Which RAID Level is Right for Me?

Transcription

1 STOR SOUTOS WT R Which R evel is Right for e? ontents ntroduction...1 R evel escriptions...1 R 0 ()...1 R 1 (irroring)...2 R 1 (Striped irror)...2 R 5 ( with parity)...2 R 5 (ot Space)...3 R 6 ( with dual parity)...3 R 10 (Striped R 1 sets)...3 R 50 (Striped R 5 sets)...4 R 60 (Striped R 6 sets)...4 R evel omparison...5 bout daptec R...5 ata is the most valuable asset of any business today. ost data means lost business. ven if you backup regularly, you need a fail-safe way to ensure that your data is protected and can be accessed without interruption in the event of an online disk. dding R to your storage configurations is one of the most cost-effective ways to maintain both data protection and access. While a number of companies offer R, not all R implementations are created equal. With over 24 years of SS development experience, only daptec offers the most robust R data protection available today, based on a hardened R code proven over years of use in demanding environments and resold by most of the top-tier computer manufacturers. To choose the R level that s right for you, begin by considering the factors below. ach one of these factors becomes a trade-off for another: ost of disk storage ata protection or data availability required (low, medium, high) erformance requirements (low, medium, high) ost boils down to the trade-off between disk capacity and added data availability or performance. or example, R 1/10 and small disk counts of R 6 are costly in terms of lost disk space (50%), but high in data availability. erformance also depends on the access pattern (random/sequential, read/write, long/short) and the numbers of users. This white paper intends to give an overview on the performance and availability of various R levels in general and may not be accurate in all user scenarios. R evel escriptions: R 0 (): Offers low cost and maximum performance, but offers no fault tolerance. single disk results in TOT data loss. usinesses use R 0 mainly for tasks requiring fast access to a large capacity of temporary disk storage (such as video/audio post-production, multimedia imaging,, data logging, etc.) where in case of a disk, the data can be easily reloaded without impacting the business. There are also no cost disadvantages as all storage is usable. R 0 usable capacity is 100% as all available drives are used. ata Stripe ogical isk O R 0 isk rray Reads or writes can occur simultaneously on every drive

2 SRVR STOR WT R Which R evel is Right for e? 2 R 1 (irroring): rovides cost-effective, high fault tolerance for configurations with two disk drives. R 1 refers to maintaining duplicate sets of all data on separate disk drives. t also provides the highest data availability since two complete copies of all information are maintained. There must be two disks in the configuration and there is a cost disadvantage as the usable capacity is half the number of available disks. R 1 offers data protection insurance for any environments where absolute data redundancy, availability and performance are key, and cost per usable gigabyte of capacity is a secondary consideration. R 1 usable capacity is 50% of the available drives in the R set. ogical isk R 1 (Striped mirroring) ombines data striping from R 0 with data mirroring from R 1. ata written in a stripe on one disk is mirrored to a stripe on the next drive in the array. The main advantage over R 1 is that R 1 arrays can be implemented using an odd number of disks. R 1 usable capacity is 50% of the total available capacity of all disk drives in the R set. ote: When using even numbers of disks it is always preferable to use R 10, which will allow multiple drive s. With odd numbers of disks, however, R 1 supports only one drive. R 1 isk rray ata is written to both disks simultaneously. Read requests can be satisfied by data read from either disk or both disks. ogical isk R 5 ( with parity): Uses data striping in a technique designed to provide faulttolerant data storage, but doesn t require duplication of data like R 1 and R 1. ata is striped across all of the drives in the array, but for each stripe through the array (one stripe unit from each disk) one stripe unit is reserved to hold parity data calculated from the other stripe units in the same stripe. Read performance is therefore very good, but there is a penalty for writes, since the parity data has to be recalculated and written along with the new data. To avoid a single drive bottleneck, the parity data for consecutive stripes is interleaved with the data across all disks in the array. R 5 has been the standard in server environments requiring fault tolerance. The R parity requires one disk drive per R set, so usable capacity will always be one disk drive less than the number of available disks in the configuration of available capacity still better than R 1 which has only a 50% usable capacity. R 5 requires a minimum of three disks and a maximum of 16 disks to be implemented. R 5 usable capacity is between 67% - 94%, depending on the number of data drives in the R set. ogical isk R 1 isk rray ata is written to two disks simultaneously, and allows an odd number of disks. Read requests can be satisfied by data read from either disk or both disks. Writes require parity update. ata can be read from each disk independently. dedicated disk is required for hot spares. ata Stripe ot R 5 isk rray

