Prior art

Analysis of Prior Art for U.S. Patent No. 11,327,669

This analysis details the prior art cited during the prosecution of U.S. Patent No. 11,327,669. Each cited reference is evaluated for its potential to anticipate the independent claims of the '669 patent under 35 U.S.C. § 102. The core invention of the '669 patent involves a data storage device with a controller that can be configured by a "storage device policy" to manage the trade-off between storage reliability and capacity.

I. U.S. Patent No. 9,665,463 B2

• Full Citation: Tadayon, Bijan, et al. System and Method for Optimizing Non-Volatile Memory Based Storage System Performance by Modifying Data Placement Policies Based on Workload Characteristics. U.S. Patent 9,665,463 B2, issued May 30, 2017. (Filed Jan. 21, 2014).
• Assignee: FADU, Inc.
• Brief Description: This patent describes a system for managing a solid-state drive (SSD) where the data placement policy is dynamically adjusted based on the characteristics of the I/O workload. It aims to optimize performance by altering how and where data is written to the non-volatile memory, considering factors like write amplification and endurance. The system can select from multiple data placement policies to best suit the current workload.
• Potential Anticipation Analysis:
  • Claim 1: This reference teaches a device controller that modifies "data placement policies." This aligns with the '669 patent's "storage device policy." The '463 patent discloses adjusting these policies to optimize performance, which inherently involves trade-offs between factors like speed (a performance metric) and endurance (a reliability metric). It describes storing content based on these policies. However, a key distinction is whether the '463 patent explicitly discloses a policy that manages the "trade-off between reliability of storage and volume of storage." While optimizing for endurance can be seen as a form of reliability, the '463 patent's primary focus is performance optimization, not a direct, user-configurable balance between raw capacity and data integrity. It also does not explicitly teach the concept of refusing a delete request based on the policy.
  • Claim 14 & 20: Similar to Claim 1, the method and system described in the '463 patent involve receiving and applying policies to storage operations. The concept of dynamically altering write strategies based on workload is present. However, the specific trade-off between reliability and volume, and the explicit function of refusing a delete request based on a policy, are not clearly articulated. Furthermore, the '463 patent does not appear to disclose storing the storage information at a separate, remote location as claimed in the '669 patent.

II. U.S. Patent No. 10,255,108 B1

• Full Citation: Bennett, James D. Write Amplification Reduction in Solid-State Drives. U.S. Patent 10,255,108 B1, issued April 9, 2019. (Filed Apr. 12, 2017).
• Assignee: Amazon Technologies, Inc.
• Brief Description: This patent addresses the problem of write amplification in solid-state drives (SSDs). It discloses a method where the SSD controller can be configured with different data placement policies. These policies are selected based on the expected usage patterns or application needs to minimize unnecessary data movement and writes, thereby improving performance and the lifespan of the drive.
• Potential Anticipation Analysis:
  • Claim 1: The '108 patent describes configuring a storage device with "data placement policies," which is analogous to the "storage device policy" in the '669 patent. The policies influence how data is physically placed on the storage medium to reduce write amplification, which directly impacts the endurance and, therefore, the long-term reliability of the SSD. This can be interpreted as managing a trade-off between performance/endurance (reliability) and how efficiently storage is used. However, the '108 patent does not appear to explicitly teach a user-defined policy to trade reliability for volume (i.e., storing data more densely at the risk of lower integrity). It also does not mention a feature to refuse delete requests based on the policy.
  • Claim 14 & 20: The methods described in the '108 patent involve applying a policy to a write request. However, the specific steps of refusing a delete request and storing metadata at a remote location are absent. The focus is on optimizing write patterns, not on providing a broad, configurable policy engine that governs a wide range of device behaviors including immutability and remote metadata management.

