Storage is the IT industry a hot topic, and the disk array (RAID) is to construct high-performance, mass storage of basic technology, but also build the basis of network storage technology. Experts believe that the performance advantages of disk arrays running in parallel due to the disk, and improving equipment operation parallelism can improve disk performance and data security. Although the disk array is a mature technology, but for most users, it is still a new thing. From the present situation, the user of the mass storage and data security needs of more and more intense, select the disk array is inevitable. Current theme focusing disk array product to introduce the new level after the introduction of RAID 6, RAID technology, product development, and describes the optional RAID products.
Recently, RAID 6 standard introduced, many companies immediately launch the product in support of RAID 6. RAID 6 to the disk array technology quiet for years once again become the focus of attention.
For many computer users, the disk array RAID (Redundant Array of Independent Disks, Redundant Array of Independent Disks) is a mature technology, already successfully applied for 20 years. University of California Berkeley in 1987 two engineers presented the concept of RAID, the first objective is to drive more than the capacity of smaller low-cost high-capacity combined into a "logical drive" or disk array, to achieve increased hard drive capacity and performance goals. 20 years, RAID launched a series of levels, including RAID 0, RAID 1, RAID 2, RAID 3, RAID4, RAID 5, and various combinations such as RAID 0 +1 and so on.
Experts believe that, RAID 6 technology is high-capacity ATA hard drive with the widely adopted and launched it again to RAID by the industry's attention. However, users expect the use of RAID technology continuously improved to ensure data security and stable operation of the disk, that since ancient times.
RAID constantly new standards
What is RAID? More simply, RAID disk array technology is the use of properties and the method of combining data striping to improve data availability and system reliability, and prevent disk failures. It makes use of RAID algorithms to calculate the loss of information data, and then, and then retrieve the data stored in the redundancy backup disk. RAID since its birth, its increasingly wide range of applications, can effectively prevent all types of system failures.
More specifically, RAID will form a common hard disk array, the host writes data, RAID controller to the host to write data into multiple data blocks, and then written to disk arrays in parallel; the host read data time, RAID controller disk array concurrent read the various scattered data on the hard disk, after regrouping them available to the host. As a result of parallel read and write operations, thereby increasing data access speed of storage systems. This average distribution of the data on the disk in various methods to prevent extensive damage to the data, combined with calibration and redundancy, we can guarantee data security.
RAID level often applied mainly RAID 0, RAID 1, RAID 5 and RAID 0 +1 are several. Experts often use a lot of mathematical knowledge to explain the principles of these standards and different.
RAID 0, also called non-redundant disk, it writes the same data as the number of disk striping, these bands are called "blocks." Realization of bands that can simultaneously access data on multiple disks, balance I / O load, increased data storage space and speed of data access speed. RAID 0 is the only one there is no redundancy of RAID technology. The performance advantages of RAID 0 is excellent, disk space utilization, low cost; less is not fault-tolerant, heat volume is huge.
RAID 1 also known as disk mirroring. The system writes data simultaneously to two hard disk duplication, but the performance of the operating system as a logical drive. So if a hard drive failure occurred, the other hard drive to retain a complete data, the system can still work as usual. System can simultaneously read data from the two hard drives, it will increase the hard drive read speed. RAID 1 has the advantage of safe, simple technology, easy management, reading and writing good performance; less than one block can not extend the hard disk capacity, data, large waste of space.
RAID 5 is a RAID 3 and RAID 4 on the basis of developments, which inherit their data redundancy and striping characteristics, and uniform data parity information stored in the array of all the hard disk. Array system can also read and write all of the hard disk, reducing the hard disk mechanical system caused by a time delay system to improve disk I / O capabilities. In addition, when the array of a drive failure, the system can be saved in other hard disk using parity information on the hard drive data recovery failure to continue to work. RAID 5 with data security, write speed, the advantages of higher utilization of space, is widely used; but the downside is that if one drive fails, the entire system performance will be greatly reduced.
Expert advice, RAID level choices are three main factors: availability (data redundancy), performance and cost. If you do not require the availability, select the RAID 0 to get the best performance. If the availability and performance is important but not a major cost factor, the number of options under the hard disk RAID 1. If the availability, cost and performance are equally important, is the usual amount of data transfer and disk options RAID3, RAID5.
