Preferred Path Selection for Distributed Data Copying
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data storage systems face challenges in efficiently performing distributed data copying operations, particularly in selecting optimal paths for data transmission between source and target devices, which can lead to performance bottlenecks and failure in ensuring data integrity in case of processor failures.
Innovation Solution
A method for determining a preferred source-target port pairing using preferred path mapping or designated groups of ports, allowing for random selection of ports if the preferred path is unavailable, and suspending discovery processing to manage requests and resume later with saved context information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed path mapping is used for data transmission, then path selection is simple and deterministic, but the system cannot adapt to path failures or performance degradation
Solution Approach 1:
The patent implements dynamic path selection by allowing the system to switch between fixed preferred paths and randomly selected paths based on operational conditions. The path mapping is no longer static but can be adjusted during runtime, enabling the system to adapt to changing network conditions while maintaining manageable complexity through structured selection algorithms.
Solution Approach 2:
The system changes the parameter of path selection from a fixed deterministic mapping to a flexible mechanism that can select from multiple paths. By introducing randomness as a selectable parameter and allowing switching between different selection modes (preferred path vs. random path), the system achieves adaptability without excessive complexity.
2Productivity
If multiple processors are used for distributed data copying, then copying efficiency increases, but the risk of performance bottlenecks on selected paths increases
Solution Approach 1:
The patent segments the data copying task across multiple processors while also segmenting the path selection into preferred paths and alternative random paths. This segmentation allows different processor pairs to use different paths, distributing the load and preventing any single path from becoming a performance bottleneck, thereby maintaining both high productivity and reliability.
Solution Approach 2:
The system dynamically selects paths for each processor pair involved in distributed copying. By allowing path selection to vary based on processor availability and path conditions, the system can optimize copying efficiency while avoiding performance bottlenecks through dynamic adaptation to current system state.
3Reliability
If discovery processing is performed continuously, then path availability is monitored, but system performance degrades due to processing overhead
Solution Approach 1:
The patent implements periodic discovery processing rather than continuous monitoring. The system performs discovery operations at intervals or on-demand based on specific triggers, reducing the processing overhead and performance impact while still maintaining adequate path availability monitoring. This periodic approach balances reliability with system performance by avoiding constant discovery operations.
Data Source
AI summary
Described is a distributed copying technique that may be used in migrating large amounts of data from one or more source devices to one or more destination devices. The data source is divided into partitions. As Fibre Channel adapters (FAs) become available, each of the FAs may copy a partition of the data. In connection with specifying paths used for the distributed copying technique, a preferred path selection (source port-target port mapping) may be made by executing code in a controlling data storage system to perform discovery processing. The preferred path selection is used for the duration of the distributed copying unless the preferred path is unable to transmit data. A target port of the preferred path may be randomly selected from all accessible target ports, and/or in accordance with a specified portion of the target ports. Preferred paths may also be specified using an API (application programming interface).