SPECIAL SECTION ON GREEN COMMUNICATIONS ON WIRELESS NETWORKS

Received January 26, 2020, accepted February 9, 2020, date of publication February 27, 2020, date of current version March 11, 2020.
Digital Object Identifier 10.1109/ACCESS.2020.2976654

A Performance-to-Cost Analysis of IEEE 802.15.4

MAC With 802.15.4e MAC Modes
NIKUMANI CHOUDHURY 1 , (Student Member, IEEE), RAKESH MATAM 1 , (Member, IEEE),
MITHUN MUKHERJEE 2 , (Member, IEEE), AND JAIME LLORET 3 , (Senior Member, IEEE)
1 Department of Computer Science and Engineering, IIIT Guwahati, Guwahati 781015, India
2 Guangdong Provincial Key Laboratory of Petrochemical Equipment Fault Diagnosis, Guangdong University of Petrochemical Technology, Maoming 525000,
China
3 Instituto de Investigación para la Gestión Integrada de Zonas Costeras (IGIC), Universitat Politecnica de Valencia, 46022 Valencia, Spain

Corresponding author: Mithun Mukherjee ([email protected])

This work was supported in part by the SERB, DST, Government of India under Grant ECRA/2016/001651.

ABSTRACT The IEEE 802.15.4 standard is one of the widely adopted networking specification for Internet
of Things (IoT). It defines several physical layer (PHY) options and medium access control (MAC) sub-layer
protocols for interconnection of constrained wireless devices. These devices are usually battery-powered and
need to support requirements like low-power consumption and low-data rates. The standard has been revised
twice to incorporate new PHY layers and improvements learned from implementations. Research in this
direction has been primarily centered around improving the energy consumption of devices. Recently, to meet
specific Quality-of-Service (QoS) requirements of different industrial applications, the IEEE 802.15.4e
amendment was released that focuses on improving reliability, robustness and latency. In this paper, we carry
out a performance-to-cost analysis of Deterministic and Synchronous Multi-channel Extension (DSME)
and Time-slotted Channel Hopping (TSCH) MAC modes of IEEE 802.15.4e with 802.15.4 MAC protocol
to analyze the trade-off of choosing a particular MAC mode over others. The parameters considered for
performance are throughput and latency, and the cost is quantified in terms of energy. A Markov model has
been developed for TSCH MAC mode to compare its energy costs with 802.15.4 MAC. Finally, we present
the applicability of different MAC modes to different application scenarios.

INDEX TERMS IEEE 802.15.4, IEEE 802.15.4e, low-power wireless personal area networks, energy
conservation, IoT.

I. INTRODUCTION schemes like synchronization [18]–[21], duty-cycling

Internet of Things (IoT) has applications in diverse areas [22]–[25], cluster-head rotation [26]–[29], etc., have been
like smart industries, smart homes, smart cities, smart grid, proposed. But, moving towards Industry 4.0, accommodating
smart health, intelligent transportation, smart agriculture etc. growing QoS requirements like latency, throughput, energy,
[1]–[9] to name a few. Industry 4.0 [10], [11] in fact refers reliability and robustness has been a major concern. Real-
to machines capable of sensing, communicating and option- ising this, the IEEE Standards Association recently came
ally taking decisions. The IEEE 802.15.4-2011 [12] is one up with IEEE 802.15.4e [30], an amendment to the existing
of the enabling standard that has been widely adopted for 802.15.4 standard. The revised standard [31] is designed for
networking of low-power, low-rate, battery-powered devices, real-time applications with latency constraints that need to
which are commonly referred to as things in IoT applica- provide better reliability and robustness.
tions. The standard is intended for applications with limited The IEEE 802.15.4e defines new MAC behaviors that
power and non-stringent throughput requirements. There- guarantee latency and enable robust communication through
fore, the protocols that have been developed mainly aim to multi-channel frequency hopping. The standard considers
minimize power consumption. Several research works have the QoS demands from various industrial applications [32].
been carried out to further enhance the energy efficiency of It defines five new MAC modes, namely, Blink Radio Fre-
the IEEE 802.15.4-2011 MAC [13]–[17]. Power-conserving quency Identification (RFID), Asynchronous Multi-channel
Adaptation (AMCA), Low-Latency Deterministic Networks
The associate editor coordinating the review of this manuscript and (LLDN), DSME, and TSCH. Blink RFID targets applica-
approving it for publication was Wei Wei . tions intended for object/personnel identification, tracking,

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/
41936 VOLUME 8, 2020
N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

and location. AMCA is a multi-channel approach used in and TSCH modes of MAC operation and their respec-
non-beacon enabled (NBE) mode for large deployments. tive challenges in implementation. The proposed Markov
DSME MAC mode supports multi-channel operation in the model for TSCH CSMA/CA and its comparison with
contention-free period (CFP) [12] to guarantee low and deter- 802.15.4-2011 MAC (to be referred as 802.15.4 MAC) is pre-
ministic latency using Guaranteed Time Slot (GTS) [12]. sented in Section V. The experimental results are described
TSCH MAC mode has received considerable attention with in Section VI. Section VII presents a summary of the sup-
the establishment of IETF 6TiSCH Working Group. The ported QoS features of different MAC modes along with their
TSCH mode supports channel-hopping and multi-channel suitable application areas. Finally, conclusions are drawn in
communication links (dedicated and shared) over a sin- Section VIII.
gle time-slot. The TSCH CSMA/CA is distinct from the
802.15.4 CSMA/CA and requires a complex operation for II. RELATED STUDY
scheduling of links over different channels. Note that the The revised IEEE 802.15.4-2015 [31] standard includes
DSME and TSCH modes support multi-hop topologies, DSME and TSCH MAC modes along with the 802.15.4 slot-
whereas modes like LLDN that support stringent timing ted CSMA/CA MAC. Support for different application spe-
requirements, operate in only star topology. cific QoS has encouraged research on the newly developed
Currently, the IEEE 802.15.4e standard is still in the early MAC modes. Several works on TSCH [54]–[61] and
days of adoption and research is actively being carried out DSME [51], [52], [62]–[65] have been carried out recently
to analyze its performance and address the gaps in imple- that focus on several aspects of their MAC behavior and
mentation. Several works have been carried out highlighting performance. The works in [54]–[57] propose scheduling
some of the limitations [33] and open issues that need to be mechanisms for TSCH networks. For example, Orches-
investigated, especially in protocol implementations. Secu- tra [54] achieves a high throughput with minimal overhead,
rity schemes and availability of supporting hardware is one whereas the adaptive static scheduling in [55] focuses on
of the primary requirements for commercial viability [33]. low and deterministic delay for the static networks. In addi-
ZigBee [34], 6LoWPAN [35], and WirelessHART [36] have tion, Wave [56] targets minimal delay by scheduling the
commercially implemented the IEEE 802.15.4-2011 as their slots based on data traffic flows. It automatically adapts
underlying standard. In addition, petroleum industries and to the available radio interfaces and channels of the sink.
refineries [37]–[39], agricultural implementations [40], [41], The Stripe [57] is a distributed scheduling mechanism that
smart city applications [42], and smart grids [43] have con- reconfigure random pre-allocated slots and later schedules
tinued with their implementation of IEEE 802.15.4 for its additional slots based on traffic. Further, authors in [59], [66]
simplicity and low complexity. Few survey [33], [44], [45] propose an adaptive channel selection mechanism for data
and recent works on LLDN [46], [47], TSCH [48]–[51], and transmission based on estimated link quality. Networks prone
DSME [51]–[53] MAC modes of operation have addressed to interference can hop over other channels and improve reli-
several existing limitations. ability by adopting channel hopping measures as described
In this paper, we present the trade-off of choosing a partic- in [58], [67]. Finally, synchronization among devices in
ular MAC protocol over others in terms of energy, latency, multi-hop networks is studied in [60]. These works aim to
and associated overhead. A Markov model for the TSCH improve the overall performance of TSCH networks by sup-
CSMA/CA is presented and compared with a similar model porting one or more QoS features.
for 802.15.4 MAC. To summarize, the contributions of this Similar to Orchestra [54], Symphony [63] proposes a new
paper are as follows. multichannel multi-time slot scheduling algorithm that inte-
• First, an analysis and comparison of the DSME and grates IPv6 Routing Protocol for Low Power and Lossy
TSCH MAC modes with the IEEE 802.15.4 MAC is Networks (RPL) over DSME. The authors in [64] presents
presented. Also, we discuss major research challenges an effective multisuperframe tuning technique that utilizes
associated with the respective MAC behaviors. CAP reduction in an effective way to improve flexibility and
• Second, we propose a Markov model to estimate the scalability while guaranteeing deterministic and low delay.
transmission time and energy consumption for trans- Reference [65] proposes a learning based beacon scheduling
mission of frames using the multi-channel approach of mechanism for IEEE 802.15.4 as well as DSME networks.
TSCH CSMA/CA. The authors in [62] proposes a DSME-based distributed
• Third, we perform the simulation and numerical analysis scheduling mechanism for mobility support. It adaptively
on the performance of 802.15.4, TSCH, and DSME assigns communication slots by analyzing the channel traffic
MAC modes. Based on these results, we outline the dis- at each node to improve the network reliability and timeli-
cussion on the choice of MAC for different applications ness. A channel access mechanism is proposed in [52] for
that have varying QoS requirements. constrained devices to reduce the packet drop rate, energy
The rest of the paper is organized as follows. Related study consumption and collisions.
is presented in Section II. Section III provides an overview of Previous works in this direction primarily focused on eval-
the IEEE 802.15.4 MAC and IEEE 802.15.4 MAC enhance- uating and improving the performance of the TSCH and
ments. Section III-C and Section III-D describe the DSME DSME MAC modes. Works like [51], [68] have compared

