CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID Which RAID Level is Right for You? Adaptec Series 8 RAID adapter
Introduction For any organization, whether it be small business or a data Pros: » Better performance than software RAID. center, lost data means lost business. There are two common » Controller cards can be easily swapped out for practices for protecting that data: backups (protecting your data replacement and upgrades. against total system failure, viruses, corruption, etc.), and RAID Cons: » More expensive than software RAID. (protecting your data against drive failure). Both are necessary to ensure your data is secure. Software RAID runs entirely on the CPU of the host computer This white paper discusses the various types of RAID system. configurations available, their uses, and how they should be Pros: » Lower cost due to lack of RAID-dedicated hardware. implemented into data servers. Cons: » Lower RAID performance as CPU also powers the NOTE: RAID is not a substitute for regularly-scheduled backups. OS and applications. All organizations and users should always have a solid backup strategy in place. How does RAID work? What is RAID? In software RAID, the RAID implementation is an application running on the host. This type of RAID uses drives attached to RAID (Redundant Array of Inexpensive Disks) is a data storage the computer system via a built-in I/O interface or a processor- structure that allows a system administrator/designer/builder/ less host bus adapter (HBA). The RAID becomes active as soon user to combine two or more physical storage devices (HDDs, as the OS has loaded the RAID driver software. SSDs, or both) into a logical unit (an array) that is seen by the In hardware RAID, a RAID controller has a processor, memory attached system as a single drive. and multiple drive connectors that allow drives to be attached There are three basic RAID elements: either directly to the controller, or placed in hot-swap back- 1. Striping (RAID 0) writes some data to one drive and some planes. data to another, minimizing read and write access times In both cases, the RAID system combines the individual drives and improving I/O performance. into one logical disk. The OS treats the drive like any other drive 2. Mirroring (RAID 1) replicates data on two drives, preventing in the computer — it does not know the difference between a loss of data in the event of a drive failure. single drive connected to a motherboard or a RAID array being presented by the RAID controller. 3. Parity (RAID 5 & 6) provides fault tolerance by examining the data on two drives and storing the results on a third. Given its performance benefits and flexibility, hardware RAID When a failed drive is replaced, the lost data is rebuilt from is better suited for the typical modern server system. the remaining drives. RAID-compatible HDDs and SSDs It is possible to configure these RAID levels into combination levels — called RAID 10, 50 and 60. Storage manufacturers offer many models of drives. Some are designated as “desktop” or “consumer” drives, and others The RAID controller handles the combining of drives into these as “RAID” or “enterprise” drives. There is a big difference: different configurations to maximize performance, capacity, a consumer drive is not designed for the demands of being redundancy (safety) and cost to suit the user needs. connected into a group of drives and is not suitable for RAID. RAID or enterprise drives, on the other hand, are designed to Hardware RAID vs. software RAID communicate with the RAID controller and act in unison with RAID can be hardware-based or software-based. Hardware other drives to form a stable RAID array to run your server. RAID resides on a PCIe controller card, or on a motherboard- integrated RAID-on-Chip (ROC). The controller handles all From a RAID perspective, HDDs and SSDs only differ in their RAID functions in its own hardware processor and memory. performance and capacity capabilities. To the RAID controller The server CPU is not loaded with storage workload so it can they are all drives, but it is important to take note of the concentrate on handling the software requirements of the server performance characteristics of the RAID controller to ensure it operating system and applications. is capable of fully accommodating the performance capabilities
1 CHOOSING THE RIGHT RAID WHITEPAPER
Approx. 400 IOPS / up to 150 MB/s 50% READ Choosing the Right RAID 2 Read performance on SSD limited as 50% of all requests go to the HDD of the SSD. Most modern RAID controllers are fast enough the missing data from the rest of the drives onto the newly- to allow SSDs to run at their full potential, but a slow RAID inserted drive. This means that your system can survive a drive WRITE controller could bottleneck data and negatively impact system failure without the complex and long-winded task of restoring performance. WRITE data from backups.
