Distributed Logical Routing Element for VM Latency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In network virtualization environments, communication between virtual machines (VMs) in different subnets leads to unnecessary network traffic and performance degradation due to the requirement for IP packets to traverse through a shared L3 router, even if the destination VM is on the same host machine, resulting in higher latency and lower throughput.

Innovation Solution

A logical routing element (LRE) with multiple logical interfaces (LIFs) is introduced, allowing each network segment to have its own LIF addressable by multiple IP addresses, enabling local L3 packet forwarding within host machines and allowing physical hosts to select designated instances for routing, thereby reducing reliance on shared L3 routers and optimizing traffic paths.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If IP packets traverse through a shared L3 router for communication between VMs in different subnets, then routing functionality is provided, but network latency increases and throughput decreases

Engineering Contradiction:
Improverouting functionalityVSAvoidnetwork latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the centralized L3 routing function into distributed L3 routing elements (LREs) deployed across multiple host machines. Each LRE handles routing for local VMs, eliminating the need for all packets to traverse a shared router and reducing latency while maintaining routing functionality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimensional approach by implementing L3 routing at the host level rather than relying solely on a centralized router. This distributed architecture adds a new layer of routing capability that operates locally, reducing the traversal path and latency for inter-subnet communication.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If IP packets traverse through a shared L3 router for communication between VMs in different subnets, then routing functionality is provided, but network throughput decreases

Engineering Contradiction:
Improverouting functionalityVSAvoidnetwork throughput
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the centralized L3 routing function into distributed L3 routing elements (LREs) deployed across multiple host machines. Each LRE handles routing for local VMs, eliminating the need for all packets to traverse a shared router and reducing latency while maintaining routing functionality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Host machines with LREs perform local routing for their own VMs without requiring external router intervention. This self-service capability allows packets to be routed locally when possible, increasing throughput by avoiding unnecessary external traversal.

Inventive Principle:
Principle #25Self-service

3Ease of operation

If a single designated instance handles L3 routing for physical hosts, then routing is simplified, but load balancing and network performance are reduced

Engineering Contradiction:
Improverouting simplicityVSAvoidnetwork performance
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent segments the single-point L3 routing function into multiple distributed LREs across different host machines. Physical hosts can select from multiple designated instances, enabling load balancing while maintaining simplified routing through the use of multiple equivalent paths.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameter of designated instance selection from a single fixed instance to multiple selectable instances. Physical hosts can choose different designated instances based on load conditions, improving performance while maintaining routing simplicity through standardized selection mechanisms.

Inventive Principle:
Principle #35Parameter changes

4Productivity

If multiple designated instances are used for L3 routing, then load balancing and network performance are improved, but routing complexity increases

Engineering Contradiction:
Improvenetwork performanceVSAvoidrouting complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements universal LREs that can function as designated instances for multiple network segments and serve both local VM routing and physical host routing. This multi-functionality reduces complexity by using a standardized component across different routing scenarios.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent uses identical LRE software instances deployed across multiple host machines, creating copies of the same routing logic. This standardization simplifies management and operation while enabling load balancing through multiple equivalent routing paths.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3629529B1Ingress ECMP in virtual distributed routing environment
Publication Date: 2022.08.31 NICIRA INC
  • EP3629529B1 patent drawingFigure 1
  • EP3629529B1 patent drawingFigure 2
  • EP3629529B1 patent drawingFigure 3

AI summary

A logical routing element (LRE) having multiple designated instances for routing packets from physical hosts (PH) to a logical network is provided. A PH in a network segment with multiple designated instances can choose among the multiple designated instances for sending network traffic to other network nodes in the logical network according to a load balancing algorithm. Each logical interface (LIF) of an LRE is defined to be addressable by multiple identifiers or addresses, and each LIF identifier or address is assigned to a different designated instance.