Huawei MapReduce Service

Exabyte-scale fully managed data warehouse (formerly ODPS) optimised for batch analytics on structured and semi-structured data, with SQL, MapReduce, and Spark computation engines, columnar storage, and tight integration with DataWorks

Serverless interactive query service that runs ANSI SQL directly on S3 data lakes with per-query pricing, Apache Iceberg and Hudi table formats, and Glue catalog integration

Managed big-data platform for running Apache Spark, Hive, Presto, Flink, Trino, and HBase across EC2, EKS, and fully serverless deployments with up to 5x faster Spark runtime and Graviton price-performance

Unified analytics service combining data warehousing and big data processing with dedicated and serverless SQL pools, Apache Spark, Data Integration pipelines, and Power BI embedded

First-party Microsoft-co-engineered deployment of the Databricks Lakehouse with native Entra ID sign-in, private networking, serverless SQL warehouses, Unity Catalog governance, Delta Lake storage, and Azure Portal billing — the sanctioned way to run Databricks inside an Azure subscription

Managed open-source analytics clusters for Hadoop, Apache Spark, Apache Hive LLAP, Apache Kafka, and Apache HBase with enterprise security via Enterprise Security Package, autoscale, and integration with ADLS Gen2 — used mainly for migrating existing OSS big-data estates into Azure

Unified data lakehouse combining the reliability of data warehouses with the scale of data lakes, built on Delta Lake with ACID transactions, schema enforcement, and time travel

Managed Apache Spark and Hadoop (formerly Dataproc and Serverless for Apache Spark), offered as per-second-billed clusters on Compute Engine or as serverless Spark batches and sessions, with Flink, Trino and other open-source components.

Managed Hadoop ecosystem clusters (Hadoop, Spark, Hive, HBase and more) for big-data processing

Managed Hadoop and Spark clusters built on Oracle Distribution including Apache Hadoop (ODH), with Hive, HBase, Spark, Trino and Ambari management, highly available master/worker/edge nodes and OCI integrations.

Serverless managed Apache Spark — submit a Spark job, OCI runs it on transient infrastructure, returns the result. No cluster to manage; pay per OCPU-second of execution.

Unified data platform, based on ForePaaS, combining connectors, an ETL/ELT processing engine, an Apache Iceberg lakehouse, a Trino-based analytics and dashboard manager, and serverless app services. Compute is billed per vCPU-hour and data scanned; there is a free tier.

Managed big-data platform running Apache Spark, Hadoop, Hive, HBase, Flink, Presto, and ClickHouse with autoscaling, Kerberos authentication, integration with COS for compute-storage separation, and Jupyter notebooks for interactive analysis