1

Simulation of RPL in IoT Network using Cooja

Simulator
Harisha K S1 and Dr Parameshachari B D2
1
PhD scholor NMIT bengaluru
2
Professor Department of Electronics and communication,NMIT Bengaluru

Abstract—The IoT, or Internet of Things, is a relatively new Objective functions, metrics, and restrictions are used by
development in the electronics and communication industries. IoT RPL to determine the optimal action to do. ETX (Expected
devices collect data through various types of sensors and send transmission count), AVG DEL (Average delay), along with
that to servers present in Cloud. IoT has made physical things
such as home appliances ,lights, fans to be connected to internet PDR (Packet delivery ratio) [3] are often used metrics for
and monitor ,control them from remote places. When many IoT assessing RPL. IoT devices Contiki OS which is lightweight.
devices are connected to internet this forms IoT network. IoT Contiki OS is written in C programming language ,hence
devices collect data from the Physical environment using sensors it is programmer friendly and fast and hardware friendly.
and forward that to the cloud. As a result, IoT networks rely .Contiki OS various Internet protocols stacks such as IPv4
heavily on the successful transmission of sensor data to the cloud.
A routing protocol is needed for this purpose. One kind of routing ,IPv6.6LoWPAN,RPL,CoAP,Telenet.Contiki OS supports mul-
protocol used in IoT networks is the Routing Protocol for Low titasking, multithreading ,GUI,networking TCP/IP suite for
Power Lossy Networks (RPL). Storage and processing capacity communication. This OS is suitable for constrained devices
in IoT devices are limited and the connections between them because it consumes less operating power ,less memory[12].
in IoT networks are often unstable. In this study, we evaluate Cooja is the network simulator supported in Contiki OS. This
RPL’s efficacy. Throughput, Latency, Network convergence time,
Packet delivery ratio are just few of the performance metrics simulator supports Graphical user interface type simulation.
that are calculated and examined. Different configurations of the Cooja simulator was implemented using c and Java. Cooja
simulated IoT networks are compared. network simulator supports different types of motes such as
Keywords: IoT (Internet of things), RPL (Routing protocol for sky, zmote. In collect view we can node information, operating
low power lossy networks), LLN Low (power lossy networks), power, number of packets received, latency ,Expected trans-
latency.
mission count(ETX). Cooja simulator is very useful tool for
analysis of IoT networks. Cooja simulator data can be saved
I. I NTRODUCTION in pcap file and analyzed using Wireshark network analyzer

I IoT is a system of interconnected devices. In an IoT large

number of Objects such as Lights, Fans, Airconditioners
,Microwave ovens, can be connected to Internet. Simplified
.Wireshark network analyzer data can be used to calculate
packet delivery ration(PDR),average latency, throughput[12]
Here, we quantify and analyse QoS metrics based on our
IoT functional stack consists of three layers: Communica- simulation of an RPL network. Section two explains the RPL
tion network layer, Things layer, along with Application and protocol. In the third part, we showcase a variety of works
analytics layer[1]. Things layer consists of sensors which from the past. The quality-of-service metrics to be used are
measure physical parameters from IoT environment for ex- detailed in Section 4. In the fifth part, we show the simulation
ample Temperature sensors. Communication network layer details and the outcomes. This section includes a comparison
transfers information collected from the things layer to its of results. In the last part, we provide a summary and outline
higher layers. The communication network layer uses various the paper’s potential future directions.
networks such as PAN,LAN,MAN,WAN etc. Communication
technologies used for data transmission includes Bluetooth, II. ROUTING PROTOCOL FOR L OW POWER LOSSY
Zigbee, WiFi,RFID [1].Communication network layer uses NETWORKS (RPL)
various protocols such as RTP,RPL,RTCP. Application and For LLNs when individual nodes have constrained resources
analytics layers provide user interface to access IoT network. (such as power, memory, and computing speed), RPL is used.
Major protocols in this layer are COAP and MQTT. These are RPL is used to construct topology in LLN. In LLNs nodes
reques response based protocols. These provide user access have mobility, which requires dynamic topology formation
to physical things[2]. When it comes to the routing protocol to provide uninterrupted service. The topology constructed
used by IoT networks, RPL is the standard. To create a in LLN is called DODAG.Topology contains many senders
directed acyclic graph (a kind of tree topology), RPL is a and one sink nodes..RPL uses distance vector algorithms
proactive protocol (DAG). In this arrangement, each terminal to construct route. DODAG is a Distance vector protocol
node decides on a parent node to serve as its gateway. where distance and direction is calculated to any node in
RPL is based on 6LoWPAN technology and operates on that network.Here next hop and cost to reach destination is com-
concept. The RPL protocol is used to route IPv6 packets. puted.A Low cost route is selected to transfer data to minimize
 2

