Switch Port Virtualization for MDC Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtualized device environments, mutually-conflicting Multitenant Devices Context (MDC) configurations on switches limit the application range of virtualized MDCs, as they are typically allocated based on chips rather than service demands, leading to potential conflicts and inefficient resource utilization.
Innovation Solution
MDCs are virtualized based on ports, with allocation of switch ports to corresponding MDCs according to service demands, using a method that maps MDC IDs and VLAN IDs to Virtual Switch Instances (VSIs) or VPN Routing and Forwarding instances (VRFs) to isolate and manage layer-2 and layer-3 packets, preventing conflicts and optimizing resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If MDCs are virtualized based on chips, then hardware resource isolation is achieved, but MDC configurations may conflict and application range is limited
Solution Approach 1:
The patent segments MDC virtualization from chip-level to port-level, dividing the switch into multiple virtual switches based on port assignments. Each virtual switch corresponds to an MDC and handles only packets from assigned ports, eliminating configuration conflicts while maintaining isolation. This segmentation allows flexible port allocation to different MDCs based on service demands.
Solution Approach 2:
The patent introduces a new dimension of virtualization by creating virtual switches as an intermediate layer between physical chips and MDCs. Instead of direct chip-to-MDC mapping, virtual switches are introduced that can be dynamically allocated to different MDCs, adding flexibility without compromising hardware isolation.
2Adaptability or versatility
If MDCs are virtualized based on ports, then resource allocation flexibility is improved, but packet forwarding complexity increases
Solution Approach 1:
The patent segments packet forwarding into distinct layers: physical port reception, virtual switch identification based on MDC ID and VLAN ID, and forwarding decision execution. This segmentation simplifies complexity by organizing the forwarding process into manageable stages with clear responsibilities at each level.
Solution Approach 2:
The patent introduces virtual switch instances (VSIs) as intermediaries between physical ports and MDCs. VSIs act as mediators that translate physical port identifiers and VLAN information into MDC-specific forwarding contexts, simplifying the overall forwarding logic while maintaining flexibility.
3Productivity
If multiple MDCs share the same chip, then resource utilization is optimized, but configuration conflicts occur
Solution Approach 1:
The patent segments MDC configurations into separate virtual switch contexts, allowing multiple MDCs to share physical chip resources while maintaining distinct configuration spaces. Each virtual switch has its own forwarding table and configuration parameters, preventing conflicts even when multiple MDCs utilize the same physical chip.
Solution Approach 2:
Virtual switches serve as intermediary layers between physical chip resources and MDC configurations. This intermediary structure allows multiple MDCs to access shared chip resources through isolated virtual interfaces, enabling high resource utilization while preventing configuration conflicts through virtualization abstraction.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A method for forwarding a packet and a switch are provided. The switch receives a packet via a port; when the packet is a layer-2 packet, determines a VSI based on an ID of a MDC to which the port belongs and an ID of a VLAN of the packet, searches for a layer-2 entry stored in a chip where the port is located and matching with the VSI and a destination MAC address of the packet, and forwards the packet based on the layer-2 entry; when the packet is a layer-3 packet, determines a second VRF based on the ID of the MDC and an ID of a first VRF to which the packet belongs, searches for a layer-3 entry stored in the chip and matching with the second VRF and a destination IP address of the packet, and forwards the packet based on the layer-3 entry.