Professional Documents
Culture Documents
Asri Samsiar, Ihsan Ibrahim, Arief Luthfi Aulia, Riri Fitri Sari
Department of Electrical Engineering, Faculty of Engineering
Universitas Indonesia
Depok 16424, Indonesia
asri.samsiar@ui.ac.id
Abstract— The Internet of Things (IoT) is connecting tons of interconnecting devices. As an alternative, Information Centric
devices connected to internet, with high dynamicity and Networking (ICN) is proposed to redesign the internet by
heterogeneity devices. IoT presents significant challenges from the replacing its host-centric model with data-centric model [9],
wireless communications perspective besides the limitation in IP [10], [11].
resource. It calls for new connection architecture based on data-
centric approaches, namely Named Data Networking (NDN). Unlike the IP-based systems, communication in ICN takes
Besides the Wi-Fi connection that widely used in IoT, NDN is place between consumer (as a user) and producer (as a publisher)
challenged by the opportunity to re-architect the latest mobile using two simple types of messages, namely Interest and Data.
network, that is Long-Term Evolution (LTE), through named- The content identification is provided by a unique name, rather
based communication. LTE could be a solution of limitation on than an IP address which is limited resource. Every node is
Wi-Fi coverage area. In this paper, the experiment is conducted to assigned to what-is-the-data (which content is expected), and not
compare the performance of NDN architecture over Wi-Fi and where-is-the-destination (which define the addresses and hosts).
LTE, in order to show the possibility of LTE running over NDN. One of the most promising information-centric based is Named
NdnSIM tool has been used to simulate the scenarios on IoT Data Networking (NDN). NDN is emerging as a potential
devices. Due to its simulation performed in a small-scale network solution to future Internet architectures. NDN can potentially
for the simulation scenario, the result shows that the collection address the limitations of the existing IP networks through
time between Wi-Fi and LTE over NDN, with varied number of
simple communication, which emphasizes on the content itself
producers, is the same. However, LTE has an advantage in further
coverage distance that Wi-Fi connection cannot achieved or
rather than its container (i.e., host) or channel (i.e., connection).
performed.
By including the content name in each packet, NDN enables
features such name-based routing, data-centric security,
Keywords— Internet of Things, LTE, Named Data Network, Wi- multipath forwarding and in-network caching [1], [4].
Fi, Wireless Networks. Furthermore, because it is natively support multicasting and
mobility, NDN is a promising solution to suit the requirements
I. INTRODUCTION of the IoT environment.
The main concept of Internet of Things (IoT) is connecting In addition, NDN offers and provides an opportunity to shift
“things” or everything from all of the world with internet, which the traditional mobile network architecture using simple and
is appeared as tons of heterogeneous devices. These massive scalable architecture. By implementing NDN protocol, wireless
number of devices have different characteristics from the current communication obtains benefits through optimized design and
internet devices, due to constrains of devices resources, efficiency of content delivery. Although Wi-Fi is fully
requirement of support for mobility and traffic type diversity [1], optimized by NDN [12], there is still a limitation that Wi-Fi has,
[2], [3]. The host-to-host communication among IoT devices is that is the coverage area. This limitation can be an obstacle when
challenging, due to its high volume of small exchanged data [4]. the IoT devices are placed on out of range area, so that it cannot
There is a need for advanced network technology especially in communicate between each other device.
wireless environment for mobile devices.
Moreover, the NDN integration with the latest mobile
In the last decade, the evolution of advanced mobile devices, network technology, i.e. LTE, that built as all-IP network, is still
such as smartphone, tablet, etc. has caused growth of IoT traffic, a challenge. Nonetheless, LTE was designed to establish high-
in both the amount of traffic and the user demand [5], [6]. To speed broadband data service with very low latency. Besides
meet the requirement, the standardization of mobile network that, numerous LTE features are provided to address the
such as Wi-Fi and LTE has been improved continuously. requirement of IoT, such as power saving, overload control,
Recently, many researchers have proposed and evaluated the signaling reduction, complexity reduction, and coverage
performance of IP-based solution with LTE enhancement [6], enhancement [7]. With those abilities, LTE can be a solution of
[7] and the coexistence of LTE and Wi-Fi [5], [8]. The limitation on Wi-Fi coverage area.
