Obviousness

For an obviousness analysis under 35 U.S.C. § 103, the focus is on whether the claimed invention as a whole would have been obvious to a person having ordinary skill in the art (PHOSITA) at the time of the invention, given the prior art. This requires identifying a motivation to combine prior art references and a reasonable expectation of success from that combination.

Based on the provided patent text for US9762692B2 and its cited prior art, here's an analysis of potential obviousness combinations for the independent claims:

General Principles for Obviousness:
An obviousness rejection requires a clear articulation of the reasons why the claimed invention would have been obvious. The rationale to modify or combine prior art does not have to be explicitly stated in the prior art; it can be reasoned from common knowledge, scientific principles, or legal precedent. Examiners are cautioned against using per se rules and must explain how the rationale applies to the facts. All claim limitations must be considered when assessing patentability.

Prior Art References (from the patent document's "Cited By" section, which also includes "Families Citing this family"):

The patent lists several "Cited By" and "Families Citing this family" references. While "Cited By" references typically point to later patents that cite US9762692, "Families Citing this family" often includes prior art that was considered during prosecution or is relevant to the field. For the purpose of obviousness, we are interested in prior art references with priority dates before April 4, 2008 (the priority date of US9762692).

Let's look at some potentially relevant prior art from the "Families Citing this family" section with priority dates before April 4, 2008:

• US7860964B2 (Priority 2001-09-28, Publication 2010-12-28, Assignee: Level 3 Communications, LLC): Titled "Policy-based content delivery network selection." This patent, assigned to the original assignee of US9762692, is highly relevant as it deals with CDNs and policy-based content delivery. It describes selecting a CDN based on various factors.
• US9167036B2 (Priority 2002-02-14, Publication 2015-10-20, Assignee: Level 3 Communications, LLC): Titled "Managed object replication and delivery." This also by Level 3, focuses on replication and delivery within a CDN.
• US8028090B2 (Priority 2008-11-17, Publication 2011-09-27, Assignee: Amazon Technologies, Inc.): Titled "Request routing utilizing client location information." Although its priority date is later than US9762692's earliest priority date, it shows the state of the art around that time regarding intelligent routing. We should prioritize prior art with earlier priority dates. Let's look for others.
• US7991910B2 (Priority 2008-11-17, Publication 2011-08-02, Assignee: Amazon Technologies, Inc.): "Updating routing information based on client location." Similar to the above.
• US7970820B1 (Priority 2008-03-31, Publication 2011-06-28, Assignee: Amazon Technologies, Inc.): "Locality based content distribution." This has a priority date close to US9762692's and is directly relevant to content distribution in a CDN based on location.
• US8601090B1 (Priority 2008-03-31, Publication 2013-12-03, Assignee: Amazon Technologies, Inc.): "Network resource identification." This also has a close priority date.
• US8321568B2 (Priority 2008-03-31, Publication 2012-11-27, Assignee: Amazon Technologies, Inc.): "Content management." This has a close priority date and is highly relevant.
• US8533293B1 (Priority 2008-03-31, Publication 2013-09-10, Assignee: Amazon Technologies, Inc.): "Client side cache management." This focuses on caching, a core aspect of CDNs.
• US8156243B2 (Priority 2008-03-31, Publication 2012-04-10, Assignee: Amazon Technologies, Inc.): "Request routing." Directly relevant to routing in a CDN.
• US7962597B2 (Priority 2008-03-31, Publication 2011-06-14, Assignee: Amazon Technologies, Inc.): "Request routing based on class." This could imply different handling for different types of content.
• US8447831B1 (Priority 2008-03-31, Publication 2013-05-21, Assignee: Amazon Technologies, Inc.): "Incentive driven content delivery."
• US8606996B2 (Priority 2008-03-31, Publication 2013-12-10, Assignee: Amazon Technologies, Inc.): "Cache optimization." Highly relevant to managing content in a CDN.

Analysis of Obviousness for Independent Claims:

Claim 1 (Method of content delivery):
This claim focuses on:

1. Obtaining a request at a first server (edge).
2. Checking for local/peer availability.
3. Determining popularity if not available.
4. If popular, first server obtains and serves.
5. If not popular, redirecting to a second server in a distinct intermediate tier, where content is partitioned, and the second server serves.

• Combination 1: US7970820B1 (Locality based content distribution) + US8321568B2 (Content management) + US8606996B2 (Cache optimization)

  • US7970820B1 describes locality-based content distribution, implying a system where content is distributed across a network (similar to a CDN with tiers) and served from locations "close" to clients. This inherently suggests a multi-server architecture with content distribution.
  • US8321568B2 details content management within such a network, which would include the storage and retrieval of content. Given the concept of "long-tail" content, managing content efficiently (e.g., deciding where to store less popular content) would be a known problem.
  • US8606996B2 explicitly addresses cache optimization, which would naturally lead a PHOSITA to consider strategies for managing cached content based on demand or popularity to maximize efficiency and minimize costs. The patent US9762692 itself states that "Content can become popular (by various measures of popularity) or fade into relative obscurity dynamically, so a content library cannot easily be explicitly partitioned. Instead, the CDN tracks popularity of certain content, and selectively migrates content toward the edge (i.e., toward the tier 1 servers) as that content becomes popular." This highlights the known need for dynamic content management based on popularity in CDNs.

