Palantir Technologies Inc.

Brian Toth

Mark Elliot

Michael Harris

Hannah Korus

Brian Schimpf

Lynn Cuthriell

Gustav Brodman

Jacob Meacham

Jonathan Hsiao

Michael Garland

Evelyn Nguyen

Greg DeArment

Bradley Moylan

David Tolnay

Jon Paek

Phillip Chen

Daniel Rothfus

Jonathon Yu

Jonathan Yu

A computer-implemented method comprises detecting, by a processor of a first host of one or more hosts in a distributed computing environment, a distributed task waiting to be started, from a replicated configuration system, the distributed task being represented by a pending tasks key. The method comprises starting, by the processor, the distributed task by performing an atomic compare and swap operation to add a started key to the replicated configuration system. The method also comprises writing a specification of the distributed task to the replicated configuration system under a new version of a current tasks key. In addition, the method comprises removing, following the writing, the pending tasks key from the replicated configuration system.

Fault-tolerant and highly available configuration of distributed services

Fault-tolerant and highly available configuration of distributed services including a computer-implemented method for role-based configuration discovery comprising receiving a request comprising an identifier of a role; identifying a first key, in a replica of a distributed configuration store, comprising a first value that matches the role identifier; identifying one or more other key-value pairs associated in the replica with the first key; and returning a response to an entity that sent the request comprising the value of at least one key-value pair that is specific to the role the service has. Also disclosed are techniques for log forwarding.

A history preserving data pipeline computer system and method. In one aspect, the history preserving data pipeline system provides immutable and versioned datasets. Because datasets are immutable and versioned, the system makes it possible to determine the data in a dataset at a point in time in the past, even if that data is no longer in the current version of the dataset.

Universal data pipeline

Fault-tolerant and highly-available configuration of distributed services

History preserving data pipeline

History preserving data pipeline system and method

Growing community of inventors

Palo Alto, CA, United States of America

Brian Toth