Analysis Seagate R&D bigwig Jan-Ulrich Thiele says the first Seagate prototype drives built with heat-assisted magnetic recording (HAMR) will arrive in late 2016 and have just 4TB capacity, according to Heise.de.
This is not quite right as what Seagate says is: "Seagate anticipates shipping a small number of HAMR qualification units to select vendors by the end of 2016. These units will be engineering samples meant for evaluation purposes only with commercial units scheduled to follow in the near term. Seagate has not publicly disclosed the capacity of these qualification units; however, we expect that commercial HAMR HDD will keep pace with the highest capacity drives on the market at time of release."
Seagate has just announced 8TB drives using existing perpendicular magnetic recording (PMR) tech. Having the first HAMR drives with the same capacity as PMR drives is disappointing, as HAMR is meant to get us past the PMR wall.
As the vertically-oriented PMR bits (collections of magnetic grains in the recording media) get smaller, their stability suffers as they become more prone to interference from neighboring bits, and read error rates increase. So much so that PMR bits can't get much smaller.
The get-out-of-jail card is to use a magnetic recording medium that is much more resistant to change and neighboring bit interference at normal operating temperatures. To write data bits, the exact right tiny area of the medium must be momentarily heated to 450oC (842oF) to lower its resistance to change, the bit written, and then the area left to cool.
A laser mounted on the read-write head is used for the heating. This makes the head's operation more complex and increases its thickness, meaning that the inter-platter gap increases in size, an unwanted result. So either the platter count gets reduced or the drive enclosure gets thicker, meaning industry-standard disk drive slots need to change to accommodate thicker drives.
It gets worse – if the proto testing completes satisfactorily in 2017, we won't see general HAMR drive availability until 2018. Even then, Seagate will ship them to hyperscale data centers and cloud data centers to apparently stem the tide of anticipated high-capacity flash drives.
It won't have escaped Seagate's attention that Samsung demonstrated a 16TB* SSD using its 3D TLC V-NAND technology at the August Flash Memory Summit. The general enterprise data center and desktop markets won't get the drives until later.
By 2018 we will surely be seeing 20-25TB SSDs if there is a continued increase in 3D NAND layers. Seagate has a heck of an engineering mountain to climb to go from 4TB HAMR proto disks in 2016 to 20-25TB HAMR drives in 2018. Seems an impossible goal.
A side note: if mass availability of HAMR drives isn't going to happen until 2019 or so, we're probably in for a tenth generation of PMR, taking us from gen 9 8TB drives to 10-11TB, if that can be managed.
A second side note: it's blindingly obvious that HGST's helium-filled drive tech is a critical advance, enabling as it does thinner platters, meaning that a HAMR-caused increased inter-platter gap can be countered. That means potential HGST HAMR drives may not need thicker enclosures, and altered disk bays are not needed either.
Seagate is between a rock and a hard place here.
A third side note: with the competitive pressure coming from 3D NAND TLC SSDs, the disk drive industry absolutely needs helium-filled HAMR drives, and probably shingled ones as well.
No helium filling means no HDD future vs 3D SSDs. That's what it looks like from here. ®
* 15TB, actually.