VOLUME 8, 2020 41937

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

the DSME and TSCH MAC mode based on the QoS features.
However, in this paper, we aim to present the trade-off of
choosing a particular MAC protocol over the others. We anal-
yse the cost (in terms of power consumption) to achieve
the desired QoS features like throughput and latency. In the
subsequent sections, we discuss the IEEE 802.15.4 MAC
followed by DSME and TSCH MAC modes of operation.

III. OVERVIEW OF 802.15.4, DSME AND TSCH MAC FIGURE 1. Superframe structure [12].
To support low-power and low-rate wireless communications
among resource-constrained devices, IEEE 802.15.4 standard
was designed. Over the last decade, it has become the most
where aBaseSuperframeDuration is defined as the number of
widely adopted standard for IP based IoT networks. Although
symbols constituting a superframe when the SO is set to zero.
new MAC behaviors have been defined, 802.15.4 MAC is still
With 0 ≤ SO ≤ BO ≤ 14 and BO = 15 implies a non-beacon
a relevant part of the current specification of the standard.
mode.
Next, we present a brief overview of the 802.15.4 MAC,
However, the 802.15.4 MAC suffers from several limita-
followed by an introduction to the new MAC modes.
tions [13], [69] like unbounded latency and low reliability.
This makes the standard unsuitable for applications having
A. IEEE 802.15.4 MAC strict QoS requirements. Available GTS are either not suffi-
Devices operating the IEEE 802.15.4 standard can either cient or may not be continuously allocated in multi-hop net-
be Fully Functional Devices (FFD) or Reduced Functional works. The transmissions over a single shared channel result
Devices (RFD). FFDs are capable of initiating a Personal in increased latency and frame loss due to contention and
Area Network (PAN) and serve as a PAN coordinator collisions, respectively. Moreover, these are also potentially
(PANC). They allow other FFDs and RFDs to associate with it vulnerable to interference with other wireless technologies
to extend the network. On the other hand, RFDs are resource- working in the same 2.4 GHz ISM band such as WLAN-
constrained and can only associate to an FFD to transmit systems of IEEE 802.11, Bluetooth, and microwave ovens.
data. It acts as an end device in the network topology. Syn- Therefore, 802.15.4 MAC is suitable for applications with
chronization between these devices is achieved with the help flexible requirements of latency and throughput. In view of
of a superframe structure [12] (shown in Fig. 1). The time this, new MAC modes are presented in the IEEE 802.15.4e
interval between two consecutive beacons is the Beacon Inter- standard that supports different QoS requirements of various
val (BI), and it consists of an active period and an optional applications. These MAC modes are expected to provide
inactive period (sleep period). Data transmissions take place data transmissions with low and deterministic latency, high
in the active period (divided into 16 equal slots), whereas reliability with dedicated communication, and multi-channel
the device enters sleep state during the inactive period. The access.
length of the active period is known as Superframe Dura-
tion (SD). Transmissions in the active period can either be
B. MAC ENHANCEMENTS IN IEEE 802.15.4e
contention-based using slotted CSMA/CA or contention-free
The enhanced version of the standard includes new
using Guaranteed Time Slot (GTS) mechanism. A maxi-
network structures and functionalities along with the slot-
mum of seven GTS slots can be allotted to the associated
ted CSMA/CA MAC to accommodate application-specific
devices in a single BI. This combination of GTS slots are
requirements in low-rate wireless personal area networks
optional and is known as Contention-Free Period (CFP).
(LR-WPAN). IEEE 802.15.4e defines five different MAC
GTS allows exclusive usage of the channel to an associated
behaviors, viz., Blink RFID, AMCA, LLDN, DSME, and
device to decrease latency in transmission. Beacon trans-
TSCH. Both RFID and AMCA are the two non-real-time
mission indicates the beginning of the Contention Access
MAC behaviors. The former targets applications intended
Period (CAP), and all the associated devices participate in
for object/personnel identification, tracking, and location.
transmitting any pending data using the CSMA/CA proce-
On the other hand, AMCA is a multi-channel approach used
dure [12]. Two parameters macBeaconOrder (BO) and
in NBE mode for large deployments. The other three MAC
macSuperframeOrder (SO) together defines the struc-
modes provide deterministic latency guarantees for time-
ture of superframe. Devices sleep in the inactive period until
critical applications. However, in LLDN mode, all the devices
the beginning of the next superframe structure, forming a
in the network are required to be directly associated with
superframe cycle. BI and SD can be determined using the
the PANC, thereby, forming a star topology-based network.
following expressions,
Therefore, among all the new MAC behaviors, DSME and
TSCH operate in beacon-enabled mode (BEM) as well as
BI = aBaseSuperframeDuration.2BO (1) support peer-to-peer connectivity to form medium to large-
SD = aBaseSuperframeDuration.2SO , (2) sized networks. Also, these two MAC modes have been

41938 VOLUME 8, 2020

N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

FIGURE 3. Example of three timeslot slotframe in TSCH [31].

FIGURE 2. DSME multi-superframe structure [31].

