Obviousness

Obviousness Analysis of US Patent 8478903 Under 35 U.S.C. § 103

This analysis evaluates the obviousness of US patent 8478903, titled "Shared content delivery infrastructure," under 35 U.S.C. § 103, considering the provided prior art references from the IPR2026-00095 petition grounds. The patent broadly covers a system where content requests are intercepted at an origin server by a "reflector" and redirected to "repeaters" in a shared content delivery network (CDN) based on dynamic selection criteria.

Identified Prior Art and Relevant Disclosures

The IPR2026-00095 petition grounds list several prior art references: US 6,178,449 B1 (Freedman), US 6,108,703 A (Leong et al.), US 5,850,523 A (Hirata et al.), US 6,434,622 B1 (Armijo et al.), and US 6,161,130 A (Chester et al.).

Upon review of available abstracts and summaries:

• US 6,108,703 A (Leong et al.) - "Global hosting system": This patent is highly relevant to the field of content delivery. It discloses a "network architecture or framework that supports hosting and content distribution on a truly global scale." The system allows a Content Provider to "replicate and serve its most popular content at an unlimited number of points throughout the world" using "a set of servers operating in a distributed manner." These servers are sometimes referred to as "hosting servers" or "ghost servers." Leong et al. explicitly teaches "intelligently routing and replicating content over a large network of distributed servers" to "move content close to the user" and manage large audiences. A significant aspect is serving a "base HTML document portion of a Web page" from the Content Provider's site while "one or more embedded objects for the page are served from the hosting servers, preferably, those hosting servers near the client machine." It also mentions optimizing memory, ensuring sufficient copies, avoiding server overload, and locating copies close to clients.

• Other Cited Prior Art (Freedman, Hirata, Armijo, Chester): Searches for abstracts of US 6,178,449 B1 (Freedman), US 5,850,523 A (Hirata et al.), US 6,434,622 B1 (Armijo et al.), and US 6,161,130 A (Chester et al.) did not yield descriptions related to computer networks or content delivery. The results for these patents indicated inventions in unrelated fields such as electro-optic devices, seismology, and drug administration. Consequently, without further context from the IPR petition explaining their specific relevance, these references cannot be effectively used in this obviousness analysis.

Obviousness Over US 6,108,703 A (Leong et al.)

The independent claims of US 8478903 (Claims 1 and 11) describe a method and system for processing resource requests in a computer network. These claims center on a "reflector" co-located with an "origin server" that intercepts client requests, decides whether to "reflect" the request to a "repeater" (a server in a CDN), selects the "best" repeater, provides the client with a modified resource identifier for the repeater, and the repeater then serves the content, caching it if necessary.

Many of these elements are present or would have been obvious in light of US 6,108,703 A (Leong et al.):

1. Distributed Content Delivery Network: Leong et al. explicitly describes a "network architecture" for "global hosting and content distribution" using "a set of servers operating in a distributed manner" to "replicate and serve its most popular content at an unlimited number of points throughout the world." This directly aligns with the concept of a CDN comprising multiple repeaters.
2. Origin Servers and Content Providers: Leong et al. discusses a "Content Provider" and content being served "from the Content Provider's site" for base HTML documents, which corresponds to the "origin server" concept in US 8478903.
3. Content Replication and Caching at Edge Servers: Leong et al. teaches that content is "replicated and served" by distributed "hosting servers" and that these servers "do not keep copies of all of the content database," implying a caching mechanism where content is fetched from the origin if not locally available. This directly covers the repeaters having a local copy or forwarding the request to the origin as recited in US 8478903.
4. Intelligent Routing and Server Selection: Leong et al. highlights "intelligently routing and replicating content" to locate copies "close to the clients that are requesting them" and to ensure "no server becomes overloaded." This directly anticipates the "best repeater selection" in US 8478903, which considers factors like network distance ("cost groups") and load.
5. Serving Referenced Resources from Distributed Servers: A key disclosure in Leong et al. is that a "base HTML document portion of a Web page is served from the Content Provider's site while one or more embedded objects for the page are served from the hosting servers, preferably, those hosting servers near the client machine." This inherently teaches the serving of referenced resources (like embedded objects) from separate servers (repeaters) distinct from the initial content source (origin server).

Motivation to Combine/Modify

A person having ordinary skill in the art (POSITA) in content delivery networks at the time of US 8478903's priority date (February 10, 1998) would have been motivated to combine or implement the concepts taught in Leong et al. using mechanisms similar to those described in US 8478903.

The explicit objectives of Leong et al. are to improve the efficiency, effectiveness, and reliability of Web content delivery, reduce the need for content providers to build massive infrastructures, and move content closer to users. To achieve these goals, implementing a mechanism to intelligently route requests for replicable content would have been an obvious design choice.

• Implementing "Reflector" Functionality: The concept of an "intelligent routing" system as described in Leong et al. would necessitate an interception and decision-making point at or near the origin server to determine whether a request for a resource, particularly an embedded object, should be handled by the origin or redirected to a distributed "hosting server" (repeater). This interception and decision-making functionality is precisely what the "reflector" in US 8478903 provides.
• Modified Resource Identifiers (URL Rewriting/Redirects): To direct a client from the origin server to a selected repeater for a specific resource, especially for embedded objects, mechanisms such as rewriting URLs within the initial HTML document or issuing HTTP REDIRECT messages were well-known and commonly employed in web technology and proxy server implementations prior to US 8478903's priority date. A POSITA seeking to implement the distributed serving of embedded content, as taught by Leong et al., would readily adopt these conventional techniques to provide the client with a "modified resource identifier designating the repeater."
• Dynamic Repeater Selection: Leong et al.'s emphasis on locating content "close to the clients" and balancing load among servers provides a clear motivation for the dynamic selection of a "best" repeater, considering factors like network proximity and current server load. These metrics and methods for their measurement and use in load balancing were also known in the art.

Therefore, the combination of a system for global content distribution and replication (Leong et al.) with conventional web technologies for request interception, redirection, and URL manipulation would have rendered the method and system of US 8478903 obvious to a POSITA. The specific components (reflector, repeater) and their interactions, including the dynamic selection process, represent obvious implementations of the broader goals and teachings of prior art like Leong et al. in the context of optimizing web content delivery.