Updated 2/4/13
One of the resources at HP (thanks Ben!) made the following comment to one of our customers and I thought it’d be a perfect post for the blog as it contains some useful information that some might not be aware of.
Here are the high-level differences between SAS and SATA disk drives:
Capacity:
- SATA (or now called NL-SAS for Nearline SAS) disk drives are the largest on the market. The largest SATA/NL-SAS drives available with widespread distribution today are 3TB.
- SAS disk drives are typically smaller than SATA. The largest SAS drives available with widespread distribution today are 600GB or 900GB.
- So, for capacity, a SATA/NL-SAS disk drive is 4X-5x as dense for capacity than SAS.
- A good way to quantify capacity comparison is $/GB. SATA will have best $/GB.
Performance:
- SATA/NL-SAS disk drives spin at 7.2k RPMs. Average seek time on SATA/NL-SAS is 9.5msec. Raw Disk IOPS (IOs per second) are 106.
- SAS disk drives spin at 15k RPMs. Average seek time on SAS is 3.5msec. Raw Disk IOPS (IOs per second) are 294.
- So, for performance, a SAS hard drive is nearly 3X as fast as SATA.
- A good way to quantify performance comparison is $/IOP. SAS will have best $/IOP.
Reliability: there are two reliability measures – MTBF and BER.
- MTBF is mean time between failure. MTBF is a statistical measure of drive reliability.
- BER is Bit Error Rate. BER is a measure of read error rates for disk drives.
- SATA/NL-SAS drives have a MTBF of 1.2 million hours. SAS drives have a MTBF of 1.6 million hours. SAS drives are more reliable than SATA when looking at MTBF.
- SATA drives have a BER of 1 read error in 10^15 bits read. SAS drives have a BER of 1 read error in 10^16 bits read. SAS drives are 10x more reliable for read errors. Keep in mind a read error is data loss without other mechanisms (RAID or Network RAID) in place to recover the data.
Here are some good links for comparing disk types:
http://www.seagate.com/docs/pdf/datasheet/disc/ds_barracuda_es_2.pdf
http://www.seagate.com/docs/pdf/datasheet/disc/ds_cheetah_15k_7.pdf
http://h18004.www1.hp.com/products/servers/proliantstorage/drives-enclosures/index.html
http://storagebuddhist.wordpress.com/2011/07/01/nearline-sas-who-dares-wins/
http://www.seagate.com/docs/pdf/whitepaper/tp_sas_benefits_to_tier_2_storage.pdf
http://www.tomsitpro.com/articles/seagate-serial-attached-scsi-disk-drive-data-storage,2-119.html
http://en.wikipedia.org/wiki/Serial_attached_SCSI#Nearline_SAS
Lewan IT Solutions Technical Blog 
Pingback: TECHNOLOGY - PAX ‘09: For the Gamers, By the Gamers - TechBlog
thanks for this very good comparison
Pingback: Our Blog – 2010 in Review «
Pingback: Nas for SMB users: an opportunity for the channel. | B2Bchannelblog.com
There are models of SAS drive that are 1 or 2TB and spin at 7200 RPMs.
As far as I know the difference between the 7200RPM SAS drives and the 7200RPM SATA drives is just the SAS/SATA interface.
Suppose you have two disk drives; identical capacity and RPMs…. one is SAS, one is SATA; I wonder if either is technologically better / how much better?
sas and sata are not identical due to some reasons
1- s a t a is serial ATA but S A S is serial attached SCSI
2- capacity and rpm performance is not the issue absolutly, but the actual difference between sata and sas is barrier limit that sas breaks
as sata breaks the drive limit of an ide interface that is two(2)
similarly sas breaks the drive limit of SCSI interface that is seven(7) with some LUN (logic unit number) sas breaks that limit barrier we can use almost 26000 drive units in an enterprise server enviornment
it’d also be good to talk about many sata drives are now marketed as ‘nearline sas’ or ‘midline sas’ where the sata protocol and drive characteristics are used but they have the same 6G/sec bandwidth available as the sas drive.
Good thing most SANs have hot spares! If the difference in BER and/or MTBF made a difference in real world business, I would think there would be a lot more businesses complaining about data loss. Putting some perspective on the that 1.2 million hours for SATA MTBF, that’s 50000 days or 137 years.
Do you have any idea about which of these are more in demand and is more scalable to meet increasing data storage requirements
@Vmax Hot spares may feel good but be a bad idea. Pulling numbers together for SAS but for SATA there’s abot a 10^14 URE level. So one in 200 million sectors will yield a URE. Given a 7 drive 2TB SATA array in RAID 5 you get 6TB of usable data space. Roughly 200 million sectors of data. So when you lose a drive in the array, what are the odds of having a URE in those 200 million sectors?
sas and sata are not identical due to some reasons
1- s a t a is serial ATA but S A S is serial attached SCSI
2- capacity and rpm performance is not the issue absolutly, but the actual difference between sata and sas is barrier limit that sas breaks
as sata breaks the drive limit of an ide interface that is two(2)
similarly sas breaks the drive limit of SCSI interface that is seven(7) with some LUN (logic unit number) sas breaks that limit barrier we can use almost 26000 drive units in an enterprise server enviornment
The difference is in the interface rather than the density. Here’s a detailed SAS vs. NL-SAS vs. SATA comparison http://www.techrepublic.com/blog/networking/how-sas-near-line-nl-sas-and-sata-disks-compare/5323
Pingback: Sas Costs | MPLS