Colossus / GFS to HDFS Mapping

Back to Infrastructure Mapping

Google InternalOpen SourceGCP
SystemColossus / GFSHDFS
ConceptDistributed file systemDistributed file system

GFS (2003) was Google’s original distributed file system. Colossus is the successor. HDFS was directly inspired by the GFS paper and became the storage layer for Hadoop. Colossus underpins nearly all Google storage services internally. No direct GCP equivalent (Colossus is the underlying infrastructure).


google-internal mapping colossus gfs hdfs