InfiniBand Routing for Ethernet Fabrics in VM Live Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud computing architectures face performance and administrative bottlenecks due to complex addressing and routing schemes in InfiniBand (IB) networks, particularly in high performance computing environments, which hinder efficient live migration of virtual machines (VMs) and resource utilization.
Innovation Solution
Implement InfiniBand routing algorithms for Ethernet fabrics, utilizing a topology provider (TP) module, routing engine (RE) module, and switch initializer (SI) module to perform discovery sweeps, assign addresses, calculate routing maps, and configure switches with forwarding tables, enabling dynamic reconfiguration and efficient resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If InfiniBand routing algorithms are implemented for Ethernet fabrics, then network performance and resource allocation efficiency are improved, but system complexity and implementation difficulty increase
Solution Approach 1:
The patent applies universality by implementing InfiniBand routing algorithms in an Ethernet fabric environment, allowing the routing engine to handle multiple protocols (InfiniBand and Ethernet) through a unified architecture. The routing engine calculates routing maps and generates forwarding tables that work across different network types, enabling single system to perform multiple routing functions.
Solution Approach 2:
The patent introduces a routing engine as an intermediary component that sits between the network switches and the management system. This routing engine acts as a mediator that performs discovery sweeps, calculates routing maps, and configures forwarding tables, thereby simplifying the overall system complexity by centralizing the complex routing logic in a dedicated intermediary component.
2Adaptability or versatility
If dynamic reconfiguration capability is added to enable live migration of VMs, then adaptability and resource utilization are improved, but overhead and processing time increase
Solution Approach 1:
The patent applies preliminary action by performing discovery sweeps and calculating routing maps in advance before live migration is needed. The routing engine proactively discovers network topology and pre-computes routing paths, storing these routing maps for quick retrieval and application during VM migration events, thereby reducing the time overhead during actual migration.
Solution Approach 2:
The patent implements dynamics by enabling the routing configuration to change dynamically during runtime. The system can recalculate routing maps and update forwarding tables on-the-fly when VMs need to migrate between hosts, allowing the network to adapt its routing paths dynamically without requiring physical reconfiguration or system downtime.
3Measurement precision
If comprehensive discovery sweeps are performed to map network topology, then routing accuracy and resource allocation are improved, but processing overhead and time consumption increase
Solution Approach 1:
The patent applies partial action by performing discovery sweeps that focus on the specific portions of the network topology that are relevant to current routing needs. Rather than continuously scanning the entire network, the routing engine performs targeted discovery sweeps to gather sufficient information for accurate routing map calculation, reducing the time overhead while maintaining routing accuracy.
Data Source
AI summary
Systems and methods for using InfiniBand routing algorithms for Ethernet fabrics in a high performance computing environment. The method can provide, at a computer comprising one or more microprocessors, a plurality of switches, a plurality of hosts, a topology provider (TP) module, a routing engine (RE) module, and a switch initializer (SI) module. The method can perform a discovery sweep, by the TP, of the plurality of hosts and the plurality of switches and assigns an address to each of the plurality of hosts and the plurality of switches. The method can calculate, by the routing engine, a routing map, based upon a routing scheme, for the plurality of hosts and the plurality of switches, the routing map comprising a plurality of forwarding tables. The method can configure, each of the plurality of switches with a forwarding table of the plurality of forwarding tables calculated by the routing engine.