Hybrid RAID Approx. 400 IOPs / up to 150 MB/s Choosing the right RAID level Hybrid RAID is a redundant50% storage READ solution that combines high There are several different RAID configurations, called “levels,” capacity, low-cost SATA or higher-performance SAS HDDs with such as RAID 0, RAID 1, RAID 10, and RAID 5. While there is low latency, high IOPs SSDs and an SSD-aware RAID adapter little difference in their names, there are big differences in their card (Figure 1). characteristics and where/when they should be used. The factors to consider when choosing the right RAID level include:
WRITE • Capacity • Performance WRITE • Redundancy (reliability/safety) • Price Approx. 25K IOPs / up to 300 MB/s 100% READ There is no one-size-fits all approach to RAID because focus on one factor typically comes at the expense of another. Some RAID levels designate drives to be used for redundancy, which Figure 1. Hybrid RAID Deployment means they can’t be used for capacity. Other RAID levels focus on performance but not on redundancy. A large, fast, highly- In Hybrid RAID, read operations are done from the faster redundant array will be expensive. Conversely, a small, average- SSD and write operations happen on both SSD and HDD for speed redundant array won’t cost much, but will not be anywhere redundancy purposes. near as fast as the previous expensive array. Hybrid RAID arrays offer tremendous performance gains over With that in mind, here is a look at the different RAID levels and standard HDD arrays at a much lower cost than SSD-only RAID how they may meet your requirements. arrays. Compared to HDD-only RAID arrays, hybrid arrays accelerate IOPs and reduce latency, allowing any server system RAID 0 (Striping) to host more users and perform more transactions per second In RAID 0, all drives are combined into one logical disk on each server, which reduces the number of servers required to (Figure 2). This configuration offers low cost and maximum support any given workload. performance, but no data protection — a single drive failure A simple glance at Hybrid RAID functionality does not readily results in total data loss. show its common use cases, which include creating simple As such, RAID 0 is not recommended. As SSDs become mirrors in workstations through to high-performance read- more affordable and grow in capacity, RAID 0 has declined intensive applications in the small to medium business arena. in popularity. The benefits of fast read/write access are far Hybrid RAID is also used extensively in the data center to outweighed by the threat of losing all data in the event of a drive provide greater capacity in storage servers while providing fast failure. boot for those servers. To learn more about Hybrid RAID, visit https://www.adaptec.com/en-us/solutions/hybrid-RAID/. Usage: Suited only for situations where data isn’t mission- critical, such as video/audio post-production, multimedia Who should use RAID? imaging, CAD, data logging, etc. where it’s OK to lose a complete drive because the data can be quickly re-copied from the source. 650,000 Any server or high-end workstation, and any computer system 619,988 620,174 Generally speaking, RAID 0 is not recommended. where constant uptime is required, is a suitable candidate for 600,000 RAID. Pros: » Fast and inexpensive. Fig3 » All drive capacity is usable. 550,000 542,380 543,285 At some point in the life of a server, at least one drive will fail. Without some form of RAID protection, a failed drive’s data » Quick to set up. Multiple HDDs sharing the data load 500,000 would have to be restored from backups, likely at the loss of make it the fastest of all arrays. RAID 0: Read RAID 6: Read some data and a considerable amount of time. Cons: » RAID 0 provides no data protection at all. LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ With a RAID controller in the system, a failed drive can simply » If one drive fails, all data will be lost with no chance be replaced and the RAID controller will automatically rebuild of recovery. 700,000 600,000 497,505 493,082 500,000 400,000 300,000 Fig4A 200,000 100,000 0 RAID 0: Write
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
23,000 22,044 22,000 21,000 20,129 20,000 Fig4B 19,000 RAID 6: Write
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,418 6,382 6,384 Fig5 5,000 4,253 4,280 4,276
0 RAID 0: Read RAID 5: Read RAID 6: Read
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,207 5,000 3,393 2,775 3,013 2,645 Fig6 2,215 0 RAID 0: Write RAID 5: Write RAID 6: Write
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,314 6,488 6,339 4,423 4,418 4,420 5,000 Fig9 0 RAID 0: Read RAID 5: Read RAID 6: Read
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,298
5,000 3,659 3,180 3,172 2,896 2,987 Fig10 0 RAID 0: Write RAID 5: Write RAID 6: Write
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,437 6,471 5,874 5,000 4,256 4,186 3,945 Fig13 0 RAID 0: Read RAID 5: Read RAID 6: Read
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ
10,000 6,310 5,000 3,275 2,853 3,079 2,480 2,834 Fig14 0 RAID 0: Write RAID 5: Write RAID 6: Write
LSI MegaRAID 9361-8i Adaptec RAID 81605ZQ CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 3
Logical Disk RAID 1E (Striped Mirroring) A B RAID 1E combines data striping from RAID 0 with data C D mirroring from RAID 1 while offering more performance than E Reads or writes can RAID 1 (Figure 4). Data written in a stripe on one drive is F occur simultaneously G on every drive mirrored to a stripe on the next drive in the array. H As in RAID 1, usable drive capacity in RAID 1E is 50% of the Data Stripe total available capacity of all drives in the RAID set.
