CAS Replication Using Active Buffer Hash Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content addressable storage systems face challenges in reducing data transfer during replication due to bandwidth limitations and latency, particularly in scenarios where source and destination storage spaces are not physically close, leading to inefficiencies in data transfer and potential for unnecessary data duplication.
Innovation Solution
The system employs internal CAS metadata to optimize replication by using hash results and addresses to identify data items, implementing an active buffer that transfers only the last item with a shared address and utilizing modified acknowledge signals to manage data transfer, thereby reducing redundant data transfers and leveraging synchronous cycles for efficient communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is replicated from source storage space to destination storage space, then data redundancy and backup are improved, but bandwidth consumption and data transfer time increase
Solution Approach 1:
The patent extracts only the hash results from the active buffer and transfers them to the destination storage space, rather than transferring the entire data items. This allows the destination to identify duplicates using hash comparison, significantly reducing bandwidth consumption while maintaining data backup reliability
Solution Approach 2:
The patent creates a simplified copy of the data identification information (hash results) rather than copying the full data items. The destination storage space uses these hash copies to determine which data items are duplicates and which need to be transferred, reducing the actual data transfer while ensuring backup integrity
2Stability of the object's composition
If all data items in the active buffer are transferred to destination storage space, then data consistency is improved, but unnecessary data transfer increases
Solution Approach 1:
The destination storage space sends an acknowledge signal back to the source storage space indicating whether data transfer is needed. This feedback mechanism allows the source to determine which data items are duplicates (based on hash comparison at destination) and avoid transferring them, maintaining data consistency while improving transfer efficiency
Solution Approach 2:
The patent performs preliminary hash comparison at the destination storage space before actual data transfer occurs. By checking if hash results already exist at the destination, the system can identify duplicates in advance and avoid unnecessary data transfer, ensuring only new data is transferred while maintaining consistency
3Measurement precision
If hash results of all data items are transferred, then duplicate detection capability is improved, but data transfer volume increases
Solution Approach 1:
The patent extracts only the essential identification component (hash result) from each data item and transfers only that to the destination. This extracted hash information is sufficient for duplicate detection purposes, providing accurate duplicate identification while keeping the transferred data volume minimal compared to transferring full data items
Data Source
AI summary
A CAS data storage system with one or more source CAS data storage spaces and one or more destination CAS data storage spaces, and a communication line therebetween, receives input data at the source storage space for local storage and for replication to the destination CAS storage space. CAS metadata is used in the replication procedure between the two separate CAS storage spaces. Thus, data at the source storage space is used to form an active buffer for transfer to the destination storage space, the active buffer holding a hash result of the respective data item and a storage address. The system detects whenever there is more than one data item in said active buffer sharing a same storage address and upon such detection transfers a respective hash result of only the last of the data items.