DCT
3:18-cv-03388
Raytheon Co v. Cray Inc
Key Events
Complaint
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Raytheon Company (Delaware)
- Defendant: Cray, Inc. (Washington)
- Plaintiff’s Counsel: THE DAVIS FIRM P.C.; Steptoe & Johnson LLP
- Case Identification: 2:16-cv-00423, E.D. Tex., 04/22/2016
- Venue Allegations: Plaintiff alleges venue is proper in the Eastern District of Texas based on Defendant’s registration to do business in Texas, its employment of sales and marketing personnel within the district, and its sales of accused supercomputer systems to customers located in the district, such as the University of Texas system.
- Core Dispute: Plaintiff alleges that Defendant’s high-performance computing systems infringe five patents related to methods for topology-aware job scheduling and backfilling in multi-node computing environments.
- Technical Context: The technology concerns software and systems for efficiently managing and allocating computational resources in supercomputers, a market critical for scientific research, energy exploration, and financial modeling.
- Key Procedural History: The complaint alleges that Defendant has been on notice of the asserted patent portfolio since at least June 30, 2015. It also references co-pending litigation between the parties involving other high-performance computing patents, filed on September 25, 2015, suggesting a pre-existing dispute over related technology.
Case Timeline
| Date | Event |
|---|---|
| 2004-04-15 | Earliest Priority Date for all Patents-in-Suit |
| 2012-12-18 | U.S. Patent No. 8,336,040 Issues |
| 2014-12-09 | U.S. Patent No. 8,910,175 Issues |
| 2015-03-17 | U.S. Patent No. 8,984,525 Issues |
| 2015-06-30 | Alleged Date of Actual Notice to Defendant |
| 2015-09-25 | Co-pending Litigation Filed Between Parties |
| 2015-11-17 | U.S. Patent No. 9,189,275 Issues |
| 2015-11-17 | U.S. Patent No. 9,189,278 Issues |
| 2016-04-22 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 8,336,040 - "System and method for topology-aware job scheduling and backfilling in an HPC environment"
The Invention Explained
- Problem Addressed: The patent’s background section describes that in conventional High Performance Computing (HPC) environments, the processing, memory, and input/output (I/O) bandwidth are often not well-balanced. This imbalance can limit parallel scalability, which is a critical factor for the performance of large scientific and engineering applications. (’040 Patent, col. 1:13-37).
- The Patented Solution: The invention proposes a system for job management where an unallocated subset of HPC nodes is identified, with each node comprising an "integrated fabric." A job is selected from a queue and executed on these nodes. By distributing switching functionality across the nodes, the system aims to provide I/O performance that is substantially equal to processing performance, thereby increasing efficiency and scalability. (’040 Patent, Abstract; col. 2:47-54).
- Technical Importance: This balanced, distributed architecture was intended to create more scalable, reliable, and fault-tolerant HPC systems compared to conventional designs that relied on more centralized switching. (’040 Patent, col. 2:58-64).
Key Claims at a Glance
- The complaint asserts independent Claim 1. (Compl. ¶14).
- The essential elements of Claim 1 are:
- Determining an original subset of unallocated nodes from a plurality of nodes, where each node comprises a "switching fabric comprising a switch integrated on the card."
- Selecting a job from a job queue.
- Executing the job on processors within the original subset of nodes.
- Determining that the job's required dimensions are greater than the topology of the original subset.
- Selecting one or more additional nodes from a second plurality of nodes that are unavailable at the time of selecting.
- Adding the selected unavailable nodes to the original subset to satisfy the job's dimensions after those nodes become available.
- The complaint does not explicitly reserve the right to assert dependent claims.
U.S. Patent No. 9,189,275 - "System and method for topology-aware job scheduling and backfilling in an HPC environment"
The Invention Explained
- The complaint does not provide the text of the ’275 Patent. Based on its identical title and its assertion alongside the related ’040 Patent, it addresses the same technical field.
