RAID Primer: What's in a number?
by Dave Robinet on September 7, 2007 12:00 PM EST- Posted in
- Storage
RAID 5
In an effort to strike a balance between performance and data redundancy, RAID 5 not only stripes data across multiple disks, but also writes a "parity" block among those disks which allows the array to recover from a single failed drive in the set. This parity block is staggered from one drive to the next, resulting in each drive having either a portion of the data that is trying to be read, or the parity block which allows the data to be reconstructed. In this way, the array gains some performance benefits in having the data striped among multiple disks, while being able to stay online after the failure of a single disk in the array.
Rather than having a dedicated drive for parity as in the less-popular RAID levels 3 and 4, the parity sharing of RAID 5 allows for a more distributed drive access pattern, resulting in improved write performance and a more even disk wear pattern than with a dedicated parity drive.
In its optimal form, RAID 5 provides substantially faster read performance than a single drive or RAID 1 configuration, but write performance suffers due to the need to write (and sometimes recalculate) the parity information for the majority of writes performed. While this write performance is still faster than a single disk configuration in most cases, the true performance benefit of a RAID 5 array is in its read ability.
It should be noted that the read performance of a RAID 5 array improves as the number of disks in the array increases. This increase in disks, however, increases the odds of a disk failure in the array due to the law of averages, which results in a performance degradation during rebuilding operations. It also increases the likelihood of the entire array being unrecoverable if a second disk in the array fails before the first failed disk is replaced. In the image shown, the array contains a "hot spare" disk - in the event of a disk failure, this "hot spare" disk would be brought into the array to replace the failed disk immediately.
RAID 5 finds a comfortable home in most "read often, write infrequently" server applications which require long periods of uptime, such as web servers, file/print servers, and database servers. Dedicated RAID 5 controllers that include large amounts of RAM can negate much of the write performance penalty, but such setups are quite a bit more costly. Note also that simplified RAID 5 controllers exist that require the CPU to perform the parity calculations, which can result in write performance that is lower than a single drive.
Pros:
RAID 6 attempts to address the most glaring of the RAID 5 issues: The comparatively large window in which the array is in a dangerous state due to a failed disk.
As in RAID 5, RAID 6 staggers its parity information across multiple drives. Its major difference, however, is that it writes two parity blocks for every stripe of data, which means that the array is capable of remaining accessible to the users even after having sustained two simultaneous drive failures. The advantage of a RAID 6 array versus a RAID 5 array with a hot spare risk is that no rebuilding is necessary to bring the last disk into the array in the event of a failure. In this sense, performance is more or less guaranteed after a single disk failure under RAID 6, whereas a significant performance hit occurs during the rebuilding required under RAID 5.
RAID 6's parity scheme is not simply multiple copies of the same parity information, but rather two different means of calculating parity information for the same data. This results in a much higher computing overhead than the already-intensive RAID 5 scheme, and a resulting increase in controller/CPU usage. This requirement for the second parity calculation and write, however, further adversely impacts the write performance of a RAID 6 array versus a RAID 5 solution.
RAID 6 is an excellent choice for both extremely mission-critical applications and in instances where large numbers of disks are intended to be used in the array to improve read performance. Because of this (and the poor write performance without special hardware), RAID 6 support is typically only included in high-end, expensive controller cards.
Pros:
In an effort to strike a balance between performance and data redundancy, RAID 5 not only stripes data across multiple disks, but also writes a "parity" block among those disks which allows the array to recover from a single failed drive in the set. This parity block is staggered from one drive to the next, resulting in each drive having either a portion of the data that is trying to be read, or the parity block which allows the data to be reconstructed. In this way, the array gains some performance benefits in having the data striped among multiple disks, while being able to stay online after the failure of a single disk in the array.
Rather than having a dedicated drive for parity as in the less-popular RAID levels 3 and 4, the parity sharing of RAID 5 allows for a more distributed drive access pattern, resulting in improved write performance and a more even disk wear pattern than with a dedicated parity drive.
In its optimal form, RAID 5 provides substantially faster read performance than a single drive or RAID 1 configuration, but write performance suffers due to the need to write (and sometimes recalculate) the parity information for the majority of writes performed. While this write performance is still faster than a single disk configuration in most cases, the true performance benefit of a RAID 5 array is in its read ability.
It should be noted that the read performance of a RAID 5 array improves as the number of disks in the array increases. This increase in disks, however, increases the odds of a disk failure in the array due to the law of averages, which results in a performance degradation during rebuilding operations. It also increases the likelihood of the entire array being unrecoverable if a second disk in the array fails before the first failed disk is replaced. In the image shown, the array contains a "hot spare" disk - in the event of a disk failure, this "hot spare" disk would be brought into the array to replace the failed disk immediately.
