Page 71 - ITU Journal Future and evolving technologies Volume 2 (2021), Issue 7 – Terahertz communications

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 7

Average of packets reception (%) 80 Flow 2 neighbours nodes ensures an average of 100% of packet
low1‑nopatch (without the algorithm). A value of 40
100
Flow 1
reception by the destination node. Indeed, the positive
impact is also clearly shown for low 2‑patch at the 130
of awake neighbours, comparing to the low 2‑nopatch,
60
where the destination node missed the packet because it
40
was asleep. The proposed algorithm shows its effect in
increasing node chances in receiving the packet if it was
20
asleep at the moment the packet reaches its zone (Fig. 17).
0
40
60
0
20
120
140
80
100
Average of awake neighbours 160 180 200 100
Fig. 14 – Percentage of arrived packets depending on neighbours 80
nodes average, for 2 lows.
for low 1 and 182 for low 2. We notice that with 60 Average of packets reception (%) 60
neighbours’ nodes (where all packets reach their 40
destination), the total number of packets sent for low 1
is 205 and 155 for low 2. 20 Flow 1-nopatch
Flow 2-nopatch
250 Flow 1-patch
Flow 2-patch
Flow 1 0 0 20 40 Average of awake neighbours 160 180 200
Flow 2
Number of sent packets 150 Fig. 16 – The impact of the retransmission algorithm.
60
80
140
120
100
200
100
50
0
0 20 40 60 80 100 120 140 160 180 200
Number of awake neighbours
Fig. 15 – Packet transmission cost depending on the average of neigh‑
bours’ nodes.

5.5 Processing at the destination zone
As explained before, the proposed sleeping mechanism
reduces the congestion problem and preserves node re‑
source consumption along the path from the source to the Fig. 17 – The retransmission algorithm applied at the destination zone
destination. But the very de inition of destination may increases the number of exchanged packets in that zone.
change depending on the application. If the destination
is de ined as an SLR address, it means that we want the
packet to reach this SLR zone and that at least one node 5.6 Average density vs awaken percentage
in this zone must receive this packet. In that case, the The node awaken interval, considered the main difference
mechanism we proposed is ef icient and can be used as between the two ways of applying a sleeping mechanism.
is. On the other hand, if we aim to reach a speci ic node at The DEDeN algorithm allows a node to estimate the
the destination zone, then more aspects have to be taken number of its neighbours. Based on this value, nodes
into consideration. When a packet arrives at the destina‑ will set their awaken interval (awaken duration) (Fig.
tion zone, the destination node by chance may be asleep 18). However, ixing an awaken percentage (e.g. 80%),
and consequently misses the packet. Different strategies means that all the nodes in the network will be awake
might be used to solve this problem, depending on the for the same duration (e.g. 80 000 fs).
node peculiarities, the local density and the application
requirement. The algorithm is well explained in Section 4. The longer the node sleeps, the lower the consumed re‑
Fig. 16 shows the impact of applying our proposed re‑ sources. The difference in nodes awaken duration is
transmission algorithm to all the nodes at the destina‑ considered a special peculiarity of the average density
tion zone. The algorithm clearly enhances the average (awakenNodes). An average density of 60 neighbours
of packet reception with the low 1‑patch compared to shows an awaken duration distribution (less resource con‑

© International Telecommunication Union, 2021 59