Dummy Journal Sequence Control for Storage Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems, transmission delays can lead to incomplete journal creation, resulting in the accumulation of unnecessary journals and the inability to discard older update numbers, which disrupts the replication process and affects system reliability.
Innovation Solution
The system introduces a dummy journal mechanism where a primary storage subsystem creates a dummy journal with a sequence number after stopping journal creation, allowing the local and remote storage subsystems to maintain data consistency and continuity even if partial journals are not created.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the system waits for complete journal creation before processing, then data consistency is maintained, but transmission delays cause operation interruptions and reduced productivity
Solution Approach 1:
The system performs preliminary actions by creating dummy journals with sequence numbers before actual data transmission is complete. This allows the remote storage subsystem to prepare for sequential processing in advance, maintaining operation continuity while ensuring data consistency when the actual journals arrive.
Solution Approach 2:
Dummy journals act as intermediaries between the primary and remote storage subsystems during transmission delays. These dummy journals with sequence numbers serve as placeholders that enable continuous operation while the actual data journals are being transmitted, resolving the contradiction between waiting for complete data and maintaining operational flow.
2Reliability
If journals are accumulated to handle transmission delays, then data completeness is ensured, but journal volume increases and older journals cannot be discarded
Solution Approach 1:
The journal management is segmented into dummy journals (containing only sequence numbers) and actual data journals. This segmentation allows the system to separate the sequencing function from the data storage function, enabling efficient management where dummy journals can be discarded after actual journals are processed, thus reducing overall journal accumulation while maintaining data completeness.
Solution Approach 2:
The system creates simplified copies (dummy journals) that contain only essential sequencing information without the full data payload. These copy journals enable the system to maintain sequence control and identify missing journals without storing complete data copies, significantly reducing the quantity of accumulated journals while ensuring data completeness when actual journals are received.
3Productivity
If the system discards older journals to manage volume, then storage efficiency improves, but sequence control is disrupted when journals are missing
Solution Approach 1:
The system implements feedback mechanisms where the remote storage subsystem continuously monitors for missing journals by checking sequence numbers in incoming journals against expected sequences. When gaps are detected, the system requests retransmission of missing journals, ensuring sequence control is maintained while allowing efficient discarding of processed journals.
Solution Approach 2:
Dummy journals with sequence numbers serve as intermediaries that enable the system to verify sequence continuity without storing complete journal data. The sequence numbers act as a lightweight reference layer that allows the system to detect missing journals and maintain proper sequencing while efficiently managing storage by discarding processed journals.
Data Source
AI summary
A primary storage system is equipped with a local storage subsystem having a mirror volume of a volume in a primary storage subsystem, and a remote storage subsystem having a mirror volume of the volume in the primary storage subsystem at a secondary site. When storing a write data from a host to a volume, the primary storage subsystem determines an SEQ# to be stored in a journal corresponding to the write request, creates a journal including the relevant SEQ# and a replica of the write data, and transmits the relevant SEQ# to the local storage subsystem, which creates a journal including the relevant SEQ# and stores it in its own journal volume. But when creation of the journal including the relevant SEQ# is stopped, the local storage subsystem creates a dummy journal including the relevant SEQ# but not including the write data.