CN100336050C - Networked mass storage device and implementation approach - Google Patents
Networked mass storage device and implementation approach Download PDFInfo
- Publication number
- CN100336050C CN100336050C CNB031091334A CN03109133A CN100336050C CN 100336050 C CN100336050 C CN 100336050C CN B031091334 A CNB031091334 A CN B031091334A CN 03109133 A CN03109133 A CN 03109133A CN 100336050 C CN100336050 C CN 100336050C
- Authority
- CN
- China
- Prior art keywords
- scsi
- fiber channel
- interface card
- node
- processing node
- 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
- 238000003860 storage Methods 0.000 title abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 51
- 239000000835 fiber Substances 0.000 claims abstract description 43
- 238000000034 method Methods 0.000 claims abstract description 11
- 239000013307 optical fiber Substances 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 12
- 238000003491 array Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 238000013500 data storage Methods 0.000 claims description 3
- 230000000977 initiatory effect Effects 0.000 claims description 2
- 238000004891 communication Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 5
- 239000007858 starting material Substances 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 239000012792 core layer Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 239000010410 layer Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
Images
Landscapes
- Power Sources (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to mass network memory equipment and an implementation method thereof, which belongs to the field of the structure of a computer storage system. The present invention is characterized in that the mass network memory equipment is a multicomputer architecture of a high performance-price ratio, which is realized by open standardized commercialized hardware, and the equipment comprises a host node as a control center, an I/O processing node used for storing data and processing a fiber channel protocol, a Myrinet interconnected network used for high-speed interconnection and a high-density magnetic disk array which are communicated with each other through an interconnected high performance network of the Myrinet so as to transmit memory requirements to each other and carry out corresponding processing. Firmware and a driver in a target mode are correspondingly developed for a commercialized fiber channel host bus adaptor (HBA) so as to realize the client/server relation between fiber channels and realize an SCSI target simulator for processing SCSI commands and SCSI messages, which is realized on the basis of software according to a complete SCSI protocol. The mass network memory equipment is the multicomputer architecture of a high performance-price ratio and high expandability, and is well compatible with communication equipment with a mass network memory of all the manufacturers.
Description
Technical field
Mass network memory devices and its implementation belong to computer memory system structure field.
Background technology
SAN memory device in the storage area network, at present both at home and abroad manufacturer adopts the mode of optical-fibre channel array control unit to realize substantially, and the optical-fibre channel array control unit is the single board computer of an integrated optical fiber channel chip, Redundant Arrays of Inexpensive Disks chip, small computer system interface or fiber channel interface chip and embedded type CPU, internal memory; It is connected with optical fiber switch, hub by fiber channel interface, is connected in the SAN network; The rear end connects SCSI hard disk or optical-fibre channel hard disk by scsi interface or fiber channel interface; Usually, the optical-fibre channel array control unit provides the firmware of oneself, and connection, the RAID dish of being responsible for optical-fibre channel such as is provided with at function.And provide various application journey DLL (dynamic link library) or agent software.The SAN memory device of realizing based on the array control unit mode exists poor expandability, poor compatibility, shortcoming such as do not open, cost an arm and a leg.
Therefore we have proposed the method for a new realization SAN memory device, it adopts based on hardware that open, standardized and realizes a high performance mass network memory devices, and the mass network memory devices is the SAN memory device of a high power capacity, enhanced scalability.We realize that mainly purpose comprises: (1) has higher performance; (2) has higher extensibility; (3) have good opening, compatibility; (4) higher system performance.To satisfy user's request in the future.
Summary of the invention
The object of the present invention is to provide a kind of whole employing new mass network storage system equipment and its implementation open, that standardized hardware system is realized.
(1) the mass network memory devices is characterized in that, it is a kind of based on the multi-computer architecture structure, is made up of host node and a plurality of I/O processing node, and it contains:
Host node adopts the PC mainboard, external interface device; It contains: Myrinet interface card and hard disk that pci bus, small-sized microcomputer system interface card, ethernet network interface card, high-speed interconnect are used, and system embeds the driver of Myrinet card;
The I/O processing node, each I/O processing node is an independently data processing unit, it contains: based on the server master board of PCI and support SMP CPU, the fiber channel interface card, the scsi interface card, the Flash disk of embedded OS is housed, the Myrinet interface card, the 512M internal memory, 2-4 INTEL Xeon, be used to handle fiber channel protocol target mode driver and I/O subsystem software, promptly on the flash memory on the I/O processing node machine, embed the driver of the target mode of fiber channel protocol, realize the SCSI middle layer agreement of target mode, pattern realizes the function of scsi disk firmware;
Internet based on Myrinet, be the interconnective system interconnection parts of I/O processing node and host node, it links to each other with host node with each I/O processing node respectively by each Myrinet interface card, be peer-to-peer between each I/O processing node wherein, and host node and I/O processing node are the client/server relations;
Compact disk array, each I/O processing node adopt the scsi interface card directly to connect the scsi disk array;
Power subsystem adopts the power-supply system with redundant power backup functionality, is responsible for to above-mentioned each parts power supply.
