# 节点的网络拓扑距离

The idea is that the bandwidth available for each of the following scenarios becomes progressively less：

1. Processes on the same node
2. Different nodes on the same rack
3. Nodes on different racks in the same data center
4. Nodes in different data centers