Home Box Office, Inc.

Jonathan David Lutz

Allen Arthur Gay

Sata Busayarat

Dylan Carney

Brandon C Furtwangler

Steven N Furtwangler

Mei Qi

Jack Song

Brandon S Furtwangler

The described technology is generally directed towards maintaining content selection graphs in an in-memory content selection graph data store in association with respective start times that indicates when the respective graphs become active. When a request to return content selection data is received, an active graph that corresponds to the request and the current time is accessed to obtain the requested content selection data. The response data can be prebuilt, e.g., in a set of active graphs for different client types, so that the response can be returned generally as is from the active graph in the set for that particular client type. A Redis cache can be used to maintain the various graph sets, including the active graph sets and graph sets that will become active at a future time.

Preloaded content selection graph for rapid retrieval

The described technology is generally directed towards coordinating the generation, validation and enabling of content selection graphs in an in-memory content selection graph data store. When a set of content selection graphs is requested, a coordinator starts the generation of the relevant graphs. Upon successful generation, the coordinator starts a validation of the generated graphs against rules for the nodes/response data in the graphs. If the generated graphs pass validation, the coordinator enables the graph set for use in an in-memory cache, whereby when a request to return content selection data is received, an active graph that corresponds to the request and the current time is accessed to obtain and return the response data as the requested content selection data.

Coordinator for preloading time-based content selection graphs

The described technology is generally directed towards generating content selection graphs in an in-memory content selection graph data store in association with respective start times that indicate when the respective graphs become active. When a content selection graph is needed for a given timepoint, a node identifier is used to obtain response data that is cached for that node identifier and timepoint. Edge data associated with the node identifier is used to determine a next node for obtaining its response data for caching, and so on. The response data can be prebuilt, e.g., generated in a set of active graphs for different client types, so that a response can be returned generally as is from the active graph in the set for the node identifier and that particular client type. For efficiency, operations can be used to avoid processing a node identifier more than once when generating a graph.

Preloaded content selection graph generation

The described technology is generally directed towards causing the generation of a content selection graph (or set of graphs) upon receipt of a notification that a new graph or replacement graph is needed with respect to a starting timepoint. For a new graph, a timepoint for that new graph indicates the need for the new graph at a given starting time. For a replacement graph, a notification (subscribed from a monitoring service) can indicate that an existing graph has changed; a replacement graph is generated with a graph identifier of the replacement graph, which then replaces the graph identifier of the existing graph in the mappings of valid graphs to start times. A Redis cache can be used to maintain the various graph sets, including the active graph sets and graph sets that will become active at a future time.

Collection of timepoints and mapping preloaded graphs

The described technology is generally directed towards garbage collecting content selection graphs and related data from in an in-memory content selection graph data store. When a set of content selection graphs expire, a more current content selection graph set becomes active, and the storage space (e.g., in a Redis cache) used by the expired content selection graphs is reclaimed via garbage collection. Some graphs can be replaced before use, referred to as orphaned graphs, and the storage space for any such orphaned graphs is also reclaimed during garbage collection. Also garbage collected is storage space including related data structures used to generate and validate graphs.

Garbage collection of preloaded time-based graph data

The described technology is directed towards locating and using a template for processing data item data from a general form into a client-specific form for returning in response to a client request. A client request includes a data item identifier and client-specific information. The data item's identifier is processed into a data type and/or identifier, e.g., a string. The client-specific information is used to determine a device class, device type and/or client platform software version. The template is found in a hierarchy of templates based upon the client-specific information and the data type or data ID string, e.g., the client-specific information may be used to determine a subset of template folders that is evaluated to find a file with a filename matching the string. The folders may be ordered from most-specific to least-specific, so as to locate the most specific template file that applies.

Templating data service responses

The described technology is generally directed towards validating content selection graphs for use in an in-memory content selection graph data store. When a content selection graph data is generated, the graph nodes can correspond to prebuilt response data. The response data for any of the nodes can be verified against rules based on the type of a node, for example, as well as based on client-specific information for types of clients. The overall validation process can validate the content selection graph for subsequent use, can result in errors or warnings being logged (which may or may not render the graph unusable, e.g., depending on severity), or can fail the content selection graph. If validated, the content selection graph can be used to rapidly return response data when content from the graph is requested.

Preloaded content selection graph validation

The described technology is directed towards returning less data than is available for a data item in response to a request to a data service. A virtual graph node is returned in response to client requests, in which the virtual node comprises a relatively lightweight set of information relative to the full set of information for the data item, e.g., maintained in a main (graph) node. A requesting client indicates that a virtual node is desired, and receives a response comprising the virtual node, generally processed from the main node's data into a reduced subset of the main node. The main node may be cached at the data service, and returned if and when requested.

Virtual graph nodes

The described technology is directed towards a data transformation pipeline architecture of a data service that processes generalized datasets into datasets (e.g., video data or graph nodes) customized for a particular client device. Described herein is maintaining a set of data transformation models at a data service, and upon receiving a client request for data, selecting a relevant subset of the transformation models and arranging the subset into a data transformation pipeline. In general, the pipeline of transformation models transforms the generalized data into the format and shape that each client device expects. The subset may be selected based upon device type, device class and/or software version information (and possibly state data) sent with each data request. The transformation models may be maintained in a hierarchical data store such as files in a file system to facilitate retrieval by searching the hierarchy for appropriate transformation models.

Data delivery architecture for transforming client response data

The described technology is directed towards a cache framework that accesses a tier of ordered caches, in tier order, to satisfy requests for data. The cache framework may be implemented at a front-end service level server, and/or a back end service level server, or both. The cache framework handles read-through and write-through operations, including handling batch requests for multiple data items. The cache framework also facilitates dynamically changing the tier structure, e.g., for adding, removing, replacing and/or reordering caches in the tier, e.g., by re-declaring a data structure such as an array that identifies the tiered cache configuration.

Multitier cache framework

The described technology is directed towards maintaining a cache of data items, with cached data items having current value subsets and next value subsets. The cache is accessed for data item requests, to return a cache miss if a requested data item is not cached, to return data from the current value subset if not expired, to return data from the next value subset if the current value subset is expired and the next value subset is not expired, or to return a cache miss (or expired data) if both subsets are expired. Cached data items are refreshed, (e.g., periodically), when a data item's current value subset is expired by replacing the data item's current value subset with the next value subset and caching a new next value subset, or caching a new next value subset when the next value subset will expire within a threshold time.

Growing community of inventors

Seattle, WA, United States of America

Jonathan David Lutz