A B Usage: Small servers, high-end workstations, and other C D E F environments with no large capacity requirements, but where the G H user wants to make sure the data is 100% recoverable in the case of a drive failure. FigureRAID 02. Disk RAID Array 0 Array Pros: » Redundant with better performance and capacity than
RAID 1 (Mirroring) RAID 1. In effect, RAID 1E is a mirror of an odd number of drives. RAID 1 maintains duplicate sets of all data on two separate Cons: drives while showing just one set of data as a logical disk » Cost is high because only half the capacity of the (Figure 3). RAID 1 is about protection, not performance or physical drives is available. capacity. Logical Disk
Since each drive holds copies of the same data, the usable Data is written to two disks A simultaneously, and allows capacity is 50% of the available drives in the RAID set. B C an odd number of disks. Read requests can be Usage: D Generally only used in cases where there is not a large E satis ed by data read from capacity requirement, but the user wants to make sure the F either disk or both disks. data is 100% recoverable in the case of a drive failure, such as accounting systems, video editing, gaming etc.
Pros: » Highly redundant — each drive is a copy of the other. A A copy B B copy C C copy » If one drive fails, the system continues as normal D D copy E with no data loss. E copy F F copy Cons: » Capacity is limited to 50% of the available drives, FigureRAID 4.1E RAIDDisk Array 1E Array and performance is not much better than a single drive. NOTE: RAID 1E is best suited for systems with three drives. NOTE: With the advent of large-capacity SATA HDDs, it is For scenarios with four or more drives, RAID 10 is recommended. possible to achieve an approximately 8TB RAID 1 array using two 8TB HDDs. While this may give sufficient capacity for many RAID 5 (Striping with Parity) small business servers, performance will still be limited by the fact that it only has two spindles operating within the array. As the most common and best “all-round” RAID level, RAID Therefore it is recommended to move to RAID arrays that utilize 5 stripes data blocks across all drives in an array (at least 3 to a more spinning media when such capacities are required. maximum of 32), and also distributes parity data across all drives (Figure 5). In the event of a single drive failure, the system reads Logical Disk the parity data from the working drives to rebuild the data blocks that were lost. A Data is written to both B disks simultaneously. RAID 5 read performance is comparable to that of RAID 0, but C Read requests can be there is a penalty for writes since the system must write both the D satis ed by data read from either disk or both disks. data block and the parity data before the operation is complete. The RAID parity requires one drive capacity per RAID set, so usable capacity will always be one drive less than the total A A copy number of drives in the configuration. B B copy C C copy D D copy
RAID 1 Disk Array Figure 3. RAID 1 Array CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 4
Usage: Often used in fileservers, general storage servers, backup Cons: » More expensive than RAID 5 due to the loss of two drive servers, streaming data, and other environments that call for good capacity to parity. performance but best value for the money. Not suited to database » Slightly slower than RAID 5 in most applications. applications due to poor random write performance. Logical Disk Pros: » Good value and good all-around performance. Writes require two parity Cons: » One drive capacity is lost to parity. A+B updates (on different drives) » Can only survive a single drive failure at any one time. C+D to survive two disk failures. E+F Data can be read from each » If two drives fail at once, all data is lost. G+H disk independently. NOTE: It is strongly recommended to have a hot spare set up with the RAID 5 to reduce exposure to multiple drive failures. Data Stripe
NOTE: While SSDs are becoming cheaper, and their improved A B C D performance over HDDs makes it seem possible to use them in E F RAID 5 arrays for database applications, the general nature of G H small random writes in RAID 5 still means that this RAID level RAID 6 Disk Array should not be used in a system with a large number of small, Figure 6. RAID 6 Array random writes. A non-parity array such as RAID 10 should be used instead. RAID 10 (Striping and Mirroring) Logical Disk RAID 10 (sometimes referred to as RAID 1+0) combines RAID Writes require parity update. A+B Data can be read from each 1 and RAID 0 to offer multiple sets of mirrors striped together C+D disk independently. Parity is (Figures 7 and 8). RAID 10 offers very good performance with distributed through the array E+F good data protection and no parity calculations. G+H to improve performance. RAID 10 requires a minimum of four drives, and usable capacity Data Stripe is 50% of available drives. It should be noted, however, that RAID 10 can use more than four drives in multiples of two. Each A B AB parity mirror in RAID 10 is called a “leg” of the array. A RAID 10 array CD parity C D E EF parity F using, say, eight drives (four “legs,” with four drives as capacity) parity G H GH will offer extreme performance in both spinning media and SSD environments as there are many more drives splitting the reads FigureRAID 5 5. Disk RAID Array 5 Array and writes into smaller chunks across each drive.
