Limelight Networks, Inc.

Amazon Technologies, Inc.

Spencer Proffit

Jesper Mikael Johansson

Tushaar Sethi

Jason Curtis Jenks

Brandon John Smith

Hanson Char

Sean Cassidy

Brandon B Low

Waylon Brunette

Jason Cetina

Soam Acharya

Ahmed A Badran

Michael Asavareungchai

Nicholas Beaudrot

Huw Morgan

A method of using a hardware security module and an adjunct application programming interface to harden tokenization security and encryption key rotation is disclosed. In various embodiments, the method comprises receiving encrypted data at a processor of a computer system, decrypting the encrypted data to cleartext in the processor, and issuing a unique token associated with the data.

Hardening tokenization security and key rotation

A method of managing resources dedicated to a content object in a content delivery network includes transcoding and storing the content object in one or more of a plurality of delivery formats, and associating a transcoding profile with the content object, the transcoding profile including records of the content object with respect to delivery of the content object to end users across the plurality of delivery formats. The method also includes periodically executing a process that determines, utilizing the transcoding profile, an elapsed time since a most recent delivery of the content object in a particular delivery format to one of the end users, and deletes a copy of the content object in the particular delivery format when the elapsed time exceeds a threshold.

Lazy transcoding and re-transcoding of media objects in an online video platform

Methods and systems generate content-categorization hierarchies and assign each content object to a set of nodes in the hierarchies. Each hierarchy includes a series of levels, each level including nodes which represent a categorization. A node in a top level is connected with multiple nodes in a lower level via a direct path. Each content object is assigned to a set of nodes, such that assignment to a lower-level node also results in assignment to each higher-level node in the lower-level node's direct path to a top of the hierarchy. Metadata of the content object is then modified to include a set of tags, each tag being a term from an assigned node. Tags are used to categorize the content object, summarize of the content object, identify recommended content objects, identify content objects responsive to a search query, and/or identify other content objects to supplement the content object.

Content tagging using multi-node hierarchical content assignments

A method and system for performing processing tasks is disclosed. At a resource, a detection is made as to when the resource is available to perform a processing task. Usage of the resource for performing processing tasks associated with each client of a set of clients is monitored. A restriction limiting which processing task is to be assigned to the resource is identified. The restriction identifies a hierarchy amongst at least two clients of the set of clients. The hierarchy is based on the monitored usage. A query identifying the restriction is generated. The query is transmitted to a remote queue in communication with a plurality of independent resources. The plurality of independent resources includes the resource. A response is received from the queue. The response identifies a processing task.

Distributing transcoding tasks across a dynamic set of resources using a queue responsive to restriction-inclusive queries

Growing community of inventors

Vashon, WA, United States of America

Spencer Proffit