Data Replication Snapshot via Remote Event Packet Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data replication techniques with snapshot capabilities are inefficient, as they require local volumes to manage and transfer snapshots, leading to excessive processing, storage, and network bandwidth usage, which delays application processing and ties up resources.
Innovation Solution
A method where a local replication service initiates a snapshot operation by flushing pending I/O operations and creating an event packet representing the local volume's state, which is then processed by a remote replication service to generate a snapshot on the replicated volume, minimizing local processing and network bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a snapshot is created on the local data source and transferred to the remote replicated source, then the snapshot captures the local volume state accurately, but it unnecessarily delays continued processing of applications and ties up network bandwidth
Solution Approach 1:
The patent extracts the snapshot creation operation from the local data source and relocates it to the remote replicated source. The local volume continues processing applications without interruption while the remote source independently creates the snapshot, eliminating the delay caused by local snapshot operations.
Solution Approach 2:
The patent introduces a replication service as an intermediary that manages snapshot operations. The replication service coordinates between the local and remote volumes, enabling the remote source to create snapshots independently without requiring local volume to pause or transfer snapshot data over the network.
2Reliability
If the local data source creates and transfers snapshots to the remote source, then the remote source maintains replicated data, but it utilizes storage on both local and remote sources and unnecessarily consumes network bandwidth
Solution Approach 1:
The patent extracts the snapshot creation function from the local data source and assigns it to the remote replicated source. This eliminates the need to transfer snapshot data over the network, as the remote source independently creates its own snapshot by replaying I/O operations, thereby conserving network bandwidth.
Solution Approach 2:
Instead of copying the snapshot from local to remote, the patent enables the remote source to recreate the snapshot independently by replaying the same I/O operations. This eliminates redundant data transfer while ensuring the remote snapshot matches the local volume state.
3Measurement precision
If applications flush pending volatile I/O operations to create a snapshot, then the snapshot reflects the current data state, but it delays continued processing of applications
Solution Approach 1:
The patent records pending volatile I/O operations in a log before the snapshot is created. This preliminary action captures the state of in-flight operations, allowing the remote source to replay these operations after the snapshot is created, ensuring consistency without requiring applications to wait for the snapshot operation to complete.
Solution Approach 2:
The patent enables applications to continue processing I/O operations continuously without interruption during snapshot creation. The system maintains a log of pending operations and ensures they are replayed on the remote source, allowing uninterrupted application throughput while still achieving consistent snapshots.
Data Source
AI summary
Techniques for data replication are presented. A snap shot request is processed within an environment associated with a local volume by flushing pending Input/Output (I/O) operations into an event packet. The event packet is sent to a replicated volume and processed against the replicated volume to create a snap shot version of the local volume.