InfiniBand vSwitch Architecture for Transparent VM Live Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing high performance computing environments face challenges in managing live migration of virtual machines due to complex addressing and routing schemes in lossless networks like InfiniBand, leading to connectivity issues and performance degradation during VM migration.
Innovation Solution
Implementing a vSwitch architecture with prepopulated LIDs and dynamic reconfiguration mechanisms to optimize resource utilization and enable transparent live migration of virtual machines, using InfiniBand technology with SR-IOV for efficient network communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional network addressing and routing schemes are used in lossless networks like InfiniBand, then network connectivity can be established, but complex addressing and routing schemes lead to connectivity issues and performance degradation during VM migration
Solution Approach 1:
The patent segments the virtual network interface into multiple virtual functions (VFs) within a single physical function (PF). Each VF can be independently assigned to different VMs, allowing granular control of network resources. This segmentation enables independent migration of individual VFs without affecting other VMs, simplifying the migration process while maintaining connectivity reliability.
Solution Approach 2:
The physical function (PF) acts as an intermediary between the virtual machine layer and the physical network infrastructure. The PF handles complex addressing and routing operations, shielding VMs from these complexities. During migration, the PF manages the transition by maintaining network connectivity while reconfiguring virtual functions, thus resolving the contradiction between reliability and complexity.
2Adaptability or versatility
If live migration of virtual machines is performed in lossless networks, then VM mobility is achieved, but network reconfiguration overhead and connectivity issues occur
Solution Approach 1:
The patent pre-configures multiple virtual functions within the physical function before migration is needed. These VFs are prepared with appropriate network settings and mappings to network resources. When migration is required, the pre-configured VFs can be quickly activated at the destination without extensive reconfiguration, significantly reducing migration time while maintaining VM mobility.
Solution Approach 2:
The system dynamically allocates and reconfigures virtual functions during VM migration. The PF can dynamically create, delete, or reassign VFs based on migration requirements. This dynamic capability allows the system to adapt to changing network conditions and VM placement requirements, enabling seamless migration with minimal downtime.
3Productivity
If SR-IOV is used for efficient network communication, then resource utilization improves, but complex device configuration and management is required
Solution Approach 1:
The system implements self-service capabilities where virtual functions automatically configure themselves based on their assigned VMs and network requirements. The PF monitors resource usage and automatically optimizes network configurations without requiring manual intervention. This self-service approach maintains high resource utilization while simplifying operations, as the system adapts automatically to changing conditions.
Data Source
AI summary
Systems and methods for supporting target groups for congestion control in a private fabric in a high performance computing environment. An exemplary method can provide, at one or more microprocessors, a first subnet, the first subnet comprising a plurality of switches, a plurality of host channel adapters, and a plurality of end nodes, including a plurality of virtual machines. The method can define a target group on one of an inter-switch link or at a port of a switch of the plurality of switches, wherein the target group defines a bandwidth limit on the at least one of an inter-switch link between two switches of the plurality of switches or at a port of a switch of the plurality of switches. The method can provide a target group repository stored in a memory of the host channel adapter where the defined target group in the target group repository is recorded.