With the latest crypto-announcement of a 128 GB, 2.5″ SATA flash drive, my patience with the flash disk folks is at an end. They aren’t shipping flash drives and, as near as I can tell, they have no intention of doing so for quite some time.
Last April Samsung announced a 32 GB flash drive, and theoretically started shipping some in an overpriced, last gen notebook. Then PQI announced a 64 GB SATA flash drive in August.
Just try to buy one. They don’t exist.
We’re being gamed
A product is announced when the vendor provides specs, availability and pricing. What we’ve been seeing are prototypes from vendors who are hoping to drum up interest from OEM customers, not real products. The vendors don’t have the vision to market these products and aren’t investing in the market.
Vendors in a daze
Flash drives cost a good deal more than disk drives. The raw chips – single level NAND – are currently about $6 per GB. Toss in the SATA interface chip, a circuit board, logic chip, assembly, connector and packaging and maybe you are talking, outside $7 a GB, and less on a larger drive. So let’s pass that through efficient distributors and they should be able to sell for $9-10 GB.
Is there a market or not?
Taiwan and China, where most of the flash thumb drives come from, excel at mass production. What we’re seeing is a kind of rain dance, where by talking about the products the vendors are hoping to get a launch customer who’ll pony up an order big enough to justify the investment required to start volume production of the products.
Maybe that is the only way. I hope not.
Early adopters are out there – if you start shipping
When USB flash thumb drives came out, six or seven years ago, they weren’t cheap either: over a dollar per megabyte. But early adopters, like me, saw the advantages and bought anyway, helping fuel the business.
I believe the same kind of dynamic is waiting to be harnessed for SATA flash drives. Execs with a $2,000 ultralight notebook would be willing to pay $500-$800 for another couple of hours of weightless battery life – and bragging rights. Plus the performance advantages, especially on boot, since Windows sleep mode has almost never worked right on any laptop I’ve had. I also think a bunch of folks would like to see what these drives could do for web-servers and database log files.
The StorageMojo take
My suspicion is that the vendors lack confidence in their wear-leveling algorithms (see RAM-based SSD’s Are Toast – Yippie ki-yay! for more on that topic), so they don’t want to go to market with a product that they can’t put a competitive (with disk drives) warranty on. C’mon guys! This is a huge opportunity to move upmarket. The margins have to be better than thumb drives.
Comments welcome, as always, especially if you’ve bought such a drive. Moderation turned on to encourage comment spammers to get a real job.
As usual, very good article 🙂
One thing I noticed is that none of the PQI’s SSD products has FCC and CE logo on their page:
http://www.pqi.com.tw/product.asp?oid=142&CATE1=151
Could it be that they are waiting FCC approval/certificate before they put the SSDs on the market? As Steve Jobs mentioned on the iPhone’s keynote “this can take 6 months” 🙂
On the PQI’s site front page there is a news article from 9 Jan 2007, and I also noticed that the SSDs are available as an option for sales questions:
(as DiskOnModule products):
http://www.pqi.com.tw/contact_sales.asp
Could StorageMogo ask them when we can expect the product to be available for purchase by Execs who hate to carry second battery? 🙂
Imho it’s quite simple, AND – no – flash HDDs as you describe will NOT fly in my opinion.
The pricing and capacity of flash is dictated by USB-sticks.
Actually it’s that what consumers can easily compare to and make judgements whether the industry wants to rip you off or not.
A consumer won’t accept price-levels of FC-15k-rpm drives !
Technically newer PCs can boot from USB or CF …
So currently – a 2GB CF-card (has IDE electronics in it) costs ~50 EURO
A USB-stick w. 2GB – ~30 EURO.
Now everybody does this Dollar/Gig calculation – and it’s IMHO BULL !
It’s more a question of ABSOULTE price.
The wear-out/warranty can be an issue – I haven’t thought on that.
to me the priorities are something like this:
1.) absolute price.
If a device of whatever capacity is less than $99 – I would give it a try.
(as long as I can revert easily to my spinning-hdd if I’m disappointed.)
2a) is it possible to install windows/linux on it for John Doe ?
2.) capacity: Does a usefull Windows/Linux fit on it ?
There is no case – if this can’t be done for the average config.
User-data still could live on a second/external stick/hdd etc., but if a typical OS won’t fit on it – and/or can’t directly boot from it – there’s no case.
3.) Reliability and widespread availability:
If you can’t buy the stuff in a tech-supermarket – no chance.
If you provide some overpriced junk like MemoryStick or xD – dead.
Before an enthusiast user would buy a HUGE (highprice) flash-gadget, he wants to test it with some low-end-part – whether it really delivers on the promises: (real-world speed, capacity acceptable ?, battery-life really longer ?)
For this – at some minor fiddling is acceptable – as long as it works.
So there is close to ZERO motivation for customers to invest 500+ bucks and not knowing whether it has a equivalent everyday benefit.
IMHO the current flash-approach of the HDD-vendors violates all those rules. – They just want to rip off the customer.
Clayton Christensen’s “The innovator’s dilemma” explains quite well, why the current HDD-vendors CAN NOT (or very unlikely) be the same guys – who will lead the FLASH or more generally the SSD transition.
So by looking at Samsung’s FLASH announcements – you’re judging the market from a wrong perspective.
I’d more look at the moves of SanDISK & the like 🙂
hirni
Robin,
I think that you are right about the wear-leveling algorithms and the resulting impact on warranty-related issues … and how to qualify these.
May also be related to modifications required to the OS … if at all possible … with the largest impact on write algorithms.
There is only ‘so much’ such algorithms can do on a Flash-based SSD without a lot of RAM …. and then we are back to cost.
hear, hear.
Your article appeals to me. I’ve been trying to get my hands on a large-sized flash drive for quite a while now – and indeed have been unable to buy one. Currently my hopes for at least some performance increase (& for me lower power usage) are on microsoft supporting their readydrive. Let us hope they have the size to start these drives shipping in masses. Then wait for the real stuff…
knijnsel
I am tired of waiting for SSD’s. I wouldn’t be so annoyed if Samsung hadn’t been shipping the Q1 for over a year now. I’m pissed.