Hypervisor Packet Routing via Guest Queue Identifiers

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing packet transmission methods in virtual systems lack efficiency in routing data packets from virtual machines to destinations, as they rely solely on hypervisor load balancing, which may not utilize the most efficient network paths or load balancing strategies.

Innovation Solution

Incorporating a queue identifier system where virtual machines can specify the transmission queue for packets, allowing them to influence the routing process and potentially use more efficient network paths or load balancing strategies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the hypervisor uses its own load balancing functions to determine which network devices to use, then the system maintains centralized control, but the transmission efficiency and path optimization are limited

Engineering Contradiction:
Improvetransmission efficiencyVSAvoidrouting control complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces queue identifiers as an intermediary mechanism between the virtual machine and the hypervisor's network transmission system. The queue identifier embedded in packet metadata serves as a mediator that carries routing information from the guest OS to the hypervisor, enabling optimized path selection without requiring complex direct communication protocols between VM and hypervisor.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The guest operating system performs preliminary routing decisions by assigning appropriate queue identifiers to packets before they are transmitted to the hypervisor. This preliminary action allows the VM to pre-determine the optimal transmission queue based on its knowledge of network conditions, application requirements, and destination characteristics, thereby improving overall transmission efficiency.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If the hypervisor manages all routing decisions, then the system maintains simplified packet structure, but the load balancing and path selection may not be optimized for specific applications

Engineering Contradiction:
Improveload balancing strategy adaptabilityVSAvoidpacket metadata complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies local quality by making the packet metadata structure adaptive rather than uniform. Different packet types and applications can have different queue identifier configurations in their metadata, allowing each application to optimize its routing according to its specific requirements while maintaining compatibility with the overall system architecture.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The queue identifier mechanism serves multiple functions: it acts as a routing label, a load balancing key, a quality of service indicator, and a network path selector. This multi-functionality allows the same metadata field to support diverse load balancing strategies and routing requirements across different applications without requiring separate mechanisms for each function.

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

3Ease of operation

If multiple transmission queues are used with different queue identifiers, then finer control over packet routing is achieved, but the system complexity increases

Engineering Contradiction:
Improverouting control precisionVSAvoidqueue management complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The system implements self-service by enabling the guest operating system to automatically select and tag packets with appropriate queue identifiers based on predefined policies or real-time conditions. This eliminates the need for manual configuration of queue assignments and reduces the operational complexity of managing multiple transmission queues, as the system adapts autonomously to routing requirements.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS10255099B2Guest-influenced packet transmission
Publication Date: 2019.04.09 RED HAT ISRAEL
  • US10255099B2 patent drawing
  • US10255099B2 patent drawing
  • US10255099B2 patent drawing

AI summary

A method performed by a physical computing system includes, with a hypervisor, receiving a first packet from a virtual machine managed by the hypervisor, the first packet for transmission to a destination, the first packet including a first queue identifier, selecting a first transmission queue from a plurality of available transmission queues to transmit the first packet, the first transmission queue being based on the first queue identifier, receiving a second packet from the virtual machine, the second packet for transmission to the destination, the second packet including a second queue identifier, and selecting a second transmission queue from the plurality of available transmission queues to transmit the second packet, the second transmission queue being based on the second queue identifier.