D. TIME-SLOTTED CHANNEL HOPPING

The TSCH networks are suitable for applications prone
incorporated in the revised IEEE 802.15.4-2015 standard.
to interference from other wireless networks. Moreover,
In what follows, we present an analysis of the two MAC
the TSCH MAC behaviour provides high reliability and time-
behaviors, namely DSME and TSCH.
critical assurance for oil/refinery industries that primarily
concern human and environmental safety. Other applications
C. DETERMINISTIC AND SYNCHRONOUS
include equipment and process monitoring like food and
MULTI-CHANNEL EXTENSION
chemical products, pharmaceutical products, water treatment,
DSME targets the time-critical applications like health mon- etc. In this mode, devices synchronize within a periodic
itoring system that requires high reliability along with low slotframe (collection of timeslots). Each timeslot is pair-
and deterministic latency. It is suitable for several industrial wise communication between two devices. A communication
applications such as factory automation, process automa- schedule is formed by setting the number of timeslots in
tion, smart metering, etc., and commercial applications like a slotframe that determines how frequently each timeslot
home automation and smart building. These applications repeats. Fig. 3 illustrates an example of a slotframe consisting
demand high scalability and robustness, which are part of of three timeslots wherein three devices A, B, and C are
the design goals of DSME. It defines a multi-superframe communicating.
structure, that is a combination of one or more 802.15.4 super- The network maintains a global count of the number times-
frames, as defined by the PANC. The cycle of one or lots that have elapsed since the beginning of the network
more superframes repeating periodically is called multi- operations. This count is the Absolute Slot Number (ASN).
superframe structure. A single channel is used in the CAP The communication links at any timeslot can either be shared
as well as to transmit an enhanced beacon (EB) [31]. The EB (CSMA/CA) or dedicated (contention free). Different com-
communicates to the associated devices about the number of munication schedules can be established by defining sev-
superframes present in the multi-superframe. DSME defines eral concurrent slotframes of different sizes. This is useful
a new parameter called multi-superframe order (MO) which is when the network is operated at different duty-cycles. The
related to superframe order (SO) [12], [31] and beacon order multi-channel communication in TSCH depends on the chan-
(BO) [12], [31] as follows. nel hopping mechanism. A link between devices is defined
0 ≤ SO ≤ MO ≤ BO ≤ 14 (3) as a pairwise assignment of directed communication. The
physical channel or frequency in a link is determined as
MD = aBaseSuperframeDuration. 2MO , (4)
follows:
where MD is the multi-superframe duration, that signifies
the length of all the individual superframes in the multi- f = F[(ASN) + Channel Offset]%Nchannels , (5)
superframe. A DSME multi-superframe structure is shown
in Fig. 2. where F is the channel Hopping Sequence list and Nchannels
Contrary to 802.15.4 MAC, the DSME superframes is the number of channels used in the current network opera-
accommodate a higher number of GTS slots using multi- tion. Communication reliability is increased through channel
channel communication in the CFP period. In a multi- hopping that mitigates the effects of interference. Also, time-
superframe structure, a coordinator can reduce the size of slotted access with dedicated links reduces collisions. This
the CAP by disabling all but the first superframe CAP, results to reduction in retransmission of frames. The most
a technique called CAP reduction. This further increases the popular feature of TSCH mode is the TSCH CSMA/CA
number of available GTS in a single superframe. The multi- algorithm and the retransmission algorithm. Devices perform
channel approach in DSME mode is equipped with channel a Clear Channel Assessment (CCA) prior to transmission.
adaptation and channel hopping techniques. Note that the If the channel is found to be idle, data is transmitted in the
link quality indicator is used to switch between channels at link; else, the device waits for the forthcoming transmission
different timeslots. On the other hand, channel hopping uses link to the destination device. The presence of dedicated com-
a pre-defined set of channels (decided by upper layers) called munication links in the multi-channel timeslots facilitates the
hopping sequence, which is followed by all the devices in the transmission of time-critical data as well as improves the
network. network robustness.

VOLUME 8, 2020 41939

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

IV. MAC COMPARISON AND CHALLENGES: DSME, TSCH,

AND 802.15.4
A. DSME AND 802.15.4 MAC
Guaranteed Time Slots has been previously used (option-
ally) in 802.15.4 MAC for transmitting in a contention-
free approach. However, DSME has mandatory CFP consti-
tuting of several GTS. It addresses the limitation on GTS
slots (seven in 802.15.4 MAC) through the multi-channel
approach. The total available DSME-GTS depends upon
the number of channels used, as a single slot can be used
for multiple communications in different frequencies. Thus,
it allows applications operating DSME MAC to schedule
transmissions with low latency and high reliability. The CAP
reduction mechanism further increases the number of GTS
slots by allotting the CAP slots (single channel) to multi-
channel GTS slots. If Nchannels channels are used in current FIGURE 4. An IEEE 802.15.4 network topology.
network operation, then the maximum possible GTS alloca-
tion in a single superframe is given by
GTSmax = (7 × Nchannels ) + CAPslots , (6) one frame per BI. To increase the contention, we increase the
where CAPslots is the total number of slots in the CAP. number of frames generated by each device to eight frames
Although all communications can take place through DSME- per BI. We compare the total time required (in terms of BI)
GTS slots, significant overhead is incurred in the GTS for transmitting a set of frames between 802.15.4 MAC and
management. DSME MAC. The associated end-devices generate and trans-
Generally, associated devices transmit pending data to the mit data frames to coordinator c1 through their respective par-
parent device within a single BI (through slotted CSMA/CA), ent coordinator. We assume preference in transmission using
whereas GTS transmissions require at least two superframes. GTS over slotted CSMA/CA by the devices. This is done
This is because GTS transmissions are preceded by GTS for ease of comparison between 802.15.4 MAC and DSME
allocation requests (in the current CAP) and response con- mode, where GTS is the primary mechanism of transmission.
trol frames that are exchanged between the parent and the The sequential steps of the 802.15.4 transmission (CAP and
associated device. If GTS slots are successfully allotted, GTS) mechanism at each coordinator is presented below.
a device transmits in the subsequent BI. Hence, DSME- 1) Step 1 (At c6). Total frames expected by coordinator
GTS may not decrease latency compared to transmissions c6 is five (one from each end-device). In the first (cur-
during CAP (as in 802.15.4 MAC) when contention in the rent) BI, end-devices request for GTS allocations. The
channel is low. However, high contention generates fre- GTS allocation requested frames are transmitted in the
quent backoffs (in slotted CSMA/CA) among the transmitting subsequent SD (next BI) after successful a allocation
devices that results to increase in transmission delay. This response by c6. Therefore, the time for receiving the
issue is addressed by the multi-channel feature of DSME last frame by c6 is within two BIs.
MAC, allowing concurrent communication through different 2) Step 2 (At c5). The total frames expected is eight
channels. Thus, in such network scenarios, DSME MAC (including five frames from c6). Similar to Step 1, all
considerably reduces the transmission latency compared to frames are transmitted (in CAP and GTSs) within two
802.15.4 MAC. BIs.
The IEEE 802.15.4 standard is designed for networking of 3) Step 3 (At c4). c4 receives frames from c5 and its
wireless devices constrained in terms of power, computation, associated end-devices. Thus, total frames expected is
and memory. The 802.15.4 MAC allows devices to enter sleep twelve (eight from c5 and four from the end-devices).
state after the active period. Duty-cycling schemes [23], [24], Here, seven frames can be transmitted through GTS
[70] optimize devices’ sleep period to prolong their battery request (in 1st BI); the rest five frames can contend in
life. However, in DSME, coordinators remain active for the the CAP period for transmission.
multi-channel GTS in the CFP, resulting in higher energy 4) Step 4 (At c3). Similar to Step 2, c3 receives fourteen
dissipation. frames in total that can be transmitted within two BIs.
Illustrative Example: Let us consider a network topology, 5) Step 5 (At c1). Finally, c1 receives frames from c3
as shown in Fig. 4 with coordinators c1, c2, c3, c4, c5 and c6. (fourteen frames), c2 (five frames) and two frames
Let BO = 5 and SO = 2 for all the coordinators in the network. from the associated end-devices within three BIs. Here,
We assume low channel contention when devices are con- if the SD is not sufficiently long for transmissions of all
tending for transmitting a fewer number of frames in a given the frames, remaining frames may be transmitted in the
period of time. Initially, let each device sense and generate subsequent SD.

41940 VOLUME 8, 2020

N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

Therefore, the aggregate time for receiving the last frame

by c1 is the sum of all the steps mentioned above, i.e.,
2 + 2 + 2 + 2 + 3 = 11 BIs. Again, let the number
of available channels in the DSME MAC mode be three.
Therefore, available GTS for each superframe is 7 × 3 = 21
GTS. For the same number of transmissions, the time required
using the GTS mechanism will be two BIs per coordinator.
That is, in aggregate, ten BIs are required. Note that we
have not considered CAP reduction, which can reduce the
total transmission time. Therefore, we observe that in low
channel contention, performance in terms of transmission
time is similar.