III. U.S. Patent No. 10,489,248 B2

• Full Citation: Frost, Gregory, et al. Dynamically Selecting Data Storage Mode to Tune Storage Performance. U.S. Patent 10,489,248 B2, issued Nov. 26, 2019. (Filed Jan. 26, 2018).
• Assignee: Hewlett Packard Enterprise Development LP.
• Brief Description: This invention relates to a storage controller that can dynamically select a data storage mode (e.g., different RAID levels) for a logical volume based on I/O characteristics. For instance, it might switch between a high-performance mode and a high-reliability mode depending on the detected workload, thereby tuning the storage system's behavior.
• Potential Anticipation Analysis:
  • Claim 1: The '248 patent clearly discloses the concept of a policy ("storage mode") that manages a trade-off between performance and reliability (e.g., RAID 1 for high reliability vs. RAID 0 for high performance). This directly addresses the "trade-off between reliability of storage and volume of storage" element of claim 1, as different RAID levels offer different balances of redundancy (reliability) and usable capacity (volume). The controller stores data according to this selected policy. However, the patent does not seem to teach the element of refusing a delete request based on the policy or the associated storage information.
  • Claim 14 & 20: The method of receiving a policy (or selecting a mode based on a policy) and storing data accordingly is taught. The core elements of managing the reliability/capacity trade-off are present. However, like the other references, it appears to be silent on the specific features of refusing a delete request and storing the associated metadata at a remote location.

IV. U.S. Patent Application Publication No. 2005/0216664 A1

• Full Citation: Chang, Robert. Method and Apparatus for Adaptive Data Storage and Management. U.S. Patent Application Publication No. 2005/0216664 A1, published Sep. 29, 2005. (Filed Mar. 24, 2004).
• Assignee: Not specified (Inventor is Chang, Robert).
• Brief Description: This application describes a system that adapts its data storage and management strategies based on various factors, including the type of data, usage patterns, and system policies. It discusses concepts like tiered storage, where data is moved between different types of storage media (e.g., fast, expensive vs. slow, cheap) based on policies to balance cost, performance, and reliability.
• Potential Anticipation Analysis:
  • Claim 1: The '664 application teaches the use of "policies" to manage data storage. It explicitly discusses balancing different storage characteristics, which includes performance, cost, and reliability. This aligns with the "trade-off" concept in the '669 patent. The controller stores data in accordance with these policies. The primary difference is the context; the '664 application appears focused on a higher-level storage system (like a network-attached storage or storage area network) managing multiple tiers of storage, rather than a single device controller's management of its internal physical media. It does not explicitly mention refusing a delete request at the device level based on a policy.
  • Claim 14 & 20: The method of using policies to govern storage is central to this application. However, its system-level perspective may differentiate it from the device-specific claims of the '669 patent. The elements of refusing a delete request and storing metadata remotely are not explicitly detailed in the abstract or a brief review.

V. U.S. Patent Application Publication No. 2013/0198424 A1

• Full Citation: Flynn, David, et al. Storage Controller with Tiered Memory. U.S. Patent Application Publication No. 2013/0198424 A1, published Aug. 1, 2013. (Filed Jan. 25, 2013).
• Assignee: Fusion-io, Inc.
• Brief Description: This document describes a storage device with a controller that manages different types of non-volatile memory (a "tiered" memory system) within a single device. The controller uses policies to determine how data is allocated and moved between these tiers to optimize for performance, endurance, or other characteristics. For example, frequently accessed data might be kept in a faster, more durable memory tier.
• Potential Anticipation Analysis:
  • Claim 1: The '424 application discloses a controller that uses "policies" to manage data placement across different storage media types within the device. This management inherently involves making trade-offs between performance and reliability (endurance of different media types). This is similar to the "trade-off" concept in the '669 patent. However, the focus is on managing tiers of memory rather than adjusting the fundamental write characteristics (like density) on a single type of medium. It also does not appear to describe a policy for refusing to delete data.
  • Claim 14 & 20: The method of using a policy to determine where to store content is present. The system stores content and associated metadata. However, the specific combination of a user-configurable policy for reliability vs. volume, the refusal of deletes, and the remote storage of metadata, as recited in the '669 claims, is not fully taught by this reference.

Summary of Prior Art Analysis

The cited prior art references establish that the concept of using policies within a storage device controller to manage data placement and optimize for performance or reliability was known in the art prior to the invention of US 11,327,669. Patents like '248 (HPE) and '463 (FADU) come close to describing the trade-off between reliability and performance/capacity.

However, none of the reviewed references appear to explicitly disclose the full combination of elements required by the independent claims of the '669 patent. Specifically, the element of the storage device policy including a provision to "refuse a delete request" based on the storage information, and the capability to store the "storage information at a remote location" (as in claim 14), appear to be distinguishing features not clearly present in this set of prior art. These elements, which provide for data immutability and disaggregated metadata management at the device level, may constitute the novel and non-obvious aspects of the invention claimed in U.S. Patent No. 11,327,669.