Home Box Office, Inc.

Jinjiang Zeng

Dexter Paul Bradshaw

Jeffrey Stamerjohn

Sata Busayarat

Allen Arthur Gay

Mei Qi

Xinkai Wang

Rick Kilgore

Fred Mameri

The described technology is directed towards maintaining and using a version-based hierarchy of software resources (e.g., file system files) to return version-specific responses to clients. A client sends its version information with each data request, and gets back a response based upon that version. Version changes are made by maintaining the current version of each software code resource and overriding the current version with a previous version for clients as needed. The technology allows updates (e.g., for new devices and new software resource versions) to be supported by inserting resources into the resource hierarchy and moving resources therein based upon versioning. A system based on deltas is also contemplated, in which only parts of a file may be changed relative to a different version, instead of overriding the entire file.

Maintaining and updating software versions via hierarchy

The described technology is directed towards presenting a server overload condition that can result from a spike in client requests. Upon receiving an incoming client request for data at a server, a request buffer condition of a client request buffer (e.g., a counter) is evaluated to determine whether the client request corresponds to an anticipated overload state with respect to a request-handling capability of the server, e.g., whether the client request buffer is full. If so, an overload response (e.g., a retry response) is returned to the client indicating so that the client retries a corresponding client request after a time delay. If not in the anticipated overload state, the request is further processed, e.g., to provide a regular response containing the requested data, such as by adding a request information entry associated with the incoming client request to the client request buffer and updating the request buffer condition.

Data service overload detection and mitigation

The described technology is generally directed towards an abstraction and file model used for subscriber service and publisher service communication. The abstraction and file model enables dynamically changing the underlying communication mechanisms and introducing optimizations such as caching (reducing latency) and batching (improving bandwidth utilization) transparently, without any change to the subscriber or publisher code. The abstraction also facilitates introducing intermediate lightweight message-oriented middleware and provides for transparently scaling communication.

Endpoint abstraction for service-to-service communication

The described technology is generally directed towards matching subscriber services to publisher services based on any number of arbitrary subscriber-specified attributes. In general, each publisher services register its capabilities as a set of attributes in a repository data store, e.g., a database. When a subscriber service wants a matching publisher service, the subscriber service specifies which attributes the subscriber service wants matched, and the repository's affiliated matching process uses the registered attributes of the publisher services to seek a match.

Service discovery using attribute matching

The described technology is generally directed towards a subscriber service using configurable data (e.g., a declarative configuration file) data to make service calls to publisher endpoints. Dynamic reconfiguration of the data changes the service call communication parameters without needing to bring down the subscriber server or its service, and/or without needing to modify/rewrite programs for the reconfiguration.

Growing community of inventors

Mercer Island, WA, United States of America

Jinjiang Zeng