Obviousness

Obviousness Analysis of US Patent 7587430 Under 35 U.S.C. § 103

This analysis identifies combinations of prior art references that would render the independent claims of US Patent 7587430 (hereinafter '430 patent) obvious to a person having ordinary skill in the art (POSA). The '430 patent addresses the challenge of efficient database synchronization between a primary site and a secondary site over a narrow-band network, particularly after events like log-less batch processing or line failures, where traditional transaction log transfer is insufficient and full database transfer is too slow and costly.

The core inventive concept of the '430 patent is the introduction of an "unnecessary information removing unit" that overwrites unused areas within updated database data with a preset character or value (e.g., '0'). This pre-processed data is then transferred via a network device equipped with a data compression unit that specifically excels at compressing successions of identical values. This combination aims to significantly reduce the amount of data transferred and thus the synchronization time over narrow-band networks, leading to lower operating costs.

Prior Art References and General Knowledge

The '430 patent itself provides several key pieces of prior art and background knowledge:

1. "Evaluation of remote backup algorithms for transaction-processing systems" by Polyzois and Garcia-Molina (1994): This reference describes "transaction log transfer" as a known method for database backup using an active (primary) and standby (secondary) system. It establishes the context of disaster recovery (DR) systems and the use of narrow-band networks for cost-effective log transfer during normal operations.
2. JP 2004-78746 A (JP '746): This patent discloses a known function in storage systems for "copying data between two storage systems (remote copy) by transferring the data via a network." This represents a general mechanism for full database transfer.
3. Known Problems in DR Systems: The '430 patent explicitly highlights the problem of synchronizing databases after "log-less batch processing" or "line failures," where full database transfer is necessitated but is impractically slow over narrow-band networks, while wide-band networks are too costly.
4. Network Device with Data Compression: The '430 patent describes a "network device" (e.g., device 140) that includes a "data compression unit" (e.g., unit 141) capable of "compressing a succession of the same value." It specifically mentions "Stacker compression" as a known method that "improves the compression ratio when the same character string (data) appears successively."
5. Database Page Management: The patent details how database pages contain "row data" and "control information," and how updates (e.g., row deletion or increased row length) can create "unused areas" within a page, leaving old data intact to maintain online performance. These unused areas may contain "non-recurring values or a string of non-recurring characters" (i.e., "unnecessary information").

Level of Ordinary Skill in the Art (POSA)

A POSA in the context of the '430 patent would be a software engineer or system architect experienced in database management systems, storage technologies, network protocols, and disaster recovery solutions. Such a person would be familiar with various data backup strategies, network performance considerations, and data compression techniques.

Obviousness Combinations and Rationale

The independent claims (1, 10, 11, and 12) of the '430 patent would have been obvious to a POSA by combining the explicit teachings and recognized problems within the patent's own background with general knowledge of data compression and database management.

Combination for Claims 1 and 10 (System and Method by DBMS)

References in Combination: Polyzois, JP '746, General knowledge of database page management, Known network compression capabilities, and the explicitly stated problem in the '430 patent.

Rationale:
A POSA, aiming to solve the acknowledged problem of slow and costly database synchronization over narrow-band networks in DR systems, would consider the existing components and their limitations.

1. DR Context and Remote Copy: The use of a log-based DR system (as implied by Polyzois) would establish the desirability of low-cost, narrow-band networks. However, the limitation that such systems require full database synchronization via remote copy (as taught by JP '746) after certain events (like log-less batch processing or failures) presents a bottleneck due to large data volumes.
2. Compression Opportunity: The POSA would be aware that network devices (like 140) often incorporate data compression units (like 141) that are particularly efficient at compressing sequences of identical characters or values (e.g., Stacker compression).
3. Exploiting Unused Areas: A POSA familiar with database page structures and update mechanisms would know that operations like row deletion or updates that increase data length can leave "unused areas" within a database page. These areas often contain arbitrary, non-uniform data, making them difficult for standard compression algorithms to reduce significantly.

Given this, the motivation to implement an "unnecessary information removing unit" (as claimed in Claim 1) or a method step of "overwriting the unused area successively with one of a predetermined character and a predetermined value" (as claimed in Claim 10) would be obvious. By deliberately clearing these unused areas with a uniform character (e.g., '0'), the POSA would predictably transform poorly compressible data into highly compressible data. This pre-processing step directly leverages the known capability of the network device's compression unit (141) to achieve a higher compression ratio, thereby reducing the "substantial transfer amount" of the database and completing synchronization in a "shorter period of time" over the narrow-band network. The placement of this unit within the DBMS (as described in the first embodiment) would be a natural design choice for application-level data manipulation.

Combination for Claims 11 and 12 (Storage System and Method by Storage System)

References in Combination: Polyzois, JP '746, General knowledge of database page management, Known network compression capabilities, and the explicitly stated problem in the '430 patent.

Rationale:
The underlying problem and the core solution remain the same as for Claims 1 and 10. However, Claims 11 and 12 specifically locate the "unnecessary information removing unit" (1400) and its method steps within the storage system's control unit (111).

A POSA would be motivated to place the "unnecessary information removing unit" in the primary storage system (as in the fourth embodiment) for several reasons:

1. Offloading Processing: Placing this function in the storage system can "prevent unnecessary information removing processing from lowering the processing performance of the primary server 100 in which the primary DBMS 101 is run." This is a common and obvious architectural optimization to distribute workload and improve overall system performance.
2. Proximity to Data Transfer: The storage system is directly responsible for managing disk I/O and initiating remote copy operations (via remote copy processing unit 113). Integrating the pre-processing step at this level means the data is optimized closer to the point of transfer across the network device (140) and its compression unit (141).

The functional outcome—generating highly compressible unused areas to speed up remote copy over a narrow-band network—is identical to the DBMS-side implementation. The choice of where to implement this data preparation (DBMS vs. storage system) would be an obvious design decision for a POSA considering factors like system architecture, performance impact, and resource allocation. Therefore, the storage system of Claim 11 and its method of Claim 12, incorporating the "unnecessary information removing unit" to cooperate with the network device's compression unit, would be an obvious variation of the concept to achieve the desired improvements in DR synchronization.