- Problem Addressed: As described in the related ’040 Patent, the technology addresses performance limitations in HPC environments caused by imbalances between processing power and I/O bandwidth, which hinders scalability. (’040 Patent, col. 1:13-37).
- The Patented Solution: The technology provides for more efficient job scheduling by being aware of the physical and logical topology of the computing nodes, allowing for dynamic allocation and "backfilling" of jobs to optimize the use of available resources. (’040 Patent, Abstract).
- Technical Importance: This approach seeks to improve overall system throughput and reduce turnaround time for computational jobs in large, production-oriented HPC environments. (’040 Patent, col. 1:33-37).
Key Claims at a Glance
- The complaint asserts independent Claim 1. (Compl. ¶22).
- The complaint does not provide the text of the ’275 Patent or its asserted claims, precluding a detailed breakdown of claim elements.
Multi-Patent Capsule: U.S. Patent No. 8,910,175 - "System and method for topology-aware job scheduling and backfilling in an HPC environment"
- Technology Synopsis: The patent describes a system for managing jobs in an HPC environment by determining an unallocated subset of nodes, each with an integrated fabric, and executing a selected job from a queue using those nodes. The technology aims to solve scalability problems in conventional HPC systems by better balancing I/O and processing performance.
- Asserted Claims: The complaint asserts independent Claim 1. (Compl. ¶30).
- Accused Features: The complaint accuses the Cray CS400-AC supercomputer and other Cray systems of infringement. (Compl. ¶30).
Multi-Patent Capsule: U.S. Patent No. 9,189,278 - "System and method for topology-aware job scheduling and backfilling in an HPC environment"
- Technology Synopsis: The patent describes a system for managing jobs in an HPC environment by determining an unallocated subset of nodes, each with an integrated fabric, and executing a selected job from a queue using those nodes. The technology aims to solve scalability problems in conventional HPC systems by better balancing I/O and processing performance.
- Asserted Claims: The complaint asserts independent Claim 1. (Compl. ¶38).
- Accused Features: The complaint accuses the Cray CS400-AC and XC40 supercomputers, as well as other Cray systems, of infringement. (Compl. ¶38).
Multi-Patent Capsule: U.S. Patent No. 8,984,525 - "System and method for topology-aware job scheduling and backfilling in an HPC environment"
- Technology Synopsis: The patent describes a system for managing jobs in an HPC environment by determining an unallocated subset of nodes, each with an integrated fabric, and executing a selected job from a queue using those nodes. The technology aims to solve scalability problems in conventional HPC systems by better balancing I/O and processing performance.
- Asserted Claims: The complaint asserts independent Claim 1. (Compl. ¶46).
- Accused Features: The complaint accuses the Cray CS400-AC and XC40 supercomputers, as well as other Cray systems, of infringement. (Compl. ¶46).
III. The Accused Instrumentality
Product Identification
- The complaint names the Cray XC40 and Cray CS-400-AC supercomputers, as well as other unspecified "systems developed by Cray," as the accused instrumentalities. (Compl. ¶¶14, 30, 38, 46).
Functionality and Market Context
- The complaint identifies the accused products as "high performance computers" and "cluster supercomputers." (Compl. ¶¶6, 13). It alleges that Defendant markets these systems to customers in various industries, including the petroleum industry, and has sold an infringing XC40 supercomputer to the University of Texas for use in its Texas Advanced Computing Center (TACC). (Compl. ¶¶6-7).
- The complaint does not provide specific technical details about the architecture, interconnect fabric, or job scheduling software of the accused supercomputers. The infringement theory is based on the general function of these systems in executing computational jobs in a high-performance, multi-node environment. (Compl. ¶¶14, 22, 30, 38, 46). No probative visual evidence provided in complaint.
IV. Analysis of Infringement Allegations
The complaint makes general allegations of infringement without providing an element-by-element mapping of the accused products to the asserted claims. The following chart summarizes the infringement theory for the lead patent based on the complaint's global allegations.
