Active-Standby PCIe Switch Redundancy for Blade Server Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing blade server systems with a PCIE switch fabric unit face reliability issues due to the failure of the switch fabric unit, causing the entire system to become inaccessible.
Innovation Solution
A PCIE switch-based server system with an active and standby switch device configuration, where the standby device synchronizes switch network configuration information from the active device, enabling seamless failover and maintaining system reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single PCIE switch fabric unit is used to interconnect server blades, then the system structure is simple and cost-effective, but the system reliability deteriorates because the entire system becomes inaccessible when the switch fabric unit fails
Solution Approach 1:
The single PCIE switch fabric unit is segmented into multiple independent PCIE switch devices (first PCIE switch device, second PCIE switch device, etc.). Each switch device connects to server blades through dedicated PCIE switch fabrics, creating modular segments. When one segment fails, others remain operational, improving reliability while maintaining manageable complexity through standardized modular units.
Solution Approach 2:
Redundant PCIE switch devices are pre-configured in the system alongside active switches. These standby switches remain ready to take over immediately upon failure of active switches, providing beforehand cushioning against failures. Configuration information is pre-synchronized to standby devices, enabling seamless failover without system downtime.
2Reliability
If redundant PCIE switch devices are deployed to improve reliability, then system accessibility is maintained during failures, but the device complexity and configuration management become more difficult
Solution Approach 1:
Multiple PCIE switch devices are merged into a unified logical switching system through standardized interconnections. The switches share common configuration protocols and management interfaces, allowing centralized configuration management despite physical redundancy. Configuration information can be synchronized across all switch devices, simplifying management while maintaining reliability.
Solution Approach 2:
Redundant PCIE switch devices are designed with universal functionality to perform both active switching and standby roles. Each switch device具备 identical hardware capabilities and software firmware, allowing any switch to take over any other's function. This multi-functionality simplifies configuration management as the same configuration template applies to all switch devices regardless of their current role.
Data Source
AI summary
A PCIE switch-based server system, switching method and device are disclosed. The system includes: an active PCIE switch device, where the active PCIE switch device includes a communication interface and a first PCIE switch module, and the first PCIE switch module includes at least two first PCIE ports; a standby PCIE switch device, where the standby PCIE switch device includes a communication interface and a first PCIE switch module, and the first PCIE switch module includes at least two first PCIE ports; where the communication interface of the active PCIE switch device and the communication interface of the standby PCIE switch device are interconnected, so that the standby PCIE switch device obtains switch network configuration information of the active PCIE switch device through the communication interface of the active PCIE switch device and the communication interface of the standby PCIE switch device.