DCT

1:22-cv-12010

Exegy Inc v. NovaSparks SA

Log In

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 1:22-cv-12010, D. Mass., 11/23/2022
  • Venue Allegations: Plaintiff alleges venue is proper in the District of Massachusetts because Defendant NovaSparks, Inc. is incorporated there, and both defendants are alleged to have regularly transacted business and committed acts of infringement in the district.
  • Core Dispute: Plaintiff alleges that Defendant’s FPGA-based financial data processing products infringe seven patents related to high-speed, hardware-accelerated processing of financial market data.
  • Technical Context: The technology domain involves using Field Programmable Gate Arrays (FPGAs) to accelerate the processing of real-time financial market data, a critical capability for ultra-low latency trading where microsecond advantages can be significant.

Case Timeline

Date Event
2006-06-19 Priority Date for U.S. Patent 7,921,046
2007-01-11 Priority Date for U.S. Patent 10,229,453
2010-07-01 Priority Date for U.S. Patents 8,762,249 & 8,768,805
2011-04-05 U.S. Patent 7,921,046 Issues
2011-12-14 Priority Date for U.S. Patent 9,047,243
2012-03-27 Priority Date for U.S. Patents 9,990,393 & 10,121,196
2014-06-24 U.S. Patent 8,762,249 Issues
2014-07-01 U.S. Patent 8,768,805 Issues
2015-06-02 U.S. Patent 9,047,243 Issues
2018-06-05 U.S. Patent 9,990,393 Issues
2018-11-06 U.S. Patent 10,121,196 Issues
2019-03-12 U.S. Patent 10,229,453 Issues
2022-11-23 Complaint Filed

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 7,921,046 - "High Speed Processing of Financial Information Using FPGA Devices"

  • Patent Identification: U.S. Patent No. 7,921,046, issued April 5, 2011.

The Invention Explained

  • Problem Addressed: The patent’s background section describes the limitations of conventional software-based systems for processing the increasingly high volume and velocity of financial market data, noting that such systems introduce unacceptable latency (’046 Patent, col. 1:25-45).
  • The Patented Solution: The invention proposes a hybrid “ticker plant” that combines software logic for control and management with firmware logic deployed on a reconfigurable device, such as an FPGA, for high-speed data processing (’046 Patent, Abstract). This architecture is designed to process streaming financial data at hardware speeds by managing a data flow that passes messages only once from the software to the firmware and once from the firmware back to the software for a given operation (’046 Patent, col. 4:1-15).
  • Technical Importance: This hybrid architecture enabled processing financial data feeds at speeds significantly greater than commodity servers running only software, a key advantage in latency-sensitive electronic trading (Compl. ¶¶8, 33).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶35).
  • Essential elements of claim 1 include:
    • A ticker plant for processing financial market data, comprising a combination of software logic and firmware logic.
    • The firmware logic is deployed on a reconfigurable logic device.
    • The software logic and firmware logic cooperate to perform a specified data processing operation.
    • The software logic is configured to manage the flow of financial market data into and out of the reconfigurable logic device.
    • Each message of the financial market data passes only once from the software logic to the firmware logic and only once from the firmware logic to the software logic.

U.S. Patent No. 8,762,249 - "Method and Apparatus for High-Speed Processing of Financial Market Depth Data"

  • Patent Identification: U.S. Patent No. 8,762,249, issued June 24, 2014.

The Invention Explained

  • Problem Addressed: The patent addresses the challenge of processing high-volume "market depth" data (also known as Level 2 data), which contains numerous individual buy and sell orders at different price levels, a task that is computationally intensive for traditional software systems (’249 Patent, col. 1:25-44).
  • The Patented Solution: The invention describes a method using a processing pipeline deployed in firmware on a reconfigurable logic device (e.g., an FPGA) to process a stream of limit order events (’249 Patent, Abstract). This hardware-based pipeline generates a "stream view of an order book," which represents a normalized stream of updates for a financial instrument, thereby accelerating the construction and maintenance of market order books (’249 Patent, col. 2:40-59).
  • Technical Importance: By accelerating the processing of market depth data, the invention allows trading systems to obtain a more accurate, real-time view of market liquidity, which is critical for making fast and informed trading decisions (Compl. ¶41).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶43).
  • Essential elements of claim 1 include:
    • A method of processing financial market data.
    • Receiving a stream of messages representative of limit order events for a financial instrument.
    • Processing the stream of messages with a processing pipeline deployed in firmware logic on a reconfigurable logic device.
    • The processing generates a stream view of an order book for the financial instrument.
    • The stream view comprises a normalized stream of updates for limit orders for the financial instrument.

