this would fit with MS talking about their SSD as virtual ram while cerny is talking about the SSD as a means of eliminating loading and streaming assets as fast as possible.I don't think a reference to access timing directly suggests DRAM vs SRAM. There's various things that affect access time.
On SRAM vs DRAM, I guess taking the patents' explanations at face value and trying to keep it simple:
SRAM might be faster/'better' for access to low-write, high-read data
DRAM might be faster/'better' for access to high-write, low-read data (at least vs a cache-miss in a SRAM setup)
The patent makes the suggestion that you might want to optimise for the former scenario in a games machine more than the latter, that it would be more advantageous to optimise for read-access to game asset data.
DRAM would I guess, also potentially have more space for cached data, at least depending on the size of your address lookup tables in use. Although there might be nothing to stop an implementation using SRAM for address lookup and block operation data, and a smaller-than-typical amount of DRAM on the side just for caching. SRAM has some other side advantages too - namely the ability to use it at a lower power-cost for standby mode operations.
As always, I wouldn't conclude for sure what the setup in either machine actually is at this point.
basically MS are planning to use the SSD as an actual additional ram which they write to a lot, while the SSD based off sony's patents will mostly read data very fast.
wonder what will developers favor more.
i do fear that writing to the SSD that much might increase the likelyhood of a SSD fault.