A city compares sFlow-like sampling with counter streaming for its leaves. How should those analytics approaches be distinguished at concept level?
Select an answer to reveal the explanation.
Short Explanation
Sampling is tasting the soup ladle by ladle; counter streaming is watching the stove dials spin in real time. Both help cook, but one profiles flows and the other tracks rates and queues. Calling them identical just confuses the kitchen.
Full Explanation
Sampled flow analytics observes a subset of packets/flows, while streamed counters export interface or queue metrics at high frequency—different city-leaf tools for different questions. They are not always the same mechanism producing identical records. Counter streaming is not limited to syslog facility codes. Sampled flow analytics is not restricted to wireless APs; it applies to DC leaves. Exam caveat: Use flow sampling for composition/top talkers and counter streaming for rate/queue time series. Operational check: For one incident, gather both a flow sample summary and counter time series to show complementary value. In civic data-center operations, record the expected CLI or collector evidence for this concept so the next on-call engineer can validate fabric behavior without re-deriving the design under incident pressure. In civic data-center operations, record the expected CLI or collector evidence for this concept so the next on-call engineer can validate fabric behavior without re-deriving the design under incident pressure. In civic data-center operations, record the expected CLI or collector evidence for this concept so the next on-call engineer can validate fabric behavior without re-deriving the design under incident pressure.