Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Define: Server Grade Hard Drive (re: time Capsule)
- Thread starter Full of Win
- Start date
- Sort by reaction score
Someone is going to need to open it up. I have no clue what Steve was talking about when he said "server grade." 
I guarantee it's just going to be a generic low bid OEM SATA drive.
I guarantee it's just going to be a generic low bid OEM SATA drive.
Someone is going to need to open it up. I have no clue what Steve was talking about when he said "server grade."
I guarantee it's just going to be a generic low bid OEM SATA drive.
Server grade HDDs are generally faster (10k or 15k rpm), they generally have more cache, and they tend to have far better components in them.
Look to SCSI drives, they were always much smaller in capacity, but had much faster access times, etc.
It will be interesting to see what they actually put in there.
What I think the time capsules will have are the enterprise level western digital hard drives. They are reasonably priced and are "enterprise" grade. They have been around for awhile and have the letters "RE2" instead of SE.
Here's a description from WD's website:
http://www.westerndigital.com/en/products/Products.asp?DriveID=238&Language=en
Here's a description from WD's website:
http://www.westerndigital.com/en/products/Products.asp?DriveID=238&Language=en
Both Seagate and Western Digital have drives that are designed to work 24/7 -- they are conventional 7200 RPM SATA or IDE drives, but their controller logic is tweaked to give more reliability and better performance in server-typical tasks, at the cost of a little bit of speed at random access tasks. Some of these drives also scale performance for lower energy consumption / less heat.
An example: usually a drive will transit the heads to a new track as fast as possible (using energy and creating higher wear). Some of these new drives calculate WHEN the next data bits are going to be needed, along with rotational latency, and only accelerate the heads fast enough to make it to the track just in time, saving juice and wear and tear.
Apple is Definitely not talking about 10K RPM or 15K RPM SAS or SCSI drives in the AEBS.
An example: usually a drive will transit the heads to a new track as fast as possible (using energy and creating higher wear). Some of these new drives calculate WHEN the next data bits are going to be needed, along with rotational latency, and only accelerate the heads fast enough to make it to the track just in time, saving juice and wear and tear.
Apple is Definitely not talking about 10K RPM or 15K RPM SAS or SCSI drives in the AEBS.
"Server Grade" would presumably mean "the same drives no server user worth their salt would take in anything less than a RAID5 config"
Nah, it's what the above posts said. There are constant-run drives with longer warranties available for a few dollars extra.
Speed isn't an issue here, as even the cheapest drives' throughput will far outstrip the abilities of a low-end CPU on a low-cost NAS to squirt traffic to and from a Gigabit (let alone wireless) LAN.
I'd be very interested to see a Small Net Builder take on the unit expecially as regards throughput.
Nah, it's what the above posts said. There are constant-run drives with longer warranties available for a few dollars extra.
Speed isn't an issue here, as even the cheapest drives' throughput will far outstrip the abilities of a low-end CPU on a low-cost NAS to squirt traffic to and from a Gigabit (let alone wireless) LAN.
I'd be very interested to see a Small Net Builder take on the unit expecially as regards throughput.
I agree. And Apple probably did that not only for the bragging rights, but also they hope not to replace too many capsule drives (and the HDD maker will pay, as the drives have extended warranties).
Could be Seagate Barracuda ES.2, Hitachi Ultrastar og WD RE2
As the Time Capsule doesn't have a fan a good choice might be the WD RE2-GP, which has significantly lower power consumption (and noise) than the competition.
Speaking of power consumption - does anyone know what it is for the Time Capsule (with disk idle/hibernating and spinning for example) ??
Someone somewhere suggested 30W, which if true would cost me 100$/year
Yeah 'someone somewhere' turns out to be the technical specifications - max.continuous power: 30W - I sure hope it idles a lot lower, but I guess we will have to wait for reviews to know for sure.
Yes, that's a lot. A normal 7200rpm 500GB drive consumes less than 10 watts in use, a little less when idling and <1 watt in stand-by.Thats annoying...30W for a 24/7 device...but i ordered it
For the WD GP drives it's more like 6/4/0.5 W, but no matter which drive is inside it seems the rest of the time capsule uses quite a bit of power - is up to 20W normal for 802.11n access points/routers ??
Hopefully the norm is a lot less!
Someone is going to need to open it up. I have no clue what Steve was talking about when he said "server grade."
I guarantee it's just going to be a generic low bid OEM SATA drive.
Server grade drives typically are designed to tolerate more environmental conditions, such as heat / vibration. Typically have better components that are rated at longer run time.
From what research I've done after my original post the drives are just meant to run in a 24/7 environment and never be turned off. I haven't seen anything special beyond the always on rating a manufacturer stamps on it.
Care to cycle power on your hard drive anyways?
If the drives are truly server grade then they may support a feature known variously as Time Limited Error Recovery or Command Completion Time Limit.
This would mean that the drives do NOT try to repair damaged sectors for 'too long' and rather report the failed sectors to a RAID controller so as not to cause the drive to completely drop offline for not being responsive...
Does the unit that contains the 'server grade' drives support a raid controller?
This would mean that the drives do NOT try to repair damaged sectors for 'too long' and rather report the failed sectors to a RAID controller so as not to cause the drive to completely drop offline for not being responsive...
Does the unit that contains the 'server grade' drives support a raid controller?