Extendible Hash Tree Directory Entry Components
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large directories in computing systems consume significant volatile memory due to the growth of hash buckets, leading to impaired performance, as prior approaches store hints for all protocols regardless of naming conventions, resulting in excessive memory usage.
Innovation Solution
Storing directory entry components in the leaves of an extendible hash tree, where each leaf covers a range of hash values and includes indications of protocols, allowing a single directory entry component to represent a directory object across multiple protocols, thus avoiding the need for hints in volatile memory and optimizing memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If hash buckets are used to store hints for all protocols, then directory objects can be accessed across multiple protocols, but volatile memory consumption increases significantly
Solution Approach 1:
The patent merges multiple protocol-specific hints into a single unified directory entry component that serves all protocols. Instead of maintaining separate hints for NFS, CIFS, and DOS protocols, the invention consolidates them into one entry that contains the inode number and name, eliminating redundant storage and reducing volatile memory consumption while maintaining multi-protocol accessibility
Solution Approach 2:
The invention extracts the essential directory information (inode number and name) from protocol-specific hints and stores it in a unified directory entry component. By removing the protocol-specific overhead and keeping only the core identifying information, the system reduces memory usage while preserving the ability to access directory objects through multiple protocols
2Adaptability or versatility
If hash buckets grow to store thousands of hints, then all directory objects can be indexed, but scanning and updating consumes substantial resources
Solution Approach 1:
The patent segments the large hash bucket structure into smaller, more manageable directory entry components. Each component contains only essential information (inode number and name) rather than full protocol-specific hints. This segmentation reduces the size of each entry and the overall structure, making scanning and updating operations more efficient while maintaining comprehensive indexing capability
3Adaptability or versatility
If separate hints are provided for each protocol, then protocol-specific naming conventions are supported, but the aggregate size of hash buckets grows proportionally to the number of protocols
Solution Approach 1:
The invention creates a universal directory entry component that serves multiple protocols simultaneously. The single entry structure with inode number and name can be used by NFS, CIFS, DOS, and other protocols without requiring protocol-specific variations. This multi-functional approach maintains support for different naming conventions while eliminating the need for separate hints for each protocol, thus preventing hash bucket size from growing proportionally with the number of protocols
Data Source
AI summary
An improved technique for managing directory objects accessible via multiple protocols includes storing directory entry components in leaves of an extendible hash tree. Each directory entry component includes an indication of each protocol to which the name conforms, thereby allowing the same directory entry component to be used for representing directory object names in multiple protocols. Efficiency is improved, especially for large directory structures, by keeping related information about directory objects together, in the same directory entry components and/or in the same leaves of the extendible hash tree.