III. L ITERATURE SURVEY

In their research paper S. Kalyani[4] described RPL and
simulated RPL protocol using Cooja network simulator. Her
simulation involves QoS parameters such as Packet delivery
ration(PDR),Latency, Throughput. Author repeated her simu-
lation for 10nodes,20 nodes and 30 nodes.The research has
Fig. 1. Networ Topology in LLNs shown PDR increases with number of nodes with increase in
latency. This paper has not proposed any QoS improvement
technique. Charles[5] has published a paper in which he
describes his experience measuring RPL’s Quality of Service
using tools like Cooja’s network simulator and Wireshark’s
network analyzer. Expected transmission cout (ETX), latency,
throughput, energy consumption, duty cycling, network con-
vergence time, and packet delivery ratio are only few of the
Quality of Service (QoS) metrics the author examined in depth
in his study. They performed packet analysis using the wire-
shark programme. Haofei Xie[6] has presented performance
evalution of reactive protocols such as AODV,DYMO and
RPL.They compared above said protocols based on average
delay.RPL outperformed against other two.M. S. Aman[7] in
his work simulated IoT network using NS2 and Cooja.He
Fig. 2. Directed Acyclic graph stated that the Cooja network simulator is user friendly due
to its GUI and simulation of WSN motes such as Sky mote
Z motes can do.In this paper study energy consumption of
source and sink is done. Results of this paper states that
delay.This increases Qos of network.RPL organizes a topology energy of source nodes increases with distance.W. Khallef
called DAG.DAG route has single destination with no outgoing [8] evaluated performance of Greedy ,k-limited and exact
edges.DAGs are further separated based on destination called algorithms to evalute performance of RPL. Here missing
as DODAGS.Network topology in LLN is as shown in Fig- nodes, Execution time, and quality of paths are considered
ure1.DAG and DODAG is shown in Figure2 and 3. RPL uses as parameters. Nguyen Thanh Long[9] in his paper presented
following identifiers a comparison of RPL with the collection tree protocol. The
1. RPL InstanceID: An Identifier used with in network to results of this paper showed that RPL has more packet
group all DODAGS which use same Objective function. reception ratio with less energy consumption. S. Solapure[10]
2. DODAGID: An Identifier used to indicate DODAG route in their paper we presented performance evalution of IoT
within instance.Instance ID and DODAH ID represents root networks with different Objective function, with respect to
from leaf to parent PDR (Packet delivery ratio), Latency, Energy consumption and
3. Rank: It is an integer which indicates position of node from network life time.X. Liu[11] in their paper has explained RPL
root. protocol, compared different simulation tools for Contiki OS
4. DIO message: This message carries RPL instance and and simulated RPL using OMNET++ tool with 500 nodes.
configuration information to node. In his article, A. R. Jadhao[12] used the cooja simulator to
5. DAO (Destination Advertisement Object):This message is compare the efficiency of RPL to that of OF0 and MRHOF
sent from sender to sink to send destination information. The (Minimum Rank Hysteresis Objective Function). He came
parent or root may acknowledge the child. to the conclusion that the MRHOF was more effective than
OF0.O. Gaddour[13] has simulated RPL and studied DAG
construction process. They concluded that DAG route should
be in the middle of surface which reduces hop count and node
should have alternate parent to avoid packet loss. P. Janani[14]
has simulated RPL using cooja simulator and analysed the
performance using hoplength and link quality. Link quality-
based RPL throughput is increased,ened to end packet latency
is reduced.

IV. RPL PERFORMANCE ANALYSIS PARAMETERS

Quality of service refers to how well a network performs
and serves its users (QoS). Quality of service is a metric that
may be used to evaluate the efficacy of a network and its
Fig. 3. Destination oriented DAG routing system. For RPL, we measure QoS using ETX, latency,
 3

throughput, power consumption, PDR and convergence time.

