http://www.techpowerup.com/188186/diablo-technologies-disrupts-latency-and-performance-metrics.html
So, they're now using the parallelized DDR3 solution to directly use Flash-based storageb by directly communicating with the CPU as a means of accessing information, for low-latency that trumps PCIe-based Flash (by 85%) and SATA/SAS (by 96%) - a lot.
Is this a good thing? Do you think that ultra-low-latency, high-bandwidth flash storage could be used as a future form-factor to replace the mSATA and NGFF (M.2) form factors, and possible even SATA standards? For example: could motherboards for consumers eventually be equipped with SO-DIMM slots to be used with Flash Memory in SFF (Small Form Factor) computers, such as Mini-ITX, in order to replace 2.5" SSDs and possibly 3.5" HDDs (once Flash Density endurance and density grows enough, and the cost falls a lot more) ?
Just a thought. Could CPUs be equipped with a small memory controller dedicated just for storage? Any takers for this idea, or is it not viable? Will it ever make it's way to the mass-consumer market, considering how Enterprise-grade SSDs and Consumer-grade SSDs have been differentiated by manufacturers?
I know intel is working on a new way of making ssd's and other storage devices store data. It is an alternative to sata and based on pcie if I'm not mistaken.
You mean Thunderbolt?
Also, Samsung just finished their 3D V-NAND Flash memory, which is going to allow them to put 20 layers of NAND Flash on top of each other. This might allow for Terabit memory chips, at much higher speeds. This means with eight chips, you now have a Terabyte SSD.
I was thinking about S-ata express. Seems like Intel hasn't got anything to do with it in particular though.
The new 1TB Samsung 840 Evo is going to be one of these said drives. except its using toggle nand instead of synchronous nand.
Sata Express might be interesting. Although I personally think that 2x PCIe 3.0 lanes via an external cable would be more than enough for storage (for this generation of computers). That's nearly 16Gbps of bandwidth, more than double that of SATA 6Gbps. Also, considering that power could be delivered via this cable, that would really help.
Although, it's much more likely that we'll just see Thunderbolt 2.0 come out with an version that doesn't include a DisplayPort connection passing through it. Thus, it would effectively replace USB 3.0, SATA 6Gbps, and many other forms. But that all depends; can a Thunderbolt 2.0 to USB 3.0 hub work, using only the power from a Thunderbolt cable? Is it going to be exclusive to Intel systems, or will it be available for ARM, AMD, and other platforms? How much is it going to cost to add it to motherboards, or put it on computer cases?
If these things aren't archieved, Thunderbolt might have the same fate as FireWire.
Sata Express isn't only interesting, it is coming. And it will be generally accepted. Thunderbolt might indeed see that same fate if prices on implementation and cables don't come down...
Consider this; DisplayPort uses two data streams, for up to 17 Gbps combined. PCIe 3.0 has 8Gbps per lane.
Put two PCIe 3.0 lanes, and you've still got less bandwidth going through copper than PCIe 3.0
And DisplayPort cables aren't that expensive. Nor HDMI for that matter.
The issue is whether or not Intel will use an open platform, and if they'll charge huge royalties. If their royalties are too high, it won't sell. They could just sell the Thunderbolt controller, and leave the connectors and cables royalty free. But that's up to them.
In the end, the technology won't decide it alone. No matter how good the specs are, if it costs too much or if the consumer doesn't use it, it won't sell.