System research Google storage I/O traces synthesised
Google just provided synthesized Google I/O traces for storage servers and discs to help researchers
I/O traces of storage devices and servers’ input/output processes are essential for studying storage behavior and performance
Designing large distributed storage systems requires representative I/O modelling
Google Cloud created Thesios to build accurate and representative I/O traces
The capacity to synthesise counterfactual I/O traces for data-driven “what-if” research is unique to Thesios
Using different workload filtering criteria to form hot and cold discs and analyzing their power consumption effects
This trace includes user-facing and internal application I/O operations
They think these traces are a unique chance to explore large-scale storage and make real progress
For more details govindhtech.com