Obviousness

Obviousness Analysis under 35 U.S.C. § 103 for US Patent 9002795

This analysis identifies combinations of prior art references that would render the claims of US Patent 9002795 obvious to a person having ordinary skill in the art (PHOSITA) at the time of the invention (priority date January 26, 2006). The focus will be on independent claim 1 and how the identified prior art, combined with common general knowledge in the field of data storage, would lead to the claimed invention.

Combination of Prior Art References

A strong argument for obviousness can be made by combining US7124272B1, US6823398B1, and US20020095546A1, along with the common knowledge of storage system design.

Claim 1 of US9002795 describes a method for allocating data attributes from an i-node to storage media by:

1. Sending a data object with i-node attributes from a host to a storage device.
2. The storage device having an object-based storage interface operable to allocate data to particular media zones based on access frequency, and storage media with multiple physically separate zones having differing storage performance attributes.
3. Receiving and providing the data object to the object-based storage interface.
4. Analyzing i-node attributes to determine their access frequencies.
5. Allocating the i-node attributes based on these frequencies, specifically:
   • Less frequently accessed portions (i-node write attributes) to one zone.
   • More frequently accessed portions (i-node read attributes) to another zone.
6. Storing the i-node attributes in the allocated zones.

Detailed Analysis of the Combination:

1. US7124272B1 (File usage history log for improved placement of files in differential rate memory according to frequency of utilizations and volatility of allocation space):

   • This reference explicitly teaches "improved placement of files in differential rate memory according to frequency of utilizations". "Differential rate memory" can be understood by a PHOSITA as storage media with differing performance characteristics, analogous to the "multiple media zones with differing zone attributes of storage performance" in US9002795. The core concept of allocating data based on its "frequency of utilizations" directly addresses the requirement in claim 1 to analyze "i-node attribute access frequencies" and allocate based on them. "Files" in this context would implicitly include their associated metadata, such as i-nodes.
   • This patent establishes the fundamental principle of utilizing usage patterns (access frequency) to intelligently place data on storage media with varying performance characteristics.

2. US6823398B1 (File system management embedded in a storage device):

   • This patent describes "file system management embedded in a storage device". This directly anticipates the "object-based storage device" (OSD) concept of US9002795, where low-level storage functions and space management are performed at the drive level rather than by the host system. This embedded file system provides the necessary "object-based storage interface" within the storage device, capable of making allocation decisions based on internal knowledge of the media.
   • A PHOSITA would be motivated to implement the usage-based allocation techniques of US7124272B1 within such an embedded file system (US6823398B1) because an embedded system has direct, intimate knowledge of the physical media's characteristics (zone attributes, defects, etc.) that a host-level file system lacks. This integration would enable more optimized and "smart decisions" regarding data placement, as described in US9002795.

3. US20020095546A1 (Method, system, and program for writing files to zone formatted storage media to improve data transfer rates):

   • This reference teaches "writing files to zone formatted storage media to improve data transfer rates". This clearly establishes the concept of "multiple media zones with differing zone attributes of storage performance" and the motivation to use these zones to enhance performance (e.g., data transfer rates).
   • A PHOSITA combining this with US7124272B1 would understand that the "differential rate memory" of US7124272B1 can be practically implemented using "zone formatted storage media" where different zones inherently offer different performance levels (e.g., outer tracks typically having higher throughput than inner tracks).

Motivation for Combination:

A PHOSITA, seeking to develop a more efficient and intelligent storage device that optimizes performance and reliability, would be highly motivated to combine these teachings:

• Improvement through device-level intelligence: US6823398B1 teaches the benefits of embedding file system management in the storage device itself, providing the drive with superior knowledge of its physical media characteristics compared to a host system. A PHOSITA would recognize that such an architecture is ideal for implementing sophisticated allocation strategies.
• Performance optimization via usage patterns: US7124272B1 provides a clear method for optimizing data placement based on "frequency of utilizations." A PHOSITA would naturally apply this principle to the device-level management provided by US6823398B1 to ensure frequently accessed data is placed on faster media, and less frequently accessed data on slower or more reliable media.
• Leveraging physical media characteristics: US200200955546A1 highlights the advantage of "zone formatted storage media" for improving data transfer rates by placing data in appropriate zones. This provides the physical mechanism for the "differential rate memory" of US7124272B1 and the "multiple media zones" of US9002795.

Specific to i-node Attributes and Read/Write Distinction:

The specific allocation of "less frequently accessed portion of the storage media comprises i-node write attributes" and "more frequently accessed portion of the storage media comprises i-node read attributes" is an obvious application of the principles taught by US7124272B1 in the context of file system metadata (i-nodes).

• It is common knowledge in the art of file systems that certain i-node attributes (like access timestamps) are updated very frequently during read operations, while others (like data block locations or creation time) are updated less frequently, primarily during write operations or initial creation. The US9002795 patent itself explicitly states: "File read operations are typically more prevalent than write operations. Therefore attributes that are updated when a read occurs are changing more often than those that change during a write operation.".
• Given US7124272B1's teaching of allocating data based on "frequency of utilizations," a PHOSITA would find it obvious to apply this principle to the different components of an i-node. The motivation, as articulated in US9002795, is to "reduce potential damage to the I-node" by separating frequently modified attributes and improving performance by placing more frequently accessed data in faster zones. This problem and its solution would be apparent to a PHOSITA optimizing storage.

Obviousness of Dependent Claims:

• Claim 2 (I-node extension for more frequently accessed portion): The concept of "I-node extensions" is discussed in US9002795 as a way to handle file growth without duplicating primary i-node information. Separating frequently accessed attributes into a smaller, distinct structure (like an "attribute node" mentioned in US9002795) that could be accommodated within an i-node extension or a similar physically separate area would be an obvious design choice for a PHOSITA attempting to implement the separation strategy.
• Claim 3 (Zone attributes in a map): Storing zone attributes in a map linked to physical addresses is a standard and well-known method for managing zoned storage media, as implied by US20020095546A1. The patent itself explicitly mentions this mapping.
• Claims 4 & 5 (Determining zone attributes by testing): Characterizing storage media performance and reliability through testing (either individual devices or samples) is a fundamental engineering practice. US9002795 explicitly describes this as the preferred method for sensing zone attributes. This is common general knowledge.
• Claim 6 (Data format conversion): Data format conversion is a known function in data channels and storage interfaces. US5963937A describes "Format conversion of storage data". Integrating this functionality into an object-based storage interface that handles data communication is an obvious design choice for a PHOSITA.
• Claim 7 (Stored information including throughput per zone and remapped sectors): Storing information like "throughput per zone" (implied by US20020095546A1's focus on data transfer rates) and "remapped sectors" (as taught by US5166936A) is essential for an embedded file system/OSD to make informed allocation decisions and manage media effectively. The patent itself confirms this capability of the disc drive.
• Claim 8 (Multiple copies of same object): Storing multiple copies for reliability is a well-established practice in data storage, often referred to as mirroring or replication, and is broadly taught in prior art and common knowledge (e.g., RAID systems). US9002795 explicitly mentions this as a method to improve reliability.

Conclusion

Based on the foregoing, independent claim 1, and by extension its dependent claims, would have been obvious to a PHOSITA at the time of the invention in light of the combination of US7124272B1, US6823398B1, and US20020095546A1. The motivation to combine these references arises from the desire to create more efficient and reliable data storage by embedding file system intelligence within the storage device, leveraging zoned media characteristics, and allocating data (including i-node attributes) based on their access frequency. The specific separation and allocation of i-node read and write attributes is an obvious application of these principles given the known differences in their access patterns and the desire to optimize performance and data integrity.