Amazon.com Inc v. Nokia Technologies Oy

1. Case Identification

• Case #: IPR2024-01507
• Patent #: 8,996,693
• Filed: October 18, 2024
• Petitioner(s): Amazon.com, Inc., Amazon.com Services LLC
• Patent Owner(s): Nokia Technologies Oy
• Challenged Claims: 1-5, 7-15, 17-20

2. Patent Overview

• Title: Method and Apparatus for Providing Dynamic Stream Processing of Data Based on Static Analytics
• Brief Description: The ’693 patent discloses methods and systems for integrating dynamic, real-time data stream processing with static, historical data analytics. The purported invention involves transferring data from a dynamic processing system to a static one for analysis and then returning the processed results to the dynamic system to generate more comprehensive and accurate outputs.

3. Grounds for Unpatentability

Ground 1: Obviousness over Foster - Claims 1-5, 7-15, and 17-20 are obvious over Foster.

• Prior Art Relied Upon: Foster et al., IBM InfoSphere Streams: Assembling Continuous Insight in the Information Revolution (2011) (“Foster”).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Foster, which was not considered during prosecution, teaches every limitation of the challenged claims. Foster describes an integrated "Big Data platform" combining IBM's InfoSphere Streams (a dynamic, real-time processing mechanism) and InfoSphere BigInsights (a static, "data-at-rest" processing mechanism). Petitioner asserted that Foster explicitly discloses the claimed method: (1) determining a "processing element" (data) within the dynamic Streams system; (2) marshalling the data for transfer (e.g., converting tuples to JSON); (3) transferring the data object to the static BigInsights system; (4) processing the data in BigInsights using static analytics (e.g., MapReduce) to enrich it or update analytical models; and (5) transferring the processed data or updated model back to Streams for use in "downstream analytics." This complete data feedback loop, which Petitioner noted was the key feature argued for patentability during prosecution, is allegedly shown in Foster’s diagrams and text.

Ground 2: Obviousness over Foster in view of Williams - Claims 1-5, 7-15, and 17-20 are obvious over Foster in view of Williams.

• Prior Art Relied Upon: Foster (a 2011 IBM publication) and Williams et al., Implementing CICS Web Services (2007) (“Williams”).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground asserts that even if Foster were found not to explicitly teach the "marshalling" and "unmarshalling" limitations, these steps were well-known and obvious to a person of ordinary skill in the art (POSITA). Williams explicitly teaches that marshalling (e.g., serialization, encoding) and unmarshalling are fundamental, conventional techniques used to translate data into a suitable format for transmission between different processes or nodes in a distributed computing environment—the exact environment described by Foster. Therefore, applying the routine marshalling techniques of Williams to the data transfers described in Foster would have been obvious.
  • Motivation to Combine: Petitioner argued a POSITA would combine Foster and Williams for several reasons. First, the references themselves suggest the combination, as Foster discloses a distributed computing system that transfers data, a scenario where the marshalling taught by Williams was a well-known and necessary step. Second, the combination would yield predictable results: applying a standard data formatting technique to a data transfer system would predictably result in formatted data suitable for transmission. Third, both references are analogous art, addressing data transfer and processing in distributed computing environments.
  • Expectation of Success: A POSITA would have had a high expectation of success because implementing marshalling is a straightforward software task involving known algorithms. Combining the teachings required only applying a conventional technique (Williams) to an existing system (Foster) to achieve a predictable, improved result without changing the fundamental principles of either reference.

4. Key Claim Construction Positions

• "processing element": Petitioner proposed this term means "data to be processed." This construction is based on the patent’s claims and specification, which repeatedly refer to the processing element being marshalled into a data object and then processed.
• "dynamic processing mechanism": Petitioner proposed this term means a "mechanism for processing data in real-time or near real-time." This is based on the patent’s description of dynamic processing as "stream processing" that operates on "fast-moving" data feeds like Twitter.
• "static processing mechanism": Petitioner proposed this term means a "mechanism for processing...historical or slow-moving data." This is based on the specification’s examples, such as "map-reduce batch jobs" operating on large databases according to predetermined schedules (e.g., daily or weekly).

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial would be inappropriate.
• Fintiv: Denial under the Fintiv factors is not warranted because no case management conference has occurred and no trial schedule has been set in the parallel district court litigation, making the PTAB the more efficient forum.
• §325(d): Denial is unwarranted because the primary prior art references, Foster and Williams, were never presented to or considered by the examiner during the original prosecution. Therefore, the petition raises new questions of patentability that the PTO has not previously considered.

6. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-5, 7-15, and 17-20 of the ’693 patent as unpatentable under 35 U.S.C. §103.