Understanding Network Flooding and Broadcast Storms: From Physical Floods to Data Center Networks

The piece begins with vivid images of rain‑induced flooding, drawing a parallel between natural floods and the concept of network flooding in Ethernet switches.

In a Layer‑2 network, when a switch receives a frame with an unknown destination MAC address, it floods the frame out of all ports except the one it arrived on; if a loop exists, the flooded frames circulate indefinitely, creating a broadcast storm that can exhaust all network resources.

The article explains that a simple tree‑topology network has no loops and thus no broadcast storms, but such a design lacks redundancy; adding backup links inevitably creates loops.

To prevent broadcast storms, two main techniques are presented: first, dividing the physical network into smaller logical broadcast domains using VLANs, which confines broadcast traffic within each VLAN; second, employing loop‑prevention protocols (referred to as “破环协议”) such as STP, RSTP, MSTP, and others, which block redundant ports under normal conditions and only activate them when a failure occurs.

The discussion then moves to traditional data‑center network architecture, describing the common three‑tier design (access layer at Layer‑2, aggregation layer providing routing, and core layer handling Layer‑3 routing) and noting that this mature structure, combined with VLANs and xSTP protocols, supports modular and reliable data‑center operations.

Finally, the article reflects on the analogy, emphasizing that just as physical floods require awareness and mitigation, understanding network flooding and broadcast storms is essential for building resilient virtualized and cloud environments, and encourages readers to also learn about real‑world flood prevention.