(2) its maximum can be supported 1-18 I/O processing node.
(3) described each I/O node maximum can connect 8 scsi disk arrays.
(4) described each I/O node contains 2-3 scsi interface card, 2 optical-fibre channel HBA.
(5) it adopts rack construction on system architecture, and system's control and these two parts of data storage are separated.
The implementation method of the mass network memory devices that the present invention proposes is characterized in that, it contains successively and has the following steps:
(1) after the power initiation system opened, the starting basic input and output system of I/O processing node server master board was contained in the embedded OS in the Flash disk;
(2) embedded OS initialization;
(3) load fiber channel protocol target mode driver, scsi target simulator, SCSI driver, Myrinet interface card driver;
(4) fiber channel protocol target mode driver waiting for server sends scsi command;
(5) fiber channel protocol target mode driver takes out the processing of I/O controll block from the request queue of fiber channel interface holding firmware;
(6) fiber channel protocol target mode driver is transmitted to the scsi target simulator with I/O controll block;
(7) the scsi target simulator obtains scsi command from I/O controll block;
(8) the scsi target simulator drives SCSI subsystem execution scsi command and obtains the result;
(9) the scsi target simulator returns the result to server by fiber channel protocol target mode driver;
(10) return step (4) (4) more set by step--handle (9).
Experimental results show that: the hardware system that this mass network memory devices adopts fully has the characteristics of high power capacity, high scalability, high compatibility, high performance price ratio, can satisfy the demand of various functions of network store system and performance, very strong practicality is arranged, excellent economy.
Description of drawings
Fig. 1. the schematic block circuit diagram of mass network equipment device equipment
Fig. 2. the program flow diagram of the implementation method of mass network equipment device equipment
The system architecture diagram of Fig. 3 .I/O processing node
Fig. 4. the rack construction block diagram of mass network equipment device equipment
Fig. 5. the software architecture diagram of mass network equipment device equipment
Embodiment
Contain following each parts among Fig. 1:
(1) host node
(2) I/O processing node
(3) compact disk array
(4) based on the internet of Myrinet
(5) power subsystem
Host node: the control center that is the mass network memory devices.It does not participate in data transmission, stores service, but its realize to system carry out the environment measuring management, to functions such as the active situation of I/O processing node monitor.Host node still is the high-performance backup engine of total system, supports tape online backup function.The hardware subsystem of host node comprises INTEL CPU, pci bus, scsi interface card, ethernet network interface card, Myrinet interface card, hard disk.Host node can move several operation systems and web server software systems such as WINDOWS2000.
I/O processing node: be to constitute the most basic, the most important parts of mass network memory devices.Each mass network memory devices maximum can be supported 1-18 I/O processing node, and the major function of I/O processing node is that data storage, fiber channel protocol are handled, the SCSI protocol processes.Each I/O processing node is an independently data processing unit.The I/O processing node is by 2-4 INTEL Xeon, and 512M internal memory, the mainboard based on pci bus, Myrinet interface card, FLASH disk, fiber channel interface card, scsi interface card are formed.Because the employing pci bus is free to select fiber channel interface card, scsi interface card.The function difference of the different processor of I/O processing node, one of them special disposal fiber channel protocol, one of them special disposal I/O subsystem.Each I/O node has the higher system performance, asks for an interview Fig. 3.
Internet based on Myrinet: be the interconnective system interconnect parts of I/O processing node and host node, replaced the High speed rear panel in the traditional design based on the internet of Myrinet.This mode has economy, advantage such as reliable.Each I/O node and host node are connected on the Myrinet network by the Myrinet interface card, are the relation of equity fully between each I/O node, and the transmission storage is asked and handled mutually.And host node and I/O node are the relations of client/server in management, and runs administrative software sends administration order on the host node, and the operation agent software is accepted administration order and handled on the I/O node.
