Dropbox Inc v. Datanet LLC

1. Case Identification

• Case #: IPR2024-00079
• Patent #: 9,218,348
• Filed: October 31, 2023
• Petitioner(s): Dropbox, Inc. and Microsoft Corporation
• Patent Owner(s): Datanet, LLC
• Challenged Claims: 1-6, 8-20, 23-31

2. Patent Overview

• Title: Automatic Real-Time File Management Method And Apparatus
• Brief Description: The ’348 patent describes a method for archiving files where an operation on an operating file triggers the creation of a backup copy. This copy is then stored and migrated through a series of storage locations: temporary, intermediate, and permanent archive storage.

3. Grounds for Unpatentability

Ground 1: Claims 1-6, 8-20, and 23-31 are obvious over Long in view of Kenley.

• Prior Art Relied Upon: Long (Patent 5,325,519) and Kenley (Patent 5,276,867).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Long taught the core event-driven archival process of the challenged claims. Long disclosed detecting a modification request for data on a hard disk, and in response, creating a "before image" backup copy. This copy is first stored in a temporary memory buffer, and upon a triggering event (e.g., buffer full), is moved to an intermediate "audit buffer" on disk. Subsequently, upon a second triggering event (e.g., a timer), the data is moved from the audit buffer to permanent external archive media. Petitioner contended that Kenley supplied the context of a hierarchical, file-based storage system that was missing from Long's data-block-level disclosure. Kenley taught a multi-tier system (primary, secondary, backing store) for storing data organized as files and using a relational database with a "system catalog" to track file attributes and locations as they are migrated. The combination of Long's event-driven, multi-stage buffering process with Kenley's file-based hierarchical system and database tracking allegedly rendered the steps of claim 1 obvious.
  • Motivation to Combine: A POSITA would combine Long's modification-driven archival approach with Kenley's file-based backup system to improve reliability. Kenley's periodic incremental backups risked data loss for modifications made between backups. Long's method of creating a backup upon each modification would solve this known problem, providing a more robust, fault-tolerant system. The references were complementary, as both described three-tier storage architectures (e.g., memory, disk, archive) and background processing. Further, Kenley's database for tracking file locations would be naturally augmented with Long's "audit header" information to track the real-time location of backup copies.
  • Expectation of Success: A POSITA would have a reasonable expectation of success because Long taught a straightforward, modification-driven archival approach applicable to known multi-tier systems. Applying this well-understood, event-driven technique to a file-based system like Kenley's was a predictable implementation.

Ground 2: Claims 1-6, 8-20, and 23-31 are obvious over Cabrera, in view of Schneider and Whiting.

• Prior Art Relied Upon: Cabrera (Patent 5,953,729), Schneider (WO 99/12101), and Whiting (Patent 5,778,395).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner asserted that Cabrera taught a method of archiving data by copying a file from local storage to a temporary "staging storage area" in response to a triggering event (e.g., expiration of a time delay). From the staging area, the file is later moved to a remote, permanent storage location. Petitioner argued Schneider improved upon Cabrera's time-based trigger by teaching the interception of a file write operation as the trigger for an automatic backup, creating a "continuously running disk backup system." This supplied the limitation of detecting an instruction and creating an archive file responsive to that instruction. Petitioner further argued that while Cabrera disclosed a "master file table" (MFT) for file attributes, Whiting provided a more robust database system. Whiting taught a "backup directory" and a "global directory database" to track where backup data files are located, enabling transparent access as files migrate through the system. This combination allegedly disclosed all steps of the claimed method.
  • Motivation to Combine: A POSITA would combine the references to create a more efficient and reliable backup system. Schneider explicitly addressed the "shortcoming" of time-based systems like Cabrera's, which could result in data loss from a failure before the next timed backup. A POSITA would therefore be motivated to replace Cabrera's time-delay trigger with Schneider's more reliable write-intercept trigger. Furthermore, a POSITA implementing Cabrera's hierarchical storage system would look to prior art like Whiting to improve the tracking of file locations. Whiting’s global directory database was designed to solve the problem of locating migrated backup files, making it a natural and necessary addition to Cabrera's system to ensure archived files could be found and accessed.
  • Expectation of Success: There was a reasonable expectation of success because POSITAs were familiar with intercepting file I/O operations to trigger other actions, as taught by Schneider. Modifying Cabrera's system to use this known technique was predictable. Similarly, incorporating a database like Whiting's to track file locations was a well-understood solution to a known problem in hierarchical storage management.

4. Arguments Regarding Discretionary Denial

• Petitioner stipulated that if the Board institutes this IPR, it will not pursue in its parallel district court proceedings the grounds raised, or which reasonably could have been raised, in this petition. Petitioner argued that based on this stipulation, under current USPTO guidance, the PTAB should not discretionarily deny institution under §314(a).

5. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-6, 8-20, and 23-31 of the ’348 patent as unpatentable.