Mid-sized LOB Containers for Disk Storage Optimization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In-memory database systems face inefficiencies when storing large objects (LOBs) due to memory waste, as LOBs that are rarely accessed consume significant memory resources and storing them in main memory slows down performance, while disk storage results in wasted disk space when objects are smaller than the page size.

Innovation Solution

A method to determine the optimal storage type for LOBs by assessing their size and access frequency, allowing LOBs to be stored in disk storage when they are too large for main memory but too small for efficient disk page allocation, using a mid-size LOB container that minimizes disk waste by storing multiple LOBs on a single disk page with unique identification and offset mapping.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If LOBs are stored in main memory, then access speed is improved, but memory consumption increases and performance slows down

Engineering Contradiction:

Improveaccess speedVSAvoidmemory consumption

Core Design Contradiction:

SpeedVSQuantity of substance

Solution Approach 1:

The patent segments LOB storage into three distinct categories based on size: small LOBs stored directly in main memory for fast access, mid-sized LOBs stored in compressed containers on disk to save space, and large LOBs stored as individual files on disk. This segmentation resolves the contradiction by allocating storage resources appropriately to different LOB sizes, improving overall system performance while managing memory consumption efficiently.

Inventive Principle:

Principle #1Segmentation

2Quantity of substance

If LOBs are stored on disk individually, then memory usage is reduced, but disk space waste increases when objects are smaller than page size

Engineering Contradiction:

Improvememory usageVSAvoiddisk space waste

Core Design Contradiction:

Quantity of substanceVSLoss of substance

Solution Approach 1:

The patent merges multiple small to mid-sized LOBs into compressed containers on disk, where multiple LOBs share a single disk page allocation. This merging approach eliminates the disk space waste that would occur if each small LOB were allocated its own page, while still keeping memory usage low by storing these LOBs on disk rather than in main memory.

Inventive Principle:

Principle #5Merging (Combining)

3Loss of substance

If mid-sized LOBs are stored in compressed containers on disk, then disk space utilization is improved, but access complexity increases

Engineering Contradiction:

Improvedisk space wasteVSAvoidstorage management complexity

Core Design Contradiction:

Loss of substanceVSDevice complexity

Solution Approach 1:

The patent implements dynamic storage management where the system automatically determines the appropriate storage location (main memory, compressed container, or individual file) based on LOB size and access patterns. This dynamic approach handles the complexity internally, providing simple access interfaces to users while optimizing storage efficiency through automated container management and compression.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS11249968B2Large object containers with size criteria for storing mid-sized large objects

Publication Date: 2022.02.15 SAP SE

AI summary

A method, computer program product and system are provided. The method, computer program product and system execute a process for storing an object in an object container that is stored in a persistency of a disk storage. The object container has size criteria whereby objects meeting the size criteria of the object container can be assigned to the object container. The object container can facilitate storing multiple objects to optimize disk storage usage by facilitating the assigning of multiple objects to the same disk storage page.