Building the internal network.
WHAT A TOP-OF-RACK SWITCH DOES
Connects the servers in one rack, and connects upward.
WHY THAT ARRANGEMENT
Cabling stays within the rack, and failure is contained.
WHAT REDUNDANCY LOOKS LIKE
Two switches per rack Two connections from each server, one to each Several paths upward
WHAT LINK AGGREGATION PROVIDES
Several physical links treated as one, for capacity and resilience.
WHAT MULTI-CHASSIS ARRANGEMENTS ADD
Aggregation across two switches, so a switch failure does not break the aggregate.
WHAT TO VERIFY
That both switches are on separate power feeds.
WHY
Otherwise switch redundancy is meaningless.
WHAT TO CONFIGURE
Loop prevention Storm control Consistent configuration across pairs
WHY CONSISTENCY MATTERS
Mismatched configuration between a pair causes intermittent faults that are extremely difficult to diagnose.
WHAT TO AUTOMATE
Configuration generation, so pairs cannot diverge.
WHAT TO MONITOR
Link status Error and discard counters Whether you are running on a single path
WHAT RISING ERRORS INDICATE
A failing cable or transceiver, before it fails completely.