communication over IP is complex on a large-scale IoT
environment, because of the addressing scheme, IP subnet, even In this paper, the research objective is to study the
routing security and protocol through all the heterogeneous and performance of NDN architecture over LTE mobile network,
The goals for the IEEE P2413 working group that developed Communication in NDN is started by receivers through the
this standard are follows: exchange of two types of packets: Interest and Data. The content
• To accelerate the growth of the IoT market by enabling identification is provided by a unique name. First, the consumer
cross-domain interaction and platform unification side will request the needed data into an Interest packet and
through increased system compatibility, interoperability sends it to the network with a unique name of Data. After that,
and functional exchangeability. the NDN router will send this interest packet to all producer for
finding the needed data. When there is a producer that has the
• To define an IoT architecture framework that covers the needed data, exactly the same with the name of the interest
architectural needs of the various IoT application packet, producer will reply a data packet that contains both the
domains. name and the content, together with a signature by the producer's
• To increase the transparency of system architectures to key which binds the two which can be seen in Figure 2. This
support system benchmarking, safety and security This Data packet travels in reverse direction following the same
assessments. path taken by the Interest packet [15].
2
requirements [6], [11]. Advanced mobile devices, which play a
major role as content demanders, continuously call for higher
bandwidth demand. There are three major technologies for
HSS
mobile wireless networks, i.e. short-range communications eNodeB MME
LTE S1-U
The mobile technologies are originally created to provide UE
Serving PDN IP
wireless data communication among telephone devices. Wi-Fi Gateway Gateway Networks
(S-GW) (P-GW)
was designed for small coverage wireless services such as home, eNodeB
office, classroom etc. To get connected to the Wi-Fi network, all E-UTRAN Enhanced Packet Core (EPC)
3
Furthermore, we use number of producers as the evaluation
Start parameter to compare the network performance, which are
shown in Table 1.
Receive a Interest
message
TABLE 1. PARAMETER SETTINGS FOR DIFFERENT SCENARIOS
no
In the Scenario I, NDN will be combined with Wi-Fi mobile
yes
Exist in FIB? Add a new PIT entry network, with the number of producers varied from 4, 8, 6 and
12 producers. In the Scenario II, the mobile network that is used
no
Send the Interest through over NDN is LTE. For LTE scenario, we used LENA, which is
the outgoing face
the ns-3 module for LTE. Both scenarios used random direction
End for the mobility model, and set the distance between each node
to 10m apart, and the other parameters are set to default.
Fig. 4. Procedures for Internet Packet Processing The results are collected based on the simulation’s collection
time that varies by different parameters. The collection time is
III. PERFORMANCE COMPARISON SCENARIO defined as the time required in a simulation scenario, from the
Interest transmission by the consumer until the expected data is
The IoT environment can be described as Wireless Sensor
retrieved.
Network (WSN), where a lot of sensors are connected via
mobile network and collect the data provided by user. In our IV. RESULT AND DISCUSSION
simulation, the sensor devices are shown as producer, which
send the Data. The user as a consumer, send the Interest and The simulation of our research and experiment about the
retrieve the Data. We assumed that the sensor devices have the performance of LTE and Wi-Fi on IoT application has been
same characteristics, i.e. the ability to move (mobility) and the conducted. This simulation represented the scenarios of our idea
power limitation (energy consumption and harvesting). that applied Wi-Fi and LTE on IoT environment separately.
Figure 6 shows our simulation visualization result. The
In our research we conduct experiments with two scenarios simulation results are shown in Table 2 with the comparison of
that will be representation of our work in comparing the collection time between IoT over LTE and Wi-Fi. We varied the
performance between Wi-Fi and LTE environments. First is the number of producers from 4, 8, 12, to 16, and used Random
Wi-Fi-based scenarios, where the consumer communicates with Direction mobility model.
the producer via Wi-Fi through a Wi-Fi Access Point. The other
scenario is LTE-based scenarios, which used the LTE eNodeB
for the interconnection. The simulation scenario is shown in
Figure 5.
LTE eNodeB
Wi-Fi AP
CONSUMER
PRODUCER
Fig. 6. The simulation result of LTE scenario in IoT implementation with NDN
using ndnSIM
4
TABLE-2. THE COLLECTION TIME OF LTE AND WI-FI WITH RANDOM over NDN architecture can be used for the implementation of
DIRECTION MOBILITY MODEL FOR IOT OVER NDN
IoT with usage in broad area such forest etc. It was also make
Number of Wi-Fi LTE faster the transmission for the producers to fulfil the consumer
Producers (seconds) (seconds) interests. In our case, the simulation result shows that both Wi-
4 7 7 Fi and LTE scenarios are nearly had the same value. It happened
8 7 7 because we use small scale network scenario. The result between
12 7 7 Wi-Fi and LTE will be differed significantly if it used large scale
16 7 7 network scenario, especially in long distance as our additional
experiment. For the future work, more complex network
As seen in Table 2, the collection time between IoT topology should be conducted to get better understanding about
application over Wi-Fi and LTE are the same i.e. 7 seconds. This the performance comparison between Wi-Fi and LTE over
result is also depicted in Figure 7 that the second line on every NDN.
