Update on mobile flash performance
Mikko Pitkanen over at the mobile development blog Delay ToleraNt posted some more tests on Nokia N800’s flash performance. He’s a doctoral candidate at the Helsinki Institute of Physics at CERN in Switzerland with a strong interest in storage.
The money quotes:
. . . the first observation is that we achieve write performance close to 1Mbit/s for small (less than 1 MB) files.
. . . the read performance is much better than for writing and is certainly enough to play movies. The write performance instead, is poor and would not allow the user to receive large files with the full bandwidth achievable by the device’s WLAN.
Mikko’s got a new Nokia N810 that he’s loving, so that will be it for the N800 data. Good data point. Mikko, if and when you get some N810 performance data please send it along. Thanks!
More flash high performance
Intel and Micron announced a very fast flash chip – 200 MB/s read and 100 MB/s write – but the press release included this big caveat:
“Micron looks forward to unlocking the possibilities with high speed NAND,†said Frankie Roohparvar, Micron vice president of NAND development. “We are working with an ecosystem of key enablers and partners to build and optimize corresponding system technologies that take advantage of its improved performance capabilities.
Translating from marketing speak: “nobody has the technology, like the translation layer, to take advantage of this chip.”
The StorageMojo take
Realizing flash’s potential will be a multi-year, multi-company effort. No one has a clear idea of what the ultimate limits will be. In the meantime the disk folks will be working to limit the damage by raising reliability, density and shock resistance. Both technologies have a place. The fight is about boundaries. And all of us consumers benefit.
Comments welcome, as always.
I thought the whole point of their ONFI initiative was to create this ecosystem. Are the buyers of these chips all holding out from going the ONFI route? Apple hasn’t given any sign of readiness to adopt the standard, right? I wrote a few thoughts on this at my site.
I don’t think so. ONFI looks like a chip-level interface, not a fake-disk interface. It is the fake-disk or translation layer that appears to be the biggest problem in getting great flash SSD performance.
As you noted in your blog, the 3 largest NAND vendors aren’t on board with ONFI and why should they be? They’re selling all they can make without it. Why would Apple care? If and when Samsung goes with it, they will too. But as the world’s largest consumer of flash Apple exists in its own ecosystem.
On a related note, I’ve been doing a BUNCH of performance testing of SSDs on my blog:
http://feedblog.org/2008/01/30/24-hours-with-an-ssd-and-mysql/
http://feedblog.org/2008/02/01/random-write-performance-in-ssds/
http://feedblog.org/2008/02/02/ssd-pbxt-crazy-suspicious/
Interesting so far but not as exciting as I would like. The random write performance is far to slow for now.
Hey Kevin (& Robin),
One of the reasons NetApp is so excited about flash is the inherent advantage WAFL’s write-anywhere architecture has for optimizing write performance to any device. Look for us to articulate this more clearly throughout 2008.
I came across an EETimes.com article on the “Ashwood memory architecture”:
http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=205604434
It seems that Mr. Ashwood has a design that allows parallel data accesses to non-volatile memory. A key challenge seems to be the “translation layer” – how do we take advantage of the potential for parallelism, especially for writes ?
Anyway, I thought this news item would be of interest to the readers of StorageMojo.
Best regards,
Manolis.
Amen, brother! Samsung and Seagate are pursuing a parallel path for these reasons. I’ve been posting on this quite a bit – here’s my latest: http://storageeffect.com/2008/01/25/bill-watkins-the-journey-to-flash/
I wonder which “science” Mikko is into, probably some humanitarian – there is no single reference which filesystem he was testing (e.g. path to files stored could give basic idea), which version of SW is running… And yes, also you Rob could take into account one interesting detail – tablets are battery-driven devices, and one of main quality metrics is survivability through battery power loss… yes, try to avoid datacenter mentality – we do NOT burn electricity like there is no tomorrow, and have to keep device surviving in not-so-professional hands (read – prone to f#¤% up poor piece of electronics)