RAID 6 (Striping with Dual Parity) Usage: Ideal for database servers and any environment with many small random data writes. In RAID 6, data is striped across several drives and dual parity is Pros: used to store and recover data (Figure 6). It is similar to RAID 5 » Fast and redundant. in performance and capacity capabilities, but the second parity Cons: » Expensive because it requires four drives to get the scheme is distributed across different drives and therefore offers capacity of two. extremely high fault tolerance and the ability to withstand the » Not suited to large capacities due to cost restrictions. simultaneous failure of two drives in an array. » Not as fast as RAID 5 in most streaming environments.
RAID 6 requires a minimum of 4 drives and a maximum of 32 Logical Disk drives to be implemented. Usable capacity is always two less than the number of available drives in the RAID set. A+B Data is striped across C+D RAID 1 pairs. Duplicate Usage: Similar to RAID 5, including fileservers, general storage data is written to each E+F drive in a RAID 1 pair. servers, backup servers, etc. Poor random write performance G+H makes RAID 6 unsuitable for database applications. Pros: » Reasonable value for money with good all-round
performance. A B » Can survive two drives failing at the same time, or C D E F one drive failing and then a second drive failing during G H the data rebuild. RAID 1 Disk Array RAID 1 Disk Array
Figure 7. RAID 10 Array – 2 Legs RAID 10 Disk Array CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 5
Logical Disk Logical Disk
A+B+C+D A+B+C Data and parity are striped across Data is striped across RAID 1 pairs. E+F+G+H all RAID 5 arrays. Read requests D+E+F Duplicate data is written to each can occur simultaneously on drive in a RAID 1 pair. I+J+K+L G+H+I every drive in an array. M+N+O+P J+K+L
A B C D A B C E F G H D E F I J K L G H I M N O P J K L RAID 5 Disk Array RAID 5 Disk Array RAID 10 Disk Array RAID 10 Disk Array RAID 10 Disk Array Figure 9.RAID RAID 50 50Disk Array Array – 2 Legs Figure 8. RAID 10 Array – 3 Legs
RAID 50 (Striping with Parity) Logical Disk RAID 50 (sometimes referred to as RAID 5+0) combines multiple A+B+C+D+E+F Data and parity are striped across RAID 10 Disk Array - 3 LEGS G+H+I+J+K+L all RAID 5 arrays. Read requests RAID 5 sets (striping with parity) with RAID 0 (striping) (Figures can occur simultaneously on M+N+O+P+Q+R every drive in an array. 9 and 10). The benefits of RAID 5 are gained while the spanned S+T+U+V+W+X RAID 0 allows the incorporation of many more drives into a single logical disk. Up to one drive in each sub-array may fail without loss of data. Also, rebuild times are substantially less than a single large RAID 5 array.
A B C D E F A RAID 50 configuration can accommodate 6 or more drives, H J L N but should only be used with configurations of more than 16 S T U V W X drives. The usable capacity of RAID 50 is 67%-94%, depending RAID 5 Disk Array RAID 5 Disk Array RAID 5 Disk Array on the number of data drives in the RAID set.