The compact disk array: each I/O node adopts the SCSI card directly to connect the scsi disk array, and each I/O node maximum can connect 8 scsi disk arrays.If each scsi disk array adopts the hard disk of 73G, mostly is 16 hard disks most, then the maximum 73*16*8*18 that supports of mass network storer is approximately the mass memory capacity about 150TB.
Power subsystem: adopt power-supply system with redundant power backup functionality.
The mass network memory devices adopts rack construction on system architecture, system's control and data store branch are separated.
As shown in Figure 4: dotted line is represented to connect by the Myrinet network, and solid line is represented to connect by SCSI bus.
Host node adopts PC device such as Tianrui of association 3130.The hardware configuration of I/O processing node is seen Fig. 3:
The mainboard of each I/O processing node adopts server master board, and as the X5DA8 of Supermicro company, X5DAE mainboard, all CPU adopt INTEL Corp. to strong serial CPU.Each I/O processing node comprises 2 fiber channel interface cards, and the QLA2310F series as QLOGIC company can realize redundancy with failover capabilities or bundle feature between them.Each I/O processing node comprises 2-3 scsi interface card, and as the 7XXX series interfaces card of ADAPTEC company, they connect highdensity disk array subsystem such as ISD PinnacleRAID 500.Power subsystem adopts the 3C3 series of the special company of power-supply system such as mountain with redundant power backup functionality, and the FLASH disk is responsible for storing various softwares, as the DOC2000 series of M-SYSTEMS company.The Myrinet interface card adopts the LANai9 series interfaces card of Myricom company.Memory system adopts the sdram memory of SAMSUNG company.
From Fig. 4 as seen, dissimilar softwares are placed apart in frame.
Goodbye Fig. 2 and Fig. 5, the fiber channel interface card of I/O processing node adopts the firmware and the driver of target mode, and wherein, the scsi target simulator is the scsi command that can dynamically adjust performance and a task processing module based on software.Application program among Fig. 5 refers to various application system programs, and as web server software, the FCP driver of server end refers in particular to the FCP driver of starter pattern.
Because the mass network memory devices adopts the fiber channel interface card, and the firmware of fiber channel interface card and driver are the starter patterns, two fiber channel interfaces can only be reciprocity fully point-to-point relations, rather than client/server relation, they directly can't transmit the SCSI agreement, therefore can't form the SAN storage system.We have proposed a new solution: be the firmware and the driver of fiber channel interface card development goal device pattern, realize the client/server relation between the fiber channel interface card.Based on the fiber channel protocol driver of fiber channel interface card realization target mode, be the key that realizes the mass network memory devices.Adopt hardware component to realize the necessary fiber channel protocol driver of realizing target mode of SAN memory device (mass network memory devices).
The scsi target simulator be one based on the treatment S CSI order that realizes according to complete S CSI agreement of software and the core layer driver of message, its simulates the function that has realized the scsi disk firmware.Because hardware system has directly adopted the scsi interface card, connect the scsi disk array by it, therefore directly adopt the various scsi commands of SCSI subsystem processes and the task requests of starter pattern.
Coupling together by the Myrinet network between a plurality of I/O processing nodes, is the relation of the relation of point-to-point, complete equity between them, and we adopt the mass network memory devices of realizing the single-point system image based on the clustered software of shared disk mechanism.They send the storage request mutually and do corresponding processing each other by the interconnected high performance network communication based on Myrinet.
The present invention has following advantage:
1, the multi-computer architecture of high performance price ratio, high scalability. Consider system enhanced scalability, high reliability, High performance price ratio, the requirements such as highly compatible, the mass network memory devices adopts at present popular mark fully on hardware system The standardization parts. Whole hardware system is made up of a plurality of I/O processing nodes and a host node. A plurality of I/O processing nodes adopt High Speed I/O path connects, and such as Myrinet, SCI etc. form multicomputer system. The mainboard of each I/O node adopts to be supported The server master board of SMP CPU links by fiber channel interface and to receive the SAN storage networking. By scsi interface link connect many Individual scsi disk array. Such massage storage system has the extensibility of height, has the good ratio of performance to price. This That the SAN memory device realized of common embedded array control unit mode is incomparable.
2, according to the hardware architecture of mass network memory devices, designed specially the software of a suitable hardware architecture System comprises the subsystem such as fiber channel protocol driver, scsi target simulator of target mode. They can with firmly Part is well in conjunction with the function that realizes network storage system.
