Why Apple's APFS won't last 30 years

When introduced at this year's WWDC, APFS - Apple's new file system - was positioned as a platform for the next 30 years. Despite its many improvements over HFS+, APFS lacks two important features to future-proof it for serious business use.
Written by Robin Harris, Contributor

The path ahead.

Photo by Robin Harris

One of Apple's key goals for APFS was to make a single file system that could be used on everything from a 12-core Mac Pro down to an Apple Watch. Compromises were inevitable.

However, APFS misses two key issues in the rapidly evolving world of storage. Here's an outline from a workstation-centric perspective.

Data integrity

APFS doesn't guarantee data integrity. Modern file systems, such as ZFS, gurantee end-to-end data integrity through checksums. APFS doesn't offer checksums, so any storage bit rot will remain undetected and uncorrected by APFS.

Storage devices typically offer error detection and (some) correction, but given the higher uncorrectable bit error rates of SSDs over disks data intensive professionals will wish for the extra protection of checksums.


APFS is block-based. But the biggest news in storage today is the advent of non-volatile random-access memory, or NV RAM. Intel's 3D XPoint is best known, but several companies - such as Crossbar - are also bringing NV RAM to market in the near future.

RAM, of course, is accessed by the byte, while hard drives and SSDs are block-based, with block sizes ranging from 512 bytes (old) to 4k bytes (new). NV RAM - which a premium vendor like Apple should adopt sooner than most - needs a file system that can handle both.

The Storage Bits take

Very few consumers care about data integrity - as the lack of complaints about the 30 year old HFS+ attests. So what if a photo from your Grand Canyon vacation gets truncated? But professionals care, and they may need a second checksummed file system for bulk storage and archiving.

Integrating NV RAM into Apple devices is the bigger issue. Some NV RAMs sport much lower power requirements than flash, a plus for mobile devices. Byte addressability also minimizes I/O latency, another desirable trait in small devices.

While APFS is a welcome and much-needed update for Apple devices, the architect's crystal ball needs some polishing. APFS may not be able to span all Apple devices for very many years, let alone decades, given the rapid changes in digital storage technologies we see coming.

Courteous comments welcome, of course.

Editorial standards