producer is LTE and the first line is Wi-Fi. These values are
retrieved from the rate trace of each scenario for every node and REFERENCES
its interest and data transactions. The results are retrieved from [1] M. Amadeo, C. Campolo, A. Iera, and A. Molinaro, “Named Data
the interest time until the interest is not satisfied again by the Networking for IoT : an Architectural Perspective,” 2014 Eur. Conf.
producers. These results had no effect when we vary the number Networks Commun., p. 1–5, Bologna, 2014.
of producers. [2] S. K. Datta and C. Bonnet, “Integrating Named Data Networking in
Internet of Things architecture,” 2016 IEEE Int. Conf. Consum. Electron.
There is another experiment that conducted to verify the LTE ICCE-TW 2016, p. 1–2, Nantou, 2016.
ability and the Wi-Fi incapability about coverage. In this [3] W. Shang et al., “Named Data Networking of Things (Invited Paper),”
experiment, the distance between each node had been set to 300 2016 IEEE First Int. Conf. Internet-of-Things Des. Implement., p. 117–
m. With that distance value, the Wi-Fi scenario couldn’t get the 128, Berlin, 2016.
result because it couldn’t connect to the other nodes in the [4] M. Amadeo, C. Campolo, and A. Molinaro, “Multi-source Data Retrieval
in IoT via Named Data Networking,” Proc. 1st Int. Conf. Information-
scenario. Even it setup in adhoc connection, the distance centric Netw., p. 67–76, Paris, 2014.
prohibited a node to reach the other and build a connection. [5] Z. Zhou, F. Teng, J. Liu, and W. Xiao, “Performance evaluation for
In our hypothesis, the collection time will be greater if the coexistence of LTE and Wi-Fi,” 2016 Int. Conf. Comput. Netw. Commun.
ICNC 2016, p. 0–5, Kauai, HI, 2016.
number of producers increased. The result of LTE scenario can
[6] J. Liu, N. Kato, J. Ma, and N. Kadowaki, “Device-to-Device
be faster than Wi-Fi because of its capabilities on highspeed Communication in LTE-Advanced Networks: A Survey,” IEEE
transmission and coverage if the distance between nodes are Commun. Surv. Tutorials, vol. 17, no. 4, pp. 1923–1940, 2015.
further apart, outside of average area of Wi-Fi coverage. For the [7] R. Ratasuk, N. Mangalvedhe, and A. Ghosh, “Overview of LTE
best scenario, LTE scenario will have the advantages for IoT enhancements for cellular IoT,” IEEE Int. Symp. Pers. Indoor Mob. Radio
implementation, especially for broad coverage area that Wi-Fi Commun. PIMRC, vol. 2015–Decem, p. 2293–2297, Hong Kong, 2015.
could not supported. Based on the analysis of the simulation, this [8] S. A. A. Termizi, S. Rahmat, and M. A. Qadeer, “LTE-U over Wi-Fi,”
system experienced an anomaly which are might be caused by IFIP Int. Conf. Wirel. Opt. Commun. Networks, WOCN, p. 1–7,
Hyderabad, 2016.
our simulation scenarios are too small to represent the real
environment and condition. [9] P. Suthar and M. Stolic, “Deploying Information Centric Networking in
LTE Mobile Networks,” IEEE Asia Pacific Conf. Wirel. Mob., p. 130–
136, Bandung, 2016.
[10] A. Morelli, M. Tortonesi, C. Stefanelli, and N. Suri, “Information-Centric
Wi-Fi LTE Networking in Next-generation Communications Scenarios,” J. Netw.
8 Comput. Appl., pp. 232–250, 2016.
7 [11] A. Gomes and T. Braun, “Load balancing in LTE mobile networks with
Information-Centric Networking,” 2015 IEEE Int. Conf. Commun. Work.
6 ICCW 2015, p. 1845–1851, London, 2015.
COLLETION TIME
Fig. 7. Comparison of Wi-Fi and LTE with Random Direction mobility model