Co-Resident Container Data Caching for Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Container environments struggle to manage stateful containers that require continued data access, as existing solutions rely on external storage systems over networks, leading to performance issues and limitations in data access speed and resilience.
Innovation Solution
Implementing a distributed object storage solution with caching mechanisms, such as the Least Recently Used Algorithm, to migrate state-essential files to co-resident hosts, allowing containers to access external storage locally and independently of container orchestrators, using direct file access capabilities and storage plug-ins.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If external storage systems over networks are used to manage stateful container data, then data accessibility across hosts is improved, but network latency and data access speed deteriorate
Solution Approach 1:
The system performs preliminary actions by copying container data from external storage systems to local storage on container hosts before the containers need to access the data. This proactive data pre-positioning eliminates network latency during actual data access operations, as the data is already locally available on the host filesystem rather than requiring network I/O during container operations.
Solution Approach 2:
The patent introduces an intermediary layer (local storage on container hosts) between the external storage system and the containers. This intermediary allows containers to access data locally with high speed while the external storage system serves as the ultimate data source, mediating between the need for fast local access and the need for centralized data management.
2Reliability
If external storage systems are used for container data, then data resilience is improved, but system complexity and dependency on external resources worsen
Solution Approach 1:
The system applies local quality by making each container host self-sufficient with local copies of the container data it needs to run. Instead of all hosts depending on the external storage system, each host has locally optimized data availability, reducing the complexity of inter-host dependencies while maintaining data resilience through local redundancy.
3Quantity of substance
If data is stored on external hosts, then resource utilization is improved, but network latency and performance worsen
Solution Approach 1:
The system performs preliminary data copying from external storage to local host storage before containers need to access the data. This pre-positioning of data eliminates network latency during container operations while still allowing efficient resource utilization, as data is transferred only when needed for container deployment or migration, not during ongoing container operations.
Data Source
AI summary
A method for providing a coresident copy of external container data, that includes identifying, via a cluster storage manager, an access request to a cluster storage for external container data, where the access request is sent from a container executing on a first host, making a determination that the external container data is not stored on the first host, and based on the determination, copying the external container data from a second host to the first host to provide the coresident copy of the external container data.