Walmart Apollo, LLC

Wal-Mart Stores, Inc.

Robert Perry Lowell

Jason Christopher Sardina

Alexei Olkhovskii

William R Eschenbruecher, Iii

Scott Melvin Harvester

Yi Lu

Nitin Chhabra, Iii

Michael Machado

Ying Zhang

Jun Yi

Robert Bruce Woods, Iii

A system can comprise one or more processors; and one or more non-transitory computer-readable media storing computing instructions configured to run on the one or more processors and perform: receiving a request for a write operation of an input record in a data store associated with a sharded database and an alternate-key-global-index (AKGI) database; generating a new optimistic lock value, the new optimistic lock value being unique in the sharded database; when the data store does not include a data record associated with the input record, creating the dummy data record in the data store; locking the data record for the write operation by setting the DROpLock of the data record to the new optimistic lock value; and performing the write operation of the input record in the AKGI database and the sharded database. In this embodiment, performing the writing operation further comprising inserting or updating an index record derived from the data record before updating the data record and locking the index record before inserting or updating the index record. Other embodiments are disclosed.

System and method for management of a database system

A method for achieving eventually-consistent data replication in a technology agnostic manner. The method can be implemented via execution of computing instructions configured to run at one or more processors and stored at one or more non-transitory computer-readable media. The method can include receiving, by a system comprising the one or more processors and the one or more non-transitory computer-readable media, a write request for an object. The write request can be received from a user device or an application server. The method further can include performing a write operation for the object, by the system, after the write request is received. Performing the write operation can comprise: (a) adding an intent record to a tracker table, the intent record comprising a primary key uniquely associated with the object and a version associated with the object; and (b) updating the object, the object including the primary key and the version, in a primary data store, after the intent record is added. The method further can include reconciling, by the system, each intent record that is stranded in the tracker table. Reconciling each intent record that is stranded in the tracker table can comprise, for a respective stranded intent record of the one or more stranded intent records, (a) attempting to obtain, from the primary data store, a primary object that is uniquely associated with a stranded primary key of the respective stranded intent record; (b) when (A) the primary object is obtained and (B) a primary object version of the primary object is equal to a stranded intent version of the respective stranded intent record, replicating the primary object from the primary data store to at least one secondary data store; (c) when the primary object is not obtained, deleting the object from the secondary data store; and (d) when at least one of: (A) the primary object is not obtained, (B) the primary object version of the primary object is not equal to the stranded intent version of the respective stranded intent record, or (C) the primary object from the primary data store is replicated to the at least one secondary data store, clearing the respective stranded intent record from the tracker table. Other embodiments are disclosed.

Technology agnostic system and method for achieving eventually-consistent data replication

Example systems and methods for integrating a sharded primary data store (e.g., a source-of-truth relational database management system), a secondary data store (e.g., external cache) and an external global index are described. The approach implements a modified read-through/write-through data access pattern that integrates read and write flows in order to support high-concurrency environments while maintaining immediate consistency between all three stores. Writes are handled using a three-phase flow that avoids concurrency-related race conditions and the need to block in the secondary store in order to maintain cross-store consistency. Reads are never dirty and will repair the secondary store as needed, presenting an immediately consistent view of data to application consumers.

Pattern for integrating primary and secondary data stores in a sharded data domain

Example systems and methods for integrating a primary data store (e.g., a source-of-truth relational database management system) and a secondary data store (e.g., external cache) are described. The approach implements a modified read-through/write-through data access pattern that integrates read and write flows in order to support high-concurrency environments while maintaining immediate consistency between stores. Writes are handled using a three-phase flow that avoids concurrency-related race conditions and the need to block in the secondary store in order to maintain cross-store consistency. Reads are never dirty and will repair the secondary store as needed without conflicting with repairs that may have been initiated by other sessions.

Integrating database management system and external cache

Growing community of inventors

Carlsbad, CA, United States of America

Robert Perry Lowell