1) CHALLENGES WITH DSME MAC

Although DSME provides a higher number of GTS compared
to 802.15.4 MAC, it poses several challenges.
• Firstly, a sophisticated slot scheduling mechanism is
necessary for allocating time and frequency slots to
multiple communicating devices.
• Secondly, for multi-hop communication, the scheduling
scheme needs to maintain the specific GTS allocations
across hops to adhere to various QoS requirements. The
difficulty arises as coordinators do not have the same
number of associated devices, and traffic flows vary
throughout the network. Thus, the scheduling mecha-
nism needs to be traffic aware and decentralized for
scalable network operations.
• Thirdly, slot management [71] in DSME is not ade-
quately addressed in the standard. This results in issues
like inconsistent slot allocation bitmap, failure of GTS
deallocation in volatile topology and collisions of slot
management handshakes in CAP.
• Finally, the absence of sleep periods in the multi-
superframe structures is the primary energy-draining FIGURE 5. TSCH CSMA-CA [31] retransmission backoff mechanism.

issue in DSME. CAP reduction minimally addresses

this issue by allowing associated devices to sleep during
DSME-GTS when they are not in either transmission in 802.15.4 MAC. It performs a single CCA before trans-
or receiving state. However, coordinators, including the mission, whereas, in 802.15.4, a device performs two CCA
PANC, has to remain active for longer duration resulting (channel free) before transmitting the frame.
in a reduction in network lifetime. Hence, further inves- Further, in 802.15.4 CSMA/CA, the retransmission mech-
tigations are required to combine energy efficiency and anism is a repetition of the entire transmission procedure,
low-latency in the DSME mode of transmissions. beginning from a random backoff wait, followed by two CCA
and finally transmission. However, the TSCH retransmission
B. TSCH AND 802.15.4 MAC for shared links is distinct from the TSCH CSMA/CA trans-
Unlike the 802.15.4 superframe structure, the concept of CAP mission mechanism. The retransmission exponential backoff
and CFP no longer exists in TSCH. Each timeslot can either is expressed in terms of the number of shared links that
be a dedicated or shared communication link. A dedicated must be skipped before attempting transmission. The backoff
link resembles to a GTS, and shared links operating the window increases for each failed transmission in a shared
TSCH CSMA/CA mechanism are allocated to more than one link. A successful transmission resets the backoff window to
pair of devices. Transmissions in the shared links may be a predefined minimum value.
prone to collisions. Recurring collisions are reduced through The superframe structure in 802.15.4 MAC ensures that
the TSCH retransmission mechanism, as shown in Fig. 5. devices enter periodical sleep periods. However, in TSCH
The TSCH CSMA/CA and 802.15.4 CSMA/CA primarily mode, devices involved in transmission remain active in their
differ in the backoff strategy followed, and the number of allotted timeslots. They may or may not enter a low-power
CCA performed. A device with data to transmit waits for mode in between slots. Even if they do, the transceiver has
the next shared link, instead of a random backoff wait as to be frequently switched between listening and sleep mode.

VOLUME 8, 2020 41941

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

TABLE 1. Comparison of the IEEE 802.15.4, DSME, TSCH MAC modes.

This is because only one pair of devices can communicate at without changing the entire schedule. To guarantee
each timeslot, and multiple communication requires several low-latency and high reliability, the schedule must
timeslots, which may not be continuous. Also, a channel allocate sufficient dedicated and shared links. Retrans-
switching overhead for different communications (between mission opportunities should be available to all devices.
different or same pair of devices) is incurred. Also, the schedule should be adaptable to variations
The transmission schedule for devices operating in in traffic flow with minimum changes. In the process
an IEEE 802.15.4 network is based on allocating non- of building such an optimal schedule, a trade-off must
overlapping slots. Distributed schedules allocate slots within be made with the energy consumption of the devices.
the 2-hop neighborhood. It is based upon the BO and SO In centralized approaches, all the devices in the network
parameters of the devices. However, in the TSCH mode of transmit their expected traffic flows along with the set
operation, computation of a schedule is highly complex [72]. of constraints to the central node, leading to a very
The TSCH scheduling task is an NP-hard problem [73]. The high transmission overhead. In the distributed approach,
schedule has to comply with various QoS demands like low each node may exchange traffic flow information with
latency and high reliability. For this, the schedule needs to its neighbors for constructing a consistent schedule.
consider the following. a) Traffic flows between the devices Further, the schedule must follow a channel hopping
and the associated deadlines for each of these flows. For sequence, defined by the higher layer in allocating chan-
instance, to achieve a given latency deadline, the last fragment nel offsets for different communication links. Finally,
transmitted must be received by the sink before the deadline. the schedule must decide an optimal number of shared
b) Ensure optimal slot allocation for each flow and consider and dedicated slots to maintain the traffic flow deadlines
link qualities along the flow path. For this, the schedule of all the devices in the network.
may consider parameters like the number of radio interfaces • Secondly, timeslots introduced in-place of the super-
per node and the number of available channels. c) Addi- frame duration of IEEE 802.15.4 do not have continuous
tionally, the schedule has to respect several constraints like sleep periods. They may or may not enter sleep state
half-duplex constraints, interference constraints, QoS con- when there are no specified communication links. Alter-
straints, buffer length constraints, and hardware constraints. nately, the radio frequently switches between active and
Table 1 summarizes the comparison between the three sleep states in TSCH mode as well as need to switch
MAC modes. between channels for different communications. Also,
the multi-channel approach in TSCH entirely depends
1) CHALLENGES WITH TSCH MAC on the channel hopping mechanism. All the devices in
The TSCH is empowered with time-slotted access along the network must be synchronized to the shared hop
with channel hopping and multi-channel capabilities. Several sequence in use. Thus, a network-wide slot synchroniza-
challenges arise with the implementation of these function- tion among the devices is required. Therefore, devices
alities, which are yet to be addressed in the IEEE 802.15.4e periodically transmit control (sync) frames that consume
standard. significant energy over a period of time. The network
• The first and foremost challenge is devising a schedul- consists of few time-source neighbors that periodically
ing mechanism that schedules the TSCH time slots transmit data or acknowledgment frames to all the neigh-
for data frames to be sent on for both dedicated and boring coordinator devices for synchronization.
shared links. The standard does not specify any pol- • Finally, the standard does not specify the criteria
icy to build and maintain the communication sched- behind the selection of a link as shared or dedicated
ule over multi-hop paths. The scheduling mechanism in TSCH CSMA/CA. The retransmission procedure in
will also control the resources allocated to each link TSCH CSMA/CA introduces a longer delay through
in the network topology. Moreover, the output schedule random backoff (in terms of links) compared to the
has to be compact, i.e., new flows may be allocated 802.15.4 CSMA/CA retransmission procedure. This is
41942 VOLUME 8, 2020
N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

FIGURE 6. Markov model for TSCH CSMA/CA and retransmission backoff mechanism.

