Cisco Systems, Inc. v. Hewlett Packard Enterprise Company

1. Case Identification

• Case #: IPR2017-01933
• Patent #: 8,478,799
• Filed: August 11, 2017
• Petitioner(s): Cisco Systems, Inc.
• Patent Owner(s): Hewlett Packard Enterprise
• Challenged Claims: 1-22 and 26-36

2. Patent Overview

• Title: Computer File System Data Structures
• Brief Description: The ’799 patent discloses a computer file system architecture that separates the logical file hierarchy from physical data storage. The system uses a "namespace file system" for directories and file metadata, and a separate "object store" that holds all data as a flat collection of opaque data "objects." Each object is identified by a globally unique "object fingerprint" derived from its content, such as via a cryptographic hash.

3. Grounds for Unpatentability

Ground 1: Obviousness over Muthitacharoen and Dabek - Claims 1-4, 7-9, 11-14, 17-22, 27, 28, and 31-35 are obvious over Muthitacharoen in view of Dabek.

• Prior Art Relied Upon: Muthitacharoen ("Ivy: A Read/Write Peer-to-Peer File System," OSDI ’02) and Dabek ("Wide-area cooperative storage with CFS," SOSP ’01).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Muthitacharoen discloses the "Ivy" peer-to-peer file system, which functions as the claimed "namespace file system," and stores all its data and metadata in a DHash peer-to-peer block storage system, which constitutes the claimed "object store." Ivy stores data in content-hashed blocks, which Petitioner contended a POSITA would understand to be "objects." The hash key for each block, generated using SHA-1, serves as the "object fingerprint" used to access the data. Muthitacharoen's system uses a "snapshot block" as an "inode map object" that maps persistent i-numbers to the content-based object fingerprints, enabling file content to change (changing the fingerprint) while the i-number remains constant. Dabek was cited to provide further detail on the DHash system, disclosing that it runs on multiple servers ("computer storage devices") scattered across the internet.
  • Motivation to Combine: A POSITA would combine these references because Muthitacharoen explicitly cited Dabek as a source for further details on the underlying DHash storage system. The references are also in the same field of peer-to-peer file systems and address analogous problems, making the combination a predictable supplementation of Ivy's system with DHash's known operational details.
  • Expectation of Success: A POSITA would have a high expectation of success, as the combination involved using the detailed description of the DHash system from Dabek to implement the file system taught in Muthitacharoen, which explicitly relies on DHash.

Ground 2: Obviousness over Muthitacharoen, Dabek, and Agrawal - Claims 5 and 6 are obvious.

• Prior Art Relied Upon: Muthitacharoen, Dabek, and Agrawal ("Design Tradeoffs for SSD Performance," USENIX ’08).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon Ground 1, adding Agrawal to teach the specific storage media recited in dependent claims 5 and 6. Agrawal described the benefits of using solid-state disks (SSDs), including NAND-flash based memory, as a high-performance, drop-in replacement for traditional hard disk drives in computer storage subsystems.
  • Motivation to Combine: A POSITA implementing the Muthitacharoen/Dabek system would have been motivated to replace the conventional hard disk drives of the time with superior SSD technology to improve performance, power consumption, and reliability. Agrawal taught that SSDs were designed to "mimic a hard disk drive," making the substitution straightforward.
  • Expectation of Success: The expectation of success would be high, as using an SSD as a replacement for a hard disk was a well-understood and advantageous design choice at the time.

Ground 3: Obviousness over Muthitacharoen, Dabek, and McKusick - Claims 10, 15, and 26 are obvious.

• Prior Art Relied Upon: Muthitacharoen, Dabek, and McKusick ("The Design and Implementation of the FreeBSD Operating System" (2005)).

• Core Argument for this Ground:

  • Prior Art Mapping: This ground added McKusick to provide specific details about the FreeBSD operating system, which Muthitacharoen stated is the environment where Ivy runs. For claim 10, McKusick disclosed the layered architecture of FreeBSD, including a "vnode layer" ("virtual file system layer") above the NFS layer, which corresponds to Ivy's implementation. This established the claimed storage stack. For claims 15 and 26, which require a "transaction log," McKusick described how FreeBSD handles file operations and reference count updates, which Petitioner argued would inform a POSITA how to implement logging for reads, writes, and deletes in the Ivy system.
  • Motivation to Combine: Because Muthitacharoen's Ivy system runs on FreeBSD, a POSITA seeking to implement or understand Ivy would have been directly motivated to consult an authoritative reference on FreeBSD's architecture, like McKusick, to understand how the systems would interface.
  • Expectation of Success: The combination represented the ordinary use of common programming and system integration techniques to implement a file system (Ivy) on a known operating system (FreeBSD) as described by McKusick.

• Additional Grounds: Petitioner asserted additional obviousness challenges based on combinations including Bunte (Patent 8,140,786) for teaching data backup to another object store for disaster recovery, and Bondurant (Patent 8,028,106) for teaching the use of a hardware accelerator to improve the performance of SHA-1 hash calculations.

4. Key Claim Construction Positions

• "namespace file system": Petitioner adopted the Patent Owner's construction from prior IPR proceedings, interpreting the term to mean "a file system that uses names." This construction simplifies mapping the Ivy system, which uses file and directory names, to the claims.
• "object": Petitioner argued that the term should be given its plain and ordinary meaning. However, Petitioner also contended that the claims are obvious even under the Patent Owner's narrower construction from prior proceedings ("a logical abstraction of variable size of data, such as a file"), noting that data structures in the prior art like files and directories are of variable size.

5. Key Technical Contentions (Beyond Claim Construction)

• A key technical contention was that at the time the prior art was created, the terminology for object-based storage was still evolving. Petitioner argued that a POSITA would understand that terms used in the prior art, such as "block" or "content-hash block" in the context of the DHash system, were functionally equivalent to the claimed term "object," as both refer to data units identified and stored based on their content.

6. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §325(d) would be inappropriate because its petition raised new arguments and prior art. Specifically, it contended that the primary references—Muthitacharoen, Dabek, Agrawal, McKusick, Bunte, and Bondurant—were never evaluated by the USPTO during prosecution or in prior IPRs against the ’799 patent. Further, the petition relied on a new expert declaration and challenged a broader set of claims than the prior IPRs.

7. Relief Requested

• Petitioner requested the institution of an inter partes review and the cancellation of claims 1-22 and 26-36 of the ’799 patent as unpatentable.