Metro-Level Distributed Caching for Lower Latency Data Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed computing systems face high computational, network, and latency costs when accessing backend databases due to inefficient local caching, particularly in serving large or global collections of audio and video files, where geographically nearby users often access similar data.
Innovation Solution
Implement metro-level distributed caching by having a designated computing system within a metropolitan area manage a proxy data server to share cached data among nearby data servers, reducing the need to access backend databases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If local caching is implemented at each data server, then access speed to frequently used data is improved, but network traffic to backend databases increases and cache utilization becomes inefficient
Solution Approach 1:
The patent merges cache resources across multiple geographically distributed data servers into a unified metro-level distributed cache. When a data server receives a data access request, it first checks its local cache, then queries other data servers in the metropolitan area through a coordinated protocol before finally accessing the backend database. This combining of cache resources reduces redundant data transfers to the backend database while maintaining fast local access.
Solution Approach 2:
The patent introduces an intermediary coordination mechanism where data servers act as intermediaries for each other's cache queries. When one data server needs data not in its local cache, it queries other data servers in the metropolitan area through a standardized protocol. This intermediary approach allows efficient utilization of distributed cache resources without requiring direct backend database access for every request.
2Loss of time
If data is cached locally at each computing system, then latency is reduced, but the system complexity and coordination overhead increase
Solution Approach 1:
The patent implements local quality by allowing each data server to maintain its own local cache with independent management policies while participating in the broader metro-level distributed cache. Each server can optimize its local caching strategy based on its specific access patterns and performance requirements, while still benefiting from the coordinated query mechanism that leverages other servers' caches. This localized approach reduces complexity compared to a fully centralized cache management system.
3Reliability
If backend database is accessed directly for every data request, then data consistency is maintained, but network costs and processing time increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-caching data at multiple data servers across the metropolitan area before actual user requests arrive. When data is accessed by one server, it is proactively cached at other nearby servers through the coordinated query mechanism. This preliminary distribution of data across the distributed cache system means that subsequent requests can be served from local or nearby caches rather than requiring direct backend database access, thereby maintaining consistency while improving throughput.
Data Source
AI summary
A distributed computing system includes a non-designated and a designated computing system. The non-designated system is configured to: receive a first data request from a user device requesting first data; determine that the first data is not stored in a first cache at the non-designated system; based on determining the first data is not stored in the first cache, send a second data request to the designated system including a request for the first data; receive the first data from the designated system; send the first data to the user device; and store the first data in the first cache. The designated system is configured to: determine that the first data is stored in a second cache at the designated system and, based on the first data being stored in the second cache, send the first data to the non-designated system.