discussed and analyzed through a proposed Markov where Tnext-linkn is the nth constant time waiting for the next
model in the next section. transmission link to destination before attempting CCA. TCCA
is the time required in CCA, Tta is the turn around time,
V. PROPOSED MARKOV MODEL FOR TSCH CSMA/CA Tl is the time for transmitting a frame of length l, ACKwait
We consider an IEEE 802.15.4e network topology with n is the time spent in waiting for acknowledgement from the
devices. We assume the existence of a TSCH transmis- coordinator, and ACKrec is time required in receiving the
sion schedule for the pairwise communication. The schedule ACK. For transmissions in dedicated links, the transmission
defines each allotted timeslot as either shared or dedicated. time is given by
In dedicated links, devices directly initiate transmission of
Tx = Tnext-link + Tta + Tl + ACKwait + ACKrec (8)
frames, whereas, in shared links, devices initially perform a
single CCA. Transmission failure in shared links is detected The energy consumption in shared transmission links is
by non-receipt of an acknowledgement. In order to reduce the Etxn = Ex (TCCA )+Eta Tta +Etx Tl + Ex (ACKwait + ACKrec )
probability of recurring collisions, the retransmission backoff
algorithm is followed, as shown in Fig. 5. CCA is independent + (n − 1)Ex (TCCA ) (9)
of the backoff stages as well as the number of retransmissions and for dedicated communication link is expressed as
previously attempted.
The Markov model for TSCH CSMA/CA and retrans- Etx = Ex Tta + Ex Tl + Ex (ACKwait + ACKrec ) (10)
mission is shown in Fig. 6. Each state in the model where Ex is energy consumed after completing a specific
can be represented with a 4-valued tuple (i, j, CCA, operation.
rnd) (i, j, CCA, rnd), where i = 0, . . . , 7 signifies the In case of a retransmission in the shared link, identified by
macMaxFrameRetries parameter, j = 0, . . . , 5 signi- i in the Markov model (macMaxFrameRetries parame-
fies the macMaxCSMABackoffs and rnd ranges from 0 to ter in TSCH backoff algorithm), the transmission time and
2BE − 1 that signifies the random number of shared links that energy consumed will be as follows
must be skipped before attempting transmission. CCA needs n
X
to be performed prior to frame transmission. It decreases after RTxn = Txn + TnBCKnext-link + nTCCA + Tta + Tl
each successful CCA and frame is transmitted when this value i=1
reaches 0. + ACKwait + ACKrec (11)
The transmission time of a frame in a shared link can be
ERtxn = Etxn + Ex (TCCA ) + Eta Tta + Etx Tl
expressed as
+ Ex (ACKwait + ACKrec )
n
X + (n − 1)Ex (TCCA ), (12)
Txn = Tnext-linkn + nTCCA + Tta + Tl
i=1 where TnBCKnext-link is the random number of shared links that
+ ACKwait + ACKrec , (7) must be skipped before attempting transmission again.

VOLUME 8, 2020 41943

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

TABLE 2. Configuration of the coordinators. TABLE 3. QoS performance measurement.

coordinators are given in Table 2. For, DSME and TSCH,

let the number of available channels be three (for ease of
A Markov model for IEEE 802.15.4 CSMA/CA was pre- computation and comparison with 802.15.4 MAC). Let each
sented in [23]. The transmission time and energy consumed device generate four frames per BI. For TSCH, we consider a
for IEEE 802.15.4 CSMA/CA frame transmission was com- timeline equivalent to four BI of c1 (coordinator with longest
puted as [23] BI). The schedules for 802.15.4 MAC, DSME and TSCH are
built using LBS [18], DSME [62] and [55]. We consider a
n
X timeslot (TSCH) to be 10 milliseconds long.
Tnmax = TBCKn + nTCCA1 + TCCA2 + Tta + Tl
We evaluate the QoS performance metric for all the
i=1
three MAC modes in MATLAB We use the expressions
+ ACKwait + ACKrec + (n − 1)TCCA2 (13)

derived from the Markov model to compute the latency and
Enmax = Ex TCCA1 + TCCA2 + Ex Tta + Ex Tl energy consumed in transmission for 802.15.4 MAC and
+ Ex (ACKwait + ACKrec ) TSCH. DSME follows a similar transmission procedure to
+ (n − 1)Ex TCCA1 + TCCA2 , the 802.15.4 MAC, but it primarily relies on DSME-GTS


(14)
for low-latency transmissions. Therefore, while building the
where TBCKn is the time spent in nth backoff state and TCCA2 transmission schedule, we allow frames to be transmitted
is the time required in CCA2 . using the available DSME-GTS and the remaining frames
From (7) and (13), it can be observed that a frame con- (if any) through the CAP. Throughput is computed based on
sumes more time in performing CCAs. The time spent wait- the number of frames received within the four BIs (equivalent
ing in the backoff stage in IEEE 802.15.4 CSMA/CA is purely time for TSCH). The amount of energy consumed for achiev-
random can either be longer or shorter than the next com- ing the desired latency and throughput is computed based
munication link to a destination in TSCH mode. However, upon the number of transmissions incurred. This includes
the TSCH retransmission backoff mechanism introduces a overhead in the transmission of control messages for schedule
longer transmission time (11) through the waiting period of computation and GTS allocation request/response, and chan-
a random number of shared links. This can be generally nel switching.
longer than a similar retransmission mechanism in IEEE Table 3 shows the measured throughput, latency and cost
802.15.4 CSMA/CA as the backoff timer is only dependent values. Next, we vary the number of nodes and record the
on the value of BE (binary exponential backoff). throughput, latency and the cost in terms of energy as shown
The energy consumption in TSCH CSMA/CA (9) is lower in Fig.7, Fig. 8 and Fig. 9. Within the considered timeline,
than IEEE 802.15.4 CSMA/CA (14) due to the additional we observe (Fig. 7) that the throughput difference between
CCA prior to transmission attempt. Also, transmissions in the MAC protocols is significant. The network complexity
dedicated communication links (10) consume considerably and channel contention increase with size, resulting in a
low energy. Retransmissions in IEEE 802.15.4 CSMA/CA longer delay in transmissions. Thus, the average throughput
is a repetition of the entire CSMA/CA transmission mech- decreases marginally in the network, degrading the perfor-
anism. However, in TSCH retransmission, energy con- mance of the MAC protocols. TSCH outperforms the other
sumption after performing CCA (12) is still lower than MAC protocols due to dedicated multichannel transmission
802.15.4 based CSMA/CA. Therefore, energy consumption links that ensure no retransmission and collision in the chan-
during transmissions is lower in TSCH mode compared to nel. Also, both 802.15.4 and DSME MAC modes utilize
IEEE 802.15.4 MAC. slotted CSMA/CA in their respective CAP period, which
induces backoff delay with the increase in channel contention.
VI. RESULTS AND DISCUSSION Fig. 8 shows the latency comparison between the MAC
In this section, the IEEE 802.15.4 MAC, DSME and TSCH protocols. Both TSCH and DSME facilitate low-latency
MAC modes are evaluated and compared based on their through the multichannel approach and the presence of dedi-
QoS performance metrics like latency, throughput and cost cated links in terms of timeslots and DSME-GTS. However,
in terms of energy. We consider an IEEE 802.15.4 network 802.15.4 MAC has limited GTS slots and the primary trans-
topology, as shown in Fig. 4 with c1, c2, c3, c4, c5 and mission mechanism is slotted CSMA/CA. With the increase
c6 acting as coordinators, while rest of nodes (21 devices) in network size, the delay time increases with the increase in
are the end-devices. The superframe configuration of the the hop number and retransmission of frames.

41944 VOLUME 8, 2020

N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

FIGURE 7. Performance of MAC protocol in terms of throughput. FIGURE 10. Performance of MAC protocol in terms of energy
consumption and latency.

FIGURE 8. Performance of MAC protocol in terms of latency.

FIGURE 11. Performance of MAC protocol in terms of energy
consumption and throughput.

TABLE 4. Simulation parameters.

FIGURE 9. Performance of MAC protocol in terms of energy consumption. functions and is suitable for applications with relaxed
throughput and latency. Fig. 10 and Fig. 11 show the
associated cost in achieving a desired level of latency and
Although TSCH and DSME achieve high throughput throughput respectively. For relaxed latency and throughput,
and low latency compared to 802.15.4 MAC, the associ- 802.15.4 MAC consumes lower energy compared to DSME
ated cost is higher in terms of energy consumption. Fig. 9 and TSCH. However, in the process of achieving low-latency
shows the cost in energy consumed by the MAC proto- or high throughput, the 802.15.4 MAC increases its active
cols. Due to the simplicity in transmission and the presence period (thereby decreasing the sleep period) to accommodate
of the sleep cycle in the 802.15.4 MAC superframe struc- more incoming frames. This results in higher cost compared
ture, devices considerably reduces their power consumption. to TSCH and DSME modes after a certain level of desired
Devices in TSCH based networks may sleep in between latency and throughput.
their transmissions links and generally remain active for Next, we conduct experiments on OMNeT++ [74] sim-
data transmissions. The 802.15.4 MAC primarily focuses on ulator and 6TiSCH [75] simulator for DSME and TSCH
energy-efficient mechanisms to operate various network based networks, respectively, to evaluate their performance.

