described memory device is from 1.92TB to 12TB, and described hot plug hard disk is 12 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 12 7200RPM;
Described processor is 1 four a nuclear 2.0GHz processor;
Described high-speed cache is the 4GB high-speed cache;
Described front end Ethernet interface is 2 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
3. cluster memory as claimed in claim 1, it is characterized in that, the capacity of described memory device is 36TB, and described hot plug hard disk is 36 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
Described processor is 2 four nuclear 2.33GHz processors;
Described high-speed cache is the 8GB high-speed cache;
Described front end Ethernet interface is 4 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
4. cluster memory as claimed in claim 1, it is characterized in that, the capacity of described memory device is 36TB, and described hot plug hard disk is 36 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 36 1TB 7200RPM;
Described processor is 1 four a nuclear 2.33GHz processor;
Described high-speed cache is the 4GB high-speed cache;
Described front end Ethernet interface is 2 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
5. cluster memory as claimed in claim 1, it is characterized in that, the capacity of described memory device is 5.4TB, and described hot plug hard disk is 12 serial SCSI technology SAS dishes, and described hard disk drive is that the serial SCSI technology SAS of 12 450GB 15K RPM drives the strive forward device;
Described processor is 2 four nuclear 2.33GHz processors;
Described high-speed cache is the 16GB high-speed cache;
Described front end Ethernet interface is 4 1Gbps front end gigabit ethernet interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
Save as the 4GB error checking and correct the ECC internal memory in described error checking and the correction ECC;
Described nonvolatile random access memory NVRAM is 512MB nonvolatile random access memory NVRAM.
6. as claim 1,2,3,4 or 5 described cluster memories, it is characterized in that described basic memory node is at least three.
7. cluster memory as claimed in claim 1 is characterized in that, described cluster memory also comprises:
With the extension storage node that described basic memory node is connected, described extension storage node comprises memory device, and described memory device comprises hot plug hard disk and hard disk drive.
8. cluster memory as claimed in claim 8, it is characterized in that, described extension storage node is connected with basic memory node by the SAS line, the capacity of described extension storage node memory device is from 6TB to 12TB, described hot plug hard disk is 12 serial line interface SATA hard disks, and described hard disk drive is the serial line interface SATA hard disk drive of 12 7200RPM.
9. cluster memory as claimed in claim 1 is characterized in that, described cluster memory also comprises:
With the accelerator node that described basic memory node is connected, described accelerator node comprises treatment facility, and described treatment facility comprises connecting port and hot plug disk in processor, high-speed cache, front network interface, the rear end; Each element is pegged graft or is arranged on the mainboard in the described treatment facility, links to each other by bus.
10. cluster memory as claimed in claim 9 is characterized in that, the processor of described accelerator node is 2 four nuclear 2.33GHz processors;
Described high-speed cache is 8GB or 32GB high-speed cache;
Described front network interface is 2 10Gbps front network interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
And, 2 hot plug hard disks;
Perhaps, the processor of described accelerator node is 1 four a nuclear 2.0GHz processor;
Described high-speed cache is the 8GB high-speed cache;
Described front network interface is 2 1Gbps front network interfaces;
Connecting port is a connecting port in 2 20Gbps InfiniBand rear ends 10,000,000,000 in the described rear end;
And, 2 hot plug hard disks.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101724549A CN102223388A (en) | 2010-05-10 | 2010-05-10 | Cluster memory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101724549A CN102223388A (en) | 2010-05-10 | 2010-05-10 | Cluster memory |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102223388A true CN102223388A (en) | 2011-10-19 |
Family
ID=44779816
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101724549A Pending CN102223388A (en) | 2010-05-10 | 2010-05-10 | Cluster memory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102223388A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102833580A (en) * | 2012-09-19 | 2012-12-19 | 南京斯坦德通信股份有限公司 | High-definition video application system and method based on infiniband |
| CN102882955A (en) * | 2012-09-18 | 2013-01-16 | 浪潮(北京)电子信息产业有限公司 | Extension method and device for cluster storage system |
| CN102929791A (en) * | 2012-10-25 | 2013-02-13 | 浪潮电子信息产业股份有限公司 | Cluster storage system expansion method based on identical expansion module |
| WO2013107029A1 (en) * | 2012-01-19 | 2013-07-25 | 华为技术有限公司 | Data processing method, device and system based on block storage |