QoS parameters are useful for assisting with the design and
implementation of a network’s infrastructure[15].
i) ETX (Expected transmission count): ETX is the total
number of packets that a node is expected to send. Equation 1
describes ETX. The chance that a given neighbour really gets
a packet is denoted by the symbol Df. Dr measures how likely
it is that a node will be acknowledged.
ET X = 1/Df × Dr (1)

ii)Latency: The delay between when a packet is delivered and

when a packet is received, is called as latency. Total latency is
the sum of the latencies of each individual packet. The average
delay is found by dividing the total latency by the number
of packets that were successfully delivered to the destination
(equation 2, 3, and 4[5]).
Latency = Packet received time - packet sent time (2)

X
T otalLatency = (P acketreceivedtime−packetsenttime) Fig. 4. Sample Simulation network
(3)

AverageLatency = T otalLatency/T otalReceivedpackets

(4)

iii)Packet delivery ratio(PDR):PDR measures how many pack-

ets a node has received in relation to how many packets
have been supplied to it. This value assures the network’s
dependability. If the number is high, you may trust it. To
calculate PDR, use eq. 5[5].
Fig. 5. Collect view available in Cooja
P DR = T otalpacketsReceived×100/T otalpacketssenttothatnode
(5)
in cooja Figure V Figure V In this simulation, we have taken
iv)Throughput: The throughput of a simulation is the number 1 UDP sink node and 10 UDP sender nodes. The default
of successfully sent packets divided by the time spent in the objective function OF0 is used in the illustrative scenario. The
simulation. Equation 6[4] describes the throughput. TX and RX efficacy rates are both fixed at 100
A. Calculation of Throughput
T hroughtput = (N umberof succesf ullpackets∗packetsize∗8)/simulationtime
To calculate Throughput we have to open the pcap file and
(6) open summary under statistics tab. We get the values of total
simulation time,Total number of packets transmitted ,number
v) Network convergence time(NCT): NCT is the time taken of bytes per second. Using this information we can calculate
to setup network by RPL.NCT is the time difference between throughput of network.
first control message and last control message from root node B. Calculation of Latency To calculate Latency we have to
to leaf nodes. NCT is given by equation 7[5]. open pcap file. The delay between when a transmission is sent
from one node and when it is received at the sink is known
N CT = LastDIOjoinedtime − F irstDIOsenttime (7)
as its latency.
C. Calculation of ETX ETX Expected transmission count
V. RESULT AND ANALYSIS which can is available in Node information tab in collect view.
In this paper we used Cooja simulator for simulation of D. Calculation of PDR Using the pcap file, the network
RPL LLN. Cooja is a GUI based network simulator available analyzer Wire shark can determine the packet delivery ratio
in Contiki 2.7 OS. In this simulator we can create motes (PDR). We need to keep track of how many packets come
with desired protocol. This simulator supports measurement from a given source and how many arrive at the sink from
of various parameters in collect view.Figure4 shows sample that source. PDR is the percentage of sent data that was
simulation network. The overall perspective of the cooja successfully received by the sink.
simulator looks as in Figure5. Figure 6 shows the collect view E. Calculation of Convergence Time The wireshark network
 4

TABLE I [11] X. Liu, Z. Sheng, C. Yin, F. Ali and D. Roggen, ”Performance Analysis
Q O S PARAMETERS M EASURED of Routing Protocol for Low Power and Lossy Networks (RPL) in Large
Scale Networks,” in IEEE Internet of Things Journal, vol. 4, no. 6, pp.
Sl no Quality of parameter/nodes 10 Nodes 20 Nodes 2172-2185, Dec.2017. Doi: 10.1109/JIOT.2017.2755980
1 PDR(%) 55 85 [12] A. R. Jadhao and S. S. Solapure, ”Analysis of routing protocol for
2 ETX 16 35 Low Power and Lossy Networks (RPL) using Cooja simulator,” 2017
3 Latency(nsec) 20.178 100.6 International Conference on Wireless Communications, Signal Processing
4 Network Convergence time(msec) 12.485 26.5 and Networking (WiSPNET), Chennai, India, 2017, pp. 2364-2368, doi:
5 Contro traffic 350 850 10.1109/WiSPNET.2017.8300183.
6 Packet Transmission Rate(bytes/sec) 883 472 [13] O. Gaddour, A. Koubâa, S. Chaudhry, M. Tezeghdanti, R. Chaari and
7 Throughtput(%) 35 87 M. Abid, ”Simulation and performance evaluation of DAG construction
with RPL,” Third International Conference on Communications and
Networking, Hammamet, Tunisia, 2012, pp. 1-8, doi: 10.1109/Com-
Net.2012.6217747.
analyzer reads the pcap file. Convergence time is calculated by [14] P. Janani, V. C. Diniesh and M. J. Auxilius Jude, ”Impact of Path Metrics
subtracting the timings of the first and final DIO messages. As on RPL’s Performance in Low Power and Lossy Networks,” 2018 Inter-
national Conference on Communication and Signal Processing (ICCSP),
can be observed in Table 1, because the network we simulated Chennai, India, 2018, pp. 0835-0839, doi: 10.1109/ICCSP.2018.8524141.
was lossless, the PDR and ETX remained unchanged through- [15] M. Asif, S. Khan, R. Ahmad, M. Sohail and D. Singh, ”Quality of
out our simulations. Throughput decrease and Convergence Service of Routing Protocols in Wireless Sensor Networks: A Review,”
in IEEE Access, vol. 5, pp. 1846-1871, 2017
time increase with increased nodes. Latency increases.

