Primos Storage Technology LLC v. Amazon.com Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Primos Storage Technology LLC (California)
  • Defendant: Amazon.com, Inc. (Delaware); Amazon Web Services, Inc. (Delaware)
  • Plaintiff's Counsel: Farnan LLP
• Case Identification: 1:25-cv-01571, D. Del., 03/06/2026
• Venue Allegations: Plaintiff alleges venue is proper in the District of Delaware because both Defendants are incorporated in Delaware and have committed acts of infringement in the District. The complaint further notes that Amazon operates at least four fulfillment centers within the District.
• Core Dispute: Plaintiff alleges that Defendant's cloud storage products and services, including Amazon S3 and other AWS offerings, infringe five patents related to transaction-based data storage systems and data coding technologies for durability.
• Technical Context: The lawsuit concerns foundational technologies for large-scale, distributed data storage, a critical component of modern cloud computing infrastructure that provides on-demand data storage and retrieval.
• Key Procedural History: The complaint alleges that Defendant Amazon has been aware of the asserted patent families for years. It specifically notes that patent applications that issued as Plaintiff's patents were cited as prior art by the U.S. Patent and Trademark Office during the prosecution of several of Amazon's own patents, beginning as early as 2014. This history is presented as the basis for Plaintiff's willful infringement allegations.

Case Timeline

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

U.S. Patent No. 7,386,663 - "Transaction-Based Storage System and Method that Uses Variable Sized Objects to Store Data," Issued June 10, 2008

The Invention Explained

• Problem Addressed: The patent's background describes the limitations of early 2000s file systems, which were typically "block-oriented," storing data in fixed-size blocks Compl. ¶ 20 '663 Patent, col. 4:4-23 This approach was allegedly slow, inefficient for writing modified data, and did not recover quickly from system crashes, making such systems ill-suited for large-scale, distributed cloud storage environments Compl. ¶¶20-21 '663 Patent, col. 4:1-11
• The Patented Solution: The invention is a "transaction-logging" data storage system that overcomes these limitations by storing variable-sized data objects sequentially in a log Compl. ¶ 27 Instead of overwriting data in place, new data and metadata are written sequentially to available storage space across multiple disks, which are organized into logical groups called "ranks" and "stripes" '663 Patent, abstract '663 Patent, col. 5:21-34 This approach is designed to improve write performance, enhance reliability, and enable features like data versioning and fast crash recovery Compl. ¶ 27
• Technical Importance: The technology represents a shift from single-disk or small-network file systems to an architecture designed for the emerging demands of cloud computing, combining concepts from transaction logging and Redundant Array of Independent Disks (RAID) to create a scalable and durable distributed file system Compl. ¶ 25

Key Claims at a Glance

• The complaint asserts independent claim 1 Compl. ¶ 62
• The essential elements of claim 1 include:
  • A data storage subsystem that stores variable-size data objects using an array of data storage units configured into one or more "ranks," which provide "stripes" for storage.
  • A "rank manager" for configuration processing of the ranks.
  • A "stripe space table" to track space usage.
  • A "mapping component" to process mapping information between object identifiers and their physical addresses.
  • A processing component that coordinates operations to implement the logging of data objects and changes to stored data.
  • The mapping component includes subcomponents to emulate common data storage abstractions (e.g., a key-value store or hierarchical file structure).

U.S. Patent No. 7,937,528 - "Transaction-Based Storage System and Method that Uses Variable Sized Objects to Store Data," Issued May 3, 2011

The Invention Explained

• The '528 Patent shares a common specification with the '663 Patent and is part of the same "Transaction-Based Storage Patents" family Compl. ¶ 15
• Problem Addressed: As with the '663 Patent, the invention addresses the speed, scalability, and reliability limitations of conventional block-oriented storage systems in the context of large-scale data management Compl. ¶¶17-18
• The Patented Solution: The '528 Patent claims a method for logging transactions in the data storage system described in the '663 Patent. The method involves storing variable-size objects in a subsystem comprised of ranks and stripes, keeping track of space usage with a stripe space table, processing mapping information, and performing coordinated operations to log new and changed data objects '528 Patent, claim 1 '663 Patent, col. 10:20-24
• Technical Importance: This patent claims the method of operating the novel file system architecture, complementing the system claims of the related '663 Patent and covering the operational process for achieving high-throughput, log-structured storage Compl. ¶ 27

Key Claims at a Glance

• The complaint asserts independent claim 1 Compl. ¶ 104
• The essential elements of method claim 1 include:
  • Storing variable-size data objects in a data storage subsystem that uses an array of storage units configured into "ranks" and "stripes."
  • Keeping track of space usage via a "stripe space table."
  • Processing, via a "mapping component," information between object identifiers and their physical addresses.
  • Performing coordinated operations among the subsystem, mapping component, and stripe space table to implement the logging of new and changed data objects.

