Oracle International Corporation

Chih-Ping Wang

Tirthankar Lahiri

Marie-Anne Neimat

John William Miller

Nagender Bandi

Chi-Kim Hoang

Paul Tuck

Mark L McAuliffe

Chi Kim Hoang

Dina Thomas

Rohan Aranha

Susan Sokeng Cheung

Susan Cheng

Dilys Thomas

Ananth Raghavan

John Raitto

Derek Taylor

Kirk Meredith Edson

Subhradyuti Sarkar

Martin Reames

Kao Makino

Sourav Ghosh

Ryder B Rishel

Varadarajan Aravamudhan

Samuel Drake

Mutsumi Kogawa

Jorge Luis Issa Garcia

David Aspinwall

Susan Cheung

Ryder Rishel

John Miller

A database is stored as a plurality of database shards in a distributed database grid comprising a plurality of grid elements, each including a mid-tier database system. A first grid element receives, from an application executing in the same memory as a mid-tier database system of the first grid element, a first database transaction including at least one database operation on specific data stored in a first database shard that belongs to the first grid element. The first grid element performs and commits the first database transaction without participation of another grid element of the plurality of grid elements. The first grid element receives a second database transaction that requires access to another database shard that does not belong to the first grid element. Multiple grid elements of the plurality of grid elements perform the second database transaction and commit the second database transaction using a two-phase commit protocol.

Direct-connect functionality in a distributed database grid

In an in-memory database management system, non-volatile random access memories (NVRAMs) are used to store database data and control data. Because this data is stored in NVRAM, the data survives system failures. Recovery from a system failure may be accomplished more quickly by, at least in part, modifying the surviving data in NVRAM, rather than loading an entire checkpoint image and applying uncheckpointed transactions needed to synchronize the database. Because in this form of recovery the database state that serves as the starting point for applying change records is the database as stored in the NVRAM, this form of recovery is referred to herein as in-memory-based recovery. Recovery, where the database state that serves as the starting point for applying change records is a checkpoint image, is referred to herein as checkpointed-based recovery. In-memory-based recovery eliminates or reduces the need to perform certain operations that are performed for checkpointed-based recovery.

Using non-volatile memory to improve the availability of an in-memory database

Techniques are described for performing dependency locking to enable parallel execution of database server processes modifying the same object. In an embodiment, a DBMS receives a request to execute an operation on its managed database. The request may include an object identifier and a version identifier for a version of a database object, which is required for the successful execution of the operation on the database. The required version of the database object may not exist and may be generated only after the execution of another, second, operation on the database. The first database server process initiates execution of the earlier received operation on the database. The first database server process queries the first database for the version of the particular database object and determines that the version of the database object does not exist yet, in one embodiment. Based on such a determination, a dependency lock is generated for the version of the object, and the first database server process is suspended, preventing the completion of execution of the earlier received operation on the database. Another, second database server process executes another operation on the database and causes the generation of the version of the object on the database. The dependency lock is thereby released for the version of the object, and the first database server process completes the execution of the first operation on the first database.

Dependency locking

In column domain dictionary compression, column values in one or more columns are tokenized by a single dictionary. The domain of the dictionary is the entire set of columns. A dictionary may not only map a token to a tokenized value, but also to a count ('token count') of the number of occurrences of the token and corresponding tokenized value in the dictionary's domain. Such information may be used to compute queries on the base table.

Column domain dictionary compression

A method, apparatus, and system for automatically determining an optimal database subsection is provided. A database subsection is selected to optimize certain benefits when the database subsection is translated, transferred, and cached on an alternative database system, which may utilize a different technology or database engine that provides certain performance benefits compared to the original database system. Algorithms such as multi-path greedy selection and/or dynamic programming may provide optimal or near-optimal results. A host for the alternative database server may be shared with or otherwise located in close physical proximity to improve latency for a database application or client layer. Once the database subsection analysis is completed, a report may be generated and presented to the user, and an implementation script may also be created to automatically configure a client host to function as a cache or replacement system according various cache size configurations described in the report.

Automatically determining an optimal database subsection

A plurality of mid-tier databases form a single, consistent cache grid for data in one or more backend data sources, such as a database system. The mid-tier databases may be standard relational databases. Cache agents at each mid-tier database swap in data from the backend database as needed. Ownership locks maintain consistency in the cache grid. Cache agents prevent database operations that will modify cached data in a mid-tier database unless and until ownership of the cached data can be acquired for the mid-tier database. Cache groups define what backend data may be cached, as well as a general structure in which the backend data is to be cached. Metadata for cache groups is shared to ensure that data is cached in the same form throughout the entire grid. Ownership of cached data can then be tracked through a mapping of cached instances of data to particular mid-tier databases.

Distributed consistent grid of in-memory database caches

A system includes an active node and a standby node and zero or more replica nodes. Each of the nodes includes a database system, such as an in-memory database system. Client updates applied to the active node are written through to the standby node, and the standby node writes the updates through to a primary database and updates the replica nodes. Commit ticket numbers tag entries in transaction logs and are used to facilitate recovery if either of the active node or the standby node fails. Updates applied to the primary database are autorefreshed to the active node and written through by the active node to the standby node which propagates the updates to the replica nodes. Bookmarks are used to track updated records of the primary database and are used to facilitate recovery if either of the active node or the standby node fails.

Database system with active standby and nodes

A replication track is a designated group of transactions that are to be replicated at a destination database in a way that, with respect to any other transaction in the replication track, preserves transactional dependency. Further, transactions in a replication track can be replicated at the destination database without regard to transactional dependency of other transactions in another track. This facilitates concurrent parallel replication of transactions of different tracks. Replicating data in this manner is referred to herein as track replication. An application may request execution of transactions and designate different tracks for transactions.

User-defined parallelization in transactional replication of in-memory database

A plurality of mid-tier databases form a single, consistent cache grid for data in a one or more backend data sources, such as a database system. The mid-tier databases may be standard relational databases. Cache agents at each mid-tier database swap in data from the backend database as needed. Consistency in the cache grid is maintained by ownership locks. Cache agents prevent database operations that will modify cached data in a mid-tier database unless and until ownership of the cached data can be acquired for the mid-tier database. Cache groups define what backend data may be cached, as well as a general structure in which the backend data is to be cached. Metadata for cache groups is shared to ensure that data is cached in the same form throughout the entire grid. Ownership of cached data can then be tracked through a mapping of cached instances of data to particular mid-tier databases.

Growing community of inventors

Palo Alto, CA, United States of America

Chih-Ping Wang