Autonomous Data Warehouse is a database which is tuned and optimized for data warehouse workloads of Oracle Database.
- It is built on Oracle Database, that has automatic Datawarehouse procedures.
- It is easy to use as all the management tasks are automated, all configuration and tuning tasks are fully automated. All data is automatically compressed and encrypted.
- It is fast since its built on Exadata and Oracle database. It also offers instant elasticity on computing and storage dimensions. When it comes to elasticity, the user can choose the exact amount of storage and CPU as needed. Later when more CPU's are required, one can Scale Up or Scale down.
- Machine Learning enables continuous optimization.
- ML in ADW delivers an excellent query performance. Since its built-on oracle database, every business intelligence and Data Integration Services that are compatible with Oracle database supports this service out of the box. For development purpose, existing tools or a newer version of SQL developer (which supports ADW) can be used.
- ADW Patches all software online at all levels (security, OS, network, database) while the system is running.
- SQL plans ‐ same types of SQL queries might run differently based on several dependencies such as data volumes, number of users on the system. ADW does not create indexes on initial instance creation. The indexes are built on the fly after auto query analysis. Index intelligence module drops and creates new indexes on the fly to continuously optimize the query performance which is Auto indexing and auto partitioning.
- It also provides an option for database administrators to override the auto decisions in business scenarios where one would want to give more priority to some critical processes ‐ such as month end close process during the beginning of each month.
Initially when provisioning; the customer will enter the details like datacenter location, CPU and memory requirements, login credentials. Once Provisioned data is loaded. The data is automatically compressed, and all the statistics for the data is created. There is no need to set up infrastructure and manage database tablespace manually. It performs all OS and SYSDBA operations. The errors in SQL are rectified and run automatically. The database is fully secured, protected and encrypted. There will be an auto recovery in case of failure such as storage, networking and compute failures.
Example: In the case of SQL query failure, recovery window is provided for manual restore.
Scale-Up and Scale-Down of compute +Storage.
ADW has the ability to scale‐up /scale down with automated steps. The pricing is based upon CPU ($/CPU/Hour) and storage ($/TB/Months).
Everything in ADW is Compressed using HCC (facts and dimensions). 3.5 to 5 times compression as compared to Amazon RDS as Amazon uses columnar instead of HCC compression. ADW enables huge results caching. Autonomous database results in fewer indexes (by up to 60%) to yield better concurrency support.
Oracle cluster health advisor continuously monitors Real Application Clusters(RAC) databases for performance and availability issues.
Memory Guard is an Oracle Real Application Clusters (Oracle RAC) environment feature to monitor the cluster nodes and prevent node stress caused by the lack of memory.
Hang Manager is an Oracle Real Application Clusters (Oracle RAC) environment feature that autonomously and reliably detects and resolves hangs and deadlocks. It keeps the resources available. Hang Manager is enabled by default.
Quality of Service
Oracle Database Quality of Service (QoS) Management adjusts the system configuration to keep the applications running at the performance levels needed by your business. It monitors and delivers Key Performance Indicators. It provides cluster-wide dashboard and phase in with the measure.
Cluster verification(Cluster Configuration)
Use configuration audit tools Oracle ORAchk and Oracle EXAchk to assess your Oracle Engineered Systems and non-Engineered Systems for known configuration problems and best practices.
Data from different sources and different formats (structured and unstructured) can be processed.
ADW workloads ‐ Data warehouse, Data Mart, Data Lake, Machine Learning