VOLUME 8, 2020 41945

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

FIGURE 12. Performance of DSME based networks in terms of latency FIGURE 15. Performance of TSCH based networks in terms of throughput
and incurred cost. and incurred cost.

in our experiment. From Fig. 12 and Fig. 13, it can be

noticed that with CAP reduction, higher throughput and lower
latency can be achieved as transmissions take place through
the DSME-GTS only. However, this results in considerably
higher energy consumption than normal DSME. Also, with
stringent requirements of latency and throughput, power con-
sumption increases. The devices have to remain active for a
longer duration to ensure transmissions to/from other devices.
Therefore, in such cases, the network needs to operate in
higher duty-cycles.
Secondly, to evaluate the performance of TSCH based
networks, we simulated three different scenarios with slot-
frame lengths of 10, 20 and 30 timeslots. Fig. 14 shows the
FIGURE 13. Performance of DSME based networks in terms of throughput increase in power consumption with lower latency guaran-
and incurred cost.
tees. The frequent slot allocation for the devices is required
to maintain strict latency throughout the network, resulting
in higher energy consumption. Similarly, high throughput
results in higher consumption of energy, as shown in Fig. 15.
Higher throughput demands an increase in the number of
frame transmission within the same amount of time. This is
achieved by scheduling frequent, dedicated links with mini-
mal empty timeslots. This also minimizes retransmission of
frames, which in turn assists in reducing latency.

VII. SUITABLE MAC MODES FOR VARIOUS

APPLICATIONS
In this section, we present QoS features considered by each
of the MAC modes and the different industrial applications
FIGURE 14. Performance of TSCH based networks in terms of latency and
suited to these MAC behaviors. The 802.15.4 MAC was
incurred cost. initially designed for all LR-WPANs and wireless sensor
networks that are typically comprised of resource-constrained
This is done due to unavailability of a single simulator devices. It is one of the widely adopted standards for real-
modeling all the three MAC modes. OpenDSME [76], an izing IP based IoT applications that have flexible through-
open-source portable implementation of IEEE 802.15.4 put and latency requirements. Network topologies operating
DSME, is imported in the OMNeT++ to realize the DSME the 802.15.4 MAC are currently used in oil/refinery indus-
MAC mode. We use an IEEE 802.15.4 network topology tries, agricultural implementation, factory automation, smart
as shown in Fig. 4. We set the same simulator parameters, city, smart home applications, etc. However, to support spe-
wherever feasible for both the simulators. Table 4 presents cific QoS requirements of applications, several new MAC
the parameter values of the simulation. behaviors were designed in IEEE 802.15.4e. The RFID
First, we use OMNeT++ to simulate our DSME based Blink mode is used for tracking and identification purposes.
network. We consider both CAP reduction and normal DSME It is also integrated with WSNs for tagging and identifying

41946 VOLUME 8, 2020

N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

TABLE 5. List of the supported features by the MAC modes.

goods [77], [78]. These networks can have very long network Finally, TSCH MAC mode is designed to serve appli-
lifetime. cations requiring high reliability and time-critical assur-
The AMCA MAC mode is suited for non-beacon PANs ances. It is suitable in sensor-actuator networks in oil and
and targets large deployments like infrastructure monitor- gas refineries where strict safety assurances are to be met
ing networks, smart utility networks, etc. These applications and maintained for both human and environmental safety.
require multi-channel and link adaptations [79] to communi- Other applications are equipment and process monitoring like
cate between several devices without compromising on net- food and chemical products, pharmaceuticals, water treat-
work performance. However, they operate in the non-beacon ment, etc. Such networks are prone to interference from
mode of operation, resulting in considerable energy dissipa- other similar networks that negatively affect the perfor-
tion. This is because the devices remain active throughout mance of the wireless devices. TSCH, with its frequency
their lifetime without the support of any synchronization hopping mechanism, mitigates the effects of such interfer-
mechanism. ence and fading link qualities, thus, improving the robust-
Further, DSME MAC was designed to cater to the require- ness of the network. Also, the absence of a long sleep
ments of applications with low and deterministic latency, period for the devices restricts network lifetime. We sum-
energy efficiency, scalability, and high reliability and robust- marize a list of supported QoS and applications suited to the
ness. Considering the criticality of exchanged data, applica- IEEE 802.15.4 MAC, RFID Blink, AMCA, LLDN, DSME,
tions like industrial automation and process control are highly and TSCH MAC modes of IEEE 802.15.4e in Table 5.
sensitive to any loss of data. Also, health-care monitoring The new MAC modes will have superior performance
systems need to guarantee low-latency for data transmis- in terms of latency, throughput, reliability, and robustness.
sions. Further, many applications like outdoor surveillance However, performance guarantees of the new MAC modes
require large and dense deployment. DSME MAC mode should not result in its application by default. Simple and
provides the solution to all such QoS requirements through ease of implementation of IEEE 802.15.4 MAC have been
the presence of a high number of GTSs, which is achieved shown to perform better in applications with non real-time
through a multi-channel approach. Also, the channel adap- requirements. For example, an application with deterministic
tation feature in DSME increases the robustness of the and low-latency requirements may consider either DSME or
network. TSCH MAC over 802.15.4 MAC. However, if the network
The LLDN mode target applications demanding central- is prone to interference and has distinct deadlines for dif-
ized control, low-latency, and robustness. For example, ter- ferent data traffic flows, TSCH is more suited than DSME.
rain survey [80] capturing large geographical areas will best Although a better performance in such a network scenario
be served with the LLDN mode of MAC behavior. This MAC is achieved, energy consumption can be higher than DSME
mode is based on star topology supporting the connectivity MAC. LLDN MAC may not be suitable if the application
of more than 100 devices to the central device. Single hop requires devices to be connected in a multi-hop scenario. Fur-
communication also helps in achieving low and deterministic ther, if the network desires for a reasonable lifetime, AMCA
latency. Data frames are re-transmitted for failed transmis- will not be a suitable option for the power-constrained,
sions. ACK frames and retransmissions increases the relia- battery-operated devices. Delayed data can either be use-
bility of networks operating MAC modes like IEEE 802.15.4, less or detrimental to the deployed geographical area. Thus,
DSME, LLDN, and TSCH. the choice of MAC plays a critical role in determining the

