Stub-State File Sharing System Reducing Network Delays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current file sharing systems with version management face network delays due to the lack of integration between on-demand copy technology and distributed caching, leading to inefficient cache queries between proxy servers, which results in delayed responses to clients.
Innovation Solution
A file sharing system that includes a master server and slave servers, where the first slave server manages copy management information to identify and query the appropriate slave server for cached data, allowing for efficient materialization of stub-state objects and reducing network delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If distributed caching technology with multiple proxy servers is implemented, then cache hit rate is improved, but network traffic increases due to cache presence/absence queries between all proxy servers
Solution Approach 1:
The system segments the file system into directories and files, and further segments caching responsibilities by introducing stub objects that represent files. Instead of all proxy servers querying about all files, each proxy server maintains stubs locally and only queries about specific files when needed, dividing the query burden across multiple servers.
Solution Approach 2:
The system performs preliminary actions by creating stub objects that contain metadata about files before actual file access occurs. When a file is accessed, the stub is materialized by querying the master server or other slave servers. This preliminary stub creation allows the system to avoid full file copying and reduces the scope of cache queries to only necessary files.
2Loss of time
If on-demand copy technology is used to reduce network communications, then network delay is reduced, but cache data sharing between servers becomes inefficient requiring more queries
Solution Approach 1:
The stub object acts as an intermediary between the client and the actual file data. Instead of direct file access or complex multi-server coordination, the stub serves as a lightweight placeholder that contains sufficient information to materialize the file when needed. This intermediary simplifies the interaction model while maintaining on-demand copy benefits.
Solution Approach 2:
The system creates simplified copies of file metadata in the form of stub objects that can be distributed across slave servers. These stubs contain essential file information without the full file content, allowing efficient distribution and querying. When materialization is needed, the actual file data is copied from the master server or another slave server that has it cached.
Data Source
AI summary
In a computer system according to an embodiment of the present invention, a first slave server includes copy management information associating a master file system identifier with an identifier of a slave server including a copy of the master file system identified by the master file system identifier. The first slave server acquires a copy of a part of a master file system. The copy includes a stub-state object (file or directory). The first slave server searches the copy management information for a slave server to be queried regarding information for materializing the stub-state object. If the slave server to be queried includes the information for materialization, the first slave server materializes the stub-state object using the information for materialization acquired from the queried slave server.