Delta Query Seeding for Live Migration Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current live migration processes of virtual machines (VMs) between datacenters often require redundant disk copying, consuming significant time and bandwidth, as they lack knowledge of existing replicated data at the destination storage disk, leading to inefficient data transfer.
Innovation Solution
The proposed solution detects and leverages existing replicated disk content at the destination host to reduce the time, bandwidth, and processing required for live migration by creating a bitmap of differences and transferring only out-of-date blocks, potentially skipping disk copy operations altogether.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If synchronous replication is not supported and full disk copying is performed during live migration, then data consistency is ensured, but migration time and bandwidth consumption increase significantly
Solution Approach 1:
The system performs preliminary replication of disk data from source to destination datacenter before the actual live migration occurs. This advance preparation ensures that when migration is needed, the data is already available at the destination, eliminating the need for time-consuming data transfer during the migration event itself.
Solution Approach 2:
The system creates a copy of the disk data at the destination datacenter through replication mechanisms. This copying is performed asynchronously in the background, allowing the original system to continue operating while the data is being prepared at the destination, thus not impacting migration timing.
2Reliability
If full disk copying is performed during live migration, then complete data transfer is ensured, but network bandwidth consumption increases significantly
Solution Approach 1:
Data replication is performed in advance through asynchronous replication mechanisms that operate independently of the migration process. This preliminary action transfers data during periods when bandwidth is more available and does not compete with migration traffic, thus reducing overall bandwidth consumption during actual migration events.
Solution Approach 2:
The system uses asynchronous replication to create copies of disk data at the destination datacenter before migration is needed. This copying mechanism operates in the background using available bandwidth resources without impacting the performance or bandwidth availability of the production system.
3Productivity
If replicated data at destination is not detected, then migration process is simple, but redundant data copying occurs wasting time and resources
Solution Approach 1:
The system implements a feedback mechanism where the destination datacenter reports back to the source about the presence and state of replicated data. This feedback allows the migration process to make informed decisions about whether full copying is necessary or if the existing replication can be leveraged, thus avoiding redundant operations.
Solution Approach 2:
The system introduces an intermediary component that manages the detection and coordination between source and destination datacenters. This intermediary handles the complexity of data detection and replication status checking, shielding the migration logic from these complexities while enabling efficient use of existing replicated data.
Data Source
AI summary
Examples perform live migration of objects such as VMs from a source host to a destination host. The disclosure exposes the contents of the storage disk at the destination host, compares the storage disk of the destination host to the source host, and during migration, migrates only data which is not already stored at the destination host. The source and destination VMs have concurrent access to storage disks during migration. After migration, the destination VM executes, with exclusive access to the storage disks.