Content uploaded by Maram Abdullah Almutairi
Author content
All content in this area was uploaded by Maram Abdullah Almutairi on Jul 31, 2022
Content may be subject to copyright.
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
212
Manuscript received March 5, 2021
Manuscript revised March 20, 2021
https://doi.org/10.22937/IJCSNS.2021.21.3.29
On the Need for Efficient Load Balancing in Large-scale RPL
Networks with Multi-Sink Topologies
Maram Abdullah1†, Ibrahim Alsukayti2††, Mohammed Alreshoodi3†††
1†2††Department of Computer Science, College of Computer, Qassim University, Buraydah, Saudi Arabia
3††† Department of Applied Science, Unizah Communit
y
Colle
g
e, Qassim Universit
y
, Bura
y
dah, Saudi Arabia
Abstract
Low-power and Lossy Networks (LLNs) have become the
common network infrastructure for a wide scope of Internet of
Things (IoT) applications. For efficient routing in LLNs, IETF
provides a standard solution, namely the IPv6 Routing Protocol for
LLNs (RPL). It enables effective interconnectivity with IP
networks and flexibly can meet the different application
requirements of IoT deployments. However, it still suffers from
different open issues, particularly in large-scale setups. These
include the node unreachability problem which leads to increasing
routing losses at RPL sink nodes. It is a result of the event of
memory overflow at LLNs devices due to their limited hardware
capabilities. Although this can be alleviated by the establishment
of multi-sink topologies, RPL still lacks the support for effective
load balancing among multiple sinks. In this paper, we address the
need for an efficient multi-sink load balancing solution to enhance
the performance of PRL in large-scale scenarios and alleviate the
node unreachability problem. We propose a new RPL objective
function, Multi-Sink Load Balancing Objective Function
(MSLBOF), and introduce the Memory Utilization metrics.
MSLBOF enables each RPL node to perform optimal sink
selection in a way that insure better memory utilization and
effective load balancing. Evaluation results demonstrate the
efficiency of MSLBOF in decreasing packet loss and enhancing
network stability, compared to MRHOF in standard RPL.
Keywords:
Low-power lossy network (LLN), Wireless Sensor Network (WSN),
RPL, Load Balancing, Multi-Sink
1. Introduction
Low-power and Lossy Networks (LLNs) consist of
resource-constrained devices with limited processing power,
memory, and energy resources. Due to the lossy nature of
radio links, the constrained devices would be
interconnected by unreliable and unstable links, with
limited bandwidth and short communication ranges. These
restrictions are challenges to developing efficient routing
solutions for LLNs. The Internet Engineering Task Force
(IETF), specifically the Routing Over Low power and
Lossy networks (ROLL) working group, has standardized
the IPv6 Routing Protocol for LLNs (RPL). It is designed
to meet the requirements of many LLN applications in a
customized and flexible manner.
RPL [1] is a distance vector routing protocol that
allows bi-directional IPv6 communication on embedded
networking devices. It supports three data traffic flows,
peer-to-peer (P2P), multipoint-to-point (MP2P), and point-
to-multipoint (P2MP) traffic. Moreover, topology in RPL is
organized as a Directed Acyclic Graph (DAG) which can
be divided into one or more Destination Oriented DAGs
(DODAGs) rooted at a sink node. The DODAGs can be
associated with one or different RPL instances. Each
Instance is established using a certain Objective Function
(OF) which specifies how routes are constructed in
DODAG and how nodes translate the metrics and
constraints into a numeric value, referred to as rank. So far,
only two OFs have been supported by standard RPL, the
Minimum Rank with Hysteresis Objective Function
(MRHOF) and Objective Function Zero (OF0). Due to the
flexibility of the RPL OF concept, a new OF can be
designed if the existing OFs in standard RPL do not satisfy
the requirements of the application. Moreover, RPL
supports two modes of operation (MoP), a storing and non-
storing mode. In the former, all joining nodes in the RPL
network store the routing information in routing table
entries for each destination in their sub-DODAG; thus, this
mode of communication requires higher memory capacity.
In the non-storing mode, only the sink has routing access to
all the network destinations, and no routing tables need to
be maintained by the non-sink nodes.