  Motivation to Combine: A PHOSITA aiming to optimize content delivery and caching in a CDN (as described in US7970820B1 and US8606996B2) would be motivated to use content management techniques (from US8321568B2) to handle different types of content, including less popular "long-tail" content. It would be obvious to differentiate content based on popularity and serve unpopular content from a less "expensive" or more centralized location (like an intermediate tier or origin) to conserve resources at the edge. The concept of partitioning content across servers in an intermediate tier to distribute load and manage storage efficiently would be a logical extension of general load balancing and data distribution techniques in a networked environment, especially when dealing with large content libraries as described in US9762692. The patent itself highlights that "the partitioning of content libraries across servers in an intermediate tier of a CDN provides a type of preemptive load-balancing such that certain parent servers are only responsible for handling a pre-defined subset of a content provider's library."

Claim 10 (Content delivery framework):
This claim describes a framework with edge servers (first tier), parent servers (second tier), and a popularity service that determines which tier handles requests based on popularity, with distinct portions of the content provider's library logically partitioned across parent servers.

• Combination 1: US7860964B2 (Policy-based content delivery network selection) + US9167036B2 (Managed object replication and delivery) + US8606996B2 (Cache optimization) + US7962597B2 (Request routing based on class)

  • US7860964B2 teaches policy-based selection in a CDN, which could encompass policies related to content popularity or tiering.
  • US9167036B2 describes managed object replication and delivery, directly addressing how content is handled and distributed within a CDN, including considerations for efficiency.
  • US8606996B2 addresses cache optimization, which strongly implies managing content placement based on its anticipated use (popularity).
  • US7962597B2 teaches request routing based on "class." A PHOSITA would readily understand "class" to include popularity (e.g., popular vs. unpopular content).

  Motivation to Combine: A PHOSITA optimizing a CDN for diverse content (including "long-tail" content, which is a known problem in the art as acknowledged by US9762692) would be motivated to combine policy-based selection and managed replication with cache optimization and class-based routing. This combination would lead to a system where content is intelligently distributed and served from appropriate tiers based on its popularity (a "class" of content). The logical partitioning of content across parent servers to manage large libraries and distribute load would be a known engineering solution in large-scale data storage and delivery systems. US9762692 notes that "the partitioning of content libraries across servers in an intermediate tier of a CDN provides a type of preemptive load-balancing". This implies that load-balancing by partitioning content was a known or obvious concept.

Claim 15 (Method of content delivery with selective redirection and hashing):
This claim details a method where a first server in any tier, finding a resource unavailable, selectively redirects the request to a second tier server based on popularity. This second server has a partitioned portion of the library, and a hash function is used to identify which specific second-tier server stores the content.

• Combination 1: US7860964B2 (Policy-based content delivery network selection) + US8156243B2 (Request routing) + US8606996B2 (Cache optimization) + any reference teaching basic hashing for data distribution (e.g., for load balancing or data partitioning, which are common in distributed systems)

  • US7860964B2 provides the framework for policy-based decisions within a CDN, which would include popularity.
  • US8156243B2 describes general request routing in a network, making the redirection of requests a known element.
  • US8606996B2 strongly motivates using popularity for caching decisions.
  • The use of hash functions for distributing data across multiple servers (e.g., for load balancing, consistent hashing, or partitioning large datasets) was a well-known technique in distributed computing and networking prior to 2008. US9762692 itself describes, "The mapping of requests/content to slots is based on some function of the object name and perhaps other information associated with the request for the object. Preferably the mapping of objects to slots is based on a hash or message digest function (such as MAD or the like) over the object name (and preferably including some parts of the query string)." This confirms that hashing for content mapping was a known technique.

  Motivation to Combine: A PHOSITA aiming to implement a robust and scalable CDN with tiered storage, as motivated by the need to handle "long-tail" content efficiently, would combine policy-based content delivery (US7860964B2) and request routing (US8156243B2) with cache optimization based on popularity (US8606996B2). Given the goal of distributing large content libraries across multiple servers in an intermediate tier to alleviate the burden on origin servers and efficiently manage storage (as described in the patent), it would be obvious to use a well-established technique like hashing to consistently and efficiently map a requested resource to its specific storage location among the partitioned servers in that tier. This ensures that redirected requests land on the correct server holding the desired content portion.