Jobs that ignore locality haul raw blocks across the data-center fabric during large shuffles. What constraint becomes visible?
Select an answer to reveal the explanation.
Short Explanation
If every truck takes the same two-lane bridge, the bridge is the bottleneck—not the warehouses. Ignoring locality floods switches with shuffle traffic, so the network shows up as a real cluster limit. Pretending Ethernet never matters in Hadoop is fiction.
Full Explanation
Distributed computing moves data among nodes during shuffles and remote reads. When jobs lack locality or shuffle heavily, network bandwidth and switch capacity constrain throughput. HDFS and MapReduce depend on the cluster network; dismissing it or blaming only public Wi-Fi misdiagnoses municipal cluster performance.