RAID 5 finds a comfortable home in most "read often, write infrequently" server applications which require long periods of uptime, such as web servers, file/print servers, and database servers. Dedicated RAID 5 controllers that include large amounts of RAM can negate much of the write performance penalty, but such setups are quite a bit more costly. Note also that simplified RAID 5 controllers exist that require the CPU to perform the parity calculations, which can result in write performance that is lower than a single drive.
Pros:
- Good usable amount of data (capacity is the sum of all but one drive in the set)
- Fault-tolerant - can survive one disk failure without impact to users
- Strong read performance
- Write performance (without a large controller cache) is substantially below that of RAID 0
- Expensive (either in terms of controller cost or CPU usage) due to parity calculations
RAID 6 attempts to address the most glaring of the RAID 5 issues: The comparatively large window in which the array is in a dangerous state due to a failed disk.
As in RAID 5, RAID 6 staggers its parity information across multiple drives. Its major difference, however, is that it writes two parity blocks for every stripe of data, which means that the array is capable of remaining accessible to the users even after having sustained two simultaneous drive failures. The advantage of a RAID 6 array versus a RAID 5 array with a hot spare risk is that no rebuilding is necessary to bring the last disk into the array in the event of a failure. In this sense, performance is more or less guaranteed after a single disk failure under RAID 6, whereas a significant performance hit occurs during the rebuilding required under RAID 5.
RAID 6's parity scheme is not simply multiple copies of the same parity information, but rather two different means of calculating parity information for the same data. This results in a much higher computing overhead than the already-intensive RAID 5 scheme, and a resulting increase in controller/CPU usage. This requirement for the second parity calculation and write, however, further adversely impacts the write performance of a RAID 6 array versus a RAID 5 solution.
RAID 6 is an excellent choice for both extremely mission-critical applications and in instances where large numbers of disks are intended to be used in the array to improve read performance. Because of this (and the poor write performance without special hardware), RAID 6 support is typically only included in high-end, expensive controller cards.
Pros:
- Fair usable amount of data (sum of all but two drives in the set)
- Provides more comfortable levels of redundancy for very large array sizes (8+ disks)
- Strong read performance
- Expensive (both in computing power, controller, and in additional "wasted" disks)
- Write performance is generally very poor compared to other RAID solutions
Facebook
Twitter
RSS
41 Comments
View All Comments
tynopik - Saturday, September 8, 2007 - link
> So I'm looking for a solution which stores my data in a "normal" way on the discs + one extra disk with the parity (somewhat like RAID 3 but without the striping).unRAID
http://www.lime-technology.com/">http://www.lime-technology.com/
tynopik - Saturday, September 8, 2007 - link
i should point out that1. it does NOT join your drives together into one volume, each drive is separate (this is basically necessary for what you want unless you go the WHS route)
2. it has to be run on a dedicated system that it turns into NAS (you can't run it on your main desktop for instance)
that said, i really like the idea, almost all of the advantages of the WHS mechanism but much more space efficient (in most cases, i assume the largest drive will always be 'lost' to parity data)
Dave Robinet - Saturday, September 8, 2007 - link
Really, you're looking for something that is several RAID 1 mirrors of single volumes.I can think of nothing off-the-shelf that fits all those needs, though "rolling your own" may help:
- Buy two drives. Create one large partition (say, D:) on drive 1. Mirror that.
- Buy two more drives. Create another large partition (say, E:) on drive 3. mirror that.
Etc, etc.
It's still the same volume, but if you do it using software, the two drives won't be dependent on each other in any way.
If you tear one of those drives out of your computer and slap it onto another one (USB connector, etc), then it'll come up just fine, with or without the mirror.
It's inelegant, and really not something I'd ever push on someone - but you've come up with a kind of oddball request, there. :) Might I ask what it's for? Maybe your criteria can be adjusted in some way.
tynopik - Saturday, September 8, 2007 - link
> but you've come up with a kind of oddball request, there. :) Might I ask what it's for? Maybe your criteria can be adjusted in some way.i understand what he's getting at
he wants protection from drive failure, so a 'parity drive' that can rebuild any one drive that fails is handy
but he's also concerned about losing more than 1 drive simultaneously
having just a plain filesystem on the disk is far more robust than any sort of striping system as worst comes to worst you can just yank any surviving drives and recover what's on them
- a series of raid 1 arrays (like what you described) works but isn't particularly flexible (need equal sized drives)
- WHS is more flexible and powerful but it still requires double the amount of storage (EXPENSIVE)
- this only requires 1 extra drive and allows it to backup any number of other drives
it comes from a desire for some protection but not being able or willing to spend enough for true duplication plus wanting something that fails gracefully (ie not raid5)
i would actually like something like that for my system, there's a chance of recovering everything, but if it hits the fan i'll be able to at least recover something plus it's not that expensive
don't forget there may be physical limitations. if you have 4 physical drives filled with data, you might have enough room and power connectors for a 5th drive, but not for 4 more
Sudder - Sunday, September 9, 2007 - link
tnx, this goes a big step in the right direction
since unraid uses slackware there should be (at least in theory) a possibility to do this with the linux "Logical Volume Manager" (allthough one would probably have to do some work so that the TOC is saved on every disc to still being able to access the data if some of the other discs are gone)
But even without, seperate volumes and the option to access them one by one, by mounting the ReiserFS, is good enough for me.
and that's a big downside.
