User`s guide

Agilent PXT Wireless Communications Test Set
User’s Guide
153
Downlink
PXT server (192.168.1.230) > iperf s –i1 –p5052 –w20m
UE Host Client PC (192.168.1.51) > iperf –c 192.168.1.230 –i1 –w20m –t300 –p5052
Parameter Descriptions
-p Specifies the port ID to use. Ideal when firewalls are in use on specific sides and allows control
over the port ID.
-P Specifies the number of concurrent parallel TCP threads/streams. This should be used in TDD TCP
or duplex TCP situations, where using a single TCP stream results in lower than expected bitrates.
Iperf reports the bitrate per parallel TCP stream and summary totals when running with P option.
T Specifies the number of seconds to run the Iperf stream.
-b Specifies the target UDP bitrate. The UDP bitrate specified (-b option: for example: b48m
generates 48mbps), must be adjusted to the current maximum theoretical bitrate supported for the
current PXT configuration. For example if the PXT UL has a limit of 25Mbps imposed, you must
specify less than or equal to 25 Mbps in Iperf.
-i Seconds between periodic bitrate reports
-w Specifies the tuning and is based on the latency in the system.
The actual achieved bitrate through the PXT is displayed on the receiver side of Iperf (Iperf Server), along
with additional information for UDP streams (packet loss, latency, and jitter). Caution should be taken to
ensure results are read from the receiver side as the transmitting side of Iperf (client) displays what it is
generating and this can differ from the actual achieved Mbps if there is packet loss and reduced pipe size
on the PXT.
Typical rates to expect for the E2E bitrate is around 1-2 Mbps less than the maximum theoretical at high
end rates by the time RF overheads are deducted such as headers, signaling, and scheduling messages. For
example, for
102 Mbps DL, we can expect an E2E receive data rate of around 99.9 Mbps. Similarly for 51 Mbps UL, we
typically achieve 48.9 - 49.9 Mbps UL. TCP & UDP achieve very similar rates with TCP sometimes slightly
less, depending on conditions.
The current version of Iperf used in the labs is 2.0.5. There are known IPv6 issues with
earlier versions.
CPU resource of the test station endpoints running Iperf should be monitored to eliminate any unexpected
performance issues related to Iperf/NIC being under-resourced. Iperf, when running at Cat3 high end rates
with typical mobile latencies of 20-30 ms and server side sequence / packet loss checking, requires a lot
of CPU. Results of Iperf, when under resourced, typically show packet loss due to Iperf/receiver side lack
of processing power, and results should be read with caution when 90-100% CPU usage is observed. If
observing high CPU usage with Iperf, try upgrading the host if possible. The PXT has been through
verification at high rates (currently up to Cat3 performance) so test equipment should be eliminated first, if
you observe unexpected packet loss or reduced rates.