Recently, I read my colleague Thom Denholm’s whitepaper on data reliability in space (scroll down below for the download link). I found the realities of space as an environment for embedded data storage to be fascinating. Equally intriguing to me was learning about the inventive ways that embedded designers tackle the problems of data integrity and reliability in such an environment.

I had to learn more.

Satellite data – a key component of our data-driven world

Many kinds of devices operate and store data in outer space. Satellites are one such device, storing and sending a huge amount of data. In the increasingly data-driven world we live in, this data plays a role in internet connections, agriculture, and the medical field – with countless more uses.

Over 3,000 satellites are currently operational and orbiting Earth. Their sensors pick up data from the space that surrounds them, as well as from Earth itself – crop health and fertilization cycles, remote care of medical patients, images captured of distant solar bodies, or changing weather patterns on other planets. All of this data can then be transmitted down to Earth, to be used in fueling complex activities vital to global societies, governments, and businesses.

But all of that is only possible if the data is reliably stored and handled first and foremost on the satellite.

An alien environment for flash media

Reliability is a particularly important consideration due to the harsh, literally other-worldly environment of outer space. The costly and carefully designed devices within a satellite are under threat from solar radiation, space dust and debris, extreme temperatures – even collisions with other inoperative satellites.

As you can imagine, all of these perils can have disastrous results on flash media. And it’s not just space anomalies that pose a threat to data, but also the physical isolation and long-distance nature inherent to space travel. With some satellites in space for decades, data retention becomes a real concern.

“Lost user data, corrupted system data, and related problems are a constant struggle for companies here on the planet,” says Thom. “If a system update can’t fix the problem, a customer can always return the device to the vendor or manufacturer for additional assistance. This is not so easy with extreme environments and distant locales.”

With potential data corruption and structural damage to the hardware in space, embedded designers need creative and reliable solutions to ensure the integrity of mission-critical data.

Achieving reliable data storage in the darkness of space

For attaining data reliability and integrity in the harsh void of space, selecting and implementing the most appropriate methods of data storage is vital. Modern satellite designs tend to use SSD flash memory to do this. Older satellites instead stored data using magnetic tape, which worked well as a cost-efficient, radiation resistant option.

Flash memory however is significantly faster than magnetic tape. Plus, due to its non-volatile and solid-state nature, flash is also more reliable. The engineering required to enable magnetic tape’s necessary moving parts is a hindrance in space, resulting in flash simply being the more reliable choice.

What this improved reliability comes with, though, is a higher price tag: flash memory is over 12 times more expensive per gigabyte than magnetic media. A tricky proposition with satellite missions typically having tight budgets. Fortunately, flash-friendly software is an excellent way to manage this increased cost while getting the absolute most out of the flash: extended hardware lifetime, the highest possible performance, and guaranteed data integrity. In the long run, quality-assured flash controllers and flash-friendly file systems make all the difference in hardware costs and storage reliability – particularly in an environment as unforgiving to data storage as outer space.

Protecting flash in space with fault tolerance

Even with the right data storage methods, failure can still happen. To protect against flash storage failures in space, hardware and software stacks can be designed for fault tolerance. This refers to a system that’s able to continue functioning even if some of its components break down or lose power. An incredibly important capability given all the unique dangers to data integrity that exist in space.

“One of the best ways to achieve fault tolerance in data storage is through redundancy,” says Thom. “It’s considered so useful that, in general, all fault-tolerant systems implement redundancy in some form. For devices in an environment like space, fault tolerance is even more important, with even the hardware components duplicated to better deal with possible errors. Among embedded systems in space, two kinds of redundancy are of particular interest: physical and functional.”

Redundancy involves storing multiple copies of data on the device, in order to ensure some version of that data will be valid. What’s also needed is a way of having the software determine which of these copies may have been corrupted by whatever cosmic catastrophe has impacted the device. Detecting and dealing with bit errors that develop on the media is a major concern and can be tackled with error correcting codes (ECC), where multiple copies of information and checking for consistency can help to ensure redundancy. If corrupted data is identified, self-repair and corrective actions can then be speedily initiated by the flash management software. Of additional importance is minimizing the impact of power interruptions, doable with atomic operations.

Whitepaper: Data reliability in space

Earlier, I mentioned Thom Denholm’s informative whitepaper on data reliability in embedded satellite designs. With satellites requiring fault tolerant hardware and software that can ensure consistency and correctness in space applications, physical and functional redundancy are needed. At the same time, budget requirements are an ongoing consideration. An excellent solution is TMR NAND flash, combined with flexible and fully supported software that matches the unique needs of the embedded design.

In this paper, Thom dives into the technical details of this topic, including some of the specific factors that makes TMR NAND flash with optimized flash management software a more solid solution to other alternatives.

Download the whitepaper here.

Final thoughts

Satellites collect, store, and send critical data while operating in a brutally harsh environment that poses unique challenges for reliably safeguarding that data. Everything from the method of data storage to the file system software must be precisely optimized to ensure the integrity of mission-critical data. Embedded designers have come up with innovative methods to manage these problems while adhering to strict budget, time, and logistics restrictions.

There exists an adage, “restrictions breed creativity”. Few environments impose such restrictions on embedded storage designers – and humanity in general – as outer space.

Maybe that’s why it fascinates us.


Learn more about fail-safe data storage in spacecraft and avionics.