3, special-purpose high-performance embedded operating system, on LINUX operating system basis, realized specially the real-time embedded OS of a suitable SAN storage array, withdraw from startup, done many improvement and adjustment aspect scheduling, little file system and the I/O subsystem in real time, improve disposal ability and the reliability of mass network storage system equipment
4, the mass network memory devices adopts 2Gb fiber channel interface card as the interface that connects the SAN storage networking, can be very The SAN communication equipment of compatible all manufacturers well. The fiber channel protocol driver of realizing not only support optical-fibre channel-2, Standards such as optical-fibre channel-3, optical-fibre channel-4, and support two kinds of patterns of starter/object machine therefore can be well and its The fiber channel interface cartoon letters of his manufacturer has good compatibility. In addition, can also well realize a plurality of optical-fibre channels The functions such as the bandwidth binding of interface card, failure recovery, the effectively reliability of expanding communication bandwidth and raising system.
The implementation method of mass network memory devices is the method for a new realization SAN network storage system, and it adopts fully Standardized hardware system, it is various to satisfy network storage system by adaptive software system architecture and functional module Demand on function and the performance. The mass network memory devices has the advantages such as high ductile, highly compatible, high performance price ratio, Therefore the implementation method based on the mass network memory devices is a method that is fit to very much industrial production and popularization. Have fine Economic worth.
Claims (6)
1, the mass network memory devices is characterized in that, it is a kind of based on the multi-computer architecture structure, is made up of host node and a plurality of I/O processing node, and it contains:
Host node adopts the PC mainboard, external interface device; It contains: Myrinet interface card and hard disk that pci bus, small-sized microcomputer system interface card, ethernet network interface card, high-speed interconnect are used, and system embeds the driver of Myrinet card;
The I/O processing node, each I/O processing node is an independently data processing unit, it contains: based on the server master board of PCI and support SMP CPU, the fiber channel interface card, the scsi interface card, the Flash disk of embedded OS is housed, the Myrinet interface card, the 512M internal memory, 2-4 INTEL Xeon, be used to handle fiber channel protocol target mode driver and I/O subsystem software, on the flash memory on the I/O processing node machine, embed the driver of the target mode of fiber channel protocol, realize the SCSI middle layer agreement of target mode, pattern realizes the function of scsi disk firmware;
Internet based on Myrinet, be the interconnective system interconnection parts of I/O processing node and host node, it links to each other with host node with each I/O processing node respectively by each Myrinet interface card, be peer-to-peer between each I/O processing node wherein, and host node and I/O processing node are the client/server relations;
Compact disk array, each I/O processing node adopt the scsi interface card directly to connect the scsi disk array;
Power subsystem adopts the power-supply system with redundant power backup functionality, is responsible for to above-mentioned each parts power supply.
2, mass network memory devices according to claim 1 is characterized in that: its maximum can be supported 1-18 I/O processing node.
3, mass network memory devices according to claim 1 is characterized in that: described each I/O node maximum can connect 8 scsi disk arrays.
4, mass network memory devices according to claim 1 is characterized in that: described each I/O node contains 2-3 scsi interface card, 2 optical-fibre channel HBA.
5, mass network memory devices according to claim 1, it is characterized in that: it adopts rack construction on system architecture, and system's control and these two parts of data storage are separated.
6, a kind of implementation method of mass network memory devices is characterized in that, it contains successively and has the following steps:
(1) after the power initiation system opened, the starting basic input and output system of I/O processing node server master board was contained in the embedded OS in the Flash disk;
(2) embedded OS initialization;
(3) load fiber channel protocol target mode driver, scsi target simulator, SCSI driver, Myrinet interface card driver;
(4) fiber channel protocol target mode driver waiting for server sends scsi command;
(5) fiber channel protocol target mode driver takes out the processing of I/O controll block from the request queue of fiber channel interface holding firmware;
(6) fiber channel protocol target mode driver is transmitted to the scsi target simulator with I/O controll block;
(7) the scsi target simulator obtains scsi command from I/O controll block;
(8) the scsi target simulator drives SCSI subsystem execution scsi command and obtains the result;
(9) the scsi target simulator returns the result to server by fiber channel protocol target mode driver;
