Juniper ACX1000 Configuration Manual page 1395

Copyright © 2017, Juniper Networks, Inc.
Throughput is the maximum rate at which the count of test frames received is
equal to the number of test frames transmitted through the network service.
You can repeat throughput tests for different frame sizes.
Latency test:
NOTE: To run latency test, you need to determine the throughput for DUT
or a network service at each of the specified frame sizes.
Starts with a stream of frames at a particular frame size through the DUT at the
determined throughput rate.
Sends an identifying tag in one frame after 60 seconds and calculate the latency
when the frame with the same tag is received by the initiator.
Is repeated for at least 20 times with the reported latency value being the average
of the recorded values.
You can repeat latency tests for different frame sizes.
Frame loss rate test:
Involves sending a specific number of frames at a specified rate through the DUT or
a network service to be tested and counting the frames that are transmitted.
Calculates frame loss rate at each point using the equation:
( ( input_count - output_count ) x 100 ) / input_count.
Runs a trial for the frame rate that corresponds to 100 percent of the configured
maximum theoretical rate.
Is repeated for the frame rate that corresponds to 90 percent of the maximum rate
used and then for 80 percent of the maximum rate until a certain trial result shows
no lost frames.
You repeat the frame loss rate tests for different frame sizes.
Back-to-back frames test:
Involves sending a burst of frames with minimum interframe gaps through the DUT
or a network service and counting the number of frames forwarded.
Is rerun with an increased length of burst frames if the count of transmitted frames
is equal to the number of frames forwarded.
Is rerun with a reduced length of burst frames if the count of forwarded frames is
less than the number of frames transmitted.
The back-to-back value is the number of frames in the longest burst that the DUT or
a network service can handle without the loss of any frames.
You can repeat back-to-back frame tests for different frame sizes.
Chapter 37: Configuring RFC 2544-Based Benchmarking Tests
1337