Netflix, Inc. v. Realtime Adaptive Streaming LLC

1. Case Identification

• Case #: IPR2018-01630
• Patent #: 9,769,477
• Filed: September 24, 2018
• Petitioner(s): Netflix, Inc.
• Patent Owner(s): Realtime Adaptive Streaming LLC
• Challenged Claims: 7, 8, 15-19, 23, 24, 28, and 29

2. Patent Overview

• Title: System for Adaptive Data Compression
• Brief Description: The ’477 patent relates to a system that compresses and decompresses data by selecting among a plurality of data compression encoders based on the actual or expected throughput of a communications channel. The system aims to solve performance "bottlenecks" by dynamically switching between different compression algorithms, including asymmetric ones, to balance compression speed and ratio.

3. Grounds for Unpatentability

Ground 1: Obviousness over Imai and Pauls - Claims 15-19, 28, and 29 are obvious over Imai in view of Pauls.

• Prior Art Relied Upon: Imai (Japanese Application Publication No. H11331305) and Pauls (European Application Publication No. EP0905939A2).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that the combination of Imai and Pauls taught all limitations of the challenged claims. Imai disclosed a system for selecting among a plurality of asymmetric audio encoders (e.g., MPEG, ATRAC) based on the measured throughput of a network channel to ensure real-time playback. Pauls disclosed an adaptive system for selecting among a plurality of asymmetric video transcoders (e.g., H.263, MPEG, MPEG2) based on network characteristics like "available bandwidth" and "bit rate." The combination rendered obvious a system using multiple asymmetric video encoders that are selected based on channel throughput. Petitioner asserted that Pauls' teaching of H.263 and JPEG, which have optional arithmetic coding modes, satisfied the "arithmetic algorithm" limitation underlying claim 28.
  • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Imai and Pauls because they address the same problem of adaptive data encoding for network transmission. A POSITA would logically apply Imai's detailed method for determining network throughput to Pauls' established video-specific transcoding system, particularly since Imai expressly stated its teachings were applicable to video signals.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success in combining the references. The combination involved applying known techniques to a predictable field, and the similarities between the systems and the well-known relationship between audio and video compression technologies would have ensured a predictable outcome.

Ground 2: Obviousness over Imai, Pauls, and Dawson - Claims 7 and 23 are obvious over Imai and Pauls in view of Dawson.

• Prior Art Relied Upon: Imai (Japanese Application Publication No. H11331305), Pauls (European Application Publication No. EP0905939A2), and Dawson (Patent 5,553,160).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground augmented the Imai and Pauls combination with Dawson to teach the limitation of using "a resolution of the data blocks" as a data parameter for encoder selection. Dawson taught a system for dynamically selecting between a lossless compression algorithm (LZW) and a lossy one (JPEG) for an image based on various parameters, including the image's resolution. The selection logic in Dawson, which checks if image size (a function of resolution) is above or below a threshold, directly taught using resolution as a selection criterion.
  • Motivation to Combine: A POSITA would combine Dawson with the Imai/Pauls system to improve and optimize the encoder selection process. Incorporating Dawson's resolution-based selection criterion was a matter of applying a known optimization technique from a related field (image compression) to enhance the video compression system, resulting in a more robust system with predictable benefits.

Ground 3: Obviousness over Imai, Pauls, and Lai - Claims 8 and 24 are obvious over Imai and Pauls in view of Lai.

• Prior Art Relied Upon: Imai (Japanese Application Publication No. H11331305), Pauls (European Application Publication No. EP0905939A2), and Lai (Patent 6,407,680).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground added Lai to the Imai/Pauls combination to teach selecting an encoder based on "a data transmission rate of the data blocks." Lai disclosed an on-demand media transcoding system that selected transcoders based on "publishing variables," which explicitly included the "bit-rate of the media content" being transmitted to the system. Petitioner contended that this incoming source bit rate, especially in a real-time streaming context, directly corresponds to a "data transmission rate" used for encoder selection.
  • Motivation to Combine: A POSITA would integrate Lai’s teachings to enhance the combined Imai/Pauls system by adding the input data's bit rate as another optimization parameter. This would allow the system to make more intelligent compression choices based not only on network conditions (from Imai/Pauls) but also on the characteristics of the source content itself (from Lai), improving the overall efficiency and performance of the real-time transcoding process.

4. Key Claim Construction Positions

• "asymmetric data compression encoder[s]": Petitioner argued this term means "an encoder(s) configured to utilize a compression algorithm in which the execution time for the compression and decompression routines differ significantly." This construction was based directly on the patent's explicit definition of an "asymmetrical data compression algorithm" and its examples (e.g., Lempel-Ziv).
• "data blocks": Petitioner proposed construing this term as "a unit of data comprising more than one bit." The rationale was that compression is only possible on data units larger than a single bit and that the patent consistently used the term to refer to units of data operated on by compression algorithms.
• "video or image data profile": Petitioner argued this term means "information used to determine which compression algorithm should be used for a video or image data type." This interpretation was derived from the patent's more general disclosure of "data profiles," which it described as maps associating data types with preferred compression algorithms.

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial would be inappropriate. The petition asserted that it was the first filed by Netflix challenging these specific claims of the ’477 patent. Petitioner also distinguished its grounds from those in a co-pending IPR (IPR2018-01413) filed by other parties, highlighting that the other petition did not rely on the Imai reference, which served as the foundational prior art for every ground asserted in this petition.

6. Relief Requested

• Petitioner requested institution of inter partes review and cancellation of claims 7, 8, 15-19, 23, 24, 28, and 29 of the ’477 patent as unpatentable.