IEEE - Aerospace and Electronic Systems - May 2020 - 9

Shi et al.

Figure 4.
Comparison of the goodput performance of LTP in transmitting a
file of 1 Mbyte over a cislunar communication channel with
highly asymmetric channel rates (with the CR of 500/1) and
various lengths of the CP timer starting from 4 s.

ANALYSIS OF CP TIMER LENGTHS FOR LTP
TRANSMISSIONS OVER ASYMMETRIC CHANNEL
For LTP transmission over asymmetric channel, if the CP
timer length is set too short, it might expire too early with
respect to the estimated length of RTT, resulting in transmission performance degradation. Therefore, for the proposed file transmission experiments over asymmetric
channel, a wide range of the CP timers are adopted to
measure their performance variation trend.
Figure 4 presents a comparison of the numerical goodput performance of LTP in transmitting a file of 1 Mbyte
with various lengths of the CP timer with highly asymmetric channel rates (with the CR of 500/1) and BER of 0
experienced. The goodput performance of LTP with five
CP timer lengths in a broad range is compared, including

4, 12, 20, 28, and 36 s. The selection of the timer starts
from 4 s because it is already experimented over symmetric channel and works well in the case of no RS delay
experienced. With the effect of the asymmetric channel
taken into consideration, the estimated RTT length is
expected to increase. For this reason, much longer CP
timers (12-36 s) are adopted.
It is observed in Figure 4 that the goodput increases
with an increase of the CP timer length until it reaches 28
s. While the performance improvement is significant in
the region of short and medium CP timer lengths (<28 s),
there is no obvious performance change observed for
transmissions with a longer timer between 28 and 36 s.
This indicates that an optimal setting of the CP timer in
the given transmission condition should be around 28 s or
higher. It further means that a CP timer length of 28 s is
able to avoid the retransmissions of the CP segments for
each block which experiences the excessive delay for the
RS due to the channel-rate asymmetry.
Figure 5 illustrate the TSG for the LTP transmission
done with the CP timer length of 4 s presented in Figure 4.
Similar to the TSG for the transmission with the symmetric channel rates in Figure 1, both the data blocks and RSs
are transmitted in a continuous manner. However, unlike
the transmissions with the symmetric channel rates, the
slope rate of the RS transmission line is obviously much
lower. This is because the transmission of the RS segments for all the LTP blocks (except the first one) experiences delay. The delay is due to the longer transmission
time of an RS for each block caused by the significantly
reduced ACK channel rate. This delay is accumulated
along with the transmission of the blocks. For this reason,
the delay experienced by the RS segments grows linearly.
Therefore, how long the resulting delay is for an individual block depends on the total number of blocks generated

Figure 5.
TSG illustrating the LTP transmission at packet level at the sending node for delivery of a file of 1 Mbyte over a cislunar communication
channel with highly asymmetric channel rates (with the CR of 500/1) and a short CP timer of 4 s.

MAY 2020

IEEE A&E SYSTEMS MAGAZINE

IEEE - Aerospace and Electronic Systems - May 2020

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - May 2020