Storage Bypass Driver for High Availability Checkpoint Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In highly available computer systems, identifying and transferring only the necessary state from a primary node to a secondary node during checkpoint operations is time-consuming, prolonging prototype design and product introduction.
Innovation Solution
Implementing special storage drivers on both nodes that redirect I/O read misses from the secondary node to the primary node's mass storage, allowing for a single mass storage image and delaying the need to determine transferred state until prototype operationality, thereby simplifying and accelerating the development process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If experimental determination of necessary state transfer is performed during prototype design, then reliability of state transfer is improved, but development time increases
Solution Approach 1:
A storage bypass driver is introduced as an intermediary component between the storage subsystem and the checkpoint system. This driver intercepts and redirects storage I/O requests, allowing the checkpoint system to transfer state without requiring experimental determination of necessary data, thereby reducing development time while maintaining reliability
Solution Approach 2:
The storage management functionality is segmented into a dedicated storage bypass driver that handles I/O redirection separately from the main checkpoint process. This segmentation allows the checkpoint operation to proceed without waiting for experimental state transfer determination, reducing development time while ensuring reliable state transfer through the specialized driver
2Reliability
If all state is transferred from primary to secondary node, then reliability of failover is improved, but checkpoint operation time increases
Solution Approach 1:
The storage bypass driver extracts and redirects only the necessary storage I/O requests related to state transfer, separating them from other storage operations. This extraction allows the checkpoint system to transfer only essential state data without transferring all node state, reducing checkpoint time while maintaining failover reliability
Solution Approach 2:
The storage bypass driver performs preliminary redirection of storage I/O requests before the checkpoint operation begins. By pre-configuring the redirection paths and identifying necessary state data in advance, the system can execute the actual state transfer more quickly while ensuring reliability, without needing to determine transferred state during the checkpoint operation itself
Data Source
AI summary
A highly available computer system has a primary compute node and a secondary compute node, both of which operate to support at least one similar process using the same instructions and the same information at substantially the same time, and in the event that the secondary compute node generates a read request that is not able to be serviced by cache memory local to it, then the secondary compute node redirects the read request to the primary compute node for servicing.