Storage Entity Replication via Chunk Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data replication methods, such as one-to-many and cascaded replication, face challenges in balancing workload across multiple destinations, leading to increased latency and resource consumption, particularly when replication speeds vary among destinations.
Innovation Solution
A method that splits a storage entity into portions and generates information for distributing these portions across multiple storage systems, allowing each system to receive the portions through different paths, with the source system verifying receipt and providing reconstruction information for complete data assembly, while also handling failures by re-routing or re-sending missing portions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If one-to-many replication is used where the source system replicates content in parallel to multiple remote sites, then the content arrives at destinations sooner, but the source system consumes more network bandwidth and computing resources
Solution Approach 1:
The patent segments the replication process by dividing content into multiple chunks and distributing replication responsibilities across multiple destination systems. Instead of the source system replicating entire content to all destinations, the source system sends content chunks to multiple destinations in parallel, and each destination system independently stores and manages its received chunks. This segmentation reduces the burden on the source system's network bandwidth and computing resources while maintaining parallel replication speed.
2Use of energy by moving object
If cascaded replication is used where content is copied sequentially through multiple systems, then network bandwidth and computing resources are conserved, but substantial latency is introduced
Solution Approach 1:
The patent segments the content into multiple chunks that can be replicated in parallel across multiple destination systems. Each chunk is independently transmitted to different destinations simultaneously, eliminating the sequential dependency inherent in cascaded replication. This segmentation enables parallel execution of replication operations, significantly reducing total replication latency while maintaining efficient resource utilization.
Solution Approach 2:
The patent implements preliminary actions by having the source system send content chunks to multiple destinations in parallel before any single destination needs to forward content to the next hop. This preliminary parallel distribution eliminates the sequential waiting time where each destination must receive and process content before the next destination can begin replication, thereby reducing overall latency.
3Adaptability or versatility
If replication to each destination is handled autonomously to adjust to the pace of each destination, then individual destination performance is optimized, but parallelism is lost and the process becomes more complicated
Solution Approach 1:
The patent segments the replication process into independent chunk-level operations that can be autonomously managed at each destination while maintaining overall coordination. Each destination system independently manages receipt, storage, and forwarding of specific content chunks, allowing adaptation to individual destination paces. The segmentation enables simple autonomous operation at each node without requiring complex inter-destination coordination, thus maintaining parallelism while adapting to varying destination capabilities.
Data Source
AI summary
A storage system, a computer program product and method for replicating a storage entity to a group of storage systems. The method may include splitting the storage entity, by a given storage system, to provide storage entity portions; generating, by the given storage system, storage entity portions information for distributing the storage entity portions between all storage systems of the group so that each storage system of the group eventually receives the storage entity portions; transmitting the storage entity portions, from the given storage system, to storage systems of the group; wherein the transmitting comprises transmitting different storage entity portions from the given storage system to different storage systems of the group; sending, by the given storage system, the storage entity portions information to at least some storage systems of the group; and attempting to verify, by the given storage system, that all the storage systems of the group received all the storage entity portions.