When I have some time I'll probably try to run it in a virtual machine (the "use a physical disc" option in VM should reduce the perfomance-penaltys significantly), but I'm not that optimistic that this will also work with the "bigger", non-free Versions that can handle more than 3 discs (e.g. handling of the registration Key, since I allready ran into some pre-boot USB Issues with VM when I tried to test the bitlocker-feature of Vista in a VM - although it just might have been my old stick or my USB-contoller ..)
yes (that's kind of a given) - the option to use discs of different sizes is a nice bonus though, since the array can now grow more "organically" over time (you just buy the disc with the best cost/gig ratio at the moment you need it without limmiting yourself to one size like with RAID 5)
with the port-multiplier Option of SATA II (up to 15 drives per cable) and an external casing I think there are ways to cope (and if you plan in advance to have X bays/connectors avalable, you just have to start a new array if the old one is full - which might be a good idea anyway as soon as you come close to double digit disc-numbers - although, that might take some time with modern disc sizes ;-) )
Again: I don't want nessecarily to being able to access my data all the time, I just want to switch from my current "DVD-storrage" to a "HD-storrage".
So what I'm looking for now is the funktionallity of unRAID (without the limitation of the drive nuber), being able to run in a VM and for free ;-).
I allready checked freeNAS and NASlite but they all seemed to be fixed either on RAID and/or JBOD without parity .. any suggestions?
tynopik - Sunday, September 9, 2007 - link
> with the port-multiplier Option of SATA II (up to 15 drives per cable)and which consumer level products support port-multipliers?
it's an optional part of the spec and most don't implement it
if you're willing to do a lot of extra work and hassle and really want offline storage you can fill a bunch of external drives with a virtual filesystem (like truecrypt for instance) and then with them all connected run par to build a par file across all your virtual disk files
disadvantages are numerous, have to be able to connect all disks at once, if you update one little piece of one drive have to recalculate the par file across all of them, etc
Sudder - Sunday, September 9, 2007 - link
most e-sata ports support it (although some controllers like the ones using JMB36X just support "Command-based Switching", e.g. all sollutions with Sil 3132 chips (e.g. many notebook S-ata - PCMCIA adapters) even support "FSI-based Switching" which works a little like SCSI (the command is sent to one disc and then the bus is free again, so you can get "close" to the theoretical 300 MB/sec with multiple discs and the bus is not blocked by one working disk (with 4 discs connected, a test showed still 40MB/sec transfer from each of the 4 parallel working discs ..)
so take AFAIKR one of the many new gigabyte mo-boards with 2 ports that can be used as e-sata, put e.g. a "Dawicontrol DC-6510 PM" on the other end of the cable (one is about 100 bucks) add a powersupply and a housing and you are good for 10 extra discs ..
look at more recent motherboards (e-sata slowly shows up on more and more boards) and you'll find that it's supported more and more
well, I'm kind of too lazy to to the par thing each and every time I just change one little file - or to be more practical, I can verry much immagine myself pushing the rebuild further and further into the future as long as I can forsee that I will add more stuff in the verry near future which then again will require a rebuild .. (if I don't find a usable sollution which does it "on the fly" I'll probably end up doing it "by hand" (evetually by adding a small RAID 5 "file-buffer" to my System to strech the write/par Intervalls) but I _really_ would prefer an automatic solution without a most likely multi-hour rebuild process (reading all discs, calculating and writing the hole par-disk) after each little change ..)
Witling - Saturday, September 8, 2007 - link
Something I don't usually see in articles on Raid is the complete lack of protection from failure due to a virus or installation of a bad driver. Both disks get corrupted.I am a home user of Raid 1 through a controller built in to the motherboard using a popular Redmond Washington operating system.
Dave Robinet - Saturday, September 8, 2007 - link
Yep - I touched briefly on this in the last part of the article.Users need to look closely at if an ARCHIVAL system (tape, etc) is better for their needs than RAID 1.
Let's face it - RAID 1 is for "(almost) always on / critically needed to be working when powered up" configurations ONLY. How many home computers fall into this category... really?
kobymu - Saturday, September 8, 2007 - link
In certain cases RAID 1 will give you better read performance.