Seagate gets hybrid SSD/HDD right

by Robin Harris on Monday, 24 May, 2010

A hybrid of SSD and hard drive that gives you the best of both worlds. That’s the theory anyway. But I won’t keep you in suspense: I think Seagate has hit a home run with their new hybrid XT architecture.

Take a standard issue 7200 rpm, 500 GB notebook SATA drive. Add 4 GB of fast and reliable single-level cell (SLC) flash. Make it look like a standard drive to the OS by keeping all the magic internal.

Give it smarts to learn about high-frequency small-block transfers. Put those blocks in the flash and voilà: super-fast small block access; leaving big sequential I/Os to the disk. The algorithm

The SSD is 2 SLC chips, each with about 75 MB/sec bandwidth. If the flash should fail – more likely than wearing out IMHO – you still have a perfectly good 7200 rpm disk. The algorithm looks to provide the most benefit on the most used apps.

Disk drives offer cheap capacity and good large read/write (R/W) bandwidth. The R/W bandwidth improves over time as drive capacities grow due to higher bit density.

Flash SSDs offer sub-millisecond access times at a high price: $2-$3/GB. SSD access times are about 150x faster than a notebook hard drive, while R/W bandwidth is only about 2x faster.

Maximum bang for the buck? SSD to store many small files – like DLLs in Windows – to minimize accesses, and let the disk handle the large R/W traffic. Why pay 40x for a 2x performance boost?

Flash capacity? 4 GB can store 4 million 1k files. That saves a lot of access time.

An early 2009 17″ unibody MacBook Pro with 4 GB RAM and 2.66 GHz Core Duo 2. Swapped the current 7200 rpm, 500 GB Hitachi for a 7200 rpm, 500 GB Momentus XT with 4 GB SLC flash.

For the first test I formatted it with HFS+ using the Mac’s Disk Utility. Then I used the excellent Carbon Copy Cloner to clone the system disk to the XT.

Boot performance sucked compared to the old drive’s ~45 second boot:

  • First boot took 2 minutes 4 seconds
  • Second took 1 minute 19 seconds
  • Third took 1 minute 6 seconds

After consulting with Seagate I did a clean install of OS X. Reformatted the XT, installed Mac OS 10.6 Snow Leopard, migrated accounts and data. Took 3 hours, but the process worked perfectly.

New boot times went to ~45 seconds again. Not what I had hoped – the SSD on my MacBook Air booted in 15-20 seconds – but competitive. Other testers have had better results so Seagate is sending me another drive. I’ll update the results later this week.

Application results were much more impressive. Because the drive learns, the first time you bring up an app it happens at disk speed. But the 2nd time!

Mail startup went from 5 seconds to 1.5. Microsoft Word startup went from about 12 seconds to less than 3. See for yourself in this 15 second clip:

And the bigger the app, the bigger the potential speed up. Look at – or not – this 1 minute clip of the 1st and 2nd time opening Final Cut Pro, the Mac pro-level video editing app:

From 45 seconds to 10 17? Way faster than my Mac Pro’s 10k system disk.

Note that rarely opened apps won’t do as well as the algorithm favors apps that get used more often. Pure SSDs don’t have that variability.

Announcement pricing:

  • 500 GB/$156
  • 320 GB/$122
  • 250 GB/$113

These prices are about double the street prices for the non-hybrid drives. Once Seagate gets some competition they should drop.

The StorageMojo take
Looks like a home run for Seagate. The architecture is clean, the performance advantages are real, and pricing is not too bad – especially compared to SSDs.

SSD adoption has stalled because flash prices have firmed up. Few are willing to pay $200-$300 for a smallish SSD in a $700 notebook. But $100? OK.

Seagate could use cheaper MLC flash in 8-16 GB caches without wearing it out. There is every reason to put this in 3.5″ drives as well. After all, if the flash does fail all your data is on the hard drive – no loss there.

By blurring the performance difference between disk and SSD, these drives will ensure that hard drives dominate for at least another decade. And they’ll put pressure on SSD prices.

But don’t count on these showing up in RAID arrays soon. Arrays already have a lot of moving parts and hybrid drives add some subtle wrinkles. Like how do you know what a small block transfer is? Or ensure you don’t have stale data? Not sure cached array controllers need this.

Courteous comments welcome, of course. No money changed hands but I’m not excited about returning the review unit. Also I sent a note off to CCC’s developer to alert him to what I saw – maybe he can figure it out. Update:As a number of commenters pointed out – thanks! – I screwed up: the original video showed a reboot from system cache. When I got the time I reshot the FCP startup after a reboot. It was still 40% of a regular drive and still much faster than a 10k drive on my Mac Pro. Sorry about that! End update.

{ 24 comments… read them below or add one }

Paul Tötterman May 24, 2010 at 6:24 am

Did you empty the disk cache between launches of applications? I don’t doubt that you did, I’d just like for it to be spelled out so that I don’t have to guess.