(10) return step (4) (4) more set by step--handle (9).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB031091334A CN100336050C (en) | 2003-04-04 | 2003-04-04 | Networked mass storage device and implementation approach |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB031091334A CN100336050C (en) | 2003-04-04 | 2003-04-04 | Networked mass storage device and implementation approach |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1447254A CN1447254A (en) | 2003-10-08 |
| CN100336050C true CN100336050C (en) | 2007-09-05 |
Family
ID=28050248
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB031091334A Expired - Fee Related CN100336050C (en) | 2003-04-04 | 2003-04-04 | Networked mass storage device and implementation approach |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100336050C (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4528551B2 (en) * | 2004-04-14 | 2010-08-18 | 株式会社日立製作所 | Storage system |
| US7577707B2 (en) * | 2004-04-21 | 2009-08-18 | International Business Machines Corporation | Method, system, and program for executing data transfer requests |
| US7535832B2 (en) * | 2004-11-22 | 2009-05-19 | International Business Machines Corporation | Apparatus and method to set the signaling rate of a switch domain disposed within an information storage and retrieval system |
| JP4404754B2 (en) * | 2004-11-30 | 2010-01-27 | 富士通株式会社 | Data storage apparatus and information processing system |
| JP4413184B2 (en) * | 2005-11-28 | 2010-02-10 | 富士通株式会社 | Data storage system and data storage control device |
| CN100442780C (en) * | 2006-01-27 | 2008-12-10 | 杭州华三通信技术有限公司 | iSCSI memory network, CD array and capacity extending method thereof |
| CN100581172C (en) * | 2006-04-19 | 2010-01-13 | 杭州华三通信技术有限公司 | Method for accessing object magnetic dish and system for extensing disk content |
| CN100451976C (en) * | 2007-07-23 | 2009-01-14 | 清华大学 | Migration management based on massive data classified memory system |
| CN101425945A (en) * | 2007-11-01 | 2009-05-06 | 刘伯安 | System or local area network implementing method for computer |
| CN105515898A (en) * | 2015-12-03 | 2016-04-20 | 浪潮(北京)电子信息产业有限公司 | Test method and system aiming at network storage system performance |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000010296A2 (en) * | 1998-08-12 | 2000-02-24 | Sc-Wireless, Inc. | Method and apparatus for network control in communications networks |
| WO2002027519A1 (en) * | 2000-09-29 | 2002-04-04 | Alacritech, Inc. | Intelligent network storage interface system and devices |
-
2003
- 2003-04-04 CN CNB031091334A patent/CN100336050C/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000010296A2 (en) * | 1998-08-12 | 2000-02-24 | Sc-Wireless, Inc. | Method and apparatus for network control in communications networks |
| WO2002027519A1 (en) * | 2000-09-29 | 2002-04-04 | Alacritech, Inc. | Intelligent network storage interface system and devices |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1447254A (en) | 2003-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1212574C (en) | End node partitioning using local identifiers | |
| CN1205562C (en) | Split transaction protocol for a bus system | |
| EP1804157B1 (en) | Data storage system and data storage control apparatus | |
| US7734778B2 (en) | Distributed intelligent virtual server | |
| US20080276032A1 (en) | Arrangements which write same data as data stored in a first cache memory module, to a second cache memory module | |
| CN1520564A (en) | Mirroring network data to establish virtual storage area network | |
| CN1723434A (en) | Apparatus and method for a scalable network attach storage system | |
| CN100336050C (en) | Networked mass storage device and implementation approach | |
| CN1270250C (en) | Method and device for raising serial storage equipment reliability and elasticity | |
| WO1999030246A1 (en) | Loosely coupled-multi processor server | |
| JP2008140387A (en) | Clustered storage network | |
| CN1746855A (en) | Method and system for backuping applied layer transparent fault-tolerant based on pseudo block | |
| CN103257941A (en) | Multi-protocol storage controller and system | |
| CN104639469A (en) | Computing and storing cluster system based on PCIE (Peripheral Component Interconnect Express) interconnection | |
| CN1581853A (en) | Method for treating group to be transmitted on network, system and programe thereof | |
| CN101031896A (en) | Low cost flexible network accessed storage architecture | |
| CN1220950C (en) | Controller for outer multi-channel network disc array and its protocol fitting method | |
| CN1205547C (en) | Method of substituting conventional SCSI disk for optical fiber disk in storage optical-fiber network | |
| CN1900931A (en) | Multiple machine file storage system and method | |
| CN1564517A (en) | Memory-network memory-magnetic disc high speed reliable storage system and its reading/writing method | |
| CN1203427C (en) | Load balance modulator possessing TCP connection fault tolerant function and its modulating method | |
| CN1829976A (en) | Integrated circuit with dynamic memory allocation | |
| CN100351766C (en) | Disk array system | |
| US20040059850A1 (en) | Modular server processing card system and method | |
| WO2010008366A1 (en) | System to connect a serial scsi array controller to a storage area network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20070905 Termination date: 20110404 |