Pipelining Paxos State Machines

A method of pipelining transactions in a distributed database comprising a plurality of replica servers at a plurality of distinct geographic locations, each replica server having one or more processors and memory, the method comprising:

• receiving, at a database management system, a read request to read data;
• determining whether a specific read timestamp is included in the read request;
• when the read request includes the specific read timestamp, determining whether the specific read timestamp is less than a minimum next new write timestamp;
• when the read request does not include the specific read timestamp, assigning, by the database management system, the specific read timestamp that is greater than a last-write timestamp of values of objects to be read and less than the minimum next new write timestamp;
• and
• reading data at the specific read timestamp when the specific read timestamp is less than the minimum next new write timestamp.

A database management system for pipelining transactions in a distributed database including a plurality of replica servers at a plurality of distinct geographic locations, the database management system comprising:

• memory;
• and
• one or more processors in communication with the memory, the one or more processors configured to:
• receive a read request to read data;
• determine whether a specific read timestamp is included in the read request;
• when the read request includes the specific read timestamp, determine whether the specific read timestamp is less than a minimum next new write timestamp;
• when the read request does not include the specific read timestamp, assign the specific read timestamp that is greater than a last-write timestamp of values of objects to be read and less than the minimum next new write timestamp;
• and
• read data at the specific read timestamp when the specific read timestamp is less than the minimum next new write timestamp.

A non-transitory computer-readable medium storing instructions executable by one or more processors for performing a method of pipelining transactions in a distributed database, the method comprising:

• receiving a read request to read data;
• determining whether a specific read timestamp is included in the read request;
• when the read request includes the specific read timestamp, determining whether the specific read timestamp is less than a minimum next new write timestamp;
• when the read request does not include the specific read timestamp, assigning the specific read timestamp that is greater than a last-write timestamp of values of objects to be read and less than the minimum next new write timestamp;
• and
• reading data at the specific read timestamp when the specific read timestamp is less than the minimum next new write timestamp.