RAID continues to introduce new standards, some standards are a wide range of applications, such as RAID 0, RAID 1, RAID 5, but followed some standard be eliminated, such as RAID4 so. However, users of disk array performance requirements are growing, RAID6 standard attendant.
RAID 6 stand out
RAID 6 is RAID in the new family member, is based on the RAID 5 expansion comes. Therefore, as with RAID 5, data and checksum are divided into data blocks, and then were stored in disk arrays each hard disk. However, RAID 6 by adding a separate check disk, it is the distribution at all disk check codes are back together, so that RAID 6 disk array to allow multiple simultaneous disk failures, which demands for high data security applications are very appropriate. In accordance with this idea, build a RAID 6 disk array requires a minimum of four drives. Experts believe that, RAID 6 and RAID 5 write performance does not improve the situation of poor, write caching can be applied only to some extent for the shortcomings of the cover, but does not fundamentally solve the problem. Because RAID 5 and RAID 6 can change depending on the application program to the data block size, so it will be the actual performance factors.
Since RAID 6 is the latest RAID redundancy, so its performance should be very good. Benefits include: random read performance is good (when using large data blocks), sustained read performance is good (when using small data blocks), can be summed up as fast read performance, greater fault tolerance.
Deficiencies included the random write performance of RAID 6: bad, because not only have to check each piece of data written to the hard disk, but also to check the hard disk in a special write data; sustained write performance generally. Can be summed up in slower write speed, RAID controller design more complex and costly.
Therefore, some experts believe that, in practical application the application of RAID 6 may not be as extensive as the other RAID standard. If you need to design to achieve this functionality generally more complex and more expensive cost RAID controller, it generally will not be integrated on the motherboard. Overall RAID 6 can tolerate two simultaneous drive failures can still recover the data, while the actual capacity of its total capacity by two hard drives. (Specifically refer to "RAID 6 can protect more of the objects")
Three Methods of RAID
The way in achieving, RAID can be divided into three categories: software implementation (software RAID), implemented in hardware RAID (hardware RAID), and the external RAID (External RAID).
In the server, the software RAID in many cases have been included in the operating system such as Windows, Netware and Linux, and become a feature of the operating system. Software RAID in charge of all operations by the processor, so the CPU resources, take up very high, which reduces the performance of the system. Software RAID does not require the addition of any hardware device because it is the user's server system by the central processing unit to provide the necessary computing resources.
In the server, hardware RAID is usually achieved by a RAID PCI card processing. RAID cards typically include processors and memory, RAID can provide the resources needed, and hardware RAID system will not take up server resources, thereby greatly enhancing system performance. Hardware RAID can connect internal hard drives, hot-swap backplane or external storage devices. No matter what kind of hard to connect, control all the RAID card, RAID is controlled by the RAID card. Typically, the hardware RAID PCI cards need to install the driver, the general operating system does not include RAID card driver. Disk array can be created before or after installation of the system, the system will be regarded as a virtual hard disk, it also has the fault tolerance and redundancy features. Disk array support for capacity expansion, just add a new hard drive and perform some simple instructions, the system can use this newly added real-time disk capacity.
External RAID is hardware RAID is a kind of difference is that RAID card will not be installed in the system, but installed in an external storage device. And this external storage devices will be connected to the system SCSI card. System does not have any RAID functionality, because it is only a SCSI card; all RAID functions will be transferred to external storage devices. Benefits of external storage is often more hard drive can be connected, not the size of the system chassis. And some advanced techniques, such as fault tolerance, it is necessary more than one server connected to an external storage external device, to provide fault tolerance.
External RAID storage system is an independent, its own CPU, cache and operating system, you can install any operating system, and therefore nothing to do with the server operating system. Because the server system where there is only one SCSI card, not a RAID card. For server systems, and SCSI card, the external type of RAID is just a large hard drive, not a particular device. The current market can be both external RAID SAN (Storage Area Network) or NAS (network attached storage), FC or iSCSI connectivity through the network, a network of storage systems can also be DAS (Direct Attached Storage).
