Pure Storage Inc v. Realtime Data LLC

1. Case Identification

• Patent #: 9,667,751
• Filed: January 30, 2018
• Petitioner(s): Pure Storage, Inc.
• Patent Owner(s): Realtime Data, LLC
• Challenged Claims: 1, 21-25, 45-48

2. Patent Overview

• Title: Data Feed Acceleration
• Brief Description: The ’751 patent describes methods and systems for providing "accelerated transmission of data." This acceleration is achieved by using data compression and decompression, where the total time to compress, transmit, and decompress a data block is less than the time required to transmit the same data block in its uncompressed form. The patent states this concept can also be applied to the storage and retrieval of data.

3. Grounds for Unpatentability

Ground 1: Claims 1, 21, 23, and 24 are obvious over Matsubara and PC Magazine.

• Prior Art Relied Upon: Matsubara (Patent 5,838,821) and PC Magazine (an article by Greg Pastrick, Jan. 28, 1992).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that the combination of Matsubara and PC Magazine taught every limitation of independent claim 1. Specifically, Matsubara allegedly disclosed a method for "analyzing content of a data block to identify a parameter, attribute, or value" by creating a histogram of byte patterns to determine the file's data type (e.g., image, text, executable). This analysis of the actual content, rather than a file header, allegedly met the claim limitation that "excludes analyzing based solely on reading a descriptor." Based on this analysis, Matsubara selected an appropriate compression encoder (e.g., JPEG, Lempel-Ziv). Petitioner contended that the final key limitation—that the time to compress and store the data is less than the time to store it uncompressed—was explicitly taught by PC Magazine. The article evaluated on-the-fly compression utilities and found that "the process of compressing and writing files composed of small records to disk is usually faster than writing the same uncompressed files to disk," providing performance charts as evidence.
  • Motivation to Combine: A POSITA would combine these teachings to improve the data-type-specific compression system of Matsubara with the known speed and efficiency benefits taught by PC Magazine. Matsubara's stated goal was to make the "most efficient use of file storage space and to reduce transmission time." Applying the techniques from PC Magazine, which demonstrated that compression could actually reduce disk-write times, directly served this goal. Petitioner asserted this was the application of a known technique to a similar system to yield predictable results.
  • Expectation of Success: A POSITA would have a high expectation of success in achieving a system where compression and storage are faster than storage alone. PC Magazine demonstrated this was already achievable with commercially available Lempel-Ziv based utilities, the same type of lossless encoder suggested by Matsubara for non-image files.

Ground 2: Claim 22 is obvious over Matsubara, PC Magazine, and Franaszek.

• Prior Art Relied Upon: Matsubara (Patent 5,838,821), PC Magazine (Jan. 28, 1992 article), and Franaszek (Patent 5,870,036).
• Core Argument for this Ground: This ground built upon Ground 1 to address dependent claim 22, which adds compressing with a "default lossless encoder" when the analysis is unable to identify the data block's type.
  • Prior Art Mapping: Franaszek was cited for its teaching of a compression method where if the data type of a block is not recognized, "a default compression algorithm is utilized." Franaszek further explained that data compression often needs to be lossless because "the loss of even a single bit can be catastrophic." This aligned with Matsubara's own teaching to use lossless Lempel-Ziv compression for files where data integrity is critical.
  • Motivation to Combine: A POSITA would be motivated to add Franaszek's default mechanism to the Matsubara/PC Magazine system to improve its robustness. Matsubara's histogram analysis might fail for unknown or new file types. Franaszek provided a known solution to this problem: apply a default lossless encoder. This would ensure that the system could handle any file type without data loss, a critical feature for a general-purpose compression system.

Ground 3: Claims 25, 45, 47, and 48 are obvious over Matsubara, PC Magazine, and York.

• Prior Art Relied Upon: Matsubara (Patent 5,838,821), PC Magazine (Jan. 28, 1992 article), and York (Patent 6,711,709).

• Core Argument for this Ground: This ground challenged system claims that corresponded to the method claims of Ground 1, with the key additional element being the implementation on a "data server."

  • Prior Art Mapping: York allegedly disclosed a system for transferring compressed data files over a network where the compressing entity is a "Unix based server." Petitioner argued that the systems described in Matsubara—such as "information retrieval systems" and "personal computer[] workstations" connected to networks—were already functionally data servers, even if that specific term was not used. York explicitly taught using a server to perform compression for efficient data transfer.
  • Motivation to Combine: A POSITA would find it obvious to implement the compression methods of Matsubara and PC Magazine in a data server as taught by York. Data servers were a natural and common platform for implementing compression techniques to save storage and reduce network bandwidth, which were the primary goals of the underlying references. Petitioner argued this was a simple matter of design choice, applying a known technique (content-aware, high-speed compression) in a well-known environment (a data server).

• Additional Grounds: Petitioner asserted an additional obviousness challenge against claim 46, combining Matsubara, PC Magazine, York, and Franaszek to teach a data server system that is configured to use a default lossless encoder when its analysis of a data block is unable to identify a data type.

4. Key Claim Construction Positions

• Petitioner argued for a construction of the term "data block(s)" to mean "a block of data ranging in size from a single bit through complete files or collections of multiple files." This construction was asserted to be consistent with the ’751 patent's specification. The proposed construction was critical to Petitioner's argument, as it allowed the "files" described in prior art references like Matsubara and PC Magazine to be equated with the "data blocks" recited in the challenged claims.

5. Relief Requested

• Petitioner requested that the Board institute an inter partes review of claims 1, 21-25, and 45-48 of the ’751 patent and find those claims unpatentable under 35 U.S.C. §103.