Google LLC v. Sandpiper CDN LLC

1. Case Identification

• Case #: IPR2025-00826
• Patent #: 9,021,112
• Filed: April 15, 2025
• Petitioner(s): Google LLC
• Patent Owner(s): Sandpiper CDN, LLC
• Challenged Claims: 1-23

2. Patent Overview

• Title: Content Request Routing and Load Balancing for Content Distribution Networks
• Brief Description: The ’112 patent discloses a method for distributing content over a Content Delivery Network (CDN). The system uses a plurality of CDN Domain Name System (DNS) servers that share a common anycast address to resolve a content hostname, routing a user request to a nearby content server to enable efficient content delivery.

3. Grounds for Unpatentability

Ground 1: Claims 1-8 and 10-23 are obvious over Glines, alone or in view of Gupta.

• Prior Art Relied Upon: Glines (WO 2001/039470) and Gupta (Patent 6,405,252).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Glines taught the core elements of the challenged claims, including a method of content delivery using multiple DNS servers sharing a common anycast address. Glines disclosed redirecting client requests to the closest replica content server by having DNS servers, addressable via a common anycast address, respond to a hostname query with the IP address of a peered content server. Petitioner asserted that Glines disclosed or suggested that each DNS server is associated with a plurality of content servers to handle failover and load balancing. To the extent Glines was not explicit on this point, or on the use of an Internet Service Provider (ISP) DNS server for the initial resolve, Petitioner argued Gupta supplied these missing elements. Gupta disclosed a network of Point of Presence (POP) servers, each with a name server and multiple content servers (WebCache servers), and explicitly taught using an ISP's client DNS server to resolve hostnames.
  • Motivation to Combine: Petitioner contended a person of ordinary skill in the art (POSITA) would combine Glines and Gupta to improve content delivery performance. Both references addressed analogous problems of efficiently delivering content over the internet. A POSITA would have been motivated to apply Gupta’s disclosure of multiple WebCache servers per name server to Glines’s system to enhance scalability, reliability, and load distribution. Similarly, applying Gupta’s use of an ISP DNS server to Glines’s system represented a known and conventional network architecture for resolving hostnames for residential internet users.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because combining the references involved applying known networking principles (e.g., server replication, ISP DNS resolution) to analogous systems to achieve predictable benefits in performance and reliability.

Ground 2: Claim 4 is obvious over Glines, with or without Gupta, in view of Chung.

• Prior Art Relied Upon: Glines (WO 2001/039470), Gupta (Patent 6,405,252), and Chung (Patent 6,470,389).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground focused on the limitation in claim 4 requiring CDN DNS servers and content servers to be "grouped as a cluster." Petitioner argued that Glines and Gupta already taught this concept by disclosing closely peered DNS and content servers that operate together. Chung was introduced to explicitly teach deploying a set of machines as a "cluster" to host a single service, where a router dispatches requests to servers within the cluster based on load.
  • Motivation to Combine: A POSITA would have been motivated to combine Chung’s explicit teaching of server clustering with the Glines/Gupta system to gain the well-known benefits of clustering, such as improved performance, scalability, and fault tolerance. Chung taught resolving a DNS request to a common cluster address, which was directly analogous to the anycast address resolution in Glines, making the technologies highly compatible.
  • Expectation of Success: Success was expected because implementing server clusters was a well-understood technique in the art for managing groups of servers, and its application to the server groups in Glines/Gupta would yield the predictable result of a more robust and scalable CDN.

Ground 3: Claims 7-9 and 19-20 are obvious over Glines, with or without Gupta, in view of Chiou.

• Prior Art Relied Upon: Glines (WO 2001/039470), Gupta (Patent 6,405,252), and Chiou (Patent 6,792,507).

• Core Argument for this Ground:

  • Prior Art Mapping: This ground targeted limitations related to content validity, such as claim 7’s step of checking "if a valid version of the content is available on the content server, serving the content; otherwise... obtaining the content from a content source." Petitioner argued that Chiou taught this exact concept for cache coherency. Chiou disclosed checking a "cache tag table" to determine if requested data is present and valid; if it is, the data is served, and if not (a cache miss), the system retrieves the data from a target/origin server.
  • Motivation to Combine: A POSITA would combine Chiou's cache validity teachings with the Glines/Gupta CDN to improve user experience by ensuring that users are not served stale or invalid content. This was a common problem in caching systems, and Chiou provided a known solution.
  • Expectation of Success: There was a high expectation of success because cache validity checking was a fundamental and compatible technique for any content caching system, including the CDNs disclosed in Glines and Gupta.

• Additional Grounds: Petitioner asserted additional obviousness challenges, including combinations incorporating Colby (Patent 6,006,264) to explicitly teach using a "content-aware flow switch" for implementing load balancing policies, as recited in claims such as 10-14, 18, and 20-22.

4. Arguments Regarding Discretionary Denial

• Petitioner argued against discretionary denial under 35 U.S.C. §314(a) based on the Fintiv factors, asserting that the co-pending district court case was in its earliest stages, with fact discovery having just opened and a Markman hearing months away. The trial was scheduled for September 2026, creating minimal overlap with the Board’s Final Written Decision timeline.
• Petitioner also argued against denial under §325(d), contending that the primary references (Glines, Chung, Chiou) were never considered during prosecution. While Gupta and Colby were cited, they were part of a list of over 500 references and were not substantively analyzed by the Examiner, meaning the core arguments in the petition were not previously considered.

5. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-23 of the ’112 patent as unpatentable under 35 U.S.C. §103.