Shared Interface for Multiple Compute Units via Virtual Tunneling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing network technologies face challenges in efficiently managing and processing network traffic across multiple compute units sharing a single network interface, as they often require complex state synchronization and fixed resource allocation, which can lead to inefficiencies and disruptions when scaling or adjusting processing capacity.

Innovation Solution

Implementing a distributed virtual appliance (DVA) system where compute units share a common interface address and use a specially configured tunneling protocol to encapsulate packets, allowing them to be distributed and processed across multiple compute units, enabling parallel processing and dynamic adjustment of resources without disrupting network traffic.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple compute units share a single network interface with complex state synchronization, then resource allocation flexibility is improved, but system complexity and processing efficiency deteriorate

Engineering Contradiction:
Improveresource allocation flexibilityVSAvoidstate synchronization complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the network interface functionality by introducing a network virtual device that separates the physical network interface from multiple virtual compute units. Each compute unit operates with its own virtual network interface, eliminating the need for complex state synchronization while maintaining resource sharing capabilities. This segmentation allows flexible resource allocation without the overhead of coordinating shared state across multiple units.

Inventive Principle:
Principle #1Segmentation

2Stability of the object's composition

If fixed resource allocation is used for compute units, then system stability is improved, but scalability and adaptability deteriorate

Engineering Contradiction:
Improvesystem stabilityVSAvoidscalability
Core Design Contradiction:
Stability of the object's compositionVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic resource allocation through virtual devices that can be created, modified, and deleted on-demand. The system allows compute units to dynamically acquire and release network interface resources without disrupting system stability. Virtual network interfaces can be allocated to compute units as needed, enabling the system to scale flexibly while maintaining stable operation through the abstraction layer provided by the virtual devices.

Inventive Principle:
Principle #15Dynamics

3Adaptability or versatility

If compute units are added or removed from a shared interface, then resource flexibility is improved, but traffic distribution stability deteriorates

Engineering Contradiction:
Improveresource flexibilityVSAvoidtraffic distribution stability
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The patent introduces a network virtual device as an intermediary between the physical network interface and compute units. This intermediary manages traffic distribution dynamically when compute units are added or removed, maintaining stability by controlling how traffic is routed to available units. The virtual device acts as a buffer that absorbs changes in compute unit composition without disrupting overall traffic flow, allowing flexible resource management while preserving distribution stability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS9602331B2Shared interface among multiple compute units
Publication Date: 2017.03.21 CISCO TECHNOLOGY INC
  • US9602331B2 patent drawing
  • US9602331B2 patent drawing
  • US9602331B2 patent drawing

AI summary

Providing a shared interface among a plurality of compute units is disclosed. A plurality of compute units is determined and a shared interface for the plurality of compute units is provided, wherein incoming traffic is received by any of the plurality of compute units. Also, the packet is received at the shared interface for a plurality of compute units. The packet is encapsulated using a first header, wherein the first header specifies one of the plurality of compute units, and wherein the one of the plurality of compute units is selected independent of an interface address associated with the shared interface.