Gah - I thought that meant a return of the...
It's back to the future for the latest Seagate large form factor disk with a low capacity. The company has introduced a 3.5-inch 1TB and 2TB disk drive spinning at 7,200rpm with a slow 6Gbps SATA interface, and called it an Enterprise Capacity drive. Remarkably, other 3.5-inch Enterprise Capacity drives of Seagate's go up to …
In which application is a 3.25" 2TB more desirable than a 4TB or 6TB or... ? Even the examples given in the article don't meet justification, especially considering Seagate has 8TB models for each scenario for a reason. A 2.5" would of made this all make sense (smallish, but cheaper and better cap than SSD).
Unless these are 50usd or less, it appears Seagate is trying to create a market for its old stock before SSDs catch up.
Off to the side, who decided 55TB to be personal usage yet magically 10 * 55TB = enterprise? I missedthe case study on this, but somehow this is presented as factual by many manufacturers.
ISTM that the biggest difference from what you would buy for you home PC is:
"Engineered for 24x7 workloads of 550 TB/yr" In many cases that would be far more important than having extra capacity or faster speeds.
I note the article wonders why it is using a 6Gbps interface rather than a 12Gbps interface. I wonder the opposite - why fit any mechanical HDD with a 12GB/s interface? The drive itself has a maximum sustained transfer rate of barely 1Gbps, and no mechanical HDD I have come across can get anywhere close to 6Gbps read or write speeds (except for short bursts to & from its RAM cache). Increasing the interface speed increases the probability of data corruption in transit and needs more expensive cables but AFAICS gives almost no advantage whatsoever.
6gbs because they're trying to make a cheap drive, SAS wouldn't be available in typical applications for this. As far as 1gbs, you might be thinking of ATA133. Which yes, 1gb was the rough limit. SCSI on the other hand...
Reliable 24x7 means what when there is also a reliable 24x7 4TB? This product seems to make sense when I read about it, but when I think about the given applications and actually implementing them, why 2TB at 3.5? Again, I think it is marketing jazz.
As far as more expensive cables, actually the opposite. A SATA cable today still costs me less than a ATA133 ribbon did 15 years ago, and SCSI cables are significantly cheaper and managable with SAS.
As far as the actual mechanical limit of a platter, I don't know. It'd be interesting to see a video of this taken solely for this determination. I have heard a 12,000rpm drive give its all, so it would be VERY interesting to hear someone force a drive to something like 24,000. Again, this would need a video to hear it as it does its best.
"RAID rebuild?" - OK, but that can be said about form factor or capacity.
> 6gbs because they're trying to make a cheap drive, SAS wouldn't be available in typical applications for this. As far as 1gbs, you might be thinking of ATA133.
No, he's thinking of the raw sustained data rate from the platter as the disk spins under the head.
If the drive has a sustained data throughput of 120MB/sec (which it might do when the head is at the outer edge, not the inner) that's about 1Gbps.
In the SMB space it's common to require cheap space & IOPS. One way to achieve this is to setup one big RAID 10 array with as many disks as you can. I run several of these setups and they work a treat.
Classic example: Dell R510, dual CPU, 64GB RAM, 12x2TB OBR10. Cheap, reliable & very capable of hosting a great many small VM's plus one large file system VM.
There is a draw back to these massive drives that are available today. When used in a raid system, and you lose an 8tb drive, it can take an extremely long time to fully recover after replacing the dead drive. And if you're running a less redundant flavour of raid then you're vulnerable until the system finishes repopulating the new drive.
Ahaha. I remember the Bigfoots. Cheap and cheerful.
Nothing can replace my old full height ST423451W. I keep it around to scare kids and win arguments.
I've had a few experiences getting Seagate to RMA a drive that failed in hardware RAID but passed their SeaTools diagnostics. This new drive would have to be pretty cheap for me to consider it.
I remember those Fujitsu Eagles! I worked on them. However, they weren't as loud as the old Maxtor 5-1/4" drives. Remember those spinning up and initializing? weeeeEEEEEP! ... weeEEP! ... weEP!weEP!weEP!weEP!weEP!weEP!weEP! You could hear that initialization from across the room with all of the HVAC blowing, it was so loud.
That Eagle thingy looks rather dainty to me.
Have a look at this beauty:
I mean the cabinet on the far left. No, that is not a washing machine (although the amount of energy being converted while spinning is probably comparable, at least it feels that way).
As others have stated towards, this makes sense in a large commodity play, scaled out environment, where you are looking for a capacity that can be rebuilt in RAID in reasonable timeframes, gives reasonable performance, and is priced at a replace-and-bin level.
Large cloud vendors would eat these things up at a sufficient pace to make it worthwhile for Seagate.
How about firmware that can not be hacked by the suite of malware that's been evolving over the years that spreads over Windows, Linux & BSD and has had aspects of it identified by experts that have called their observations, BadBios, BadUSB, Stuxnet, Duqu, Flame & most recently Shamoon?
Its only a simple ask for hardware than needs a physical intervention to short some pins or jumpers of graphics cards, CMOS chips, usb connected devices, hard drives, printers and more, to flash the firmware whilst employing bug free code.
If we cant have that, then enjoy your repression if you are even clever enough to spot it, as you go about deferring your judgement in all aspects of your life to so called experts.