3 SRVR STOR WT R Which R evel is Right for e? 3 R 5 (ot Space): rovides the protection of R 5 with higher /Os per second by utilizing one more drive, with data efficiently distributed across the spare drive for improved /O access. R 5 distributes the hot-spare drive space over the +1 drives comprising the R-5 array plus standard hot-spare drive. This means that in normal operating mode the hot spare is an active participant in the array rather than spinning unused. n a normal R 5 array adding a hot-spare drive to a R 5 array protects data by reducing the time spent in the critical degraded state. owever, this technique does not make maximum use of the hot-spare drive because it sits idle until a occurs. Often many years can elapse before the hotspare drive is ever used. or small R 5 arrays in particular, having an extra disk to read from can provide significantly better read performance. or example, going from a 4-drive R 5 array with a hot spare to a 5-drive R 5 array will increase performance by roughly 25%. One downside of R 5 is that the hot-spare drive cannot be shared across multiple physical arrays as with standard R 5 plus hot-spare. This R 5 technique is more costefficient for multiple arrays because it allows a single hot-spare drive to provide coverage for multiple physical arrays. This configuration reduces the cost of using a hot-spare drive, but the downside is the inability to handle separate drive s within different arrays. This R level can sustain a single drive. R 5 useable capacity is between 50% - 88%, depending on the number of data drives in the R set. R 5 requires a minimum of four disks and a maximum of 16 disks to be implemented. R 6 ( with dual parity): ata is striped across several physical drives and dual parity is used to store and recover data. t tolerates the of two drives in an array, providing better fault tolerance than R 5. t also enables the use of more cost-effective, but less reliable, T and ST disks for business critical data. This R level is similar to R 5, but includes a second parity scheme that is distributed across different drives and therefore offers extremely high fault tolerance and drive- tolerance. R 6 can withstand a double disk. R 6 requires a minimum of four disks and a maximum of 16 disks to be implemented. Usable capacity is always 2 less than the number of available disk drives in the R set. ote: With less expensive, but less reliable ST disk drives in a configuration that employs R 6, it is possible to achieve a higher level of availability than a ibre hannel rray using R 5. This is because the second parity drive in the R 6 R set can withstand a second during a rebuild. n a R 5 set, the degraded state and/or the rebuilding time onto a hot spare is considered the window at which the R array is most vulnerable to data loss. uring this time, if a second disk occurs, data is unrecoverable. With R 6 there are no windows of vulnerability as the second parity drive protects against this. ata Stripe ogical isk Writes require two parity updates (on different drives) to survive two disk s. ata can be read from each disk independently. ata Stripe ogical isk R 5 isk rray Writes require parity update. ata can be read from each disk independently. s are distributed through the array, improving performance. R 6 isk rray R 10 (Striped R 1 sets): ombines R 0 striping and R 1 mirroring. This level provides the improved performance of striping while still providing the redundancy of mirroring. R 10 is the result of forming a R 0 array from two or more R 1 arrays. This R level provides fault tolerance up to one disk of each may fail without causing loss of data.

