US 8,019,882 B2 · Content identification for peer-to-peer content retrieval

Abstract

Described is a technology in which client content requests to a server over a wide area network (WAN) are responded to with hash information by which the client may locate the content among one or more peer sources coupled to the client via a local area network (LAN). The hash information may be in the form of a segment hash that identifies multiple blocks of content, whereby the server can reference multiple content blocks with a single hash value. Segment boundaries may be adaptive by determining them according to criteria, by dividing streamed content into segments, and/or by processing the content based on the content data (e.g., via RDC or content/application type) to determine split points. Also described is content validation using the hash information, including by generating and walking a Merkle tree to determine higher-level segment hashes in order to match a server-provided hash value.

Claims18

Dependent claims

In a computing environment, a method comprising:

receiving a request for content from a network client coupled via a wide-area network;
and
returning hash information to the client in response to the request, the hash information comprising segment information including at least one segment hash that is useable to identify at least one content block within a segment of blocks that includes the content, at least some of the content being maintained at a peer machine coupled to the client over a local area network, the at least one segment hash being further useable with a content hash of a first content block to validate the first content block, the first content block being maintained at the peer machine, the at least one segment hash being based on content of the segment, the content hash being based on content of the first content block.

One or more computer storage media having computer-executable instructions, which when executed perform steps, comprising:

requesting content from a server coupled over a wide-area network;
and
receiving hash information in response to the request, the hash information containing segment information, the segment information including a segment hash for a segment of content;
using the segment information to locate a source of the content that is coupled to a requester of the content via a local area network;
receiving the content from the source, the content including a first content block, the segment including the first content block;
and
validating the content using the segment hash and a first content hash of the first content block, validating the content including:
(a) generating a Merkle tree having a root corresponding to a server segment hash provided in the segment information,
(b) computing a first block hash of the content;
(c) using the first hash and a second hash related to the content to determine a current higher-level segment hash;
and
(d) using the current higher-level segment hash for validation, including:
(i) if the current higher-level segment hash does not correspond to a root level, using the current higher-level segment hash and another segment hash at a same level in the tree to determine a new higher-level segment hash, setting the current higher-level segment hash to the new higher-level segment hash, and returning to step (d), or
(ii) if the current higher-level segment hash does correspond to the root level, comparing the current higher-level segment hash with the server segment hash to determine whether the content is valid.

11 more elements

Dependencies are inferred from the claim text.

Content identification for peer-to-peer content retrieval

Life of this patent

Content identification for peer-to-peer content retrieval

Life of this patent

Abstract

Claims18