PTAB
IPR2026-00174
Microsoft Corp v. Sandpiper CDN LLC
Key Events
Petition
Table of Contents
petition
1. Case Identification
- Case #: IPR2026-00174
- Patent #: 9,762,692
- Filed: December 17, 2025
- Petitioner(s): Microsoft Corporation
- Patent Owner(s): Sandpiper CDN, LLC
- Challenged Claims: 1, 2, 4-15, 17, 18
2. Patent Overview
- Title: Method and Framework for Content Delivery
- Brief Description: The ’692 patent discloses a Content Delivery Network (CDN) with a hierarchical, multi-tier server architecture. The system manages content delivery by storing popular content on first-tier "edge" servers for faster access, while storing less popular content on distinct second-tier servers deeper in the network.
3. Grounds for Unpatentability
Ground 1: Obviousness over Seed in view of Swildens - Claims 1, 2, 4-15, 17, and 18 are obvious over Seed in view of Swildens.
- Prior Art Relied Upon: Seed (Application # 2006/0015574) and Swildens (Application # 2004/0221019).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that Seed taught the core architecture of the ’692 patent, including a multi-tier CDN with a first tier of edge servers and a second tier of parent servers. Seed disclosed a method where, upon receiving a request at a first-tier server for a resource not in its cache, the server determined if the resource was popular. If popular, the server would obtain and serve it; if not popular, it would direct the client to a second-tier parent server. Petitioner asserted this mapped to the primary limitations of independent claims 1 and 15. However, Seed only implicitly suggested distributing different library portions across parent servers. Swildens was argued to explicitly supply the limitation of partitioning the library into distinct portions across the second-tier servers (claim 1[d][ii]). Swildens described using a hash function to map content identifiers to specific servers in its intermediate tier, ensuring each stored a distinct portion of the total library. The combination also addressed key dependent claims, as Swildens’ hashing algorithm taught the logical partitioning of claim 2 and the use of hashes for redirection (claim 7), while Seed’s teachings met limitations for content types (claim 4) and popularity thresholds (claim 8).
- Motivation to Combine: A POSITA would combine Swildens’ partitioning method with Seed’s CDN architecture for two primary reasons. First, it would reduce storage costs on parent servers by eliminating the need to store the entire content library on each, a direct economic benefit taught by Swildens. Second, it would improve the efficiency of locating unpopular content by replacing Seed’s inefficient, recursive redirection process with Swildens’ direct, hash-based redirection, leading to faster service for users.
- Expectation of Success: A POSITA would have a reasonable expectation of success because Seed already envisioned distributing portions of content across its parent tier. Swildens provided a well-known, efficient, and predictable method (hashing) to purposefully implement this distribution, representing a straightforward and easily implementable modification.
Ground 2: Obviousness over Swildens in view of Seed - Claims 1, 2, 4-8, 10-15, 17, and 18 are obvious over Swildens in view of Seed.
- Prior Art Relied Upon: Swildens (Application # 2004/0221019) and Seed (Application # 2006/0015574).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner contended that Swildens, as the primary reference, taught most limitations of the independent claims. Swildens described a CDN with a first tier of edge servers that determined content popularity. For unpopular content, Swildens directed the client to a second tier of origin servers. Crucially, Swildens already taught using a hash function to distribute distinct portions of the content library across these second-tier origin servers. This system also mapped to framework claim 10, where Swildens’ own alternative embodiment of a centralized "auto migration server" functioned as the claimed popularity service. Petitioner argued that adding an intermediate "parent" server tier, as taught by Seed, between Swildens' edge and origin tiers was a well-known and obvious modification. This change would align Swildens' architecture with claim limitations requiring the second tier to be intermediate between the first tier and a main origin server (claims 1[d][iii], 14). This combination also rendered the dependent claims obvious, as Swildens' hashing taught logical partitioning (claim 2) and the use of popularity thresholds (claim 8).
- Motivation to Combine: A POSITA would be motivated to incorporate an intermediate "parent" server tier from Seed into Swildens' architecture as a routine and predictable design choice based on cost-performance tradeoffs. This modification would improve CDN performance by creating caches of unpopular content closer to the edge, reducing latency, and distributing request loads away from the main origin servers. This was a well-known method for scaling a two-tier CDN, making it an obvious improvement to Swildens' system.
- Expectation of Success: A POSITA would have a high expectation of success because implementing an intermediate parent server tier was a common practice for enhancing CDN performance. Swildens already envisioned adjusting server arrangements based on design considerations, and Seed provided an express framework for structuring such a tier to accomplish the same goals.
4. Key Claim Construction Positions
- Conditional Claiming: Petitioner highlighted a prior Board finding that independent claims 1 and 15 use conditionals ("if...then...otherwise") that encompass at least two mutually exclusive methods: one for popular content (e.g., claim 1[c]) and one for unpopular content (e.g., claim 1[d]). Petitioner argued that since the prior art taught both scenarios, the claims were obvious regardless of whether only one conditional path was required.
- "distinct portions": Petitioner addressed the potential indefiniteness of this term, which Patent Owner had previously interpreted as allowing for "some overlap." Petitioner argued that the prior art, particularly Swildens' use of a hash function, suggested a method that would result in non-overlapping portions. This non-overlapping distribution indisputably fell within any reasonable construction of "distinct portions," rendering the claims obvious even if the term’s outer bounds were unclear.
5. Relief Requested
- Petitioner requested institution of an inter partes review and cancellation of claims 1, 2, 4-15, 17, and 18 of the ’692 patent as unpatentable.
Analysis metadata