Mia May 24, 2010 at 6:51 am

Maybe you should mention if you did reboot between the tests or not. Without rebooting it might still be in the systemcache.

Chris Ferebee May 24, 2010 at 7:06 am

Robin, what version of Carbon Copy Cloner did you use? Older versions are not aware of the transparent HFS+ compression introduced with Snow Leopard, and cloning a 10.6 install with them will leave you with an installation that takes up several more GB on disk than the original – something that would probably hit this drive’s caching scheme quite hard.

If it was the current version of CCC, it would be interesting to see whether SuperDuper has the same problem. Again, it would have to be the latest version.

The easiest way to determine whether files are compressed on disk is probably with Amit Singh’ hfsdebug command-line utility, as Snow Leopard hides the compression rather effectively.

hotdog003 May 24, 2010 at 9:53 am

“This hard drive is great because apps start up faster the second time you launch them!” <– YOUR DOING IT WRONG

Of course apps start up faster the second time–they're still in your cache! Next time, reboot between successive launches for a better picture.

Gonzague May 24, 2010 at 1:23 pm

I have the same question as Paul (and I think the answer is no) regarding apps caches.

I *think* the boost in term of performances doesnt just come from the drive but also from Mac OS that caches a lot of thing the first time you launch an app which improves dramatically launch time for any further application … launch.

Matt Spradley May 24, 2010 at 2:23 pm

Windows 7 is optimized for hybrid drives. OSX needs some work in that department from what I have heard.

Anonymous May 24, 2010 at 2:36 pm

Unless your vids are edited, it seems you are measuring cold v.s. hot startup, where Mac OS itself caches most of the application data in memory. As a simple example — I just tried running Firefox. Cold startup took over 10 seconds, whereas hot startup was just under 2 seconds.

Jason Ozolins May 24, 2010 at 6:19 pm

This could have been a much more interesting product if they had gone just a bit further and held back a hundred or so MB as a write accelerator. This would be great for journaling/log-structured filesystems that actually try to give you some guarantee that synchronous writes really made it out to stable storage. If the FS journal is laid out sequentially, the write accelerator could give low latency by aggregating small writes into something more like a streaming workload.

Of course, this would pose a bigger problem w.r.t. wear levelling than their chosen scheme, especially when there is not that much flash to level writes over. Having a tiny supercap that would provide enough emergency power to allow writeout of dirty data from the drive’s RAM-based cache to flash would give some of the same effect, if the drive ignored barrier sync and flush dirty cache commands, and the additional write load would be miniscule.

This is not so big a problem for a laptop because you don’t usually lose power when under a write load – you can just enable write-back caching on the drive, tell your FS not to bother issuing barrier sync and flush cache commands, and cross your fingers that the battery doesn’t die at the wrong time – but it would be a much more interesting device for a desktop system or as a JBOD component.

Kebabbert May 25, 2010 at 6:19 am

ZFS gives this for free. Add a SSD disk as a cache, by typing one command. Done. This is much more flexible, you can swap disks to larger or faster ones in the future, or if the disk krasches. I dont see the point of having a 4GB SSD disk hybrid drive? It is too small, not flexible. ZFS is more tested, and flexible and free. Less expensive than special tailored solutions which are jack of all trades, master of none.

Nathan May 25, 2010 at 8:04 am

How does a hybrid drive compare in terms of power consumption? That would be my biggest concern when using the drive in a laptop.

I’d guess that, depending on your usage, it would offer the same efficiency as a traditional HDD. Better efficiency for light users, worse for heavy multitaskers.

Giorgio Regni May 25, 2010 at 12:35 pm

Finally those hybrids have arrived, it’s about time. This is a 1st generation so expect bugs and other speed limitations but this opens a way to much more efficient storage on a laptop, much cheaper than going full SDD.

It could be useful in the server room too.

Taylor May 25, 2010 at 12:49 pm

I think the tone of this article is borderline ridiculous given that a) your boot tests showed no improvement, and b) your timings for program launches are completely useless because you have no reference data from the non-hybrid drive (and, like others have said, the system cache is likely responsible for what you thought were improvements due to the SSD). I expect better from this column.

Robin Harris May 25, 2010 at 6:52 pm


The first run in the video was the first startup on the XT. That should reflect what that drive does with no SSD caching. It is also in line with what I’ve observed – but did not video – on my Hitachi 7200, 500 GB drive system drive. I re-ran the FCP startup test after re-booting the system to clear the caches. Instead of 10 seconds it took 17 – still much faster than even my 10k VelociRaptor.

Other sites will detail timing and power. My approach is to look at the technology and what it delivers to people who don’t read tech sites like AnandTech. While the XT certainly has room for improvement it is the first hybrid that delivers improvements that most civilians will appreciate.


Taylor May 25, 2010 at 8:51 pm

Thanks for clearing that up. To be honest I am fairly excited about this drive.

