Cisco Systems Inc v. UMBRA Technologies Ltd

1. Case Identification

• Case #: IPR2024-00344
• Patent #: 11,146,632
• Filed: January 29, 2024
• Petitioner(s): Cisco Systems, Inc.
• Patent Owner(s): UMBRA Technologies Ltd.
• Challenged Claims: 1-7

2. Patent Overview

• Title: DATA BEACON PULSER(S) POWERED BY INFORMATION SLINGSHOT
• Brief Description: The ’632 patent relates to a distributed network storage system. The claimed topology involves a first node with a read queue, write queue, and parallel file system, where the first node writes data to a second node's parallel file system, which is subsequently read into the second node's read queue.

3. Grounds for Unpatentability

Ground 1: Obviousness over Agarwala - Claims 1-7 are obvious over Agarwala

• Prior Art Relied Upon: Agarwala (Patent 9,582,421, which incorporates by reference U.S. Provisional Application 61/739,685).
• Core Argument for this Ground:

  • Prior Art Mapping: Petitioner argued that Agarwala, which teaches a distributed multi-layer caching storage system, discloses all limitations of the challenged claims. Petitioner mapped Agarwala’s interconnected “storage nodes” (e.g., 102A, 102B) to the claimed first and second nodes. It was asserted that a person of ordinary skill in the art (POSITA) would understand Agarwala’s “write cache 208” and “read cache (DOCL) 210” within each storage node to be the functional equivalents of the claimed write and read queues, as they temporarily hold data for writing to or after reading from persistent storage. Agarwala's "StorFS" system, which partitions and stripes file data across multiple distributed nodes with a global namespace, was alleged to be the claimed “parallel file system.”

    Petitioner contended that the core process of claim 1 is disclosed by Agarwala’s asynchronous replication mechanism. In this process, data from the write cache (first write queue) of a first storage node is replicated to the persistent storage (second parallel file system) of a second mirror node. Subsequently, if there is a read cache miss on the second node, that replicated data is read from its persistent storage into its read cache (second read queue).

    For dependent claims, Petitioner argued:

    • Claim 2 (symmetric data flow): Agarwala discloses that any storage node can write to any other node in the network, rendering the reverse data flow obvious.
    • Claim 3 (carrier file): The data transfer between nodes in Agarwala would necessarily use network packets, which a POSITA would recognize as a carrier file comprising a header, body (payload), and footer.
    • Claim 5 (set frequency): Agarwala’s “flusher” module operates “periodically” to move logged data from high-speed to persistent storage, meeting the limitation of writing additional data at a set frequency.
    • Claim 6 (changed data only): Agarwala’s disclosure of a deduplication module that checks a hints table to avoid writing duplicate data to persistent storage was argued to teach writing only information that has changed.
    • Claim 7 (third node): Agarwala teaches replication to multiple mirror nodes (e.g., 302B and 302C), directly disclosing a third node with the same components.

  • Motivation to Combine (for §103 grounds): This ground is based on a single reference. The Petitioner’s argument is that a POSITA would have understood the components and functions described in Agarwala to be directly applicable to the claimed invention, as Agarwala addresses the same problem of efficient data management in a distributed storage system. The motivation for implementing dependent claim features, such as using network packets (claim 3), was presented as a standard and obvious design choice for internode communication.

  • Expectation of Success (for §103 grounds): Petitioner asserted that a POSITA would have had a high expectation of success because Agarwala describes a complete and functional distributed storage architecture, and its disclosed components were designed to work together predictably.

4. Key Technical Contentions (Beyond Claim Construction)

• Functional Equivalence of Terminology: A central contention is that the terminology used in Agarwala describes structures functionally identical to those claimed in the ’632 patent. Specifically, Petitioner argued that a POSITA would understand that Agarwala’s “write cache” and “read cache” perform the same functions as the claimed “write queue” and “read queue,” and that its distributed, striped “StorFS” system is a type of “parallel file system.”

5. Arguments Regarding Discretionary Denial

• Petitioner argued that the Board should not exercise its discretion to deny institution. Under 35 U.S.C. §325(d), denial was asserted to be improper because the Agarwala reference was never cited or considered by the examiner during prosecution and is not cumulative of the art of record.
• Under 35 U.S.C. §314(a) and the Fintiv factors, Petitioner argued for institution, stating that the co-pending district court litigation is in a very early stage with no trial date set and minimal investment. Petitioner further noted the exceptional strength of the invalidity case based on Agarwala’s parallel disclosure and confirmed it would provide a Sotera-style stipulation, agreeing not to pursue any IPR ground that is instituted in the district court litigation.

6. Relief Requested

• Petitioner requests institution of an inter partes review (IPR) and cancellation of claims 1-7 of the ’632 patent as unpatentable.