Memory // Storage

Unexpected Power Loss and How SSD Can Help

By Kristen Hopper - 2017-02-15

Similar to planning for winter, by stocking up on batteries, canned food and fuel, it is also important for embedded manufactures to ensure they are using the correct SSD for their application. Storage drives are at risk in power loss events. Unplanned power interruptions can present perhaps the greatest risk to embedded computing systems. When power is lost during system operation, drives can be corrupted and data ruined, resulting in downtime as drives must be reformatted, operating systems reinstalled, or products returned. This downtime greatly impacts overall reliability, dependability, customer goodwill and total cost of ownership.

Embedded SSDs are designed with reliability as a key priority due to various system applications typically operating 24/7, having long product deployments and often operating in harsh environments with exposure to extreme temperatures. Power fail protection is a differentiator for embedded SSDs and many vendors tout solutions.

Power fail protection technology is undoubtedly one of the most important added features and a significant differentiator in embedded SSDs. Truly rugged SSDs must be engineered at the hardware and firmware architecture level specifically to withstand aggressive power cycling.  

This Micron SSD power loss protection white paper details Micron’s data protection scheme employment to mitigate asynchronous power loss.  In a “clean” power-down scenario, once power is shut off in the case of a SATA-interface drive, the ATA STANDBY IMMEDIATE (STBI) command is the last command sent by the host and acknowledged by the SSD. This host data physically stored by the non-volatile media and confirmed to the host is referred to as “data-at-rest”.  In the event of an asynchronous power loss, failure to receive and/or acknowledge this command can result in a longer than normal time-to-ready (TTR) sequence during the ensuing power-on because the drive was not able to complete write operations prior to the power loss such that “data-in-flight”, consisting of host computer writes not yet stored and/or acknowledged by the media, as well as addressing information, may have been lost and now has to be reconstructed.  While data-at-rest protection is considered “good enough” for Client SSDs, of which automotive and industrial SSDs are a subset, Micron provides additional methods to increase a Client-class drive’s robustness against data-in-flight loss due to asynchronous power-down—additional NAND hold-up capacitance and proprietary firmware optimizing interaction of microcontroller features and on-board voltage regulation. Once again, Micron demonstrates that “good enough” is not enough by offering superior SSD value through incorporation of innovative and cost-effective feature enhancements.

Kristen Hopper

Kristen Hopper

Kristen Hopper is Senior Manager, Managed NAND Design at Micron.  An IEEE Senior Member with >25 years’ global semiconductor industry experience, she received her M.S.E.E. from the University of California at Berkeley and B.S.E.E. with Distinction from Cornell University.