Let’s not candy-coat things. Modern society faces a serious and increasingly complex challenge: the insatiable, darn near Atshenesque, demand to create and store vast amounts of data. This issue has propelled the resurgence of once considered outdated ideas like “nearline” storage solutions back into the mainstream. In this corner of the market, Seagate’s Exos line has been consistently delivering capacity, reliability, and value options tailormade to enterprise demands… and thus has been the linchpin of many a data center architect’s “grand design” for many years now. This year, however, is somewhat different than previous years. Instead of releasing a refreshed Exos X with a moderate ~2TB total increase in capacity, Seagate has decided it is time to push the boundaries of what traditional HDDs can do. Yes, Seagate has finally deemed its HAMR technology ready for primetime. Thus, the release of the all-new Exos M 30TB. With 30TB of space and an average online asking price of ~600USD (or ~2 cents per GB… aka what the Exos X24TB price per GB on launch day was), this new Exos “M” 30TB drive marks a pivotal moment in Seagate’s illustrious history. One that will either retroactively be seen as the savior of the industry or the architect of its demise.
Without a crystal ball, all we can say is… Yes! Finally! After many years of delays, Seagate’s take on Heat-Assisted Magnetic Recording (HAMR) technology is now widely available. As the name suggests, this long-awaited innovation uses a nanophotonic laser to heat a small portion of the disk’s surface (~30nm) during writing, allowing data to be packed into much smaller, yet stable, magnetic grains (5’ish nm). This breaks through the physical limits of traditional magnetic recording and allows for a Great Leap Forward from 2.4TB per platter to 3TB in one single generation. Although it represents a whopping generational 25% aerial density increase, to non-storage enthusiasts, an extra 6TB per drive may not sound all that impressive. Especially given the nosebleed asking price. However, in the coming years, HAMR will quickly scale well beyond previous CMR limitations to 40… 50… even 60TB capacity options. All the while staying within the familiar 3.5-inch form-factor and the associated limitations of said form-factor.
All of which is a siren song tailor-made for Enterprise clients… as this directly translates into great rack density – aka more data stored in the same amount of rack space. Which in turn helps reduce the total cost of ownership by minimizing floor space, power consumption, and cooling demands per terabyte stored – a mere 0.32watts per TB. This, however, is just the start. Backstopping this new tech is a new Read Write head configuration that builds upon previous Exos X designs to ensure data integrity and reliability. In the coming pages, we will go over it, but in a nutshell, these new heads may share a similar ancestry with the ‘X’, but are entirely new beasts with some rather tasty tech baked into them.
In the meantime, that new r/w head tech is backstopped by a new Iron/platinum “super lattice” platter technology, which is uniquely designed to handle the insane stresses HAMR will place on it. We are not joking when we say that the engineering problems of creating new R/W heads combined with needing to invent a new ‘super lattice’ platter tech were the major contributing factors to why HAMR(time) was delayed… and delayed… and delayed. So delayed we thought we would not only get GTA 6 first… but MC Hammer would have shuffled off into the next life long before Seagate’s HAMR was released.
Thankfully, Seagate’s software (aka firmware) team was not idle during these unforeseen delays. Instead of idly cashing a paycheck while the hardware engineers sweated, the software engineers actually took the time to create an entirely new MTC (multi-tier caching) variant that builds upon all previous MTC options… such as “borrowing” the subdivision of said cache from SKYHAWK and working it into meeting the needs of Enterprise consumers. Not satisfied with just creating an entirely new subcategory of MTC, they also took the time to tightly integrate it with the HAMR media’s unique characteristics (thermal, performance, response time, etc.). Thus, creating the most sophisticated firmware any mere HDD has ever used. One where a complex internal dance of cache flushing and data migration constantly takes place so as to maximize throughput and minimize heat-related stress being placed upon the new iron/platinum lattice.
All of which sounds great. On paper. However… Enterprise clients are extremely cynical when it comes to buying into the hype surrounding any labeled as “new and improved”. More than once in the past, “new” has been synonymous with ‘finicky’, ‘unreliable’, and even ‘prima donna’. As such, this new breed of Exos M will have to prove itself a worthy successor to the various Exos X generations in use today. Let’s see if they could indeed pull all that off.