U.S. Patent No. 8,768,805 - "Method and Apparatus for High-Speed Processing of Financial Market Depth Data"

  • Patent Identification: U.S. Patent No. 8,768,805, issued July 1, 2014.
  • Technology Synopsis: The patent addresses the high-speed processing of financial market depth data. It describes using data processing pipelines implemented in firmware on reconfigurable logic to create a ticker plant capable of accelerating the speed with which this data is processed (Compl. ¶¶48-49).
  • Asserted Claims: Independent claim 1 (Compl. ¶51).
  • Accused Features: The Accused Products are alleged to be ticker plants that process financial market depth data with firmware logic (Compl. ¶¶49-50).

U.S. Patent No. 9,047,243 - "Method and Apparatus for Low Latency Data Distribution"

  • Patent Identification: U.S. Patent No. 9,047,243, issued June 2, 2015.
  • Technology Synopsis: The patent relates to distributing processed financial market data to consumers with low latency. The described method uses direct memory access (DMA) write operations to deliver consumer-specific data streams into shared memory (Compl. ¶¶56-57).
  • Asserted Claims: Independent claim 24 (Compl. ¶59).
  • Accused Features: The Accused Products, which are alleged to process and distribute financial data at high speeds (Compl. ¶¶56, 58).

U.S. Patent No. 9,990,393 - "Intelligent Feed Switch"

  • Patent Identification: U.S. Patent No. 9,990,393, issued June 5, 2018.
  • Technology Synopsis: The patent describes offloading data processing to an "intelligent feed switch" located upstream or downstream from a trading platform. This switch contains co-resident switching logic and a processor (such as an FPGA) that analyzes and repacketizes message data into outgoing, consumer-specific data packets (Compl. ¶¶64-65).
  • Asserted Claims: Independent claim 1 (Compl. ¶67).
  • Accused Features: The Accused Products, which are alleged to function as intelligent switches for processing financial market data (Compl. ¶66).

U.S. Patent No. 10,121,196 - "Offload Processing of Data Packets Containing Financial Market Data"

  • Patent Identification: U.S. Patent No. 10,121,196, issued November 6, 2018.
  • Technology Synopsis: The patent describes an intelligent feed switch with co-resident switching logic and a processor (e.g., FPGA) for offloading data processing. The processor performs a packet mapping operation to determine the financial market data feed associated with an incoming packet, access its metadata, and associate that metadata with the packet (Compl. ¶¶72-73).
  • Asserted Claims: Independent claim 1 (Compl. ¶75).
  • Accused Features: The Accused Products, which are alleged to function as intelligent feed switches that offload processing operations (Compl. ¶74).

U.S. Patent No. 10,229,453 - "Method and System for Low Latency Basket Calculation"

  • Patent Identification: U.S. Patent No. 10,229,453, issued March 12, 2019.
  • Technology Synopsis: The patent teaches a data processing platform with a basket calculation engine implemented on an FPGA. A hardware pipeline processes streaming financial market data to compute net asset values (NAVs) for baskets of financial instruments using a delta calculation approach (Compl. ¶¶80-81).
  • Asserted Claims: Independent claim 1 (Compl. ¶83).
  • Accused Features: Defendant's "ETF Calculator" and other Accused Products are alleged to be data processing platforms that compute basket values for financial instruments (Compl. ¶¶13, 82).

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are the NovaTick Ticker Plant, NovaTick Appliance, NovaTick Stand-Alone PCIe Card, NovaLink API, NovaSparks API, NovaSparks Wire Format (NSWF), NovaTick NBBO Calculator, and ETF Calculator products (collectively, the “Accused Products”) (Compl. ¶13).

