and Overview (Sid Wagle) ● Ozone bucket types (Ethan Rose) ○ Legacy, Object Store (OBS), and Filesystem Optimized (FSO) ○ Diverse workloads and when/how to use? ○ Demo ● Ozone Snapshots (Prashant ✓ ofs (Hadoop compatible) ✓ S3 ✓ • Path normalization handled by config key: ozone.om.enable.filesystem.paths • No directories, only prefixes • Existing buckets automatically inherit this layout 25

sec) HDFS 0.572 Ozone With FSO Layout 0.854 Ozone With OBS Layout 12.219 Hive 删除表(Rename操作) ● FileSystem delete on table directory path ● Moves table data to trash 举例: fs.delete(“/catelog_sales”)

sec) HDFS 0.572 Ozone With FSO Layout 0.854 Ozone With OBS Layout 12.219 Hive 删除表(Rename操作) ● FileSystem delete on table directory path ● Moves table data to trash 举例: fs.delete(“/catelog_sales”)

Ozone 可以被 Hadoop 生态中的应用，如 Apache Hive、Apache Spark 和 Mapreduce 无缝对接。Ozone 支持 Hadoop Compatible FileSystem API (aka OzoneFS)。通过 OzoneFS， Hive，Spark 等应用不需要做任何修改，就可以运行在 Ozone上 。Ozone 同时支持 Data Locality，使得计算能够尽可能的靠近数据。