Graceful IPv6 Address Migration for VM Mobility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Virtual machine mobility in networks with different switch devices leads to sub-optimal forwarding, as the switch device prefix for the moved virtual machine still points to the original egress switch device, preventing direct packet forwarding to the new egress switch device.
Innovation Solution
Implementing a method for graceful address migration by provisioning a device prefix on the egress device and installing it in the hardware of ingress devices, using IPv6 address formats to support subnet prefix aggregation, and employing FIB CAM tables for efficient packet forwarding, allowing packets to be switched directly between ingress and egress switch devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtual machine mobility is enabled across different switch devices, then network flexibility is improved, but packet forwarding performance deteriorates due to sub-optimal routing paths
Solution Approach 1:
The system performs preliminary actions by detecting VM movement before it causes forwarding problems, and pre-establishes the optimal forwarding path by updating the ingress switch's FIB CAM table to point to the new egress switch. This proactive approach ensures that when packets need to be forwarded, they immediately take the optimal path rather than following stale routing information.
Solution Approach 2:
The system implements feedback mechanisms where the new egress switch sends notifications to the ingress switch about the VM's new location. This feedback loop allows the network to dynamically adapt routing information based on actual VM positions, ensuring that forwarding paths are continuously optimized as VMs move between switches.
2Device complexity
If FIB CAM table size is limited, then hardware resource constraints are reduced, but the ability to support large numbers of virtual machines deteriorates
Solution Approach 1:
The system segments the forwarding information by separating VM-specific routing entries from switch-specific routing entries. By organizing FIB CAM table entries to store device prefixes at the ingress switch and maintaining fewer, more general entries, the system reduces the burden on individual switch hardware while still supporting a large number of VMs across the network.
Solution Approach 2:
The system merges routing information at the network level rather than requiring each switch to maintain complete individual routing tables. By combining forwarding logic across multiple switches and using centralized coordination for FIB CAM table updates, the system achieves support for over a million VMs without requiring large FIB CAM tables on individual hardware devices.
Data Source
AI summary
Systems and methods are provided to detect when a virtual machine moves across a switch device boundary, keep existing application sessions alive, and provide optimal forwarding for new application sessions with graceful address migration. Virtual machine movement can be detected on the new switching device by comparing the Device ID in the Source Address of the ND-NA with the Device ID of the new switching device. A temporary hardware forwarding sub-optimal path is maintained for existing application sessions. A new IP address is assigned to the moved virtual machine and an optimal forwarding path is programmed in hardware for new application sessions. The temporary sub-optimal hardware forwarding path is removed after all old application are terminated.