'040 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| A method comprising: determining... an original subset of a plurality of nodes, the original subset comprising nodes currently unallocated to a job, each node in the plurality of nodes comprising a switching fabric comprising a switch integrated on the card... | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 7:59-8:7 |
| selecting a job from a job queue; | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 11:40-42 |
| executing the selected job using one or more processors of one or more nodes of the original subset; | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 12:65-67 |
| determining that dimensions of the selected job are greater than a topology of the original subset; | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 15:8-13 |
| selecting one or more nodes from a second plurality of nodes... wherein the selected one or more nodes from the second plurality are unavailable at the time of selecting; and | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 15:8-13 |
| adding the nodes selected from the second plurality to the original subset to satisfy the dimensions of the selected job after the nodes selected from the second plurality become available. | The complaint alleges that Cray’s supercomputer systems, such as the XC40, perform topology-aware job scheduling and therefore practice the claimed method. | ¶14 | col. 15:13-17 |
- Identified Points of Contention:
- Technical Questions: The complaint does not provide factual support detailing how the accused Cray systems perform the specific steps of Claim 1 of the ’040 Patent. A central question will be whether Cray's job schedulers implement the claimed "backfilling" method, which requires selecting specific nodes that are currently unavailable and reserving them for a job to be executed after they become available. Evidence of this specific reservation-and-execution sequence will be critical.
- Scope Questions: A potential dispute may arise over the scope of the term "switching fabric comprising a switch integrated on the card." The infringement analysis will depend on whether the specific hardware architecture of Cray's interconnect technology (e.g., the Aries interconnect used in the XC series) meets this claim limitation.
V. Key Claim Terms for Construction
- The Term: "switching fabric comprising a switch integrated on the card"
- Context and Importance: This term appears in Claim 1 of the ’040 Patent and defines a specific hardware architecture for the nodes. The case may turn on whether the architecture of the accused Cray supercomputers, which may use different interconnect technologies, falls within the scope of this term. Practitioners may focus on this term because it appears to be a key point of novelty distinguishing the invention from systems with centralized switches.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification's summary states that an advantage of the invention is "at least partially reducing, distributing, or eliminating centralized switching functionality." (’040 Patent, col. 2:47-50). This stated purpose could be used to argue for a construction that covers a variety of distributed switching architectures, not just the exact one depicted.
- Evidence for a Narrower Interpretation: The detailed description and figures repeatedly illustrate this concept with a specific embodiment: an integrated switch (345) on a blade or motherboard (315). (’040 Patent, Fig. 3A; col. 8:1-7). This explicit depiction of a "switch" on a "card" could support a narrower construction limited to that hardware arrangement.
VI. Other Allegations
- Indirect Infringement: The complaint alleges induced infringement, stating that Defendant knowingly and intentionally encouraged others to use and sell the accused systems in a manner that constitutes infringement. (Compl. ¶15). The complaint does not, however, plead specific facts supporting this allegation, such as references to user manuals or marketing materials that instruct on infringing use.
- Willful Infringement: The complaint alleges willful infringement based on Defendant's alleged "actual notice" of the asserted patents since "at least about June 30, 2015," and its subsequent decision "to continue to infringe." (Compl. ¶¶13, 17). The allegation is grounded in specific pre-suit knowledge allegedly conveyed to Defendant.
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of technical implementation: does the job scheduling software in the accused Cray supercomputers perform the specific "backfilling" methodology recited in Claim 1 of the ’040 Patent, which requires selecting currently unavailable nodes for a job and adding them to the allocation only after they become free, or does it use a different, non-infringing scheduling algorithm?
- A second central question will be one of hardware scope: can the claim term "switching fabric comprising a switch integrated on the card" be construed to read on the architecture of Defendant’s proprietary interconnects (e.g., the Aries interconnect), or is there a fundamental structural difference that places the accused systems outside the scope of the claims?
- Given the general nature of the infringement allegations, an early procedural question may be one of pleading sufficiency: does the complaint provide sufficient factual detail under the "Twombly/Iqbal" standard to make a plausible claim for infringement, or is it vulnerable to a motion to dismiss for failure to state a claim?