you:
Thanks for your summary - I still hadn't the time to listen to
the call. I think higher packaged DRAM is a good topic, but
I'm wondering, how hard it is to crack this "community"?
me:
Well if you believe what was said at the presentation there is supposedly a 60% TCO advantage for a typical 5k pizza box search engine server farm. The guy from Frost and Sullivan broke out those costs advantages somewhat near the end of the presentation. The presentation was undoubtedly about raising awareness, but DELL and HP were mentioned a couple of times. I'm sure the presentation we saw has been seen by a lot of potential partners but have no idea what the acceptance/commitment level is. From a getting product into the data center quickly point of view, I think it's probably more important to have a big name such as DELL or HP on the server. Getting those guys on board is going to be way more than half the battle. Still, were only talking about 5% to 10% of the server market and I'm not sure how enticing that would be to a DELL/HP. To SPSN however, getting that business would probably result in a doubling of quarterly revenues.
you:
I'm still questioning the speed of these devices. I really can't imagine, how these could perform (nearly) on par with conventional DRAM.
me:
In the presentation they gave, no mention was made of the type of DRAM used. However, using identical machines except for the memory, the EcoRAM machine used 19% of the energy the DRAM machine used while running at 80% of the speed of the DRAM machine. Later they mentioned that this system was a couple of months old and newer EcoRAM ran faster. Personally, if they can get the EcoRAM to run at DRAM speeds the other advantages(lower cooling costs, green advantages, higher density memory, better reliability, lower floor space costs, memory persistency, etc.) will make the product highly advantageous.
you:
The next question is, how could this be transformed into other types of applictions like SSDs or changes in the NOR vs. NAND (cell phones) environment.
me:
The distinct impression I got was that SPSN isn't interested in competing in the commodity markets. To me that means they wont try to compete with serial access products (memory cards, MP3 players, etc.). MirrorBit's advantages seem to be mostly in the Random Access arena. Still there are areas where combinations of MirrorBit and ORNAND such as Eclipse can apparently compete, such as wireless. At least until SPSN gets on a more secure financial footing I don't think they will trying competing in markets where they don't have a perceived technical advantage. Q4 should be really interesting as that will be the time when all these Eclipse based products are scheduled to hit the market. |
