Vehicle Network Accelerator Virtualization for Fewer Physical Ports
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automotive network architectures face challenges in integrating efficient network accelerators for high-bandwidth switching and routing with minimal area overhead, particularly in managing a reduced number of physical ports while supporting a high number of virtual machines, leading to increased complexity and cost.
Innovation Solution
A network architecture that maps physical ports and virtual machine interfaces onto shared virtual machine ports, supporting virtualization with reduced gate count and complexity, and offering flexible configuration of virtual machines per AXI Master port.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a reduced number of physical ports is used in the switch, then area overhead and gate count are minimized, but supporting a high number of virtual machines becomes challenging
Solution Approach 1:
The patent introduces a virtualization dimension by mapping multiple virtual machine interfaces to a reduced set of physical ports through virtual switch ports. This allows the system to support many VMs (high adaptability) while using fewer physical ports (low area overhead) by adding the virtualization layer that maps virtual interfaces to physical interfaces many-to-one or many-to-few.
Solution Approach 2:
Each physical port is designed to handle multiple virtual machine interfaces simultaneously, making the physical port universal in its function. The switch fabric and egress queues are configured to serve multiple VMs through a single physical port, allowing one physical port to perform the work of multiple dedicated ports while maintaining support for high numbers of VMs.
2Adaptability or versatility
If more virtual machine interfaces are mapped to physical ports, then virtualization capabilities are enhanced, but device complexity increases
Solution Approach 1:
The patent segments the switch functionality into distinct components: ingress port processing, virtual switch port mapping layer, switch fabric, and egress queue management. This segmentation allows complex virtualization mapping to be handled in a dedicated layer without overwhelming the entire device, reducing overall complexity while enhancing virtualization capabilities.
Solution Approach 2:
The patent introduces virtual switch ports as an intermediary layer between physical ports and virtual machine interfaces. This intermediary abstracts the complexity of mapping many VMs to fewer physical ports, providing a clean interface that simplifies the overall device architecture while enabling advanced virtualization capabilities.
3Ease of manufacture
If physical ports are reduced to lower cost, then manufacturing cost decreases, but bandwidth availability for virtual machines is limited
Solution Approach 1:
The patent creates virtual copies of network interfaces through virtual switch ports that map to physical ports. Multiple VMs can have their own virtual interface copies that share the physical port bandwidth, allowing cost-effective reduction in physical ports while maintaining adequate bandwidth availability through virtualization and bandwidth sharing mechanisms.
Solution Approach 2:
The patent merges multiple virtual machine traffic streams through shared physical ports using a packet buffer and switch fabric. Multiple VMs' bandwidth requirements are combined and managed through shared physical infrastructure, reducing the total number of physical ports needed while maintaining sufficient bandwidth availability through efficient traffic aggregation and switching.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The bandwidth of SOC interfaces is exploited while minimizing the number of physical ports via a networking accelerator (2000) for use on board a vehicle, for instance, that comprises: media access control, MAC controller circuitry (2001) configured to provide a MAC port layer to control exchange of information, wherein exchange of information comprises data flow transmission to virtual machine ports, VMPs over a data link (Ethernet/CAN/LIN), virtual machine transmission, VM Tx bridge circuitry (2003) configured to handle transmission data flow to the virtual machine ports, VMPs, transmission router/switch circuitry (2002) configured to route/switch data flow from the MAC controller circuitry (2001) to the VM Tx bridge circuitry (2003), and queue handler circuitry (2006A, 2006B) configured to provide queue management for data flow between the MAC controller circuitry (2001) and the VM Tx bridge circuitry (2003). The VM Tx bridge circuitry (2003) comprises virtual destination address circuitry (2003C) configured to implement router/switch virtualization in said transmission router/switch circuitry (2002) with a virtual machine transmission descriptor (2003C) based on a virtual machine port, VMP tag (3000E) indicative of a physical resource in the queue handler circuitry (2006A, 2006B) selectable for data flow transmission, in combination with a virtual machine extended identifier, VMEID (3000D).