What NASA’s Mars Rover Tells Us About SSD Life

The Mars Exploration Rover Opportunity has likely died. NASA has exhausted its efforts trying to make contact with the rover, and will wrap its 14-year mission that lasted longer than anyone at Jet Propulsion Laboratory ever expected.

A dead battery is the likely culprit. NASA believes Opportunity’s solar panels became caked with dust, preventing recharge. This comes four years after NASA was able to salvage the rover from a near-death experience related to its flash memory solid-state drive.

Solid-State Storage Drives on Opportunity

Then in 2015, NASA called it “amnesia”— Opportunity was having problems with its onboard data storage drive, resulting in nightly loss of data after its system powers down. This came in year ten of the originally planned three month mission, and experts recognized the glitch as an effect of old age.

After ten years of exploring Mars’ Meridiani Planum region in sub-freezing temperatures, Opportunity’s hardware seemed to have deteriorated.

NASA Project Manager John Callas suspected the failure wasn’t due to the harsh Martian environment, however.

“Flash memory has a limitation on how many times you can read and write to it,” Callas told Discovery News, indicating that the daily writing and re-writing of expansive sets of data to Opportunity’s memory banks are the real toll-takers.

NASA engineers drove the rover during daylight hours, gathering telemetry data about Mars’ surface. To conserve energy at night, the rover powered down, storing the day’s data in the non-volatile flash memory banks. Unlike volatile memory like RAM, this type of solid state memory does not require constant power to store data.

The Opportunity has seven onboard flash memory banks. Engineers at JPL pinpointed that one of the banks was defective. As a result, when Opportunity tried and failed to copy data to the broken memory bank, its software forces a reboot and loses the data. What’s worse, commands from mission control tripped the system into a cycle of continuous reboots, essentially shutting down the mission for hours at a time.

JPL deployed software that programs that enabled Opportunity to ignore the defective bank. It prompted the rover to write to one of six banks still functioning.  Initially NASA hoped it was enough to keep Opportunity driving for a little longer.

To everyone’s surprise, Opportunity lasted another four years.

Lifespan of commercial solid-state memory

When comparing SSD vs HDD in computer systems, durability is a huge selling point for solid-state storage. SSDs use an integrated circuit assembly which is more resilient compared to the spinning platters in mechanical disk storage. But flash isn’t perfect. Like any drive, SSDs have a life expectancy.

We think of SSD lifespan in program and erase (P/E) cycles. Mainstream solid-state drives we use in computers generally wear out after being programmed and erased 10,000 times. Manufacturers calculate their warranty offering using a measure called ‘Total Bytes Written (TBW) or a length five-years, whichever comes first.

For example, the Samsung 860 EVO warranty shapes out like this:

• 250 GB drive is rated 150 TBW
• 500 GB drive is rated 300 TBW
• 1 TB drive is rated 600 TBW
• 2 TB drive is rated 1,200 TBW
• 4 TB drive is rated 2,400 TBW

If you do the math, you would need to write and erase 320 GB per day to exceed TBW ratings before the warranty period. As ridiculous as it seems, SSD research and development aims at extending lifespan further.

Several years ago, engineers at the Taiwanese semiconductor company Macronix created self-healing flash memory that they claim can survive more than 100 million program/erase cycles. That is likely to outlast human civilization. This experimental design charges circuits with bursts of extreme heat (1,472 degrees Fahrenheit) which reverses the oxidation that wears out flash transistors with each cycle.

Why larger SSD drives are generally more durable

There are lengthy performance testing reports to browse on the Web, where SSDs are put through rigorous paces of continuous writing and erasing. In many reports, like this one from an independent German SSD tester, larger drives (500 GB) kept under a 10 percent failure rate for right around 10 years—an interesting congruence with what we saw with the NASA rover.

Smaller drives did not fare as well. The smallest drive tested (32 GB) reached a 10 percent failure rate after just 1.5 years.

With this in mind, it is remarkable that Opportunity’s memory banks held out as long as they did. “The rover has been amazingly healthy considering how much we’ve used it,” Callas said. He likened the rover to an aging parent, who may be in good health, but “could have a massive stroke right in the middle of the night—so we’re always cautious something could happen.”

Related Smart Buyer articles

• 2019 SSD Buying Guide
• How to Install an SSD into an old laptop
• HDD or SSD? How to Buy Disk Storage in 2019
• Best Selling SSDs for Business Systems (2018)

This article was originally published Jan. 8 2015, and updated Feb. 13, 2019. 

Summary

Article Name

A Good Run: What NASA’s Mars Rover Tells Us About SSD Life

Description

The NASA Mars rover has a faulty flash memory bank, and provides a good illustration of what commercial flash memory users can expect for SSD life span.

Author

Adam Lovinus