Multi-Patent Capsule: U.S. Patent No. 10,599,344

• Patent Identification: U.S. Patent No. 10,599,344, "Transaction-Based Storage System and Method that Uses Variable Sized Objects to Store Data," Issued March 24, 2020.
• Technology Synopsis: The '344 Patent is also part of the "Transaction-Based Storage Patents" family and shares the same specification as the '663 and '528 patents Compl. ¶ 15 It claims a storage system comprising a file system that uses variable-size objects, a plurality of ranks providing stripes, and at least one stripe buffer for appending objects before they are written as a unit into the ranks Compl. ¶ 142
• Asserted Claims: Independent claim 1 and dependent claim 2 Compl. ¶ 142
• Accused Features: The complaint accuses products like Amazon S3 and Hadoop of infringing by, for example, storing data in a buffer before writing it across multiple storage servers (R2D2s or DataNodes) Compl. ¶¶146-148 Compl. ¶¶156-157

Multi-Patent Capsule: U.S. Patent No. 8,078,944

• Patent Identification: U.S. Patent No. 8,078,944, "Systems, Methods and Computer Program Products Including Features for Coding and/or Recovering Data," Issued December 13, 2011.
• Technology Synopsis: The '944 Patent, one of the "Data Coding Patents," addresses the problem of ensuring data durability on storage systems with high hard drive failure rates Compl. ¶ 30 It discloses a method of "erasure coding," which transforms "W" data inputs into "H" output data streams using Galois field operations and a generator matrix, allowing the original data to be recovered from any "W" of the "H" streams Compl. ¶ 34 The method also involves creating "H-sized intermediaries" to improve performance and applying checksums to ensure data integrity Compl. ¶¶35-37
• Asserted Claims: Independent claim 1 and dependent claim 2 Compl. ¶ 188
• Accused Features: The complaint accuses Amazon S3 and Hadoop/EMR of using erasure coding to divide data objects into "shards" (W data inputs) and generate "parity shards" (creating H output streams) for durable storage across multiple disks Compl. ¶¶189-191 Compl. ¶¶196-198

Multi-Patent Capsule: U.S. Patent No. 8,312,356

• Patent Identification: U.S. Patent No. 8,312,356, "Systems, Methods and Computer Program Products Including Features for Coding and/or Recovering Data," Issued November 13, 2012.
• Technology Synopsis: The '356 Patent, the other "Data Coding Patent," shares a common specification with the '944 Patent and claims a system for performing the erasure coding method Compl. ¶ 28, fn. 39 It describes a system with non-transitory computer-readable media containing instructions for a processor to perform the transformation process, including producing H-sized intermediaries and distributing the resulting output data streams across processing components Compl. ¶ 219
• Asserted Claims: Independent claim 17 and dependent claim 20 Compl. ¶ 219
• Accused Features: The complaint alleges that Amazon S3 and Hadoop/EMR are systems that implement the claimed erasure coding process, breaking data into chunks, creating H-sized intermediaries, and distributing the resulting shards across multiple disks and servers Compl. ¶¶220-224 Compl. ¶¶227-231

III. The Accused Instrumentality

• Product Identification: The complaint identifies two categories of accused products: "Transaction-Based Storage Patents Accused Products" and "Data Coding Patents Accused Products" (Compl. ¶¶40; Compl. ¶ 54). Specific examples include Amazon Simple Storage Service (S3), Apache Hadoop on Amazon Elastic MapReduce (EMR), Amazon Elastic Compute Cloud (EC2) instances running the XFS file system (including on Amazon Elastic Block Store (EBS)), and Amazon FSx for Lustre Compl. ¶ 40
• Functionality and Market Context:
  • The accused instrumentalities are core components of Amazon's cloud computing platform, marketed as providing highly durable, scalable, and high-performance data storage (Compl. ¶ 41).
  • Amazon S3: An object-based storage service that implements versioning to maintain a log of every version of a stored object (Compl. ¶ 42). It is alleged to employ "striping techniques," distributing data objects as "shards" across arrays of storage devices to enhance performance (Compl. ¶ 43; Compl. ¶ 66). S3 is also alleged to use erasure coding and checksums to achieve its marketed "99.999999999% data durability" (Compl. ¶¶55-56). An architectural diagram provided in the complaint illustrates S3's use of "sharding" to spread data across physical drives Compl. p. 28
  • Hadoop on Amazon EMR: An offering that integrates the open-source Hadoop Distributed File System (HDFS) framework for large-scale data processing (Compl. ¶ 45). HDFS is alleged to implement a "write-ahead log" to record file system modifications as transactions and to employ striping by splitting files into blocks written across multiple storage units (DataNodes) (Compl. ¶¶46-47). The complaint includes an architectural diagram showing how HDFS distributes and replicates data blocks across different "Datanodes" organized into "Racks" Compl. p. 34

