Sign in with
Sign up | Sign in

Acceleration Through Prediction

Seagate: Winning the Battle of the Boot Drives
By
Brought to you by What's this

Users’ data patterns are predictable. The trick, then, is to predict them. While some very complex algorithms can help accomplish this, many of the clues used to perform acceleration derive from the user’s actions. Chances are high that if a user accesses the same file two or three times in a short while, he or she will need it again in the near future. Those files should probably be copied from disk and cached in the faster NAND media. This is a large part of what Seagate Adaptive Memory™ technology does. Adaptive Memory is a self-learning system that works with the operating system to monitor file use and leverage the drive’s solid state memory to optimal benefit. This caching encompasses both the OS files needed for booting and hibernation as well as the data used most frequently during operation.

The following diagram illustrates how Adaptive Memory technology identifies which data to cache in NAND.

Adaptive Memory technology is specific to Seagate SSHDs. Other hybrid vendors will have their own approaches to NAND utilization. Regardless of the specific data handling schemes, though, this study by Seagate should make it clear that relatively little flash memory is needed to accommodate the acceleration needs of most consumers and business workers.

However, there could be a catch. The NAND cells comprising the SSD side of a hybrid drive will wear out over time. All NAND cells do. And the smaller the amount of NAND, the faster those cells will degrade, all other things being equal. Thus when Seagate announced that it would use 8GB of NAND in its hybrids—a far cry from the 64GB to 256GB now found in mainstream consumer SSDs—many worried that the drives might fail prematurely. Seagate determined to answer the question once and for all.