RPL was originally designed with consideration of
memory limitations. However, memory overflow would
occur in dense topologies as a result of the limited hardware
capabilities of LLN devices. This leads to the problems of
node unreachability and routing failures. Even in the case
of having multi-sink setups, RPL fails to efficiently support
load balancing among multiple sinks. The absence of an
efficient mechanism of load-balancing among the sinks in
large-scale scenarios would degrade network stability and
reliability. Therefore, in this paper, we propose a novel RPL
objective function that addresses such problems. It relies on
a practical solution that incorporates existing concepts, such
as multi-sink load balancing, memory-overflow avoidance,
peer-to-peer communication, and DAG selection. Our
mechanism supports multi-sink RPL networks.
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
213
The remainder of this paper is structured as follows;
Section 2 discusses the related work of existing proposed
solutions for this problem; Section 3 describes the RPL load
balancing problem; Section 4 thoroughly describes the
proposal; in Section 5 we evaluate the performance; and,
finally, Section 6 presents the conclusion and proposals for
future work.
2. Related Work
Various enhancements have been achieved to the
standard RPL in recent years. The focus has been on
introducing several OF and routing metrics to advance its
basic functionalities. For instance, [2] proposes a load
balancing solution, called QU-RPL, by avoiding queue
overflows. It allows each node to use two metrics, queue
utilization and hop distance to the sink, during the parent
selection process. By evaluating it in a real testbed, the
authors have proven the practicality of their proposed
protocol over RPL. Moreover, the work in [3] extends the
RPL protocol to improve network lifetime and decrease
packet loss. This was based on considering the remaining
queue and energy level of candidate parents besides the
ETX metric in the parent selection mechanism. Similarly,
[4] proposed Quell, an alternative objective function which
relies on both the queue length and ETX metrics for the
selection of parent node.
An energy-balanced RPL was also introduced in [5],
which distributes fairly the energy consumption among all
the bottleneck nodes by supporting multipath structure.
[6]and [7] introduced enhancements to RPL reliability,
energy consumption, and load distribution. [6]combined
child node count and other primary metrics, such as ETX
and hop count in parent selection mechanism. [7] presented
ETXPC-RPL which was based on the ETX and parent
account metrics.
[8], [9], and [10] have each proposed an OF which
prevents overloaded nodes from being rapidly drained,
resulting in an increase in network lifetime. Authors in [8]
proposes an objective function for efficient routing (OF-ER)
to increase the lifetime of all the nodes. It was based on
avoiding congested paths by diverting network traffic
through the paths with less-congested nodes and fewer
energy-constrained nodes. On the other hand, the proposed
solution in [9] was based on restricting the selection of
overloaded nodes during parent selection and enabling
nodes to detach from overloaded parent nodes.
Other solutions were proposed based on sink
selection instead of parent selection mechanism in single-
sink load balancing. For instance, in [11], the sink-selection
process is based on tree size to facilitate tree balancing
across the network sinks. However, the greedy selection of
the least tree size may give rise to the overloading of
bottleneck nodes, since not all the nodes are equal in
hardware capabilities. On the other hand, [12] presented a
dynamic and distributed load balancing algorithm for
multiple sinks based on 6LoWPAN, named MLEq. It was
based on distributing the traffic of intersection-area nodes
across the sinks to reduce the traffic congestion and satisfy
the capacity fairness. Furthermore, [13] introduces different
OFs and metrics such as hop count, available bandwidth,
delay, buffer occupancy, and ETX; and analyzes the impact
of duty-cycling, the number of sinks, and the data traffic
load on RPL performance in multi-sink scenarios. The dual
objective of single-sink and multi-sink load balancing in
WSN is addressed in [14]. The Reactive and Adaptive
Load-Balancing (RALB) algorithm is proposed to combine
multiple path metrics as well as sink conditions to balance
traffic loads in large-scale scenarios.
It can be seen that most of the proposed OFs are
specific to the single-sink scenarios without sufficient
consideration of multi-sink load balancing. Furthermore,
the issue of memory overflow in large-scale scenarios has
not been effectively addressed in the context of multi-sink
topologies. It is evident that there is a need to address
efficient load balancing in conjunction with memory-
overflow avoidance. Thus, our proposed work fills this gap
and provides an effective OF that satisfies such a need while
conforming to the standard RPL.
3. Problem Statement
In standard RPL networks, in particular with the storing
mode, each node stores the reachability information in its
routing table entries for each RPL node in its sub-DODAG.