VI. CONCLUSION A ND FUTURE SCOPE

In this study, we have modeled the operation of a Sky Motte-
based IoT network using the RPL protocol. Throughput, net-
work convergence time, packet delivery ratio, and other QoS
metrics are all assessed with varying numbers of simulated
nodes. It has been discovered that as the number of nodes
grows, so do the PDR, the latency, the time it takes for the
network to converge, and the throughput it can achieve. The
goal for the future is to create efficient algorithms that can
decrease latency even as the number of nodes increases. Indi-
cators of interest include power consumption and portability.

R EFERENCES
[1] K. Ashton, That “Internet of Things”, RFiD Journal (2009)
[2] Luigi Atzori Antonio Iera, Giacomo Morabito, The Internet of Things:
A survey, Computer Networks, Volume 54, Issue 15, 2010,Pages 2787-
2805,ISSN 1389-1286, https://doi.org/10.1016/j.comnet.2010.05.010.
[3] A. Nauman, Y. A. Qadri, M. Amjad, Y. B. Zikria, M. K. Afzal and
S. W. Kim, ”Multimedia Internet of Things: A Comprehensive Sur-
vey,” in IEEE Access, vol. 8, pp. 8202-8250, 2020, doi: 10.1109/AC-
CESS.2020.2964280.
[4] S. Kalyani and D. Vydeki, ”Measurement and Analysis of QoS Pa-
rameters in RPL Network,” 2018 Tenth International Conference on
Advanced Computing (ICoAC), Chennai, India, 2018, pp. 307-312, doi:
10.1109/ICoAC44903.2018.8939052.
[5] Charles, A.s. Joseph and Palanisamy, Kalavathi. (2018). QoS Measure-
ment of RPL using Cooja Simulator and Wireshark Network Analyser.
International Journal of Computer Sciences and Engineering. 06. 283-291.
[6] H. Xie, G. Zhang, D. Su, P. Wang and F. Zeng, ”Performance evaluation
of RPL routing protocol in 6lowpan,” 2014 IEEE 5th International
Conference on Software Engineering and Service Science, Beijing, 2014,
pp. 625-628. Doi: 10.1109/ICSESS.2014.6933646.
[7] M. S. Aman, K. Yelamarthi and A. Abdelgawad, ”A comparative analysis
of simulation and experimental results on RPL performance,” 2017 IEEE
8th Annual Ubiquitous Computing, Electronics and Mobile Communica-
tion Conference (UEMCON), New York City, NY, 2017, pp. 483-487.
Doi: 10.1109/UEMCON.2017.8248996.
[8] W. Khallef, M. Molnar, A. Bensliman and S. Durand, ”On the QoS routing
with RPL,” 2017 International Conference on Performance Evaluation
and Modeling in Wired and Wireless Networks (PEMWN), Paris, France,
2017, pp. 1-5, doi: 10.23919/PEMWN.2017.8308028.
[9] Nguyen Thanh Long, N. De Caro, W. Colitti, A. Touhafi and K. Steenhaut,
”Comparative performance study of RPL in Wireless Sensor Networks,”
2012 19th IEEE Symposium on Communications and Vehicular
[10] S. Solapure, H. Kenchannavar and U. P. Kulkarni, ”Analysis of Var-
ious RPL Protocol Objective Functions for Quality of Service Param-
eters,” 2022 IEEE International Conference on Blockchain and Dis-
tributed Systems Security (ICBDS), Pune, India, 2022,pp. 1-5, doi:
10.1109/ICBDS53701.2022.9935956