PTAB

IPR2021-01408

Intel Corp v. StreamScale Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Accelerated Erasure Coding Using Parallel Processing
  • Brief Description: The ’296 patent discloses methods for accelerating matrix-based erasure coding, such as Reed-Solomon coding, on a computing system. The purported invention uses a "parallel multiplier" to concurrently multiply multiple data entries by a single factor and implements specific data access patterns to improve processing efficiency.

3. Grounds for Unpatentability

Ground 1: Obviousness over Macy and Plank - Claims 23, 24, 26, 27, 29-32, 34, 35, and 37-39 are obvious over Macy in view of Plank.

  • Prior Art Relied Upon: Macy (Patent 7,343,389) and Plank ("A Tutorial on Reed-Solomon Coding for Fault-Tolerance in RAID-like Systems," Technical Report CS-96-332, 1997).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Macy discloses the core elements of the ’296 patent’s method, including an accelerated error-correcting code (ECC) system on a standard computer architecture. Macy teaches using a "parallel multiplier" via Single Instruction, Multiple Data (SIMD) instructions (like PMODMUL or PSHUFB) to speed up the modular multiplication central to erasure coding. Macy further discloses ordering these matrix operations through specific data access patterns ("column-by-column" and "row-by-row") to enhance performance, directly corresponding to limitations in independent claim 23. To the extent Macy's "forward error control" was not explicitly erasure coding, Petitioner asserted that Plank, a foundational report on Reed-Solomon coding, supplies this context. Plank details encoding original data into data and check matrices, and subsequently decoding to recover lost data, which Petitioner mapped to the limitations of dependent claims covering data recovery (e.g., claims 26, 27, 29-32).
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Macy and Plank to improve the performance of RAID storage systems, a field common to both references. Petitioner contended that Macy provides techniques to "speed up modular multiplication" for data recovery in RAID systems but does not offer specific implementation details for erasure coding itself. A POSITA would have looked to a well-known reference like Plank, which provides those exact implementation details for erasure coding in RAID-like systems. The combination represents applying Macy's known acceleration techniques to Plank's conventional and compatible erasure coding framework.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success because both references operate in the same technical field (RAID systems) and use the same underlying mathematical principles (Galois Field multiplication). The combination involved using a known optimization method from Macy to improve a similar, well-understood system from Plank to achieve predictable performance gains.

Ground 2: Obviousness over Macy, Li, and Plank - Claims 25 and 36 are obvious over Macy and Li, optionally in further view of Plank.

  • Prior Art Relied Upon: Macy (Patent 7,343,389), Li ("Parallelized Network Coding With SIMD Instruction Sets," 2008), and Plank (1997 technical report).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground builds upon the Macy/Plank combination to address claims 25 and 36, which add limitations requiring a "plurality of processing cores" and scheduling data processing across them. Petitioner argued that Li explicitly teaches accelerating matrix-based coding using multi-core processors. Li discloses partitioning a calculation into multiple threads and assigning them among multiple CPUs to "fully utilize" the processing cores. This directly maps to the claim limitations of dividing the data and check matrices into pluralities and assigning them to different processing cores for concurrent generation of check data.
    • Motivation to Combine: Petitioner argued a POSITA, having combined Macy and Plank to create an efficient erasure coding system, would be motivated to incorporate Li's teachings for further performance improvements. Li states that "modern commodity processors are routinely multi-core" and provides a method for parallelizing the exact type of table-based multiplication (using the PSHUFB instruction) disclosed in Macy. The motivation was to leverage the common and known technique of multi-core processing to further accelerate the Macy/Plank system, a predictable and logical design choice.
    • Expectation of Success: Success was expected because Li's system is highly compatible with Macy's, using the same SIMD instructions for the same purpose. Tailoring the Macy/Plank implementation for Li's disclosed dual-core environment would have been a straightforward application of a known parallelization technique to improve a similar system, yielding predictable results.

  • Additional Grounds: Petitioner asserted that claims 23, 24, 34, and 35 are obvious over Macy alone, arguing Macy inherently teaches or suggests all elements. It also presented an alternative ground for claims 25 and 36 over just Macy and Li, relying on the same multi-core processing arguments.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under Fintiv would be inappropriate. The core arguments were:
    • The petition was filed expeditiously, just eight months after the initial complaint in the parallel district court litigation.
    • The district court case was in a very early stage, with no trial date set and fact discovery not yet commenced, making it likely that a Final Written Decision (FWD) from the IPR would issue before any trial.
    • Petitioner stipulated that, if the IPR is instituted, it would not pursue the same invalidity grounds in the district court, thus mitigating concerns of duplicative efforts and conflicting decisions.
    • The petition presents strong grounds for unpatentability, which weighs in favor of institution to promote the AIA's goal of efficiently weeding out invalid patent claims.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 23-27, 29-32, and 34-39 of the ’296 patent as unpatentable.