Standby/Failover Rack deployments are a common way to compensate for downtime risk from the TOR Single Point of Failure (SPOF). But having so much redundancy creates power issues and cost. Lack of preparation for a TOR failure can lead to unacceptable levels of service.
Microsoft has developed a new Dual TOR implementation using Credo HiWire SWITCH AECs controlled by SONIC’s Dual TOR Management Container. This new implementation can be seamlessly implemented at ant time, eliminating the need for rack redundancy, thus reducing power and cost (or repurposing racks to increase capacity).