The Future of Enterprise Drives

By | March 16, 2011

Strangely enough I found Seagate’s recent line-up of enterprise drives very interesting.  To me, it was significant in emphasising and playing a part in two important trends –

  1. The move to 2.5-inch form factor drives (often referred to as SFF)
  2. The slow death of the 15K drive

Both of the above points were highlighted as part of the announcement regarding the Savvio® range of drives.

The rise of the 2.5”

The Savvio® range now has a 2.5-inch 900GB 10K drive with a 6Gbps SAS interface.  That’s impressive.  At the time of writing this is the highest capacity 2.5-inch 10K drive on the market.

It seems like only yesterday when 1TB 7.2K 3.5-inch drives were considered to be huge (and slow).  Oh wait a minute, it was only yesterday 😉

With the above in mind I can’t help but get a little excited that we are at, or extremely close to, an inflection point where 2.5-inch drives surpass 3.5-inch for performance, power, and now capacity.  Seagate are leading the charge.

Annoyingly, the one area where 2.5-inch drives still lag behind their 3.5-inch cousins is at the bottom end, the 7.2K SATAII type drives.  Even the Seagate announcement of their Constellation® drives are still based on the ugly and againg 3.5-inch form factor.  Bit of a bummer really, because until they convert to 2.5-inch form factor we will have to maintain an portion of 3.5-inch form factor shelves in our storage arrays – nobody wants that, not customers or vendors.

Oh and lets be honest, once you've clapped your eyes on the 2.5-inch form factor, 3.5-inch drives just look soooo yesterday –

sff drives

The slow and painful death of the 15K drive

The other interesting point is that while the Savvio® 10K drive has a whopping 900GB capacity, its sister, the 15K drive only sports a miserly 300GB 🙁

I’ve argued for a while now that the value of the 15K drive (with the additional R&D requirements for the manufacturers) has been eroded by the rise of SSD.  The performance uplift from a 10K to 15K drive is insignificant compared to the uplift from 10K to SSD.  So I argue there is no need longer a need for the 15K drive!

Think about it….. the number 3 seems to be the magic number when it comes to storage tiering.  For most people storage tiering looks, or will look, like the following –

Tier 1 – SSD

Tier 2 – 10K SAS

Tier 3 – 7.2K NL-SAS

I’m sure that the Barry Burkes of this world would love us to think we can get away with just two tiers, SSD and SATA, or Flash and Trash as some call it, but it just doesn’t seem like we’re quite there yet.

If you don’t think that 15K drives are harder to develop than 10K, look at the fact that the 10K drive has 3 platters and the 15K only 2 platters……

Are Western Digital Up For the Fight

In my previous post I posed the question as to whether or not Western Digital (who recently announced the purchase of HGST – the Hitachi disk drive division) would be up for a battle with Seagate in the enterprise space. 

By “enterprise space” I mean the 10K and 15K drive space.

You see, there are undertones in some of the talk around the recent Seagate drive announcement that smack of the type of things WD may not want to get involved in.  For example, there was talk around Seagate having the bragging rights when it comes to areal density (506Gb per square inch in the 900GB Savvio® for those like me who care – note that this is denser than the 3TB Constellation® they announced).

This type of talk smacks of the vendor cold wars, where EMC edged ahead of Hitachi, only for Hitachi to then edge ahead of EMC, only for EMC to….  Think Intel vs AMD.

Do WD want to get involved in that?  Especially when SSD and NL-SAS are destined to win!  And win at the expense of 10K and 15K.

It seems like the EMC’s and Hitachi’s of this world think that the future of enterprise disk arrays is Flash and Trash, then it is probably safe to assume that they are building the architectures of the future for a world where 10K and 15K drives no longer exist.  So why would WD bother investing in 10K and 15K drives when their biggest customers are building a future where they don’t exist?!

Oh and there was talk of the 3TB Constellation(r) having increased RAID rebuild capabilities, but so far nobody from Seagate has answered my questions as to what that amounts to.

Just my thoughts, courteous comments very welcome.

You can debate this and other techie stuff with me on Twitter @nigelpoulton

One thought on “The Future of Enterprise Drives

  1. Steven Pemberton

    To quote Karl (the Marxist storage admin) – "From each according to his ability, to each according to his need".
    It seems inevitable that lower cost 7k SATA drives will provide the bulk capacity, and SSD (and/or Flash + cache) will provide the performance. Also, wide-striping across many SATA spindles tends to compensate for lower individual drive performance (vs. traditional array layouts).
    In the medium-long term, I don't see how an intermediate tier of 15k drives can be justified given that their IOPS performance is so "SATA-like" when compared to SSD/Flash.

Leave a Reply

Your email address will not be published. Required fields are marked *


*

You can add images to your comment by clicking here.