It should be noted that you can have more than two legs in a Figure 10. RAID 50 Array – 3 Legs RAID 50. For example, with 24 drives you could have a RAID 50 of two legs of 12 drives each, or a RAID 50 of three legs of eight RAID 60 Disk Array - 3 LEGS drives each. The first of these two arrays would offer greater RAID 60 (Striping with Dual Party) capacity as only two drives are lost to parity, but the second array RAID 60 (sometimes referred to as RAID 6+0) combines multiple would have greater performance and much quicker rebuild times RAID 6 sets (striping with dual parity) with RAID 0 (striping) as only the drives in the leg with the failed drive are involved in (Figures 11 and 12). Dual parity allows the failure of two drives the rebuild function of the entire array. in each RAID 6 array while striping increases capacity and Usage: Good configuration for cases where many drives need to performance without adding drives to each RAID 6 array. be in a single array but capacity is too large for RAID 10, such as Like RAID 50, a RAID 60 configuration can accommodate 8 or in very large capacity servers. more drives, but should only be used with configurations of more Pros: » Reasonable value for the expense. than 16 drives. The usable capacity of RAID 60 is between 50%- » Very good all-round performance, especially for 88%, depending on the number of data drives in the RAID set. streaming data, and very high capacity capabilities. Note that all of the above multiple-leg configurations that are Cons: » Requires a lot of drives. possible with RAID 10 and RAID 50 are also possible with » Capacity of one drive in each RAID 5 set is lost to parity. RAID 60. With 36 drives, for example, you can have a RAID 60 » Slightly more expensive than RAID 5 due to this lost comprising two legs of 18 drives each, or a RAID 60 of three legs capacity. with 12 drives in each. Usage: RAID 60 is similar to RAID 50 but offers more redundancy, making it good for very large capacity servers, especially those that will not be backed up (i.e. video surveillance servers handling large numbers of cameras). CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 6
Pros: » Can sustain two drive failures per RAID 6 array within Random data is generally small in nature (i.e., small blocks), the set, so it is very safe. with a large number of small reads and writes making up the data » Very large and reasonable value for money, considering pattern. This is typified by database-type data. this RAID level won’t be used unless there are a large Streaming data is large in nature, and is characterized by such number of drives. data types as video, images, general large files. Cons: » Requires a lot of drives. While it is not possible to accurately determine all of a server’s » Slightly more expensive than RAID 50 due to losing data usage, and servers often change their usage patterns over more drives to parity calculations. time, the general rule of thumb is that random data is best suited Logical Disk to non-parity RAID, while streaming data works best and is most cost-effective on parity RAID.
A+B+C+D Data and parity are striped Note that it is possible to set up both RAID types on the same E+F+G+H across all RAID 6 arrays. controller, and even possible to set up the same RAID types on I+J+K+L Reads can occur simultaneously on every drive in an array. the same set of drives. So if, for example, you have eight 2TB M+N+O+P drives, you can make a RAID 10 of 1TB for your database-type data, and a RAID 5 of the capacity that is left on the drives for your general and/or streaming type data (approximately 12TB). Having these two different arrays spanning the same drives will
A B C D not impact performance, but your data will benefit in performance E G from being situated on the right RAID level. M O Drive size performance F H I J K L While HDDs are becoming larger, they are not getting any faster N P — a 1TB HDD and a 6TB HDD from the same product family
RAID 6 Disk Array RAID 6 Disk Array will have the same performance characteristics. This has an impact when building/rebuilding arrays as it can take a long time to write Figure 11. RAID 60 Array – 2 Legs all of the missing data to the new replacement drive. RAID 60 Disk Array Conversely, SSDs are often faster in larger capacities, so an 80GB SSD and an 800GB SSD from the same product family will have Logical Disk quite different performance characteristics. This should be checked carefully with the product specifications from the drive vendor A+B+C+D+E+F Data and parity are striped to make sure you are getting the performance you think you are G+H+I+J+K+L across all RAID 6 arrays. Reads can occur simultaneously getting from your drives. M+N+O+P+Q+R on every drive in an array. S+T+U+V+W+X With HDDs it is generally better to create an array with more, rather than fewer, drives. A RAID 5 of three 6TB HDDs (12TB capacity) will not have the same performance as a RAID 5 array made from five 3TB HDDs (12TB capacity). A B C D E F G I K With SSDs, however, it is advisable to achieve the capacity S U W required from as few as drives possible by using larger capacity SSDs. These will have higher throughput than their smaller H J L M N O P Q R counterparts and will yield better system performance. T V X
