Obviousness

Obviousness Analysis of US Patent 10701173 under 35 U.S.C. § 103

This analysis identifies combinations of prior art references that would render the claims of US Patent 10701173 obvious to a person having ordinary skill in the art (POSA). The core innovation of US10701173 resides in the "late binding" of caching policies, where the caching policy is determined at request time rather than at fill time, to ensure the latest policy is applied without forcing invalidations.

Prior Art References

The patent US10701173 itself lists the following as prior art:

• U.S. Published Patent Applications Nos. US 2013-0159472 and US 2013-0159473, both filed Dec. 12, 2012; and No. US 2014-0344399 A1, filed Jun. 17, 2014; and U.S. Pat. No. 6,185,598, filed Feb. 10, 1998.

Combinations and Motivation for Combination

A POSA in content delivery networks would possess knowledge of existing caching mechanisms, content delivery strategies, and methods for managing content policies. The motivation to combine references would stem from a desire to improve efficiency, ensure content freshness, and reduce operational overhead in CDNs.

Combination 1: US 6,185,598 in view of general knowledge of dynamic content and policy management

• US 6,185,598 (Farber et al.): This patent describes a computer network with origin servers and repeaters (caches). It explicitly states that when a repeater gets a client's request, it responds by returning the requested resource. If the repeater has a local copy, it returns that copy; otherwise, it forwards the request to the origin server to get the resource and saves a local copy. This demonstrates a fundamental caching mechanism within a network. The patent also mentions that "static" and "dynamic" are relative terms, and "static resources may change at some regular, albeit long, interval."

• Obviousness Argument: A POSA, aware of the teachings of US 6,185,598 regarding caching and the dynamic nature of content, would be motivated to ensure that cached content remains fresh and adheres to current policies. While US 6,185,598 describes a basic caching process where a repeater serves a local copy if available, it doesn't explicitly detail how caching policies are dynamically applied at request time to account for potential updates.

  The motivation for a POSA to combine this with general knowledge of dynamic policy management would be to address the problem of stale content and inefficient invalidation mechanisms. If a caching policy (e.g., TTL) for a resource changes, simply relying on the initial fill time (as implied by US 6,185,598's description of serving a local copy) could lead to serving outdated content. A POSA would understand the need to re-evaluate the caching validity more frequently or dynamically. Therefore, a POSA would be motivated to introduce a mechanism to check the current caching policy at the time of a request to determine if the locally cached content is still valid. This would inherently lead to the concept of "late binding" of caching policies to avoid serving stale content without immediately resorting to expensive invalidations across the entire CDN.

Combination 2: US 2013-0159472 / US 2013-0159473 / US 2014-0344399 A1 in view of US 6,185,598 and existing CDN architectures.

• US 2013-0159472, US 2013-0159473, US 2014-0344399 A1: While the full text of these applications is not provided in the prompt, US10701173 itself states that "Exemplary CDNs are described in U.S. Published Patent Applications Nos. US 2013-0159472 and US 2013-0159473... and No. US 2014-0344399 A1". This indicates that these documents describe established CDN architectures and services, which would inherently include aspects of content delivery, request handling, and potentially content caching, even if the specifics of dynamic policy application are not detailed.

• Obviousness Argument: Assuming these references disclose conventional CDN operations, including the use of caching nodes and handling of content requests, a POSA would look to optimize their performance and content freshness. Combining the general CDN frameworks described in these applications with the basic caching principles of US 6,185,598 would be a natural step.

  The explicit problem addressed by US10701173 is the challenge of "invalidating cached content... every time a caching policy for that content is set" and instead determining "the caching policy of requested content at request time rather than at fill time." This problem-solution approach, in the context of known CDN architectures (from the 2013 and 2014 applications) and basic caching (from US 6,185,598), suggests an obvious modification. A POSA encountering the inefficiencies of widespread invalidation in a complex CDN would naturally seek a more granular and dynamic way to manage caching. The idea of "late binding" the policy at the time of request is a logical step to achieve this, allowing for immediate reflection of policy changes without the overhead of full invalidations. This would involve storing policies separately (e.g., in a database as described in US10701173) and querying them at request time, which is a common approach for dynamic configurations in distributed systems.

Conclusion

The core inventive step of US10701173—determining the current cache policy at request time (late binding) to avoid frequent invalidations—appears to be an obvious combination of existing caching techniques (as shown in US 6,185,598) and general knowledge of dynamic content management within the context of Content Delivery Networks (as broadly described by the other cited prior art applications). A person having ordinary skill in the art would be motivated to improve the efficiency and content freshness of CDNs by implementing a dynamic policy evaluation mechanism at the point of content delivery, rather than relying solely on static or fill-time policies. The separate storage and retrieval of policies (e.g., from a database/RuleBase as described in US10701173) is a conventional method for managing dynamic configurations in distributed systems and would be readily adopted by a POSA to implement late binding.