PTAB

IPR2013-00596

Apple Inc v. PersonalWeb Technologies LLC

Key Events

Petition

petition

Title: Content-Based Identifiers
Brief Description: The ’310 patent discloses data storage systems that use "content-based identifiers" to manage data items. These identifiers are generated by applying a mathematical function (e.g., a message digest) to the contents of a data item, allowing the system to perform functions like eliminating duplicate files across a network.

Prior Art Relied Upon: Browne (a February 1995 University of Tennessee technical report titled "Location-Independent Naming for Virtual Distributed Software Repositories").
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that Browne disclosed every limitation of the challenged claims. Browne’s system used "Location Independent File Names" (LIFNs) to identify files, where a core component is a "signature" computed by applying an MD5 message digest function to the file's contents—identical to the ’310 patent’s "content-based identifier." A client computer requests a file from an LIFN server using this LIFN; the server then determines if it has a corresponding location (URL) for that LIFN in its database. If a mapping exists, the server provides the location, thereby selectively permitting access to the file. If no mapping is found, access is not provided.

Prior Art Relied Upon: Browne (as described above) and Stefik (Patent 7,359,881).
Core Argument for this Ground:
- Prior Art Mapping: Petitioner asserted that Browne taught the core invention of using a content-based identifier (the LIFN) to request and locate a data item. Stefik was introduced to supplement Browne’s teaching on "selective access." Stefik addressed the specific problem of preventing unauthorized access to digital works by allowing access only to licensed or authorized parties. Petitioner contended that Stefik’s explicit teachings on authorization and selective access were well-known at the time.
- Motivation to Combine: A POSITA would combine Browne’s content-based naming system with Stefik’s robust access control methods to solve the known problem of managing access to digital content in a distributed network. Implementing a more formal authorization system, as taught by Stefik, would have been a predictable improvement to the simpler location-based access control in Browne.
- Expectation of Success: A POSITA would have a reasonable expectation of success in combining these systems, as both operated in networked computer environments and addressed related problems of file identification and access. Using a known unique identifier (like Browne's LIFN) within a known authorization framework (like Stefik's) was a conventional design choice.

Prior Art Relied Upon: Woodhill (Patent 5,649,196).
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that Woodhill’s distributed storage management system anticipated the claims. Woodhill disclosed using "Binary Object Identifiers" (BOIs) to uniquely identify data objects for backup and restore operations. The BOI was generated from the content of the binary object using CRC, LRC, and hash values, making it a content-based identifier. In a restore operation, a local computer sends a request with BOIs to a remote backup server. The server compares these identifiers to its own database to determine if the objects match. It selectively permits access (i.e., transmits the data) only for non-matching or missing objects, thereby preventing redundant data transfer.
Additional Grounds: Petitioner asserted additional challenges, including that claims 24, 32, 70, 81, 82, and 86 are anticipated by a 1991 newsgroup post by Langer and by the Farber publication (WO 96/32685). Further obviousness grounds were asserted based on Langer in view of Stefik and Woodhill in view of Stefik, which relied on similar motivations to combine content-based identification with more explicit access control.

"content-based identifier" / "content-dependent name" / "digital identifier": Petitioner argued that these related terms, which appear in the challenged claims but not all in the specification, should be construed synonymously to mean "an identity for a data item generated by processing all of the data in the data item, and only the data in the data item, through an algorithm that makes the identifier substantially unique." This construction was asserted to be consistent with the ’310 patent’s disclosure and a district court ruling.
"selectively permitting" / "selectively allow": Petitioner proposed these terms should be interpreted to mean "permitting based on whether a condition is met." The condition, in the context of the claims, was argued to be whether the provided content-based identifier corresponds to a known value in a database, which determines if access is granted or denied.

Priority Date Entitlement: A central contention was that the ’310 patent was not entitled to its claimed priority date of April 11, 1995. Petitioner argued that the challenged claims require a function that is either a "message digest" or a "hash function." The term "hash function" was first introduced in the application that led to the ’310 patent and does not appear in the priority documents. Petitioner asserted this addition constituted new matter, as the priority documents failed to provide a written description for the distinct "hash function" species. This argument, if successful, would make references like Browne, Woodhill, Langer, and Farber available as prior art under §102(b).

Petitioner requests institution of an inter partes review and cancellation of claims 24, 32, 70, 81, 82, and 86 of the ’310 patent as unpatentable.