Seagate Hits 1 Terabit Per Square Inch

MrSeb was one of several readers to submit news that drive manufacturer Seagate has announced (and demoed) the first hard drive to squeeze a terabit into each square inch of platter. "'Initially this will result in 6TB 3.5-inch desktop drives and 2TB 2.5-inch laptop drives, but eventually Seagate is promising up to 60TB and 20TB respectively. To achieve such a huge leap in density, Seagate had to use a technology called heat-assisted magnetic recording (HAMR). Basically, the main issue that governs hard drive density is the size of each magnetic 'bit.' These can only be made so small until the magnetism of nearby bits affects them. With HAMR, 'high density' magnetic compounds that can withstand further miniaturization are used. The only problem is that these materials, such as iron platinum alloy, are more stubborn when it comes to writing data — but if you heat it first, that problem goes away. With HAMR, Seagate has strapped a laser to the hard drive head; when it wants to write data, the laser turns on. Reading data is still done conventionally, without the laser. In theory, HAMR should allow for areal densities up to 10 terabits per square inch (magnetic sites that are just 1nm long!), and thus desktop hard drives in the 60TB range."

Re:Wondering

[fuzzyfuzzyfungus]

Depends on your definition of 'current'; but it shouldn't be an issue(density strictly speaking, isn't even meaningfully visible to the motherboard, except in the broad terms that denser=greater capacity from whatever number of platters is viable).

That said, there are probably still a large number of motherboards that will be questionably bootable from the greater-than-2-terabyte drives that these platters are presumably intended for(some ghastly MBR thing); but anything new enough for 48-bit LBA and a modern OS should, at least, support perfectly normal OS use of the drive once everything is booted.

Re:Wondering

[fuzzyfuzzyfungus]

Also wondering, will this set back SSD by 5 years?

Probably not: This advance(while definitely helpful to the HDD, and no doubt some very impressive engineering work from the R&D team) is a reinforcement of exactly the same virtues that HDDs have historically had and of virtually no value in addressing their historical weaknesses:

1. Capacity/dollar: Once the production is tooled up, the cost/gigabyte for HDDs can be expected to continue to decline.
2. Linear read/write speed: Because of their high areal density and fairly swift rotation, HDDs can read or write like a bat out of hell as long as they don't have to do much seeking. Seeky or random I/O tanks them because of the need to physically move the head around and possibly wait the better part of a platter rotation for the spot you want.

It will continue to be the case that HDDs are cheap for the capacity, and fast as hell for nice, linear, streaming operations; but SSDs can churn out the random I/O without breaking a sweat and are available in physically smaller and more shock-resistant packages(the economical range for HDDs is basically defined in multiples of the volume of a 2.5inch HDD, and don't drop them, SSDs start at BGAs the size of your fingernail and scale in multiples of those until your wallet explodes...

Re:Wondering

[fuzzyfuzzyfungus]

It's a legacy thing, not an intentional-crippling thing:

The BIOS' handling of block devices dates back to when booting your OS off a floppy wasn't considered deviant behavior, and a 5MB HDD was some pretty serious gear. The details are kind of messy [wikipedia.org]...

Most reasonably contemporary stuff should at least do 48-bit LBA; but there are still a lot of systems in the wild that still need MBR, at least on the boot disk(which limits you to 2TB partitions).

Re:100% shark jokes

[fuzzyfuzzyfungus]

To the best of my understanding, there is one major difference: the magneto-optical drives(minidisc and others) used the laser to do reads as well as to heat sectors to lower their coercivity so that the magnetic head could rewrite them. This HDD-derived technology does magnetic reads; but incorporates a laser for heating during writes, allowing you to use high-coercivity materials(which allow smaller sectors to remain stable over time; but would be prohibitive to rewrite in their normal state).

Re:Downsides?

[Anonymous Coward]

The media heats and cools in 100-200 picoseconds; the laser turns on and off much faster than that. No addition to latency. Laser lifetime and reliability will be an engineering hurdle, but not a showstopper by the time a production drive is approved for release. The spot size of the laser on the media is much less than 100 x 100 nm (probably less than 50 x 50 nm) so the total heat added to the drive from the laser light itself is quite small. More heat will be added from the electronics, so thermal management of the drive environment my be more critical. However, caveats to all of these statements are that this is an early demo, not a production-ready drive, and in fact is likely not actually a real HDD like you would put in your PC. These demos are done in a lab environment with lab electronics, and lab mechanical systems to stay on-track. Still, this is a very significant step in showing that HAMR is on track for product plans later this decade.

Re:Wondering

[petermgreen]

Computers tend to measure things with fixed size binary numbers since these are by far the most efficient format for them to handle and process. When chosing the size of these numbers there is always a compromise between efficiency and future proofing. Usually the margin left by the designers is enough to last a while.

However for long lived standards as the years pass that margin is eaten up. Eventually it reaches the point where all the margin is eaten up and things have to be redesigned . The most recent one we hit was that the conventional MBR partition table has a limit of 2^32 sectors (=2TiB assuming standard size sectors). Making things worse is the fact that MS refuses to support the combination of a GPT partition table on the boot drive with conventional BIOS booting so the motherboard may be able to see and access the large drive but it if doesn't support UEFI you can't use the whole drive as a windows boot drive.

Afaict the next barrier we will hit is the LBA48 limit of 2^48 sectors (=128PiB assuming standard size sectors). So a 60TB drive shouldn't be any more problematic than a 3TB one.

Re:Wondering

[qubezz]

As areal density increases on hard drives, so does the transfer rate. The linear density of a track increases, and the amount of data that passes under the head in one rotation of the disc increases. This is how the 5400rpm discs of today have 120MB/s transfer rates compared to the 10MB/s transfer rates of the same rotation speed ten years ago.

Imagining some system you don't own and benchmarks that exist only in your head is not a practical measure of what consumers will own in the future, and rotational media will continue to occupy the same place it does now for the next several years, as the mainstream consumer PC storage product, and as the main data (blu-ray rip) storage and backup media for enthusiasts with SSD operating system drives.

My next system will have a killer refresh rate with a P6 chip. Triple the speed of the Pentium. RISC architecture is gonna change everything. That's too much machine for you.

Re:Wondering

[BattleApple]

You go right ahead and wait for 2 fucking hours for your 50 GB Bluray image to be copied/processed on your mechanical toaster; I'm sticking with my 1 minute with complete silence and low power consumption.

Just in case you've been wondering - this is why you are still a virgin.

Re:Cost? Reliability? Water Resistance?

[rtaylor]

Speed on spinning disk is a function of density as much as anything else.

The tighter you pack the bits, the more bits pass under the head in a given time frame, which makes it faster.