If any node has no free memory for a new destination, as a
result of the limited hardware capabilities of typical LLN, it
rejects the route advertisement (DAO message) and
becomes unable to forward traffic for that particular
destination. In the common RPL implementations, such as
in ContikiOS and TinyOS, the DAO message is silently
dropped. As a result, the path would be partially built, since
the destination remains unreachable to the sink.
Consequently, the sink is left with no choice but to drop
incoming packets destined to that destination.
To illustrate the impact of this problem in P2P traffic,
where the sink is neither source nor destination, we consider
the following scenario depicted in Fig. 1. Assuming that
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
214
node A is unable to act as a packet forwarder due to its lack
of memory, therefore node R becomes unreachable to node
A and to all higher nodes in the DODAG, including the sink
S. If node G sends a data packet destined to node R, the
packet is first transmitted to R’s preferred parent, then to the
common ancestor node A. While there is no routing entry
corresponding to node R in A’s routing table, node A
forwards the packet to its parent node. This process
continues until the data packet reaches sink S. Eventually,
the packet is dropped as the sink S fails to redirect it to the
destination node R
without a corresponding routing entry in
its routing table. Hence, all the packets sent to unreachable
nodes are dropped after they have been forwarded upward
to the common ancestor, which is, ultimately, the sink.
Figure 1: Unreachability problem example
Therefore, this unreachability problem leads to high
packet drops which adversely affects the protocol reliability.
This is more significant in the cases of large-scale
topologies where large sub-DODAGs can be segregated
and remain unreachable. In the simple example scenario
represented in Fig. 1, the three nodes constituting the sub-
DODAG of node R would also become out of reach to all
the other nodes.
In addition to memory overflow, there are other causes
of packet drops in RPL networks. These include link loss,
which happens if the packet exceeds the retransmission
limit; and, queue loss, which is caused by buffer overflows.
Load balancing is a common approach to address packet
loss resulted from limited link capacity and queue size.
Nevertheless, less attention was paid to routing loss by RPL
sink nodes due to the memory overflow problem.
Considering that it may happen earlier than link and queue
losses and would occur in even light traffic scenarios,
addressing routing loss is perceived more critical to
minimize packet loss. According to the results in [15], the
packet loss percentage caused by a routing loss in the
storing mode is extremely high, compared with link and
queue losses. More specifically, RPL with storing mode
suffers from high packet drops, routing failures, and
unreachable destinations due to having no memory space to
store the routing entries for all child nodes in nodes’ sub-
DODAG.
In standard RPL networks with multiple-sink
topologies, memory overflow is not recognized by OF0 and
MRHOF. With the existence of memory-
overflowed ancestor nodes and continuous occurrence of
routing loss, affected RPL nodes do not change their sinks.
To the best of our knowledge, neither the standard RPL OFs
nor existing proposed OFs take into consideration a sink
selection based on minimizing routing loss and avoiding the
unreachability problem. To overcome this problem,
adopting new routing metrics and applying an effective
DODAG selection strategy specified for large-scale
scenarios is recommended.
4. The Proposed Load Balancing Objective
Function in Multi-Sink
In this section, we present the proposed Multi-Sink
Load Balancing OF (MSLBOF) to address the load
balancing problem across multiple sinks. It is based on a
dynamic update of RPL according to the DAG status in
order to achieve effective load balancing across the sinks.
MSLBOF aims to extend MRHOF with other parameters
along with ETX in a way that helps RPL to distribute the
load more evenly across the sinks and avoid unreachability
problems. ETX usually offers a channel status estimation
that is related to link loss but may not reflect the actual
status of the network [2].
Furthermore, routing loss may happen earlier than link loss
in resource-constrained RPL networks with hundreds of
nodes having limited memory capacity. Therefore, we
combine node rank with memory utilization indicators that
take into consideration the DODAG status during the sink-
selection process.
4.1 Memory Utilization Metrics
In MSLBOF, Memory utilization is modeled by two
sink’s metrics, that provide composite indicators of the
DAG status. The first one is the DODAG size metric which
allows balancing the sink’s load by considering the number
of nodes for each sink. The DODAG size plays a key role
in the unreachability problem, as increasing DODAG size
leads to memory overload in high-level located nodes. This
results in some destinations being unreachable to the sink,
which triggers routing drops. MSLBOF allows each sink to
count the number of nodes in its DODAG and advertise it.