VOLUME 8, 2020 41947

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

overall network performance and safety of the application. [4] W. Wei, Z. Sun, H. Song, H. Wang, X. Fan, and X. Chen, ‘‘Energy
On the other hand, for a network with relaxed throughput balance-based steerable arguments coverage method in WSNs,’’ IEEE
Access, vol. 6, pp. 33766–33773, 2018.
and latency requirements, the IEEE 802.15.4 MAC will be [5] M. Zhu, L. Chang, N. Wang, and I. You, ‘‘A smart collaborative routing
better suited than the DSME or TSCH modes in terms of protocol for delay sensitive applications in industrial IoT,’’ IEEE Access,
energy consumption. This is also observed from Fig. 10 and vol. 8, pp. 20413–20427, 2020.
[6] I. You, G. Pau, V. M. Salerno, and V. Sharma, ‘‘Special issue ‘Internet of
Fig. 11. Applications like industrial monitoring and control Things for smart homes,’’’ Sensors, vol. 19, no. 9, p. 4173, Sep. 2019.
have dynamic traffic requirements at different phases of oper- [7] A. K. Sangaiah, D. V. Medhane, G. Bian, A. Ghoneim, M. Alrashoud, and
ation. The duty-cycle of the devices and the transmission M. S. Hossain, ‘‘Energy-aware green adversary model for cyber physical
security in industrial system,’’ IEEE Trans Ind. Informat., to be published.
schedule needs to be adapted for optimal power consumption.
[8] W. Li, C. Zhu, V. C. M. Leung, L. T. Yang, and Y. Ma, ‘‘Performance com-
Overhead in duty-cycling followed by updating the transmis- parison of cognitive radio sensor networks for industrial IoT with differ-
sion [18] schedule is comparatively lower in 802.15.4 MAC ent deployment patterns,’’ IEEE Syst. J., vol. 11, no. 3, pp. 1456–1466,
Sep. 2017.
than DSME and TSCH based networks. Networks that pri-
[9] R. Lacuesta, J. Lloret, M. Garcia, and L. Peñalver, ‘‘A secure protocol for
marily require to operate under very low duty-cycle have spontaneous wireless ad hoc networks creation,’’ IEEE Trans. Parallel
benefited through the use of 802.15.4 MAC, allowing long Distrib. Syst., vol. 24, no. 4, pp. 629–641, Apr. 2013.
network lifetime. [10] L. D. Xu, E. L. Xu, and L. Li, ‘‘Industry 4.0: State of the art and future
trends,’’ Int. J. Prod. Res., vol. 56, no. 8, pp. 2941–2962, Mar. 2018.
[11] V. Sharma, G. Choudhary, Y. Ko, and I. You, ‘‘Behavior and vulnerability
VIII. CONCLUSION AND FUTURE WORK assessment of drones-enabled industrial Internet of Things (IIoT),’’ IEEE
In this paper, we have presented an analysis of DSME Access, vol. 6, pp. 43368–43383, 2018.
and TSCH MAC and primarily compared against the IEEE [12] IEEE Std 802.15.4: Wireless Medium Access Control (MAC) and Phys-
ical Layer (PHY) Specifications for Low-Rate Wireless Personal Area
802.15.4 MAC. The multi-channel approach in DSME and Networks (WPANs), Standard P802.15.4a/D7, 2007. [Online]. Available:
TSCH allows dedicated and reliable communication between http://www.ieee.org/Standards/
devices. These MAC modes are suitable for networks hin- [13] G. Anastasi, M. Conti, and M. Di Francesco, ‘‘A comprehensive analysis
of the MAC unreliability problem in IEEE 802.15.4 wireless sensor net-
dered with interference from other wireless networks as well works,’’ IEEE Trans Ind. Informat., vol. 7, no. 1, pp. 52–65, Feb. 2011.
as suited for applications with deterministic or low latency [14] L. Wu, S. Zhuo, Z. Wang, and Z. Wang, ‘‘PQueue-MAC: An energy effi-
requirements. However, it introduces the design of complex cient hybrid MAC protocol for event-driven sensor networks,’’ Int. J.
Distrib. Sensor Netw., vol. 11, no. 2, Jan. 2015, Art. no. 160167.
synchronization schedules as well as GTS management in [15] L. Tang, Y. Sun, O. Gurewitz, and D. B. Johnson, ‘‘PW-MAC: An energy-
multi-hop networks to maintain low latency QoS require- efficient predictive-wakeup MAC protocol for wireless sensor networks,’’
ments. Frequent channel adaptations for multiple commu- in Proc. IEEE INFOCOM, Apr. 2011, pp. 1305–1313.
[16] N. Choudhury, R. Matam, M. Mukherjee, and L. Shu, ‘‘Beacon synchro-
nications and lack of a complete sleep period may increase nization and duty-cycling in IEEE 802.15.4 cluster-tree networks: A
the energy consumption of such a network setup. In addition, review,’’ IEEE Internet Things J., vol. 5, no. 3, pp. 1765–1788, Jun. 2018.
a Markov model for TSCH CSMA/CA and retransmission is [17] A. N. Alvi, S. H. Ahmed, M. A. Yaqub, N. Javaid, S. H. Bouk, and
D. Kim, ‘‘An improved IEEE 802.15.4 superframe structure with minimum
proposed to estimate energy consumption and transmission
delay and maximum CFP link utilization,’’ Adhoc Sensor Wireless Netw.,
time. Finally, we present a summary of the supported QoS vol. 35, no. 2, pp. 151–171, Dec. 2017.
features of different MAC modes along with their suitable [18] N. Choudhury, R. Matam, M. Mukherjee, and J. Lloret, ‘‘LBS: A beacon
application areas. A trade-off of choosing a particular MAC synchronization scheme with higher schedulability for IEEE 802.15.4
Cluster-Tree-Based IoT applications,’’ IEEE Internet Things J., vol. 6,
mode over others is to be made based on the QoS require- no. 5, pp. 8883–8896, Oct. 2019.
ments of an application. [19] A. Koubâa, A. Cunha, M. Alves, and E. Tovar, ‘‘TDBS: A time division
The IEEE 802.15.4-2015 standard has been recently beacon scheduling mechanism for ZigBee cluster-tree wireless sensor
networks,’’ Real-Time Syst., vol. 40, no. 3, pp. 321–354, Oct. 2008.
ratified and do not specify the implementations of several [20] P. Muthukumaran, R. de Paz, R. Spinar, and D. Pesch, ‘‘Meshmac:
mechanisms, like DSME-GTS allocations and TSCH link Enabling mesh networking over IEEE 802.15.4 through distributed beacon
scheduling. Also, the challenges highlighted in the paper scheduling,’’ in Proc. Int. Conf. AdHoc Netw., Sep. 2009, pp. 561–575.
[21] B. Nefzi, D. Khan, and Y.-Q. Song, ‘‘TBoPS: A tree based distributed
may delay the adoption of 802.15.4e as the de-facto com- beacon only period scheduling mechanism for IEEE 802.15.4,’’ in Proc.
munication standard for future IoT applications. Neverthe- IEEE 8th Int. Conf. Distrib. Comput. Sensor Syst., May 2012, pp. 341–346.
less, the revised standard aims to achieve highly reliable [22] M. S. Akbar, H. Yu, and S. Cang, ‘‘TMP: Tele-medicine protocol for
slotted 802.15.4 with duty-cycle optimization in wireless body area sensor
and efficient communication for applications with specific networks,’’ IEEE Sensors J., vol. 17, no. 6, pp. 1925–1936, Mar. 2017.
QoS requirements. As a future work, we intend to develop [23] N. Choudhury, R. Matam, M. Mukherjee, and L. Shu, ‘‘Adaptive duty
a DSME-GTS as well as TSCH slot scheduling mechanism cycling in IEEE 802.15.4 cluster tree networks using MAC parameters,’’
for multi-hop networks operating with different traffic flows. in Proc. 18th ACM Int. Symp. Mobile Ad Hoc Netw. Comput. (Mobihoc),
Jul. 2017, pp. 37–38.
[24] C. H. S. Oliveira, Y. Ghamri-Doudane, and S. Lohier, ‘‘A duty cycle self-
REFERENCES adaptation algorithm for the 802.15.4 wireless sensor networks,’’ in Proc.
[1] J. A. Stankovic, ‘‘Research directions for the Internet of Things,’’ IEEE Global Inf. Infrastruct. Symp. (GIIS), Oct. 2013, pp. 1–7.
Internet Things J., vol. 1, no. 1, pp. 3–9, Feb. 2014. [25] P. Park, S. C. Ergen, C. Fischione, and A. Sangiovanni-Vincentelli, ‘‘Duty-
[2] B. Unlu, B. Ozceylan, and B. Baykal, ‘‘DIVVY: An efficient shared cell cycle optimization for IEEE 802.15.4 wireless sensor networks,’’ ACM
scheduling method and algorithm for 6TiSCH-based IoT networks,’’ IEEE Trans. Sensor Netw., vol. 10, no. 1, pp. 1–32, Nov. 2013.
Trans. Green Commun. Netw., vol. 3, no. 1, pp. 170–179, Mar. 2019. [26] H. Tavakoli, J. Misic, V. B. Misic, and M. Naderi, ‘‘Energy-efficient
[3] H. Dou, Y. Qi, W. Wei, and H. Song, ‘‘Carbon-aware electricity cost min- cluster-head rotation in beacon-enabled IEEE 802.15.4 networks,’’
imization for sustainable data centers,’’ IEEE Trans. Sustain. Comput., IEEE Trans. Parallel Distrib. Syst., vol. 26, no. 12, pp. 3371–3380,
vol. 2, no. 2, pp. 211–223, Apr. 2017. Dec. 2015.

