Interconnect Isolation via Virtual Link Mapping for VM Error Containment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic systems with multiple virtual machines (VMs) face issues where errors in one VM can propagate, causing the entire system to crash, necessitating a system-level reset, which compromises robustness.
Innovation Solution
An interconnect isolation system that maps VMs to independent virtual links, enabling recovery mechanisms for individual links without affecting others, thus improving robustness and reducing hardware requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple VMs share the same interconnect, then hardware area and power consumption are reduced, but error propagation causes system-level crashes
Solution Approach 1:
The shared interconnect is segmented into multiple independent virtual links through mapping VMs to virtual links. Each virtual link operates independently, so errors in one VM are contained to its specific virtual link and cannot propagate to other VMs. This segmentation achieves isolation without requiring separate physical interconnects for each VM.
Solution Approach 2:
Virtual links act as intermediaries between VMs and the physical interconnect. The mapping mechanism translates VM communications into virtual link transactions, providing a layer of isolation. This intermediary structure enables error containment while maintaining shared hardware resources.
2Reliability
If dedicated physical links are used for each VM, then error isolation is achieved, but hardware area and power consumption increase
Solution Approach 1:
Multiple virtual links share the same physical interconnect infrastructure, making the physical interconnect universal for all VMs. This multi-functionality allows a single physical interconnect to serve multiple VMs with error isolation, eliminating the need for dedicated physical links for each VM and reducing overall hardware area and power consumption.
Solution Approach 2:
Instead of copying the entire physical interconnect for each VM, the patent creates virtual copies (virtual links) that map to the shared physical interconnect. This allows multiple VMs to have independent error isolation while sharing the same physical hardware, reducing resource duplication.
3Object-generated harmful factors
If system-level reset is implemented, then error propagation is handled, but system robustness is compromised due to frequent restarts
Solution Approach 1:
The interconnect is segmented into independent virtual links, allowing individual error recovery without system-level resets. When an error occurs in one VM, only its specific virtual link needs recovery, not the entire system, thereby maintaining robustness while handling error propagation.
Solution Approach 2:
The virtual link mapping mechanism enables self-service error recovery where each virtual link can be independently recovered without external intervention or system-wide restart. This self-service capability improves system robustness by eliminating frequent system-level restarts.
Data Source
AI summary
An interconnect isolation system includes a shared memory, at least one processing circuit, and an interconnect. The at least one processing circuit is arranged to execute a plurality of virtual machines (VMs). The interconnect is shared by the plurality of VMS, and includes at least one interconnect node. The at least one interconnect node includes at least one transmission circuit, wherein each of the at least one transmission circuit comprises a plurality of virtual links, the plurality of VMs are mapped to the plurality of virtual links, and each of the plurality of VMs transmits a transaction to the shared memory through a mapped virtual link.