This information is then utilized by each node to achieve
optimal DAG size distribution across the sinks in the
network. Accordingly, the sink with the least number of
nodes will be elected by the node as a preferred sink.
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
215
Since not all nodes are equal in respect to hardware
capabilities; some nodes may have smaller routing table
sizes than other nodes in the same network. Thus, the
greedy selection of the lesser loaded sink may lead to
overloading some critically placed nodes and give rise to
the unreachability problem and routing loss [14]. To
overcome this problem, the DODAG-size metric operates
in conjunction with another memory utilization metric
called reachability metric. It requires each sink to record
dropped data packets sent to unreachable destinations, in
P2P traffic networks and propagates this information via
DIO control messages. This enables a practical way to
monitor the reachability of the DODAG nodes by detecting
the routing loss that occurs in sink nodes as a result of
memory overflows.
Therefore, combining Memory Utilization metrics with
ETX provides a better understanding of the DAG status.
This allows RPL nodes to perform optimal sink-selection
towards achieving effective multi-sink load-balancing.
4.2 Propagation of Load Balancing Information
In MSLBOF, each sink distributes its LB information
in transmitted DIO messages to its DODAG nodes. There
are several approaches to implementing Information
propagation in standard RPL. These include using a Metric
Container within the DIO message or adding a new DIO
option, then modify the OF to utilize that information
during the processing of the DIO messages. Since, both use
the existing control message type (i.e., DIO message) and
within standard RPL scope, we adopted the second
approach for MSLBOF. A new standard ICMP option
named Sink Load-Balancing Option (SLBO) is included in
each DIO control message. When a sink generates a DIO
message, it scans its routing drop counter to get the number
of routing drop packets. It then updates the number of nodes
that are joint to its DODAG (DODAG size) in order to inject
them into the SLBO option as depicted in Fig. 2. This
information has to be updated and recorded in each DIO;
accordingly, all the nodes have the latest updates. In this
way, we utilize the transmitted DIO message with the 4-
byte overhead required to carry the SLBO option and evade
extra overhead that can be caused by adding a new control
message for the entire network.
4.3 Sink Selection Mechanism
In standard OFs, the sink-selection process is based on
the best rank regardless of other metrics that give
information about the stability and reliability of the sinks.
However, MSLBOF goes into action when a node n that is
associated with current DAG s
1
receives a DIO message
from a parent in a different DAG s
2
. Hence, the sink-
selection process starts by adding the Memory Utilization
Metrics, including The Dag size metric dag_size() and
reachability metric routing_drop() to the Rank rank() in
order to select the best sink. The main equations of
MSLBOF are shown in (1) and (2):
(1)
Where α
a
+
α
b
+ α
c
= 1, these coefficients are used as weight
factors that are defined based on the application
requirements. The normalized values rank(), dag_size(),
routing_drop() are combined as a weighted sum operation.
Afterward, the node n selects the best sink as
(2)
Algorithm 1: SLBOF Sink-Selection Algorithm
Thus, MSLBOF allows each node to select the sink
node having the lower number of nodes in its DODAG and
fewer routing drops instead of only Rank-based selection.
The details of the sink-selection process are illustrated in
Alg. 1. Each node runs a separate instantiation of this
algorithm, upon the reception of DIO messages from
different DODAGs.
4.4 Avoiding the Herding Problem
Most existing load balancing mechanisms suffer from the
instability problem as a result of the continuous switching
to reach the load balancing across the network’s sinks. This
problem is named as Herding Effect and resulted from
weakly implemented load-balancing OF. It leads to a high
switching rate trying to achieve equal load distribution
regardless of the topological stability.
Figure 2: Sink Load-Balancing Option (SLBO)
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
216
Therefore, an efficient load balancing mechanism
implies providing the protocol with a fair distribution
among sinks as well as ensuring network stability. To
mitigate unnecessary DAG switches, we have taken an
approach called stability condition. As each time the node n
that is joint to DODAG s
1
receives a DIO message from
different DODAG s
2
it decides to switch to g
2
if
(3)
where µ is a stability constant added to sink-selection
equation (3) in order to minimize inefficient and useless
DODAG switches.
5. Evaluation
5.1 Evaluation Setup
To evaluate the proposed solution, we carried out
experimental simulations with the Cooja Simulator. It is an
emulation and simulation tool available as a part of Contiki
OS and enables working its RPL implantation. Contiki 3.0
was utilized in all the experimental setups. In order to
analyze the performance of MSLBOF, we compare it
against MRHOF in ContikiRPL. The configuration
parameters given in Tab.1 were considered for all the
carried out experiments.