4 SRVR STOR WT R Which R evel is Right for e? 4 Usable capacity of R 10 is 50% of available disk drives. ogical isk ogical isk ata is striped across R 1 pairs. uplicate data is written to each drive in a R 1 pair. ata and parity are striped across all R 5 arrays. Read requests can occur simultaneously on every drive in an array.,,,,,,, O, R 1 rray R 50 (Striped R 5 sets): ombines multiple R 5 sets with R 0 (striping). helps to increase capacity and performance without adding disks to each R 5 array (which will decrease data availability and could impact performance when running in a degraded mode). R 50 comprises R 0 striping across lower-level R 5 arrays. The benefits of R 5 are gained while the spanned R 0 allows the incorporation of many more disks into a single logical drive. Up to one drive in each may fail without loss of data. lso, rebuild times are substantially less then a single large R 5 array. R 10 isk rray R 1 rray R 5 rray R 50 isk rray R 60 (Striped R 6 sets): ombines multiple R 6 sets with R 0 (striping). ual parity allows the of two disks in each R 6 array. helps to increase capacity and performance without adding disks to each R 6 array (which would decrease data availability and could impact performance in degraded mode). ogical isk [...] 0 ata and parity are striped across all R 6 arrays. Reads can occur simultaneously on every drive in an array. [...X] 0 R 5 rray Usable capacity of R 50 is between 67% - 94%, depending on the number of data drives in the R set. O Q R S T U V W X R 6 rray R 60 isk rray R 6 rray

5 SRVR STOR WT R Which R evel is Right for e? 5 R evel omparison eatures R 0 R 1 R 1 R 5 R 5 R 6 R 10 R 50 R 60 inimum # rives ata rotection o rotection Two-drive Up to one disk in each Up to one disk in each Up to two disk s in each Read erformance igh igh igh igh igh igh igh igh igh Write erformance igh edium edium ow ow ow edium edium edium Read erformance (degraded) Write erformance (degraded) / edium igh ow ow ow igh edium edium / igh igh ow ow ow igh edium ow apacity Utilization 100% 50% 50% 67% - 94% 50% - 88% 50% - 88% 50% 67% - 94% 50% - 88% Typical pplications igh end workstations, data logging, real-time rendering, very transitory data Operating system, transaction databases Operating system, transaction databases ata warehousing, web serving, archiving ata warehousing, web serving, archiving ata archive, backup to disk, high availability solutions, with large capacity requirements ast databases, application arge databases, file, application ata archive, backup to disk, high availability solutions, with large capacity requirements Types of R Types of R Software-ased ardware-ased xternal ardware escription est used for large block applications such as data warehousing or video streaming. lso where have the available U cycles to manage the /O intensive operations certain R levels require. ncluded in the OS, such as Windows, etware, and inux. ll R functions are handled by the host U which can severely tax its ability to perform other computations. est used for small block applications such as transaction oriented databases and web. rocessor-intensive R operations are off-loaded from the host U to enhance performance. attery-back write back cache can dramatically increase performance without adding risk of data loss. onnects to the server via a standard controller. R functions are performed on a microprocessor located on the external R controller independent of the host. dvantages ow price Only requires a standard controller ata protection and performance benefits of R ore robust fault-tolerant features and increased performance versus software-based R OS independent uild high-capacity storage systems for highend bout daptec R daptec s R technology is a thoroughly tested, proven, and trusted software solution that has been deployed in millions of business-critical installations around the world. s such, our hardened R code is the most robust and reliable data protection technology available today. n fact, daptec R ships inside the most popular name-brand more than 1.5 million each year! daptec, nc. 691 South ilpitas oulevard ilpitas, alifornia Tel: (408) ax: (408) iterature Requests: US and anada: 1 (800) or (408) World Wide Web: re-sales Support: US and anada: 1 (800) or (408) re-sales Support: urope: Tel: (44) or ax: (44) opyright 2005 daptec nc. ll rights reserved. daptec and the daptec logo are trademarks of daptec, nc., which may be registered in some jurisdictions. ll other trademarks used are owned by their respective owners. nformation supplied by daptec nc., is believed to be accurate and reliable at the time of printing, but daptec nc., assumes no responsibility for any errors that may appear in this document. daptec, nc., reserves the right, without notice, to make changes in product design or specifications. nformation is subject to change without notice. /: rinted in US 10/ _1.11