HDD Industry Trends - Part Two: To the Future

Posted by Mark Teter, Chief Technology Officer
March 26, 2013

In our last blog post, we shared <some history of HDD>, which brings us up to the present. This next post shares some insights into the future direction of HDD.

To improve recording density in the future, the HDD industry (mainly Seagate) has been developing Thermal Assisted Magnetic Recording (TAMR), also called Heat Assisted Magnetic Recording (HAMR). However, this will be the first integration of optical and magnetic heads. This technology induces magnetization reversal using laser-emitted thermal energy to balance the trade-off between ensuring thermal stability and the difficulties in writing to magnetic disks. In these processes, data is written to a magnetic disk by focusing a laser beam on very small spots in order to heat only very specific, small sections of the disk.

Microwaves are considered to be another supplementary measure and are applied in Microwave Assisted Magnetic Recording (MAMR) technology. Here, the spin-torque oscillator—an ultra-small microwave generator—is used to send microwaves to tiny sections of a magnetic disk, eventually lowering the required amount of magnetic energy necessary to write data. In theory, this technology can also achieve a magnetic recording density of 3 Tbit/in2. (SMR, HAMR & BPM = 5Tb/in2; PRM = 1TB/in2, which is reaching its super paramagnetic limit due to constraints of grain size and its already ½ media thickness).

Meanwhile, HGST (Western Digital subsidiary) is approaching the same problem with a different solution—filling a drive with low-friction helium gas and adding more platters so that, for example, a 4TB 4-platter drive becomes a 6-7 platter drive in the same enclosure with capacity ranging between 6TB and 7TB, assuming 1TB/platter technology. HGST could well start small, so to speak, and introduce a 5-platter 5TB drive. SMR drives take longer to rewrite data because tracks overlapped by the new data track have to be reconstructed before the new data is laid down, thus creating a rewrite time penalty. Helium gas-filled drives won't suffer from this disadvantage.

HDD will be eased out of traditional markets, but the “bulk” data storage market will become immense and ruled by HDD.

Flash, on the other hand (called ‘flash’ by its inventor at Toshiba who thought that the erasure process of the memory contents was like the flash of a camera), we can get a density of 10TB per rack unit giving us petabytes-scale SSD deployments. But SLC, with only 1 bit per single level cell, is very expensive storage.  MLC has two (or three) bits per cell and can easily halve the cost of SLC NAND storage. The tradeoff of 10x the endurance for 2x the price led most enterprise applications to adopt SLC.

Also ,the consumer market which currently drives NAND manufacturing depends on MLC technology for cameras, video recorders, and USB sticks. As such, MLC volumes are significantly higher than SLC and hence, the cost of manufacturing MLC parts is considerably cheaper. Consequently, the industry is using MLC and leveraging advanced controller functionality to increase its durability. (With flash, it’s especially important to isolates writes—to avoid this asymmetrical issue—separate reads to flash, and writes to DRAM.)

However, the rub with flash is slow random writes speeds. There’s no rewrite capability—it must erase before a write, and erase is very slow as it’s done on large blocks. In fact, as the number of writes increases, the read performance decreases. Writes are problematic because for a flash based storage device to write data, it must first erase the old data in the location to which it is about to write. Unlike a hard disk, when flash erases data, it does so by writing zeros to that cell (and it might first have to copy the old data to another cell in the interim). This is known as write amplification. The effect is that, after a flash device fills up, the number of writes per write is doubled.

As a result, there are seven ways to introduce flash into our traditional HDD architecture:

  1. SSDs Slotted into HDD Slots
  2. Flash in Array Controller Cache
  3. Newly Architected Arrays using Flash/DRAM
  4. All-Flash Arrays
  5. SSDs Slotted into Server
  6. Flash Used Strictly as Cache, and
  7. Storage Appliances (VSA).

Stay tuned for more details on these new storage architectures.

About Mark Teter Before he retired from ASG in 2013, Mark Teter was Chief Technology Officer (CTO) and the author of 'Paradigm Shift: Seven Keys of Highly successful Linux and Open Source Adoptions.' As CTO, Mark regularly advised IT organizations, vendors, and government agencies, and he frequently conducted seminars and training programs.

Filed Under: Data Storage

0 Responses to 'HDD Industry Trends - Part Two: To the Future'

Leave a Comment

Please copy "3eB9OOhWZ1ZgWPDn9j29MjWXLIyjZFu0" into the field labeled "Uncaptcha"