Table 1: General configurations in experiments
In our physical topology, we consider a uniform 2D-
grid within an area of size 110m x 110m. The network
consists of 100 nodes, each of which is Z1 motes, with two
sinks that are placed in the top-left and top-right corners of
the topology. The Z1 Zolertia platform is equipped with a
16 MHz MSP430F2617 microcontroller, 8 Kbytes RAM,
and 92 Kbytes ROM for storage.
Each sink has two node lists: the sending and receiving
lists. Nodes in the sending list continuously generate an
IPv6 data packet of 127 bytes in size. The packets are sent
to the nodes in receiving list in order to satisfy P2P
communication. Each experiment takes 2100 sec. and the
simulation results are averaged over 10 simulations.
5.2 Results and Discussion
The comparison was made according to different
network measures and considering critical performance
aspects. These include Packet Delivery Ratio (PDR), loss
ratio, and network overhead.
As depicted in Fig. 3, MSLBOF shows significant
improvement in PDR. MSLBOF demonstrates higher PDR,
which is about 16% more, compared to MRHOF in standard
RPL. MSLBOF helps each common node located between
two DODAGs to effectively select the best sink considering
memory utilization metrics and avoid the unreachability
problem.
Figure 3: Packet Delivery Ratio
Fig. 4 shows the routing loss ratio for each sink in the
cases of MRHOF and MSLBOF. As MSLBOF applies the
memory utilization metrics to evenly distribute the nodes
across the sinks, it shows a high reduction in routing loss
ratio. We observed in MRHOF that sink1 suffered from
higher routing loss due to the memory overflow problem. It
is evident that MSLBOF succeeded in balancing the load
across the sinks and enhancing P2P communication.
Figure 4: Routing Loss Ratio per sink
Parameter Configuration
Simulation Duration 2100 sec.
Simulation area 110 x 110 m
2
Network Scale 100 nodes (including 2 sinks)
Mote Emulated Z1 mote
Radio Model Unit Disk Graph Medium (UDGM)
Node Spacing 10 m
Transmission Range 13 m
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
217
Considering the limitations in embedded memories of
IoT devices, the majority of the packet loss in MRHOF is
routing loss compared to other losses (i.e. queue loss, link
loss) as shown in Fig. 5. Higher routing loss resulted in
inappropriate DODAG selection in MRHOF. MSLBOF
handled this problem by using proper metrics for effective
DAG load balancing to avoid overloaded- memory and
unreachability problems. In general, MSLBOF achieved a
lower packet loss ratio compared to MRHOF.
Figure 5: Total Loss Ratio
Network stability is highly affected by DODAG
switches, as increasing the number of switches means extra
control overhead to the network and less network instability.
Fig. 6 presents the number of transmitted DIO and DAO
messages. MSLBOF has fewer transmitted control
messages, which means that the topology tends to be more
stable than MRHOF. Furthermore, MSLBOF showed a
reduction in the number of DODAG switches, up to 76%,
due to applying the stability-condition approach.
Figure 6: Control message overhead
6. Conclusion
Node unreachability is a common problem in RPL
networks, particularly with large-scale setups. It results in
critical routing failures due to memory overflows. The
proposed solution in this paper enables RPL with an
effective objective function, MSLBOF, to alleviates such a
problem in RPL multi-sink setups. It allows each node to
select its sink according to newly introduced memory
utilization metrics. MSLBOF demonstrated significant
improvements compared to MRHOF-RPL. It was able to
achieve higher PDR and lower packet loss while
maintaining better network stability. Expanding the scope
of this work in terms of applying other parent selection
mechanisms and adopting a different set of metrics will be
the primary focus of future work.
References
[1] A. Brandt, J. Hui, K. Pister, J.P. Vasseur, and R. Alexander,
'"rfc6550,", -03.
[2] Hyung-Sin Kim, Hongchan Kim, Jeongyeup Paek and
Saewoong Bahk, '"Load Balancing Under Heavy Traffic in
RPL Routing Protocol for Low Power and Lossy Networks,"
TMC, vol. 16, no. 4, Apr 01, pp. 964-979.
