VXLAN Multicast Load Distribution via Dynamic BIDIR Rearrangement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current VXLAN solutions face issues with traffic concentration on few links due to separate multicast trees for tenants and within tenants, leading to operational challenges in distributing network load effectively across the VXLAN topology.

Innovation Solution

Preconfiguring multiple potential Rendezvous Points (RPs) across the network and using a Management and Provisioning Entity (MPE) to dynamically rearrange bidirectional groups (BIDIRs) based on traffic thresholds, creating new BIDIRs and moving VMs to distribute traffic load more evenly across the VXLAN system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If separate multicast trees are provided for each tenant and within tenants, then multicast functionality is achieved, but traffic concentrates on few links across the VXLAN

Engineering Contradiction:
Improvemulticast functionalityVSAvoidtraffic concentration
Core Design Contradiction:
Adaptability or versatilityVSObject-generated harmful factors

Solution Approach 1:

The patent segments the single multicast tree into multiple separate multicast trees, each associated with a specific Rendezvous Point (RP). By dividing the tenant groups across multiple RPs and trees, traffic is distributed across different network paths and links, preventing concentration on a single link while maintaining multicast functionality for each segment.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an additional dimension of organization by creating hierarchical multicast trees at different levels (VXLAN-level trees and tenant-level trees). This multi-dimensional tree structure allows traffic to be distributed across multiple RPs and trees simultaneously, spreading load across more links while preserving the ability to provide targeted multicast service.

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

2Productivity

If multiple RPs are preconfigured, then load distribution is improved, but system complexity increases

Engineering Contradiction:
Improveload distribution efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent preconfigures multiple RPs and establishes multiple multicast trees in advance before traffic arrives. This preliminary setup allows the system to immediately distribute traffic across multiple paths when needed, improving load distribution efficiency without requiring complex real-time decision-making during operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces RPs as intermediary nodes that manage and control multicast traffic distribution. Each RP acts as a mediator for its associated multicast tree, handling the complexity of tree management and traffic routing centrally, which simplifies the overall system architecture while enabling efficient load distribution across multiple RPs.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS9515935B2VXLAN based multicasting systems having improved load distribution
Publication Date: 2016.12.06 DELL PROD LP
  • US9515935B2 patent drawing
  • US9515935B2 patent drawing
  • US9515935B2 patent drawing

AI summary

A virtual extensible local area network (“VXLAN”) system includes servers having VXLAN tunnel end points (“VTEP”) and virtual machines (“VMs”), multiple routers to transmit network data using bidirectional groups (“BIDIRs”) and rendezvous points (“RPs”), and an associated management and provisioning entity (“MPE”) adapted to arrange BIDIRs dynamically based upon network traffic such that traffic is more evenly distributed across the VXLAN. Some of the routers can be preconfigured as potential RPs, and the MPE can activate and deactivate these routers to and from RP status as needed. The MPE creates and rearranges BIDIRs based upon volume thresholds of traffic being reached in BIDIRs, such as by splitting overloaded BIDIRs. Network traffic can be monitored by VTEPs and RPs.