On September 1, 2009, Oracle announced general availability of the Oracle Database 11g Release 2 version downloadable from the Oracle Technology Network. As the Oracle Essentials: Oracle Database 11g book was published just as Oracle Database 11g Release 1 became generally available, this blog entry will provide you with some quick highlights of a selected number of the new features and capabilities that are in Release 2 and not covered in the book. Of course, in future editions of the book, we'll add these capabilities to the overall context of how the features are typically deployed and used. Nevertheless, I hope those of you familiar with Oracle Database 11g will find this to be a useful list.
Real Applications Clusters Improvements
Oracle Real Application Clusters (RAC) added server pooling for databases and applications where server resources are dynamically assigned to run specific workloads. The resources are policy managed (e.g. minimum and maximum servers and by relative importance) and unassigned servers go into a free pool.
A new intelligent installer requires 40 percent few steps to install RAC. Nodes can be dynamically added or removed from a cluster.
Oracle RAC One Node was introduced as a new Oracle Database Option. RAC Node One is designed to deliver most of the redundancy and availability in RAC to single node instances deployed on a Grid infrastructure. It will typically be deployed in scenarios such as server consolidation and standardized database environments and it supports live migration across the Grid, implementation of rolling patches, server failover within the Grid, and on-line upgrades to multi-node RAC instances.
High Availability and Related Improvements
On-line application upgrades (that don't interrupt availability of those applications) will be enabled through an interesting new capability called edition-based redefinition. "Edition" is a new non-schema object type introduced in Oracle Database 11g Release 2. Application code changes are installed in the privacy of a new edition and changes to data are safely made since Oracle writes them only to the new columns or new tables not seen by the old edition of the application. A cross-edition trigger propagates data changes between editions.
Active Data Guard enhancements in Oracle Database 11g Release 2 include standby query Service Level Agreements configurable from zero to "n" seconds and automatic bad block repair.
Backup and Recovery enhancements include new RMAN compression, tablespace point-in-time recovery, and an Oracle Secure Backup Cloud Module. The Total Recall Option adds schema evolution support.
Data Warehousing and Exadata Improvements
Data warehousing improvements include the addition of an ANSI compliant recursive WITH clause and new analytical functions (Nth value for inter-row calculations and LISTAGG aggregate function).
The Oracle Database optimizer now can derive the degree of parallelism for an individual query (based on the cost of scan operations) using Automated Degree of Parallelism. Parallel statement queuing can decide when to execute a statement given availability of needed parallel servers and is monitored in Enterprise Manager. Automated Degree of Parallelism can also be used to trigger in-memory parallel execution (new in memory capabilities are described below).
New to the Exadata Storage Server Software (first introduced in Oracle Database version 184.108.40.206) is hybrid columnar compression. The columnar organization transparently organizes and stores data by table column increasing compression ratios, typically by 10X for queries, and thus improves scans. Individual row organization is self-contained in compression units so minimal I/O is required to retrieve an entire row. In archival mode, compression ratios of 15-50X are typical. During direct load operations, data is transformed into columnar format and compressed.
Other Exadata Storage Server Software improvements include smart scan offload for the aforementioned hybrid columnar compression, smart scan offload for tables with more than 255 columns, smart scan offload for encrypted tablespaces and columns, and offload to storage of data mining scoring. Enterprise Manager Performance Diagnostics adds SQL monitoring support and I/O performance graphs for Exadata Storage.
Other Performance Improvements
Automatic Storage Management (ASM) will place infrequently accessed data on inner rings of physical disks and places frequently accessed data on outer rings, thus improving performance. ASM is extended to support a general-purpose cluster file system (ACFS) and supports Oracle database files, Oracle binaries, and non-Oracle related files. Read-only snapshots for ACFS are supported (with up to 64 point-in-time copies of file system data).
Oracle Database 11g Release 2 introduces a new In-memory Database Cache Grid that are connected by peer-to-peer communications to each other. Data that can be cached in middle tier resources includes database tables and subsets of rows and columns, making this solution especially ideal where high transaction throughput is required.
Advanced Compression Option improvements include SecureFiles compression with 3x lower CPU utilization and Data Guard Redo Transport compression.
Streams performance was improved providing faster transformations and optimized for one to many propagations. A new extended Streams API (Xstreams) was added.
Data Masking Improvements
Oracle Database 11g Release 2 added a Data Masking Pack that provides out-of-the box mask formats. Supported through Oracle Enterprise Manager, available masking formats provided include condition-based masks, compound masks, integrated clone and mask workflow, and deterministic masks.