Http://baike.baidu.com/search?word=%E5%8C%97%E4%BA%AC%E5%88%98%E5%BF%83%E8%AF%AD%E4%B8%B0%E8%83%B8%E8%8C%B6
From a technical point of view, NAND flash flash media, although there is a problem of the number of erasure, but with software algorithms, system-level data protection technology, NAND flash life is no longer a difficult problem, and NAND Flash's more advantages have become an important driver of the technology's push forward. SSDs built with NAND flash are rapidly replacing traditional disks, have been widely used in the Internet, and continue to penetrate into the enterprise-class market that demands data reliability. Beijing Liu Language Breast Enhancement tea
SSDs have several very unique features compared to traditional disks. The first feature is that SSD has very good read and write performance, especially in the random read and write performance far more than the traditional disk, second, SSD has a high storage density, in the unit volume, SSD storage density is higher than the traditional disk, This year many manufacturers of 2.5-inch single-disk SSD capacity can reach 8GB, exceeding the capacity of the same volume disk; Thirdly, SSD has better controllability in data reliability, NAND Flash storage medium service life is controllable and predictable, so SSDs can combine storage media life and data characteristics with system-level application software to maximize the reliability of your data. On the disk, the life expectancy of the storage medium is very difficult; The SSD is a semiconductor storage, and as the market size advances, the price of the SSD itself will be much lower than the disk, although the current SSD price is higher than the disk, but its price drops far higher than the disk, The price of future SSDs is less than the disk is no suspense.
Compared with the traditional disk, the SSD is once again approaching the CPU in performance, which makes the performance price scissors in the computer architecture become smaller. However, compared with DRAM memory technology, there are some performance gaps in SSDs. When the CPU accesses the cache (SRAM), there is an NS level of access latency, while accessing memory (DRAM), there is a 10ns level of access latency, but there is a 10us level of access latency when accessing the SSD. Although the MS-level access latency of the disk is reduced by an order of magnitude, there is an order of magnitude performance gaps in SSD and DRAM access latencies. How to reduce this performance gap is an issue that the industry needs to consider.
Non-volatile memory technology NVDIMM