Traffic Replication Device for Instant Routing Engine Switchover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet routers experience significant delays during switchover between master and backup routing engines, leading to packet losses and protocol timeouts due to the time-consuming process of re-enumerating PCIe connections and loading drivers.
Innovation Solution
The implementation of a traffic replication device that replicates control traffic and transmits it to both routing engines via dedicated Ethernet connections, allowing for instantaneous switchover by selecting control signals from the backup engine upon detection of a fault in the master engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional PCIe connection re-enumeration and driver loading methods are used for switchover, then routing engine redundancy is achieved, but switchover time increases to many seconds causing packet losses and protocol timeouts
Solution Approach 1:
The patent segments the control traffic path from the data traffic path. Control traffic is forwarded to a traffic replication device via a first link, while data traffic is forwarded to the routing engine via a second link. This segmentation allows control plane switchover to occur independently and instantly without affecting the data plane, resolving the contradiction between maintaining redundancy and minimizing switchover time.
Solution Approach 2:
The traffic replication device serves as an intermediary between the packet forwarding board and the routing engines. It replicates control traffic to both the master and backup routing engines simultaneously, enabling the backup engine to be pre-loaded with control plane state. This intermediary mechanism allows instantaneous switchover by simply redirecting control traffic to the backup engine without requiring PCIe re-enumeration or driver reloading.
2Loss of time
If control traffic is forwarded to both routing engines via dedicated Ethernet connections, then instantaneous switchover is enabled, but device complexity increases due to additional links and replication device
Solution Approach 1:
The traffic replication device performs multiple functions: it receives control traffic from the packet forwarding board, replicates the control traffic to both routing engines, and can select control signals from either engine based on fault detection. This multi-functionality consolidates what would otherwise require separate components, managing complexity while enabling instantaneous switchover.
Solution Approach 2:
The system performs preliminary action by continuously replicating control traffic to both routing engines and maintaining control plane state in the backup engine beforehand. When a fault occurs, the backup engine is already prepared with the necessary control information, eliminating the need for time-consuming initialization during switchover. This preliminary preparation reduces switchover time without requiring complex real-time computation during the fault event.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A disclosed computing device capable of instantly switching over between routing engines may include (1) a packet forwarding board configured to (A) forward control traffic via a first link to a traffic replication device and (B) forward data traffic via a second link to a first routing engine, (2) the traffic replication device configured to (A) replicate the control traffic received from the packet forwarding board and (B) select control signals received from the first routing engine, (3) the first routing engine configured to receive control traffic from the traffic replication device, and (4) a second routing engine configured to receive control traffic from the traffic replication device. Various other apparatuses, systems, and methods are also disclosed.