| CN105912275A (en) * | 2016-04-27 | 2016-08-31 | 华为技术有限公司 | Method and device for establishing connection in nonvolatile memory system |
| CN108664420A (en) * | 2017-03-28 | 2018-10-16 | 慧与发展有限责任合伙企业 | The locking and access based on capacity of shared permanent memory |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079897A (en) * | 2007-06-22 | 2007-11-28 | 西安交通大学 | A concurrent storage system construction method for convenient expansion of storage node quantity |
| US20090249013A1 (en) * | 2008-03-27 | 2009-10-01 | Asif Daud | Systems and methods for managing stalled storage devices |
| CN201699750U (en) * | 2010-05-10 | 2011-01-05 | 北京月新时代科技有限公司 | Colony memory |
-
2010
- 2010-05-10 CN CN2010101724549A patent/CN102223388A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079897A (en) * | 2007-06-22 | 2007-11-28 | 西安交通大学 | A concurrent storage system construction method for convenient expansion of storage node quantity |
| US20090249013A1 (en) * | 2008-03-27 | 2009-10-01 | Asif Daud | Systems and methods for managing stalled storage devices |
| CN201699750U (en) * | 2010-05-10 | 2011-01-05 | 北京月新时代科技有限公司 | Colony memory |
Non-Patent Citations (2)
| Title |
|---|
| SHAI HARMELIN.ET: "《Isilon IQ Scale-out NAS for High-Performance Applications-Whitepaper》", 31 July 2009, ISILON SYSTEMS, article "Optimizing Performance with Isilon IQ Storage" * |
| 编辑部: "Isilon重视中国市场推出三款新品", 《中国通信》, no. 6, 30 June 2009 (2009-06-30) * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013107029A1 (en) * | 2012-01-19 | 2013-07-25 | 华为技术有限公司 | Data processing method, device and system based on block storage |
| CN103329111A (en) * | 2012-01-19 | 2013-09-25 | 华为技术有限公司 | Data processing method, device and system based on block storage |
| CN103329111B (en) * | 2012-01-19 | 2015-05-27 | 华为技术有限公司 | Data processing method, device and system based on block storage |
| US9507720B2 (en) | 2012-01-19 | 2016-11-29 | Huawei Technologies Co., Ltd. | Block storage-based data processing methods, apparatus, and systems |
| CN102882955A (en) * | 2012-09-18 | 2013-01-16 | 浪潮(北京)电子信息产业有限公司 | Extension method and device for cluster storage system |
| CN102833580A (en) * | 2012-09-19 | 2012-12-19 | 南京斯坦德通信股份有限公司 | High-definition video application system and method based on infiniband |
| CN102833580B (en) * | 2012-09-19 | 2016-03-02 | 南京斯坦德云科技股份有限公司 | Based on HD video application system and the method for infiniband |
| CN102929791A (en) * | 2012-10-25 | 2013-02-13 | 浪潮电子信息产业股份有限公司 | Cluster storage system expansion method based on identical expansion module |
| CN105912275A (en) * | 2016-04-27 | 2016-08-31 | 华为技术有限公司 | Method and device for establishing connection in nonvolatile memory system |
| WO2017185996A1 (en) * | 2016-04-27 | 2017-11-02 | 华为技术有限公司 | Method and device for establishing connection in nonvolatile memory system |
| US10866911B2 (en) | 2016-04-27 | 2020-12-15 | Huawei Technologies Co., Ltd. | Method and apparatus for establishing connection in non-volatile memory system |
| CN108664420A (en) * | 2017-03-28 | 2018-10-16 | 慧与发展有限责任合伙企业 | The locking and access based on capacity of shared permanent memory |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11580041B2 (en) | Enabling use of non-volatile media—express (NVME) over a network | |
| US11860813B2 (en) | High level instructions with lower-level assembly code style primitives within a memory appliance for accessing memory | |
| US11704059B2 (en) | Remote direct attached multiple storage function storage device | |
| US20150127691A1 (en) | Efficient implementations for mapreduce systems | |
| KR101340176B1 (en) | Smart scalable storage switch architecture | |
| US20230273889A1 (en) | Compute sled providing high speed storage access through a pci express fabric between compute nodes and a storage server | |
| KR101744465B1 (en) | Method and apparatus for storing data | |
| US20170177222A1 (en) | Methods and systems for data storage using solid state drives | |
| CN112115090A (en) | Multi-protocol support for transactions | |
| KR20120087980A (en) | Multi-interface solid state disk(ssd), processing method and system thereof | |
| WO2016196766A2 (en) | Enabling use of non-volatile media - express (nvme) over a network | |
| US10540307B1 (en) | Providing an active/active front end by coupled controllers in a storage system | |
| CN102209103A (en) | Multicasting write requests to multiple storage controllers | |
| CN102223388A (en) | Cluster memory | |
| CN101169703A (en) | Solid state hard disk memory card based on RAID technology | |
| US20170220506A1 (en) | Modular Software Defined Storage Technology | |
| JP2017531856A (en) | Active storage units and arrays | |
| Poremba et al. | There and back again: Optimizing the interconnect in networks of memory cubes | |
| CN107632923A (en) | A kind of tandem type hard disk and its alarm method | |
| CN102937997A (en) | Data processing system | |
| CN103036958A (en) | Server cluster high-speed memory system | |
| CN201699750U (en) | Colony memory | |
| CN105204787B (en) | A kind of data of PCIE SSD array reads method, system and reading/writing method | |
| CN110471627A (en) | A kind of method, system and device of shared storage | |
| CN202206413U (en) | iSCSI storage node and architecture |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20111019 |