[3] S. Taghizadeh, H. Bobarshad and H. Elbiaze, '"CLRPL:
Context-Aware and Load Balancing RPL for Iot Networks
Under Heavy and Highly Dynamic Load," Access, vol. 6, pp.
23277-23291.
[4] M. Eshghie and N. Yazdani, '"Quell: Lightweight Load-
Balancing Scheme for 6TiSCH IoT networks," ISTEL, pp.
23-27.
[5] O. Iova, F. Theoleyre and T. Noel, '"Exploiting multiple
parents in RPL to improve both the network lifetime and its
stability," ICC, pp. 610-616.
[6] B. Ghaleb, A. Al-Dubai, E. Ekonomou, W. Gharib, L.
Mackenzi and M. Bani Khala, '"A New Load-Balancing
Aware Objective Function for RPL's IoT Networks," HPCC,
pp. 909-914.
[7] H.S. Altwassi, Z. Pervez, K. Dahal and B. Ghaleb, '"The RPL
Load Balancing in IoT Network with Burst Traffic
Scenarios," SKIMA, pp. 1-7.
[8] P. Singh and Y. Chen, '"RPL Enhancement for a Parent
Selection Mechanism and an Efficient Objective Function,"
JSEN, vol. 19, no. 21, Nov 01, pp. 10054-10066.
[9] C. Ji, R. Koutsiamanis, N. Montavont, P. Chatzimisios, D.
Dujovne and G.Z. Papadopoulos, '"TAOF: Traffic Aware
Objective Function for RPL-based Networks," GIIS, pp. 1-5.
[10] M. Mamdouh, K. Elsayed and A. Khattab, '"RPL load
balancing via minimum degree spanning tree," WiMOB, pp.
1-8.
IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021
218
[11] P. Kulkarni, S. Gormus and Zhong Fan, '"Tree Balancing in
Smart Grid Advanced Metering Infrastructure Mesh
Networks," greencom, pp. 109-115.
[12] Minkeun Ha, Kiwoong Kwon, Daeyoung Kim and Peng-
Yong Kong, '"Dynamic and Distributed Load Balancing
Scheme in Multi-gateway Based 6LoWPAN," ithings, pp.
87-94.
[13] M.O. Farooq, C.J. Sreenan, K.N. Brown and T. Kunz,
'"Design and analysis of RPL objective functions for multi-
gateway ad-hoc low-power and lossy networks," Ad hoc
networks, vol. 65, Oct, pp. 78-90.
[14] Ye Miao, S. Vural, Zhili Sun and Ning Wang, '"A Unified
Solution for Gateway and In-Network Traffic Load
Balancing in Multihop Data Collection Scenarios," JSYST,
vol. 10, no. 3, Sep, pp. 1251-1262.
[15] M. Mahyoub, A. Mahmoud, M. Abu-Amara and T. Sheltami,
'"An Efficient RPL-based Mechanism for Node-to-Node
Communications in IoT," JIoT, Nov 16, pp. 1.
Maram Abdullah received the B.S. degree in Software
Engineering (SE) from Hail university, Saudi Arabia in 2014. She
is currently working toward a master’s degree in Computer
Science (CS) at Al-Qassim university.
Ibrahim S. Alsukayti received the B.S. degree in Computer
Science, from Qassim University, Buraydah, KSA, in 2006. He
then received the M.S. degree in Computer and Information
Networks from the University of Essex, Colchester, UK, in 2010
and the Ph.D. in Computer Networks from Lancaster University,
Lancaster, UK, in 2014. Currently, he is an Associate Professor
with the Computer Science Department, College of Computer,
Qassim University, Buraydah, KSA. He is also a team member of
two ongoing funded research projects and the director of a research
group targeting Internet of Things technologies & applications.
His research interests include network routing, Wireless Sensor
Networks, networking protocols, network security, and IoT.
Mohammed Alreshoodi received the B.S. degree in computer
science from Qassim University, Buraydah, KSA, in 2004 and the
M.S degree in computer networks from University of Essex,
Colchester, UK, in 2011. He also received a Ph.D. degree in
computer networks from the University of Essex, Colchester, UK,
in 2011. He is currently an Associate Professor in the department
of applied science in Unizah community college at Qassim
University, KSA. His research interest includes computer
networking, wireless networks, WSN, IoT, networks security. He
is a member of the WSN research group and Cyber Security
research group at College of Computer, Qassim University, KSA.
