Another day goes by and we have another story about hard drive technology. Yesterday, Western Digital held a media event (I wasn’t invited) to announce MAMR, the latest technology for improving hard drive density. MAMR stands for Microwave Assisted Magnetic Recording and is a process to increase the stability of recording with very small grain sizes and thus high densities. (as a background see this recent post where I talk about some of the technologies).
Microwave Assisted Magnetic Recording isn’t a new technology. Check out this presentation at IDEMA in 2007, which talks about the technology and provides some of the physics around how it works. The science looks complicated but can be summarised quite easily. As we reduce the grain size on a hard drive platter, the ability to accurately set the magnetic state of each grain becomes much harder. Smaller grains require a smaller write head, which causes problems in providing enough energy to the head to set the magnetic state in a stable manner. MAMR reduces the coercivity (the resistance to being magnetised) of the grains and so allows smaller grains to be written much more reliably.
You can find much more information on MAMR and the recording process in the further reading at the end of this post. However, the benefit of using MAMR over other techniques like HAMR is the reduced amount of energy needed to set state, plus reliability of the components themselves, which can be manufactured in existing factories. It seems that the current announcement talks about Western Digital’s ability to reliably implement MAMR using a spin torque oscillator within the write head.
The 100TB HDD
Western Digital is claiming MAMR could get platter densities to 4 terabits/in², compared to 1.3 terabits/in² with SMR. This translates into drive capacities of 40TB by 2025. However, Seagate was claiming HAMR could reach 5 terabits/in², so the two technologies seem on a par with each other. I don’t think WD is claiming better scalability with MAMR, but rather that the technology is more reliable and easier to bring to market. This may simply mean we see WD continuing to beat Seagate on the “world’s highest capacity drive” metric for some years to come.
The Architect’s View
Good, so doing the micro/macro view thing, at the micro-level we see another roadblock being removed in the inexorable push for larger drive capacities. Hard drives get to live another day. At the macro level, $/GB ratios will continue to drop and hard drives will increasingly be for archive purposes because the transfer rates won’t get proportionally better. Why? Because we scale capacity in two dimensions but read/write in one (circular) dimension and rotational speed is constant.
This problem makes me wonder whether the price reductions we will see with these drives are really worth the benefit. If data can’t be transferred easily on/off a drive, will we end up leaving redundant data on there, because it’s too time-consuming to re-use the space? Remember the “Disks for Datacenters” and SNIA report (see this post from last year)? Perhaps it would be nice to see some progress from the HDD vendors on how they are addressing these requirements, rather than the quest for larger capacities. We will have to wait and see.
- Dude, Where’s My 100TB Hard Drive?
- Unreliable Disks for Better Scale-out Storage
- Microwave Assisted Magnetic Recording (IDEMA PPT, retrieved 12 October 2017)
- MAMR Breakthrough Exceeds Expectations (WD Innovation website, retrieved 12 October 2017)
- The Need for Energy-assisted Recording Technology (WD Innovation website, retrieved 12 October 2017)
- Western Digital Unveils Next-Generation Technology to Preserve and Access The Next Decade of Big Data (WD Press Release, retrieved 12 October 2017)
Comments are always welcome; please read our Comments Policy. If you have any related links of interest, please feel free to add them as a comment for consideration.
Copyright (c) 2007-2020 – Post #A6D8 – Brookend Ltd, first published on https://blog.architecting.it, do not reproduce without permission.