Re:Big data is dead (2023)
by giantrobot on Monday May 27, 2024 @01:00PM
(#40492555)
They're likely record based formatting rather than file based. At the high level the code is just asking for a record number from a data set. The data set is managed including redundancy/ECC by the hardware of that storage device.
I wonder how they dealt with common storage issues like backups and disks having bad sectors.