Four interfaces RAID
RAID Card RAID Interface is supported hard disk interface, the key there are four: IDE interfaces, SCSI interfaces, SATA interface, and FC interfaces.
1. IDE Interface
IDE (Integrated Drive Electronics, Integrated Drive Electronics), it's hard drive to hard disk controller integrated with the plate body, reducing the number of hard disk interface cable length, the reliability of data transmission has been enhanced, hard drive manufacturing is more easily , hard drive manufacturers are no longer worried about whether their hard drive controller compatible with other manufacturers; to the user, the hard disk to install them is also more convenient. IDE interface technology has been evolving since its inception, the performance is also rising, its own low-cost, compatibility and features for creating other types of hard disk status can not substitute.
IDE hard drive represents a type, but in practical applications, people also used to refer to the earliest use IDE IDE type hard drive ATA-1, this type of interface, along with the development of interface technology has been eliminated, and its More after the type of hard disk interface, such as ATA, Ultra ATA, DMA, Ultra DMA IDE hard drives, etc. are all interfaces. As the IDE port is parallel port, so in order to distinguish and SATA I hard drives, IDE port hard disk drives, also called PATA port.
2. SCSI Interface
SCSI (Small Computer System Interface) - is completely different with the IDE interface. IDE interface is a standard interface for common PC, but SCSI is not designed specifically for hard disk interface, is a widely used small machines on the high-speed data transmission technology. SCSI interface has wide application, multi-task, wide bandwidth, CPU occupancy rate is low, and support hot swap, etc, but the higher prices make it difficult to 濡?as popular IDE hard drives, SCSI hard drive, therefore mainly used in high end servers and high-end workstations. SCSI hard drives and ordinary IDE hard drive interface, compared to many advantages such as speed and performance of the hard disk itself is relatively high, hard drive speed fast, large capacity cache, CPU occupancy rate, scalability is much better than IDE hard drives, and supports hot swap .
3. SATA Interface
Use the SATA (Serial ATA) hard drive also known as serial port hard drive is the future trend of the mainstream PC with a hard drive. By the Intel, APT, Dell, IBM, Seagate, Maxtor composed of several major vendors Serial ATA standards established by the Commission from the Serial ATA 1.0 specification has been developed to a Serial ATA 2.0 specification. Serial ATA Serial connection, Serial ATA bus with embedded clock signal, with a stronger error correction capability than in the past the biggest difference is able to transmit commands (not just data) to check if found errors automatically corrected, which greatly improved the reliability of data transmission. Serial interface also has a simple structure to support the hot swap benefits.
Serial hard disk is a completely different parallel ATA hard disk interface, a new type of serial transmission of data due to known. Is compared to parallel ATA, Serial ATA continuous serial transmission of data, one can only transmit a data. This can reduce the number of SATA interface pins, so that fewer number of connecting cables, efficiency will be higher. Indeed, Serial ATA 4-pin can only do all the work, were used to connect cable, connect ground wire, send data and receive data, while the structure of such systems can reduce energy consumption and reduce system complexity.
Serial ATA from a higher potential for development more, Serial ATA 1.0 data transfer rate up to the definition of 150MB / s, than the latest parallel ATA (or ATA/133) can achieve 133MB / s maximum data transfer rate also high, and in the Serial ATA 2.0 data transfer rate of 300MB / s, SATA will eventually achieve 600MB / s maximum data transfer rate.
4. Fibre Channel
Fibre Channel (FC) and SCIS interfaces, like the initial design and development is not for the hard disk interface technology, is designed specifically for network systems, but as the demand for storage systems of speed, gradually applied to the hard disk system. Fibre Channel hard disk drives for more storage systems to improve the speed and flexibility was developed, which greatly improved the appearance of multi-drive system, the communication speed. The main characteristics of Fibre Channel are: hot swap of high-speed bandwidth, remote connection, connection and large number of devices. Fibre Channel is like a server system environment such as multi-drive designed to meet the high-end workstations, servers, mass storage sub-network between peripherals through hubs, switches, and two-way point to point connection, serial data communications system for high data rate requirements.