IV. Analysis of Infringement Allegations

'663 Patent Infringement Allegations

'528 Patent Infringement Allegations

• Identified Points of Contention:
  • Scope Questions: A central point of contention may be whether the highly abstracted and proprietary components of Amazon's cloud architecture (e.g., "R2D2s," "C3PO," "Skynet," "blindexes") can be mapped onto the more concrete architectural elements recited in the claims, such as "rank", "rank manager", and "stripe space table". The defense may argue that Amazon's system is architecturally distinct and does not meet these limitations. The complaint's use of a diagram from an AWS presentation showing "Data shards" and "Parity shards" suggests Plaintiff will argue that Amazon's own technical descriptions align with the patent's concepts Compl. p. 108
  • Technical Questions: The infringement theory hinges on whether the accused products' functionalities perform the specific actions required by the claims. For instance, a question may arise as to whether Amazon S3's "versioning" feature (Compl. ¶ 63) constitutes the "transaction-logging" system claimed in the '663 Patent, or if there are fundamental operational differences. Similarly, for the Data Coding patents, a question may be whether Amazon's erasure coding implementation uses the claimed method of creating and combining "H-sized intermediaries" Compl. ¶ 192

V. Key Claim Terms for Construction

• The Term: "rank"

  • Context and Importance: This term is foundational to the claimed system architecture in the Transaction-Based Storage Patents. Its definition will determine whether Amazon's alleged groupings of servers ("R2D2s" in S3 or "DataNodes" in HDFS) fall within the scope of the claims. Practitioners may focus on this term because the complaint's theory requires mapping Amazon's complex, distributed server pools to this specific claim element.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specification describes a rank as including an "array of disks" '663 Patent, col. 5:42-43 and notes that the system is implemented using "arrays disks that are arranged in ranks" '663 Patent, abstract, which could support a broad interpretation covering any logical or physical grouping of storage devices.
    • Evidence for a Narrower Interpretation: The detailed description frequently discusses ranks in the context of RAID configurations '663 Patent, col. 3:1-65, which might support an argument that a "rank" must possess specific RAID-like properties, potentially narrowing its scope.

• The Term: "stripe space table"

  • Context and Importance: This element is a specific mechanism for resource management in the claimed transaction-logging system. The infringement analysis will depend on whether the methods used by Amazon S3's "C3PO" servers or HDFS's "NameNode" to track available storage (Compl. ¶ 68; Compl. ¶ 117) are equivalent to the claimed "stripe space table".
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The claim language requires a table that "keeps track of space usage" Compl. ¶ 62 This functional language may support a broad construction that covers any data structure or process that achieves this purpose, regardless of its specific implementation.
    • Evidence for a Narrower Interpretation: The patent depicts the "stripe space table" with specific categories such as "EMPTY," "ALMOST EMPTY," and "NOT EMPTY" '663 Patent, FIG. 11 A defendant may argue that the term should be limited to a structure with these or similar discrete categorical states, rather than any generic space-tracking mechanism.

VI. Other Allegations

• Indirect Infringement: The complaint alleges that Amazon induced infringement by actively encouraging customers to use the accused products through marketing materials, user guides, and technical assistance (Compl. ¶ 99; Compl. ¶ 137). Contributory infringement is alleged on the basis that the accused products are material components of the invention, are not staple articles of commerce, and were known by Amazon to be especially adapted for infringing use (Compl. ¶ 100; Compl. ¶ 138).
• Willful Infringement: Willfulness allegations are based on Amazon's alleged long-standing, pre-suit knowledge of the asserted patent families (Compl. ¶¶93-96). The complaint specifically alleges that applications corresponding to the asserted patents were cited as prior art during the prosecution of Amazon's own patents, forcing Amazon to "study the specification" of the patents-in-suit to overcome rejections (Compl. ¶¶89-91; Compl. ¶ 127-129).

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

• A core issue will be one of definitional scope: can the architectural terms of the patents, such as "rank", "rank manager", and "stripe space table", which are described in the context of RAID-like disk arrays, be construed to cover the highly abstracted, proprietary, and massively distributed server and software components of Amazon's modern cloud storage infrastructure?
• A central evidentiary question will be one of functional equivalence: does the accused S3 "versioning" system or HDFS "write-ahead log" operate in substantially the same way as the claimed "transaction-logging" system, and does Amazon's erasure coding algorithm practice the specific claimed steps of generating and combining "H-sized intermediaries"?
• A key question for damages and willfulness will be one of knowledge and intent: what was the extent of Amazon's knowledge of the asserted patents based on the cited prosecution history of its own patent applications, and what steps, if any, did it take to assess the risk of infringement?