Abstract: A real-time transaction processing system and method include the provision of a real-time transactional data store having an importer subsystem embedded therein. The importer subsystem is configured to connect directly to a real-time data source without using an intermediate subsystem. Events are received via the real-time data source at the importer subsystem, which transacts the events into the real-time transactional data store to create one or more records in the real-time transactional data store corresponding to the events.
Type:
Grant
Filed:
December 13, 2016
Date of Patent:
April 23, 2019
Assignee:
VoltDB, Inc.
Inventors:
Stefano M. Santoro, Anish Khanzode, Ryan Betts
Abstract: Methods and systems are disclosed for real-time transactional database transformation implemented as part of a real-time transactional database management system.
Abstract: Database replication is performed by tracking sequence numbers associated with binary change logs. At a producer cluster, a binary change log is generated based on a write transaction made to a database partition in the producer cluster. Included with the binary change log is a sequence number corresponding to the write transaction. The binary change log is transmitted from the producer cluster to a consumer cluster, where a determination is made whether to apply the binary change log based on the included sequence number. The binary change log is then applied to a database partition in the consumer cluster. The sequence number is stored at the consumer cluster by combining the sequence number with a numerical range having a first number and a second number, where the second number has a value corresponding to the most recent binary change log applied to the database partition in the consumer cluster.
Type:
Grant
Filed:
September 14, 2017
Date of Patent:
September 4, 2018
Assignee:
VoltDB, Inc.
Inventors:
Ning Shi, Walter Weiss, Yang Lu, Rui Shen, Manju James
Abstract: A distributed shared-nothing database provides serializable isolation for transactions and includes a mechanism for adding storage and processing capacity to the database without stopping the database from processing transactions.
Type:
Grant
Filed:
November 20, 2014
Date of Patent:
May 2, 2017
Assignee:
VOLTDB, INC.
Inventors:
Ariel D. Weisberg, Ning Shi, Steven Z. Cooper, Stefano M. Santoro
Abstract: Methods and systems are disclosed for detecting data divergence or inconsistency across replicas of data maintained in replica nodes in a shared-nothing distributed computer database system. The replica nodes communicate with a coordinator node over a computer network. The method includes the steps of: (a) receiving an operation at the coordinator node; (b) transmitting the operation to the replica nodes to be executed by each replica node to generate an operation result and a hash representation of the operation or of the operation result; (c) receiving the operation result and the hash representation generated by each of the replica nodes; and (d) determining whether the operation resulted in data divergence or inconsistency by detecting when the hash representations received from the replica nodes are not all the same.
Abstract: A distributed shared-nothing database provides serializable isolation for transactions and includes a mechanism for adding storage and processing capacity to the database without stopping the database from processing transactions.
Type:
Application
Filed:
November 20, 2014
Publication date:
July 2, 2015
Applicant:
VoltDB, Inc.
Inventors:
Ariel D. Weisberg, Ning Shi, Steven Z. Cooper, Stefano M. Santoro
Abstract: Methods and systems are disclosed for real-time transactional database transformation implemented as part of a real-time transactional database management system.
Abstract: Methods and systems are disclosed for detecting data divergence or inconsistency across replicas of data maintained in replica nodes in a shared-nothing distributed computer database system. The replica nodes communicate with a coordinator node over a computer network. The method includes the steps of: (a) receiving an operation at the coordinator node; (b) transmitting the operation to the replica nodes to be executed by each replica node to generate an operation result and a hash representation of the operation or of the operation result; (c) receiving the operation result and the hash representation generated by each of the replica nodes; and (d) determining whether the operation resulted in data divergence or inconsistency by detecting when the hash representations received from the replica nodes are not all the same.