abfs

ABFS, short for Azure Blob File System, is a Hadoop-compatible file system driver that enables applications to access data stored in Microsoft Azure Blob Storage using the Hadoop FileSystem API. It provides a bridge between distributed processing frameworks such as Hadoop MapReduce and Apache Spark and Azure Storage, allowing scalable data processing without moving data into a separate shared filesystem.

The ABFS connector is part of the Hadoop ecosystem and is implemented as a FileSystem under the package org.apache.hadoop.fs.azurebfs. It supports two URI schemes: abfs:// for Azure Blob Storage accounts and abfss:// for ADLS Gen2 storage accounts with a secure, authentication-enabled path.

Authentication is supported via multiple methods, including Azure Active Directory OAuth 2.0 (with service principals or managed identities), storage account keys, and Shared Access Signatures. ABFS can be configured in Hadoop and Spark clusters on Azure HDInsight, Azure Databricks, and other environments that run Hadoop-compatible software.

Use cases include data lakes on Azure, large-scale ETL pipelines, and analytics workloads that require direct access to data in Blob Storage without copying to HDFS. Performance and semantics vary with storage tier and account type; ABFS emphasizes compatibility and scalability with cloud storage rather than full POSIX-like filesystem behavior.

See also: Azure Blob Storage, Azure Data Lake Storage Gen2, Hadoop FileSystem, HDInsight, Apache Spark.