Though, I wonder how it response to defragmenting of the filesystem. Hopefully, if you defrag after months of use, it doesn’t take more months of use for it to adjust to new “hot spot” locations.

Gonzague May 26, 2010 at 2:20 am

Robin : we get your point but i think the videos are quite misleading, one could think the perfomance gain from cold to hot startup time for the applications you launch in them is only the result of the drive’s technology when its .. not :-s

The best videos – imho – when it comes to showing the performances a drive can bring are done by comparing two identical setups w/ different drives 🙂

PS : i’m not certain that a reboot cleans / empties the cache but Onyx does that pretty well –>

Greg Roody May 27, 2010 at 6:32 am

I don’t see the concern here. Putting cache on a storage device is a tried and true means of increasing performance. Every modern storage array does this.

If data is in cache, the response time is orders of magnitude better than if you have a cache miss and have to go get it off the spinning brown stuff. The algorithms that support the cache are the stuff of multi-million dollar patent fights. Things like LRU, MRU, and pre-fetch separate cache from mere buffer.

Whether an app on a particular mac starts in 5 seconds or 6 isn’t the point.

The design is unique, and it can’t help but improve performance for certain workloads. Like SSD’s however, the challenge I think will be knowing which workloads gain most from it.

For laptops, it’s just a no-brainer though as long as the costs are not out of line.

Jacob Marley May 29, 2010 at 12:58 pm

Making a lot of assumptions/simplification, if a user’s average daily footprint of disk block reads is less than 4GB then the read cache will accelerate majority of IO.

If the footprint is larger then like any cache overflow situation, the cache will be thrashed. In this case a lot more writes to the cache, which leads to the question of are they doing any wear leveling or not?

Do you still have the drive?
Can you use it as your daily system drive and let us know how things pan out over a few weeks of use?


darkfader June 7, 2010 at 4:28 pm

Nice review, but…

are you aware your system is crammed full?

I’m wondering if my girlfriend’s OS X desktop is still the worst I ever saw. But yours appears still a bit sorted 😉

Anyway, I’m just waiting for seagate to bring the same hybrid option for a combo of 2.5″ 15k drives with 8GB SLC. The price difference is so nicely low and I don’t think we’ll see cheaper, easier self-tuning storage tiering for a long time to come.


james braselton July 6, 2010 at 12:10 pm

hi there wow ssd 150 times faster then hard drives why has sony not gone ssd for ps3? it is soo slow with the hard drive i hope that this hybrid technology can be applied to a raptor or cheeta hdd at 10,000 or 15,000 rpms. that’d be awsome.

james. braselton August 28, 2010 at 7:54 pm

hi. there. darkfader. i. am. too waiting for the. 2.5 inch. 15,000. rpm. 8. gb. ssd. flash. drive. hybrid. be. great. for. my. ps3.

Brad Edwards September 2, 2010 at 1:40 pm

Seagate has a video comparing four different technologies (SSD, Hybrid, 10K, 7200) on the same laptop with a clean boot straight to a test package. There’s certainly a grain of salt to be taken with the video, but it was interesting just the same. I imagine power consumption is somewhat more than the stock HD, but still they make a compelling sales pitch if you can get near SSD performance with standard HD capacity for only a relatively small premium.

Mike September 6, 2010 at 12:45 pm

Oracle uses this type of technology in the Exadata machine. Logically, there’s disk, flash storage and memory. Depending on the need for the data (statistics again), the data is copied onto flash or taken into memory for speedy access.
Nice to see that similar things are coming to the consumer market.

Greg Reiter January 10, 2011 at 11:44 pm

Hi Robin,

I put one of these drives in my MBP i7, not as a boot drive, but as a replacement to my optical drive for use in storing virtual instrument libraries. I installed an OCZ Agility 240gb drive as my boot drive. It was a real toss up as to what would serve me best. What I really wanted was a 500gb SSD drive to serve as my optical drive replacement, but the prices are still too crazy right now. I have discovered that the boot time is er, ok. As I have some of my virtual instrument libraries stored on the SSD boot drive, I have to constantly battle having enough space on that drive. My optical replacement: Momentus XT 550gb has been filled up to 90% since day one.

What I have experienced as the best performance gain from having my OCZ 240 gb SSD boot drive is application start up, but I could get that same performance gain from having a second Momentus XT 500gb drive serving as my boot drive, which is why I have just purchased a second one recently. I’m in the process of getting my energy level up again to want to tear apart my MBP to go through this tedious process all over again!

My next experiment will be using two Momentus XT 500gb drives in my MBP i7 and using my OCZ Agility 240 SSD as an external 1394b drive for virtual instrument libraries.

Yup, I shoulda waited on the OCZ Agility drive, but it does run cooler than the Momentus XT.

David Jonas February 6, 2011 at 2:21 pm

are the controllers on these swappable without needing any soldering or identical firmware and/or logicware ?

Leave a Comment

{ 3 trackbacks }

Previous post:

Next post: