Viasat, Inc.

David Lerner

Peter Lepeska

Devin Toth

Douglas C Larrick

Eric E Prouty

Eric Prouty

Douglas Larrick

Embodiments seek to protect privacy of potentially sensitive client resources in web transactions using crowd-disambiguation. Crowd-disambiguation machines can aggregate information about resources from multiple clients as resource fingerprints, and can use the fingerprints to provide crowd-sourced services in a privacy-protected manner. For example, embodiments can communicate a resource fingerprint as a fully ambiguated resource instance (FARI) and a partially disambiguated resource instance (PDRI). When one (or few) clients communicates the resource fingerprint, the identity of the resource remains obfuscated from the crowd-disambiguation machine. As more clients communicate fingerprints for the same resource (e.g., identified by the matching FARIs), respective, differently generated PDRIs of those fingerprints enable the crowd-disambiguation machine to resolve further portions of the resource, ultimately permitting the resource to be revealed and considered non-private (e.g., for use in hint generation or other crowd-sourced services).

Machine-driven crowd-disambiguation of data resources

Avoiding redundant data transfers using delta coding techniques when reliably and opportunistically communicating data to multiple user systems. According to embodiments, user systems track received block sequences for locally stored content blocks. An intermediate server intercepts content requests between user systems and target hosts, and deterministically chucks and fingerprints content data received in response to those requests. A fingerprint of a received content block is communicated to the requesting user system, and the user system determines based on the fingerprint whether the corresponding content block matches a content block that is already locally stored. If so, the user system returns a set of fingerprints representing a sequence of next content blocks that were previously stored after the matching content block. The intermediate server can then send only those content data blocks that are not already locally stored at the user system according to the returned set of fingerprints.

Opportunistic content delivery using delta coding

Systems and methods are described for avoiding redundant data transfers using delta coding techniques when reliably and opportunistically communicating data to multiple user systems. According to embodiments, user systems track received block sequences for locally stored content blocks. An intermediate server intercepts content requests between user systems and target hosts, and deterministically chucks and fingerprints content data received in response to those requests. A fingerprint of a received content block is communicated to the requesting user system, and the user system determines based on the fingerprint whether the corresponding content block matches a content block that is already locally stored. If so, the user system returns a set of fingerprints representing a sequence of next content blocks that were previously stored after the matching content block. The intermediate server can then send only those content data blocks that are not already locally stored at the user system according to the returned set of fingerprints.

Embodiments seek to improve web page loading time using server-machine-driven hint generation for based on client-machine-driven feedback. For example, client computers having page renderers are in communication with content servers and hinting processors. The hinting processors can use hinting feedback from multiple page rendering instances to automatically generate hints for optimizing loading and/or rendering of those pages. In some implementations, in response to page requests from the page renderers, content servers can request hints from hinting processors and send those hints to the requesting page renderers for use in improving the page loading experience. In other implementations, in response to page requests from the page renderers, content servers can instruct the requesting page renderers to contact an appropriate hinting processor and to retrieve appropriate hints therefrom for use in improving the page loading experience.

Server-machine-driven hint generation for improved web page loading using client-machine-driven feedback

Embodiments seek to improve prefetch hinting by using automated browsing clusters to generate and update hinting models used for machine-generated hints. For example, hinting machines can include browsing clusters that autonomously fetch web pages in response to update triggers (e.g., client web page requests, scheduled web crawling, etc.) and generate timing and/or other hinting-related feedback relating to which resources were used to load the fetched web pages. The hinting machines can use the hinting feedback to generate and/or update hinting models, which can be used for machine-generation of hints. Some embodiments can provide preliminary hinting functionality in response to client hinting requests, for example, when hinting models for a requested page are insufficient (e.g., unavailable, outdated, etc.). For example, without having a sufficient hinting model in place, the hinting machine can fetch the page to generate preliminary hinting feedback, which it can use to machine-generate preliminary hints.

Hint model updating using automated browsing clusters

Embodiments use crowd disambiguation techniques to protect the privacy of potentially sensitive client resources in web transactions. Crowd disambiguation servers can aggregate information about resources, such as URLs, accessed by clients, in the form of resource fingerprints submitted by the clients. Said resource fingerprints can be used to provide crowd-sourced services in a privacy-protected manner. For example, in some embodiments a fingerprint of a URL visited by a client can be communicated to the server as both a fully ambiguated resource instance (FARI) and a partially disambiguated resource instance (PDRI). When only one client, or a limited number of clients, has communicated a certain resource fmgerprint, the underlying identity of the resource, in this case the URL, remains obfuscated from the crowd disambiguation server, which lacks sufficient information to reconstruct it. As more clients communicate fmgerprints for the same resource (as identified, for example, by the FARIs), the corresponding PDRIs, which are different from client to client, enable the crowd disambiguation server to gradually reconstruct further portions of the resource, ultimately permitting the entire resource to be reconstructed. In that case, the resource is considered non-private, and can be further used e.g., in hint generation or other crowd-sourced services.

Growing community of inventors

Newton, MA, United States of America

David Lerner