Maybe can somebody enlighten why didn't do two actuators on same plates, even second only reading the hdd performance could be easily increased? Because current solution looks like put two hard drives in single enclosure and connected like raid0.
It’s gonna be a difficult task to develop a firmware for the drive to benefit from this advanced hardware design.
In the current-generation spinning HDDs, the firmware has no idea about a file system. It doesn’t even know which sectors are holding data and which ones are empty. The firmware maintains a simple mapping between logical blocks and physical sectors, just a linear mapping + small lookup table for relocated sectors.
For performance reasons, you’d want to balance the data between two halves of the drive, and maintain that balance.
Current-gen SSDs solve somewhat similar problem because they do wear leveling. They do that with custom-designed chips. Just reusing the chip will ruin the performance of this dual-head spinning HDD, because SSDs don’t account for slow head movement latency, flash memory access speed is almost independent of the physical location of the block being accessed.
Technically, it might be easier to solve the problem on the higher level of the stack, i.e. in the file system implementation. But Seagate is not in a position to offer a solution on that level, they sell hard drives, not computers or operating systems.
2 comments
[ 3.5 ms ] story [ 17.8 ms ] threadIn the current-generation spinning HDDs, the firmware has no idea about a file system. It doesn’t even know which sectors are holding data and which ones are empty. The firmware maintains a simple mapping between logical blocks and physical sectors, just a linear mapping + small lookup table for relocated sectors.
For performance reasons, you’d want to balance the data between two halves of the drive, and maintain that balance.
Current-gen SSDs solve somewhat similar problem because they do wear leveling. They do that with custom-designed chips. Just reusing the chip will ruin the performance of this dual-head spinning HDD, because SSDs don’t account for slow head movement latency, flash memory access speed is almost independent of the physical location of the block being accessed.
Technically, it might be easier to solve the problem on the higher level of the stack, i.e. in the file system implementation. But Seagate is not in a position to offer a solution on that level, they sell hard drives, not computers or operating systems.