Hybrid Distributed Logical Router for Cloud VM Mobility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flexible Software Defined Networking (SDN) in public cloud infrastructure faces limitations due to restricted Layer 2 (L2) capabilities, which hinder the movement of Virtual Machines (VMs) across hosts and proper routing of traffic, as the cloud provider does not self-learn L2 addresses and lacks information for routing traffic to moved VMs.
Innovation Solution
A Layer 3 (L3) underlay routing solution that manages route tables in a data center by generating and transmitting route tables from a control plane to all hosts, updating entries for underlay routers to route traffic to the host containing the VM, ensuring next-hop routing for logical and underlay networks, even when VMs are moved across hosts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If SDN uses flexible L2 capabilities for VM movement, then VM mobility is improved, but cloud provider infrastructure compatibility deteriorates due to L2 address restrictions
Solution Approach 1:
The patent transitions from L2-based routing to L3-based routing by introducing an underlay network layer. Route tables are updated with L3 routing entries that map VM IP addresses to host IP addresses, enabling routing decisions at the network layer rather than relying on L2 address learning. This dimensional shift resolves the contradiction by making routing independent of L2 address restrictions.
Solution Approach 2:
The control plane acts as an intermediary that manages route tables and distributes routing information to all hosts. It receives VM placement information and generates appropriate L3 routing entries, mediating between the overlay network requirements and the underlay network constraints. This intermediary function ensures accurate routing information is maintained without relying on cloud provider L2 capabilities.
2Extent of automation
If cloud provider restricts L2 address self-learning, then infrastructure control is improved, but VM traffic routing deteriorates when VMs move across hosts
Solution Approach 1:
The patent replaces the mechanical L2 address self-learning mechanism with a control-plane-driven L3 routing system. Instead of relying on automatic L2 address propagation through switch learning, the system uses centralized route table management that distributes routing information based on VM placement. This substitution eliminates dependency on L2 self-learning while maintaining VM relocation capability.
3Productivity
If SDN maintains flexible VM placement, then compute resource utilization is improved, but network routing accuracy deteriorates without L2 self-learning
Solution Approach 1:
The control plane implements a feedback mechanism where it receives VM placement information from hosts and uses this feedback to generate and update route tables. This closed-loop system ensures routing information remains synchronized with actual VM locations, preventing information loss even as VMs are dynamically placed across hosts for optimal resource utilization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The disclosure provides an approach for overcoming the limitations of a cloud provider network when a data center with software-defined network and multiple hosts, each with multiple virtual machines, operates on the cloud provider network. Single-host aware routers and a multiple-host aware distributed router are combined into a hybrid router in each host. The hybrid router receives a route table from the control plane of the data center and updates the received table based on the locations of VMs, such as edge VMs and management VAs on each of the hosts. An agent in each host also updates a router in the cloud provider network based on the locations of the virtual machines on the hosts. Thus, the hybrid routers maintain local routing information and global routing information for the virtual machines on the hosts in the data center.