RAID 6 Disk Array RAID 6 Disk Array RAID 6 Disk Array Size of array vs size of drives It is a little-known fact that you do not need to use all of your drive capacity when creating a RAID array. When, for example, creating FigureRAID 12. 50 RAID Disk Array 60 Array- 3 LEGS – 3 Legs the RAID array in the controller BIOS, the controller will show you the maximum possible size the array can be based on the When to use which RAID level drives chosen to make up the array. We can classify data into two basic types: random and streaming. During the creation process, you can change the size of the array As indicated previously, there are two general types of RAID to a lesser size. The unused space on the drives will be available arrays: non-parity (RAID 1, 10) and parity (RAID 5, 6, 50, 60). for creating additional RAID arrays. CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 7
A good example of this would be when creating a large server Of course, rebuild times on SSD arrays are dramatically quicker and keeping the operating system and data on separate RAID due to the fact that the drives are smaller and the write speed of arrays. Typically you would make a RAID 10 of, say, 200GB for the SSDs are much faster than their spinning media counterparts. your OS installation spread across all drives in the server. This would use a minimal amount of capacity from each drive. You can Default RAID settings then create a RAID 5 for your general data across the unused space on the drives. When creating a RAID array in the BIOS or management software, you will be presented with defaults that the controller This has an added benefit of getting around drive size limitations proposes for the RAID settings. The most important of these is for boot arrays on non-UEFI servers as the OS will believe it is the “stripe size.” While there is much science, math and general only dealing with a 200GB drive when installing the operating knowledge involved in working out what is the best stripe size system. for your array, in the vast majority of cases the defaults work best, so use the 256kb stripe size as suggested by the controller. Rebuild times and large RAID arrays The more drives in the array, and the larger the HDDs in the SSDs and read/write cache array, the longer the rebuild time when a drive fails and is In an SSD-only RAID array, disabling the read and write cache replaced or a hot-spare kicks in. While it is possible to have will improve performance in a vast majority of cases. However, 32 drives in a RAID 5 array, it becomes somewhat impractical you may need test whether enabling read and write cache will to do this with large spinning media. improve performance even further. Note that it is possible to For example, a RAID 5 made of 32 6TB drives (186TB) will disable and enable read and write cache on the on the fly without have very poor build and rebuild times due to the size, speed affecting or reconfiguring the array, or restarting the server, so and number of drives. In this scenario, it would be advisable testing both configurations is recommended. to build a RAID 50 with two legs from those drives (180TB capacity). When a drive fails and is replaced, only 16 of the drives (15 existing plus the new drive) will be involved in the rebuild. This will improve rebuild performance and reduce system performance impact during the rebuild process. Note, however, that no matter what you do, when it comes to rebuilding arrays with 6TB+ SATA drives, rebuild times will increase beyond 24 hours in an absolutely perfect environment (no load on server). In a real-world environment with a heavily- loaded system, the rebuild times will be even longer. CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 8
RAID Level Comparison
Features RAID 0 RAID 1 RAID 1E RAID 5
Minimum # Drives 2 2 3 3
Data Protection No Protection Single-drive failure Single-drive failure Single-drive failure
Read Performance High Medium Medium High Medium (depends on data Write Performance High Medium Medium type) Read Performance (degraded) N/A Medium Medium Low
Write Performance (degraded) N/A Medium Medium Low
Capacity Utilization 100% 50% 50% 67% - 94%
Non-mission-critical data, Cases where there is not a Small servers, high-end Fileservers, general storage such as video/audio post- large capacity requirement, workstations, and other servers, backup servers, production, multimedia but the user wants to make environments where no large streaming data, and other imaging, CAD, data logging, sure the data is 100% capacity requirements, but environments that call for etc. where it’s OK to lose a recoverable in the case where the user wants to good performance but Typical Usage complete drive because the of a drive failure, such as make sure the data is 100% best value for the money. data can be quickly re-copied accounting systems, video recoverable in the case of a Not suited to database from the source. editing, gaming etc. drive failure. applications due to poor GENERALLY SPEAKING, random write performance. RAID 0 IS NOT RECOMMENDED.
Fast and inexpensive. All drive Highly redundant – each Redundant with better Good value and good all- capacity is usable. Quick to drive is a copy of the other. performance and capacity around performance. Pros set up. Multiple HDDs sharing If one drive fails, the system than RAID 1. In effect, RAID the data load make it the continues as normal with no 1E is a mirror of an odd fastest of all arrays. data loss. number of drives.