From a holistic point of view, hard disk interface is divided into IDE, SATA, SCSI and Fibre Channel 4 kinds, IDE interface hard disk are used for household products, are also part of the application server; SCSI interface drives are mainly used in server market; and fiber channel only in the high-end servers, expensive; SATA hard drive is kind of new type of interface, it is in the market penetration stage, in the home market has a broad prospect. IDE and SCSI in the large category, but also the separation of a variety of specific interface types, they each have different specifications, with different speeds, such as ATA100 and SATA; Ultra160 SCSI and Ultra320 SCSI represent a specific hard disk interface, the speed difference between the respective bigger.
Application from the server, storage to PC
RAID technology has two characteristics: first, speed second is security. The user seems to form a disk group is like a hard drive, users can partition it, format and so on. The operation of the disk array with a single hard drive exactly the same. The difference is that the storage disk array is much higher speed than a single hard disk, and can provide automatic data backup. Data backup function is damaged in the event of user data, the use of corrupted data backup information can be restored, thereby safeguarding the security of user data. Since these two advantages, RAID technology has been applied to the server early in the SCSI interface hard disk system. With the development of computer technology in recent years, PC's CPU speed machine has entered the era of GHz. IDE interface hard disk to be outdone, have launched a ATA66 and ATA100 hard drive. This makes RAID technology is used in middle and low personal PC, server or even possible.
RAID technology since its inception, the field of high-end server applications on a broad, critical data for the escort industry, although in recent years has begun marketing to the personal PC, but it is still mainly used in high-end computing applications.
In recent years the growing popularity of high-performance computing applications, as global economic integration and a highly developed information technology, trades, especially large-scale scientific applications and Internet applications, data show exponential growth, vast amounts of critical data security become a core part of high-end server applications, RAID technology in the server space becomes more and more important. Thus, the current server technology, have invested in RAID a lot of energy. Server manufacturers in the design of server time, that will basically RAID technology as a product of the standard parameters.
With the development of network storage, RAID has been widely used in various storage systems. Introduced a variety of enterprise storage systems is varied. However, we can summary of the characteristics of their development. First, rising speed, FC and iSCSI in the performance of the competition, so that the performance of storage systems continue to rise; Second, costs are reduced, unit storage costs continue to decrease; security and manageability, improved. These are the development of network storage has laid a solid foundation.
Little information: RAID advantages
鈼?expanded storage capacity could be composed of multiple hard disk capacity of a huge storage space.
鈼?reduce the unit cost of capacity on the market's highest capacity hard drive capacity per trillion price, much higher than the universal type drive, so a more universal type disk array consisting of, the unit price is much lower.
鈼?improve the speed of a single hard disk storage to improve the speed of each period are subject to technical conditions, is due to further often difficult. The use of RAID, you can also share data on multiple hard disk read or write, so the overall speed increase exponentially.
鈼?reliability of RAID systems can use two hard drives, mirrored memory synchronization is complete, such security measures for the network server is the most important as this.
鈼?fault-tolerant RAID controller is a key function of fault tolerance. Fault-tolerant single-block disk array, if an error does not affect the continued use of the whole array, advanced RAID controller also has a save function.
鈼?support for IDE RAID There are still a feature that is supported ATA/66/100. RAID is also divided into two types of SCSI RAID and IDE RAID, IDE RAID of course much cheaper. If the host motherboard does not support ATA/66/100 hard drive, then through the RAID card, you can use a new hard disk ATA/66/100 function.
相关链接:
FTP Clients commentsVBScript DateDiff FunctionWrite-once HD-DVD discs will be in volume production to wait for final approvalVIDEOFORMAT ipodeBiz.scm Introductiongoogle Video formate-cology in the Pan Micro Series 31Thunder look at Real-time on-demand player in the player to build pan-J2ME learning Reviewing the Old PostsNew Ad BlockersTools AND Editors WizardReview Animation ToolsGateway Bank Of China Has Become The OddsDomain name will be the first time Chinese and other non-Latin alphabet and opening support 11 langu.vob filesFu Chong: the general trend of those who know great thingsXbox 360 rmvb