Unified Storage Layer for Multi-Version Database Servers

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current database machines face inefficiencies in consolidating multiple database server versions due to the need for separate storage layer partitions, which limits the utilization of shared resources and restricts the ability to leverage smart functionality across different versions.

Innovation Solution

Implementing a storage layer with offload servers grouped into offload groups, allowing cell servers to service requests from various database server versions by delegating smart requests to appropriate offload servers based on version compatibility, and enabling user-defined offload groups for diagnostics and testing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the storage layer is vertically partitioned into separate partitions for each database server version, then each database server can communicate with its corresponding storage layer partition, but the utilization of shared resources becomes ineffective

Engineering Contradiction:
Improveversion compatibilityVSAvoidresource utilization
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The storage layer is designed with a unified architecture that can service multiple database server versions simultaneously. Offload servers are configured to handle smart requests from different database server versions through a single storage layer interface, eliminating the need for separate partitions for each version and enabling effective resource sharing while maintaining version-specific compatibility

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Productivity

If a single storage layer services multiple database server versions, then resource utilization improves, but version-specific communication protocols and data structures become incompatible

Engineering Contradiction:
Improveresource utilizationVSAvoidversion compatibility
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

Offload servers act as intermediaries between the storage layer and database servers of different versions. These offload servers translate and adapt communication protocols and data structures, allowing the unified storage layer to service multiple database server versions without direct version-specific configurations

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If separate storage layer partitions are created for each database server version, then version-specific functionality is maintained, but the ability to consolidate multiple database servers into a single database machine is limited

Engineering Contradiction:
Improveversion-specific functionalityVSAvoidconsolidation capability
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent merges multiple version-specific storage layer functions into a single unified storage layer. By using offload servers to handle version-specific smart requests, the system consolidates multiple database server versions onto one database machine while maintaining the necessary version-specific functionality through the offload server abstraction layer

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS9898490B2Systems and methods for supporting multiple database server versions on a database machine
Publication Date: 2018.02.20 ORACLE INT CORP
  • US9898490B2 patent drawing
  • US9898490B2 patent drawing
  • US9898490B2 patent drawing

AI summary

Techniques are described herein for supporting multiple versions of a database server within a database machine comprising a separate database layer and storage layer. In an embodiment, the database layer includes compute nodes each hosting one or more instances of a database server. The storage layer includes storage nodes each hosting one or more instances of a storage server, also referred to herein as a “cell server.” In general, the database servers may receive data requests, such as SQL queries, from client applications and service the requests in coordination with the cell servers of the storage layer.