RAID 0 provides no data Capacity is limited to 50% Cost is high because only half One drive capacity is lost protection at all. If one drive of the available drives, and the capacity of the physical to parity. Can only survive a Cons fails, all data will be lost with performance is not much drives is available. single drive failure at any one no chance of recovery. better than a single drive. time. If two drives fail at once, all data is lost. CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 9
RAID Level Comparison — continued
Features RAID 6 RAID 10 RAID 50 RAID 60
Minimum # Drives 4 4 6 8
Data Protection Two-drive failure Up to one drive failure in Up to one drive failure in each Up to two drive failure in each sub-array sub-array each sub-array
Read Performance High High High High
Write Performance Low High Medium Medium
Read Performance (degraded) Low High Medium Medium
Write Performance (degraded) Low High Medium Low
Capacity Utilization 50% - 88% 50% 67% - 94% 67% - 94%
Typical Usage Similar to RAID 5, including Ideal for database servers Good configuration for cases RAID 60 is similar to RAID 50 fileservers, general storage and any environment with where many drives need to be but offers more redundancy, servers, backup servers, many small random data in a single array but capacity making it good for very large etc. Poor random write writes. is too large for RAID 10, such capacity servers, especially performance makes RAID as in very large capacity those that will not be backed 6 unsuitable for database servers. up (i.e. video surveillance applications. servers handling large numbers of cameras).
Pros Reasonable value for Fast and redundant. Reasonable value for the Can sustain two drive failures money with good all-round expense. Very good all-round per RAID 6 array within performance. Can survive performance, especially for the set, so it is very safe. two drives failing at the same streaming data, and very high Reasonable value for money, time, or one drive failing and capacity capabilities. considering this RAID level then a second drive failing won’t be used unless there during the data rebuild. are a large number of drives.
Cons More expensive than RAID 5 Expensive as it requires four Requires a lot of drives. Requires a lot of drives. due to the loss of two drive drives to get the capacity Capacity of one drive in each Slightly more expensive than capacity to parity. Slightly of two. Not suited to large RAID 5 set is lost to parity. RAID 50 due to losing more slower than RAID 5 in most capacities due to cost Slightly more expensive drives to parity calculations. applications. restrictions. Not as fast than RAID 5 due to this lost as RAID 5 in streaming capacity. environments. CHOOSING THE RIGHT RAID WHITEPAPER
Choosing the Right RAID 10
Types of RAID
Types of RAID Software-Based Motherboard-Based Adapter-Based
Description Included in the OS, such as Windows®, Processor-intensive RAID operations Processor-intensive RAID operations and Linux. are off-loaded from the host CPU to are off-loaded from the host CPU a RAID processor integrated into the to an external PCIe adapter. All RAID functions are handled by the motherboard. host CPU which can severely tax its Battery-back write back cache can ability to perform other computations. dramatically increase performance without adding risk of data loss.
Typical Usage Best used for large block applications Inexpensive. Best used for small block applications such as data warehousing or video such as transaction oriented streaming. Also where servers have databases and web servers. the available CPU cycles to manage the I/O intensive operations certain RAID levels require.
Pros Lower cost due to lack of RAID- Lower cost than adapter-based RAID. Offloads RAID tasks from the host dedicated hardware. system, yielding better performance than software RAID. Controller cards can be easily swapped out for replacement and upgrades. Data can be backed up to prevent loss in a power failure.
Cons Lower RAID performance as CPU also No ability to upgrade or replace More expensive than software powers the OS and applications. the RAID processor in the event of and integrated RAID. hardware failure.
May only support a few RAID levels.
About Adaptec RAID
While many companies offer RAID, not all RAID implementations are created equal. With nearly 35 years of protecting, accelerating, and conditioning data as it moves through the I/O path, Adaptec by PMC offers the most robust RAID data protection available today, including support for RAID levels: 0, 1, 1E, 5, 6, 10, 50, and 60, plus Hybrid RAID 1 and 10.
PMC-Sierra, Inc. World Wide Web: www.adaptec.com 1380 Bordeaux Dr. Pre-Sales Support: US and Canada: 1 (800) 442-7274 or (408) 957-7274 or [email protected] Sunnyvale, CA 94089 USA UK: +44 1276 854 528 or [email protected] Tel: +1 (408) 239-8000 Australia: +61-2-90116787 Germany: +49-89-45640621 or [email protected] Singapore: +65-92351044
© Copyright PMC-Sierra, Inc. 2015. All rights reserved. PMC, PMC-SIERRA and Adaptec are registered trademarks of PMC-Sierra, Inc. “Adaptec by PMC” is a trademark of PMC-Sierra, Inc. Other product and company names mentioned herein may be trademarks of their respective owners. For a complete list of PMC-Sierra trademarks, see www.pmc-sierra.com/legal. WP_CHOOSERAID_063015_US Information subject to change without notice.