TCP Bulk Tests
Description
In this section the results of the executed TCP bulk tests will be discussed.
Again the Iperf
V. 1.6.5 traffic generator has been used to create the TCP flows. Multiple,
parallel flows can be generated relatively light-weighted by using the so called
pthread library. As often in these type of tests they are executed as a
function of the # parallel flows and of the TCP window size which will be
varied such that the sum of the TCP window over the # parallel flows will
be kept constant.
Results
In the following plots the sum TCP throughput has been displayed as a function
of the sum TCP window sizes, taken over the # parallel flows, and of those
parallel flows. The maximum used TCP window size per flow was 64 MBytes.
The test duration of each flow was 60 s with interval bandwidth reports of
4 s. The sum TCP throughput values have been calculated by taking the
median from the sum of the interval bandwidth reports which were re-sampled to
exact 4 s. intervals such that their sum could be taken in a proper way.
In
these results are presented for the test direction with rate limiting
Amsterdam Science Park -> StarLight and in
for the reverse direction without rate limiting.
| . |
|
Sum TCP bulk throughput values, taken over the
# parallel flows, as a function of the sum TCP window size, also taken
over the # parallel flows, and of the # parallel flows for the
test direction with rate limiting Amsterdam Science Park ->
StarLight. |
| . |
|
Sum TCP bulk throughput values, taken over the
# parallel flows, as a function of the sum TCP window size, also taken
over the # parallel flows, and of the # parallel flows for the
test direction without rate limiting StarLight ->
Amsterdam Science Park. |
In
the average TCP bulk throughput per flow has been displayed as a function of the
TCP window size per flow for the test direction with rate limiting
Amsterdam Science Park -> StarLight. The results for
some selected # parallel flows have been represented by their
corresponding plot traces. In
these data have been presented for the reverse direction without rate limiting.
| . |
|
Average TCP bulk throughput values per flow as a
function of the TCP window size per flow for the test direction with rate
limiting Amsterdam Science Park -> StarLight. The
results for some selected parallel flows have been represented by their
corresponding plot traces. |
| . |
|
Average TCP bulk throughput values per flow as a
function of the TCP window size per flow for the test direction without rate
limiting StarLight -> Amsterdam Science Park. The
results for some selected parallel flows have been represented by their
corresponding plot traces. |
Conclusions
From the TCP bulk tests presented in the
the following conclusions can be drawn:
-
In the test direction with rate limiting,
Amsterdam Science Park -> StarLight, the obtained
throughput completely collapses. In fact the throughput is independent from
the used TCP window but it is proportional with the # flows. A possible
explanation is that reserved memory which is related with the configured
50 KBytes peak-rate and 50 KBytes burst size (see the
"Setup" section) is too
small.
-
The results in the test direction without rate limiting,
StarLight -> Amsterdam Science Park, are as could
be expected.