Recently, an article about the reliability of SSD solid state hard drive data is very popular. It is said that an engineer from Seagate Alvin Cox found that the solid state hard disk will be data without power. As time goes by, and temperature has a very large impact on the reliability of SSD, finally concluded that the reliability of most SSDs is only 2 years, it is recommended that important data should not be saved with SSD. Authoritative experts + terrible consequences + alarmist writing — — This is simply the perfect QQ space and the current WeChat circle template, Nima, when did this spread to the DIY circle? Although everyone now knows the benefits of SSD, one problem that SSD can't always avoid is data reliability. Congenital P/E operation not only gives SSD super performance, but also brings hidden dangers & mdash; — data life is limited, upgrade with TLC will lead to reduced life, upgrade to a more advanced process will also lead to reduced life, people who do not know the truth will inevitably have to doubt the life of SSD, which would have been reasonable, but mentioned earlier Is the rumor that the SSD is not powered to cause data loss is true? A successful rumor will never be 100% wrong, a clever lie will always be a three-pointer, and the previous rumor is also the same. — SSD does not power will lose data and temperature has SSD life The impact is correct, but it’s not right to exaggerate the impact to such an extent. Anandtech’s website now has a sneak peek. The first is the identity of the previous report author, Alvin Cox. This report is not from Seagate, but from the JEDEC organization. Alvin Cox was the chairman of the JC-64.8 standard group (also known as the SSD committee), but one thing is correct. ——Alvin Cox is indeed a senior engineer at Seagate, but this has nothing to do with his report. The JEDEC organization itself is a joint organization of vendors. It is normal for members to come from different vendors. The current JC-64.8 standard group is Led by Frank Chu of HGST. Before answering the previous question about SSD data retention, let's first understand how vendors determine SSD reliability. First, the SSD must guarantee a certain amount of capacity, so it is impossible to use a lot of blocks for the OP space, which will reduce the available capacity of the user. Second, the SSD must meet the specified UBER and FFR specifications. The following table explains the meaning of UBER and FFR. The former is the proportion of bytes in the SSD read byte, and the FFR is the proportion of functional failure. Consumer and enterprise. The SSD has different requirements on these two indicators. The UBER of the consumer SSD is 10-15, which is the byte read failure rate of 1/1015, and the FFR cannot be higher than 3%. The last point is that the SSD is not powered, to achieve the JEDEC data retention rate, the consumer SSD is 30 & deg; the C temperature is 1 year, the enterprise level is 40 & deg; C is 3 months. The above indicators are all represented by TBW, which is the meaning of the data write amount in the SSD logo we see, such as 100TB data write life, which means that the index can meet the above requirements when writing 100TB data. . This table from Intel temperature will also affect the SSD data storage, from the consumer-supplied SSD shutdown temperature and the storage time under the active temperature, 40 & deg; C activity temperature, 30 & deg; C shutdown temperature SSD The data retention time is 52 weeks, which is one year. The higher the shutdown temperature, the shorter the SSD data retention time, the active temperature is 25-30°C, but the shutdown temperature is 55°C, the time for the SSD to save the data is only 1 week, which is the previous rumor> The time is only a few days, the source of this, it does happen, but it is not representative at all. In reality, the shutdown temperature of 55°C is almost impossible (PS: If there is such a situation, I feel that I should care about my personal safety rather than data security at this time), because most SSDs are still used at room temperature, temperature. Usually within 30°C. The temperature at the time of the activity is usually around 40 & deg; C, and other parts of the system will also generate heat. Finally, technically explain the problem of saving this data. The conductivity of semiconductors is closely related to temperature, which is not good news for NAND flash memory, because the electronics should not move after power-off, otherwise the charge of the Cell unit will change. In other words, as the ambient temperature rises, electrons will escape from the floating gate more quickly, eventually causing the voltage state of the Cell cell to change, resulting in unreadable data. (that is, the SSD can no longer save data) is another thing in the active state, higher temperature will have higher silicon-based conductivity, higher current during P/E operation, and less pressure on the tunnel oxide layer. This will increase the reliability of the Cell unit because SSD reliability is primarily limited by the ability of the tunnel oxide layer to hold electrons in the floating gate. (The NAND read and write principle can also refer to our previous article.) In short, in a typical SSD usage environment, we do not need to worry about the SSD data retention problem. The previous table has only passed the reliability test, and the new SSD data retention rate will be higher, which is usually 10 years for MLC NAND flash memory. If you buy an SSD and don't use it, then the possibility of it being directly eliminated is actually more likely than the SSD can't retain the data. This article comes from [system home] www.xp85.com Related articles recommended: 1. Faster, larger solid state drive 2. How to choose SSD? How to see the solid state hard disk is good or bad 3. The difference between the benefits of the solid state hard disk and the ordinary hard disk. 4. The rookie must see the solid state hard disk partition graphics detailed