SAN Virtualizer Write Coordination for Mirrored Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Storage Area Networks (SANs), there is a challenge in determining the appropriate point for generating a transfer ready command when handling write operations to mirrored storage volumes, as existing methods are unclear on how to coordinate data transfer across multiple mirrored storages efficiently.
Innovation Solution
A virtualizer in the SAN receives write commands and sends them to mirrored storages, buffering transfer ready messages until all storages are ready, ensuring data is sent only after confirmation from all mirrored storages, thereby managing data transfer efficiently across multiple mirrored volumes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the virtualizer waits for transfer ready messages from all mirrored storages before sending data, then data reliability is improved, but write operation speed deteriorates
Solution Approach 1:
The virtualizer sends write commands to all mirrored storages in advance and buffers incoming data until transfer ready messages are received. This preliminary coordination of storage readiness ensures data reliability while minimizing waiting time for actual data transfer.
Solution Approach 2:
The virtualizer acts as an intermediary between the initiator and mirrored storages, managing the coordination of transfer ready messages and data transfer. It buffers data and controls the timing of transfers to multiple storages, resolving the contradiction between reliability and speed.
2Productivity
If the virtualizer sends data to all mirrored storages simultaneously, then write operation speed is improved, but data consistency deteriorates
Solution Approach 1:
The virtualizer uses feedback from transfer ready messages received from each mirrored storage to determine when to send data. This feedback mechanism ensures that data is sent only when all storages are ready, maintaining data consistency across the mirrored volume.
3Reliability
If the virtualizer buffers data until all mirrored storages are ready, then data loss prevention is improved, but write operation latency increases
Solution Approach 1:
The virtualizer performs preliminary actions by sending write commands to all mirrored storages and receiving transfer ready messages before buffering and sending actual data. This advance coordination reduces the overall latency by preparing the storage system in advance.
Solution Approach 2:
The virtualizer maintains continuous useful action by buffering data and immediately transferring it to all mirrored storages once readiness is confirmed, minimizing the time data remains buffered and reducing overall write operation latency.
Data Source
AI summary
In one embodiment, a method is provided comprising: receiving, at a virtualizer, a write command from an initiator in a storage area network, wherein the storage area network includes the initiator and a plurality of mirrored storages; sending, from the virtualizer, a write command to the plurality of mirrored storages; receiving, at the virtualizer, a transfer ready message from a first of the plurality of mirrored storages; sending a transfer ready message from the virtualizer to the initiator in response to the receiving of the transfer ready message from the first of the plurality of mirrored storages; receiving, at the virtualizer, a data message from the initiator; and sending, from the virtualizer, a data message to the plurality of mirrored storage once transfer ready messages have been received from each of the plurality of mirrored storages.