Functionality and Market Context

The complaint alleges the Accused Products are “FPGA-based high performance and ultra-low latency trading solutions for financial markets” (Compl. ¶10). Their function is to provide “normalized market data, including book building and consolidation in less than one micro-second” by using FPGA processors for high-speed processing of financial information (Compl. ¶¶11, 12). The complaint provides a series of screen captures from Defendant's website setting forth details of the Accused Products that process real-time financial market data feeds using FPGAs (Compl. ¶13). It further alleges that the Accused Products compete directly with Plaintiff's patented products and services (Compl. ¶16).

IV. Analysis of Infringement Allegations

The complaint references claim-chart exhibits (Exhibits 9-15) that are not provided in the document. The infringement theory for each lead patent is summarized below in prose.

'7,921,046 Infringement Allegations

The complaint alleges that the Accused Products directly infringe at least claim 1 of the ’046 Patent (Compl. ¶35). The infringement theory is that the Accused Products function as a "ticker plant" that processes financial market data using a combination of software and firmware logic implemented on reconfigurable hardware (FPGAs) (Compl. ¶¶32-33). It is alleged that this architecture accelerates data processing to hardware speeds and satisfies each limitation of the claim (Compl. ¶36).

'8,762,249 Infringement Allegations

The complaint alleges that the Accused Products directly infringe at least claim 1 of the ’249 Patent (Compl. ¶43). The infringement theory is that the Accused Products are configured as a ticker plant to process financial market depth data using firmware logic. This use of firmware pipelines is alleged to greatly accelerate the speed of processing and to meet every limitation of the claim (Compl. ¶¶40-41, 44).

Identified Points of Contention

  • Scope Questions: A potential issue for the court will be the construction of the term "ticker plant" as used in the ’046 and ’249 Patents. The dispute may turn on whether Defendant’s products, described as “FPGA processors” and solutions, constitute a complete "ticker plant" as defined by the patents, or if they are merely components of such a system.
  • Technical Questions: For the ’046 Patent, a key technical question is whether the accused products’ architecture adheres to the "passes only once" limitation for data flow between software and firmware. For the ’249 Patent, a central question will be whether the processing performed by the Accused Products constitutes the claimed step of "generat[ing] a stream view of an order book", which may require a specific set of data transformations beyond simple filtering or forwarding.

V. Key Claim Terms for Construction

’046 Patent, Claim 1: "passes only once"

  • Context and Importance: This term is central to the claimed efficiency of the invention. Infringement will depend on whether the data flow architecture in the Accused Products for a given market data message constitutes a single round-trip between the software and firmware logic, as opposed to a more iterative or chatty communication protocol.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent's focus is on achieving hardware-level speed, which may suggest that the "single pass" refers to the core data payload for a processing operation, potentially allowing for ancillary control signals to make multiple trips without defeating the purpose of the claim.
    • Evidence for a Narrower Interpretation: The explicit language "only once from the software logic to the firmware logic... and only once from the firmware logic to the software logic" could be interpreted strictly to mean that for any given message-processing task, the entire associated data and control exchange is limited to a single, discrete handoff in each direction.

VI. Other Allegations

Willful Infringement

The complaint alleges that Defendants' infringement has been willful for all seven asserted patents (Compl. ¶¶38, 46, 54, 62, 70, 78, 86). The complaint does not, however, plead any specific facts to support pre-suit knowledge of the patents, such as prior correspondence, litigation, or public statements.

VII. Analyst’s Conclusion: Key Questions for the Case

  • A core issue will be one of definitional scope: can the term "ticker plant," which implies a complete data processing system, be construed to cover Defendant's products, which are marketed as FPGA-based "processors" and "solutions" for financial data? The case may depend on whether the Accused Products are found to be infringing components or entire infringing systems under the patent claims.
  • A key evidentiary question will be one of technical operation: what is the precise data-flow architecture between the software and hardware elements within the NovaTick products? Discovery will likely focus on whether this architecture meets the specific "passes only once" limitation of the ’046 Patent, a functional requirement central to the patent’s claim of novelty and efficiency.