Solidata Technology

What Happens Inside SSDs When the Power Goes Down?

What Happens Inside SSDs When the Power Goes Down?

Solidata Technology

Will data get corrupted after sudden power loss? The risk of data corruption from power cycling isn't a random, unforeseeable event. It's a direct result of choices made (or not made) when that solid-state drive (SSD) was designed.

This article looks at what happens inside various types of SSDs when the power goes down, which is an area where products differ a lot. Some of the technical theory that constrains the freedom of SSD designers will be explained.

It may be surprising to learn that the ideal state for an SSD is when it's in the powered up state. This is when it's at its most reliable. A well-designed SSD should look after itself and its data when its powered up and down, which is when the SSD needs to be checked to see if it comes with 'Integral Power Loss Protection'.

What's Power Loss Protection?

When the power has no any protective action, if the SSD unexpectedly meets a power loss, the data in the buffer will be fully lost, resulting in a data loss or data corrupted.

In order to avoid this case, a protection mechanism in the SSD is needed, which is called 'Power loss protection', to prevent the data loss in the buffer.

Researchers agree the impact of power loss on flash memory reliability by theoretical deduction and laboratory test. Once the power goes down, the SSD can still work, but if the power loss has caused permanent damage to the NAND Flash, the data will become unreliable.

What's integral and non-integral power loss protection?

i)Generally, the buffer data include firmware (F/W), temporary user data and the mapping table. The F/W data size is very small; the mapping table data size is very big; the temporary operational data is varied and the data size is between F/W size and mapping table size.

If the F/W is lost, the drive can't be detected. If the mapping table is lost, the data can't be found. Usually, the temporary user data is not very important for user. If there is no power loss protection design, The F/W loss will lead to 'the drive can't be detected' (after a certain number of times of power loss).If the power supply of the capacitor is enough for F/W data writing back, but it's not enough for mapping table writing back, the drive can be detected but the operating system (OS) will be lost and even corrupt.

ii)Thus, the integral power loss protection means that F/W data and mapping table data can be totally written back to the NAND from the double data rate (DDR).

iii) When a non-integral protection SSD works as a second drive, there will be no problem and risk; but when it works as a master drive, there will be a big risk of losing OS data unexpectedly, especially when the equipment maintenance is not very convenient for some mission-critical applications.

Why some SSDs have to be with non-integral Power Loss Protection?

i) There is no big capacitor for their selection, no good countermeasures for their hardware design.

ii) If the mapping table data is stored in NAND, it will increase the NAND FLASH load (refresh frequency is very big), which is even unacceptable for single-level cell (SLC) Nand Flash.

How to design an Integral Power Loss Protection SSD?

For the super capacitor selection, it must meet three conditions:
i) The super cap must be provided with enough capacitance, to ensure the F/W and mapping data can be totally written back to the NAND from the DDR.

ii) The super cap can work at industrial temperature for industrial and military systems.

iii) The protection circuit ensures no liquid-leaking to avoid short circuits on current surge

The Solidata Rana SSD is designed with an industrial grade 3F capacitor, which supports an Integral Power Loss Protection in extreme environments. It support at least three seconds of capacitance so as to ensure that all system and user data will be written back to the NAND.

What happens in the SSD when the power goes down??

For a SSD with big capacitor design, it takes some time for the super cap to discharge after the power goes down. For example, it takes about two seconds for a 120GB SLC Solidata SSD to go from power on to ready time. During this discharging period, The SSD can't be detected temporarily because it can't communicate with the host. After the discharging period, the disk can be detected and work well again.

Summary

As you can imagine, it needs to be ensured that all data in flight is controlled and monitored. It can be done, if there is control of all the F/W data, temporary user data and the mapping table data inside the SSD. That depends on what happens next and the skill of the SSD's designer.