TLDR Tech / Data
Scaling Grab's Data Lake: Our journey to Apache Iceberg adoption
Grab successfully migrated petabyte-scale Hive Parquet tables to Apache Iceberg, resulting in a 10x query speedup, a 95% reduction in daily S3 API costs for operations, and 50% compute savings on ML pipelines. They also built a custom UnifiedSparkCatalog to abstract various data lake formats behind a single, compatible interface.
Read the full story