Sap Se

Google Inc.

Sybase, Inc.

Simhachala Sasikanth Gottapu

Colin Florendo

Sarika Iyer

Amit Pathak

Nandan Marathe

Anant Agarwal

Prasanta Ghosh

Prateek Agarwal

Jay Gorakhnath Sudrik

Rahul Mittal

Yahong Wang

Mihnea Andrei

Robert Schulze

Anjan Kumar Amirishetty

Yanhong Wang

Adrian Dragusanu

Anthony Hsu

Anup Ghatage

Randall Hammon

Niranjan Nilakantan

Junjie Liang

Aspects of the disclosure are directed to a parallel recovery mode that applies log records while allowing read queries on read-replica databases. The parallel recovery mode can include applying log records in log sequence number (LSN) order for a block or for multiple blocks, and managing log records affecting multiple blocks. The parallel recovery mode can further manage dependency between different log records and maintain transactional consistency on read queries.

Technique for parallel recovery on read replica

Disclosed herein are system, method, and computer program product embodiments for adaptive caching for hybrid columnar databases with heterogeneous page sizes. An embodiment operates by scanning one or more pools comprising one or more pages of the same size in a buffer cache. The embodiment determines an increment of a reuse rate for the pools in the buffer cache within a time interval. The embodiment determines a cumulative reuse rate that is the sum of the increments of the reuse rate over several time intervals. The embodiment determines a gliding average reuse rate of the cumulative reuse rate over several time intervals. The embodiment compares the average reuse rates of the plurality of the pools to a threshold to dynamically determine whether a pool should reuse memory from the existing pages of the same pool or rebalance memory from one or more victim pools.

Adaptive caching for hybrid columnar databases with heterogeneous page sizes

Disclosed herein are system, method, and computer program product embodiments for non-blocking backup for tertiary initialization in a log replay only node. An embodiment operates by performing a standard log replay on a secondary server and briefly suspending the standard log replay in response to tertiary initialization. Further, the secondary server may determine backup block information and perform a page-aligned backup process from the secondary server to a tertiary server. Additionally, the secondary server may determine log replay block information, and perform a modified log replay concurrently with the backup process based on the backup block information.

Non-blocking backup in a log replay node for tertiary initialization

Disclosed herein are system, method, and computer program product embodiments for adaptive caching for hybrid columnar databases with heterogeneous page sizes. An embodiment operates by receiving a request to load a new page of memory from a disk in a buffer cache. The embodiment scans one or more pools comprising one or more pages of the same size in a buffer cache. The embodiment determines an increment of a reuse rate for the pools in the buffer cache within a time interval. The embodiment determines a cumulative reuse rate that is the sum of the increments of the reuse rate over several time intervals. The embodiment determines a gliding average reuse rate of the cumulative reuse rate over several time intervals. The embodiment compares the average reuse rates of the plurality of the pools to a threshold to dynamically determine whether a pool should reuse memory from the existing pages of the same pool or rebalance memory from one or more victim pools.

Disclosed herein are system, method, and computer program product embodiments for performing fair prefetching. An embodiment operates by splitting a data vector into a first subrange and a second subrange. The embodiment performs a first chance prefetch operation on the first subrange based on a fixed number of pages, thereby loading a set of pages of the first subrange into a main memory. The embodiment performs the first chance prefetch operation on the second subrange based on the fixed number of pages, thereby loading a first set of pages of the second subrange into the main memory. The embodiment performs a second chance prefetch operation on the second subrange based on the performing the first chance prefetch operation on the second subrange, thereby loading a second set of pages of the second subrange into the main memory. The embodiment then executes the query.

Fair prefetching in hybrid column stores

A system for an in-memory row storage architecture can be provided. In some implementations, the system performs operations including processing a database statement comprising a first database transaction affecting one or more rows in an in-memory row store, registering the database statement at a start of execution of the database statement, the start of execution occurring at a first time, unregistering the registered database statement at an end of execution of the database statement, determining a second time at which execution of an oldest currently registered database statement was started, assigning a garbage collection thread to a second database transaction committed at a third time and affecting at least one row of the plurality of rows, and activating the garbage collection thread to reclaim memory within the in-memory row store when the third time is less than the second time. Related systems, methods, and articles of manufacture are also described.

Garbage collection for in-memory row storage

A transaction descriptor associated with a vertical chain of row versions is received. The vertical chain of row versions is traversed. The vertical chain is part of a grid structure formed by a number of vertical chains intersected with a number of horizontal chains. A link to a current row version is terminated. A link from the current row version to an older row version in a horizontal chain is locally stored and terminated. The older row version is set as ready for garbage collection. The current row version is set as ready for garbage collection. A link from the current row version to a next row version in the horizontal chain is locally stored and terminated. The next row version is appointed as current.

Non-blocking parallel memory mechanisms

Growing community of inventors

Dublin, CA, United States of America

Simhachala Sasikanth Gottapu