41948 VOLUME 8, 2020

N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

[27] N. Choudhury, R. Matam, M. Mukherjee, and J. Lloret, ‘‘A non-threshold- [50] A. Elsts, X. Fafoutis, J. Pope, G. Oikonomou, R. Piechocki, and
based cluster-head rotation scheme for IEEE 802.15.4 cluster-tree net- I. Craddock, ‘‘Scheduling high-rate unpredictable traffic in IEEE
works,’’ in Proc. IEEE Global Commun. Conf. (GLOBECOM), Dec. 2018, 802.15.4 TSCH networks,’’ in Proc. 13th Int. Conf. Distrib. Comput.
pp. 1–6. Sensor Syst. (DCOSS), Jun. 2017, pp. 3–10.
[28] W. R. Heinzelman, A. Chandrakasan, and H. Balakrishnan, ‘‘Energy- [51] G. Alderisi, G. Patti, O. Mirabella, and L. L. Bello, ‘‘Simulative assess-
efficient communication protocol for wireless microsensor networks,’’ in ments of the IEEE 802.15.4e DSME and TSCH in realistic process automa-
Proc. Syst. Sci. 33rd Annual Hawaii Int. Conf., Jan. 2000, pp. 10–18. tion scenarios,’’ in Proc. IEEE 13th Int. Conf. Ind. Informat. (INDIN),
[29] A. B. Bomgni, M. L. F. Sindjoung, A. B. Bomgni, E. T. Fute, Jul. 2015, pp. 948–955.
G. Chalhoub, and C. T. Djamegni, ‘‘ISCP: An instantaneous and [52] P. Sahoo, S. Pattanaik, and S.-L. Wu, ‘‘A reliable data transmission model
secure clustering protocol for wireless sensor networks,’’ Netw. Protocols for IEEE 802.15.4e enabled wireless sensor network under WiFi interfer-
Algorithms, vol. 10, no. 1, pp. 65–82, Apr. 2018. ence,’’ Sensors, vol. 17, no. 6, p. 1320, Jun. 2017.
[30] IEEE Standard. 802.15.4e, Part. 15.4: Low-Rate Wireless Personal Area [53] P. Sahoo, S. Pattanaik, and S.-L. Wu, ‘‘A novel IEEE 802.15.4e DSME
Networks (LR-WPANs) Amendment: MAC Sub-Layer, Standard 802.15.4e- MAC for wireless sensor networks,’’ Sensors, vol. 17, no. 12, p. 168,
2012, 2012. [Online]. Available: http://www.ieee.org/Standards/ Jan. 2017.
[54] S. Duquennoy, B. Al Nahas, O. Landsiedel, and T. Watteyne, ‘‘Orchestra:
[31] IEEE Standard for Low-Rate Wireless Networks, Standard
Robust mesh networks through autonomously scheduled TSCH,’’ in Proc.
IEEE 802.15.4:2015, Apr. 2016, pp. 1–709. [Online]. Available:
Conf. Embedded Netw. Sensor Syst., Nov. 2015, pp. 337–350.
http://www.ieee.org/Standards/
[55] X. Fafoutis, A. Elsts, G. Oikonomou, R. Piechocki, and I. Craddock,
[32] J. Wu, S. Guo, J. Li, and D. Zeng, ‘‘Big data meet green challenges: Big ‘‘Adaptive static scheduling in IEEE 802.15.4 TSCH networks,’’ in Proc.
data toward green applications,’’ IEEE Syst. J., vol. 10, no. 3, pp. 888–900, IEEE 4th World Forum Internet Things (WF-IoT), Feb. 2018, pp. 263–268.
Sep. 2016. [56] R. Soua, P. Minet, and E. Livolant, ‘‘Wave: A distributed scheduling algo-
[33] H. Kurunathan, R. Severino, A. Koubaa, and E. Tovar, ‘‘IEEE 802.15.4e in rithm for convergecast in IEEE 802.15.4e TSCH networks,’’ Trans. Emerg.
a nutshell: Survey and performance evaluation,’’ IEEE Commun. Surveys Telecommun. Technol., vol. 27, no. 4, pp. 557–575, Oct. 2015.
Tuts., vol. 20, no. 3, pp. 1989–2010, 3rd Quart., 2018. [57] I. Juc, O. Alphand, R. Guizzetti, M. Favre, and A. Duda, ‘‘Stripe: A dis-
[34] Zigbee Specification, Alliance, ZigBee, and Others. [Online]. Available: tributed scheduling protocol for 802.15.4e TSCH networks,’’ INRIA,
http://www.zigbee.org/download/standards-zigbee-specification/ France, LIG Res. Rep. RR-LIG-54 HAL, Sep. 2017. [Online]. Available:
[35] Z. Shelby and C. Bormann, 6LoWPAN: The Wireless Embedded Internet, https://hal.inria.fr/hal-01585904
vol. 43. Hoboken, NJ, USA: Wiley, 2011. [58] C.-F. Shih, A. E. Xhafa, and J. Zhou, ‘‘Practical frequency hopping
[36] D. Chen, M. Nixon, and A. W. Mok, Real-Time Mesh Network for Indus- sequence design for interference avoidance in 802.15.4e TSCH networks,’’
trial Automation. Cham, Switzerland: Springer, Apr. 2010, pp. 1–6. in Proc. Int. Conf. Commun. (ICC), Sep. 2015, pp. 6494–6499.
[37] F. Chraim, Y. B. Erol, and K. Pister, ‘‘Wireless gas leak detection and [59] P. H. Gomes, T. Watteyne, and B. Krishnamachari, ‘‘MABO-TSCH: Mul-
localization,’’ IEEE Trans Ind. Informat., vol. 12, no. 2, pp. 768–779, tihop and blacklist-based optimized time synchronized channel hopping,’’
Apr. 2016. Trans. Emerg. Telecommun. Technol., vol. 29, no. 7, p. e3223, Aug. 2017.
[60] T. Chang, T. Watteyne, K. Pister, and Q. Wang, ‘‘Adaptive synchronization
[38] M. Gaafar and G. G. Messier, ‘‘Petroleum refinery multiantenna propa-
in multi-hop TSCH networks,’’ Comput. Netw., vol. 76, pp. 165–176,
gation measurements,’’ IEEE Antennas Wireless Propag. Lett., vol. 15,
Jan. 2015.
pp. 1365–1368, 2016.
[61] A. K. Raghu, A. E. Xhafa, Z. Jianwei, R. Vedantham, and X. Lu, ‘‘Slot
[39] A. Shukla and H. Karki, ‘‘Application of robotics in onshore oil and gas skipping techniques for reduced power consumption in time slotted chan-
industry—A review Part I,’’ Robot. Auto. Syst., vol. 75, pp. 490–507, nel hopping MAC protocol,’’ U.S. Patent 9 854 516, Dec. 26, 2017.
Jan. 2016. [62] Y.-S. Lee and S.-H. Chung, ‘‘An efficient distributed scheduling algorithm
[40] M. Bacco, A. Berton, A. Gotta, and L. Caviglione, ‘‘IEEE 802.15.4 air- for mobility support in IEEE 802.15.4e DSME-based industrial wireless
ground UAV communications in smart farming scenarios,’’ IEEE Commun. sensor networks,’’ Int. J. Distrib. Sensor Netw., vol. 12, no. 2, Jan. 2016,
Lett., vol. 22, no. 9, pp. 1910–1913, Sep. 2018. Art. no. 9837625.
[41] N. Ahmed, D. De, and I. Hussain, ‘‘Internet of Things (IoT) for smart [63] H. Kurunathan, R. Severino, A. Koubâa, and E. Tovar, ‘‘RPL
precision agriculture and farming in rural areas,’’ IEEE Internet Things J., over DSME: A technical report,’’ CISTER-Res. Centre Realtime
vol. 5, no. 6, pp. 4890–4899, Dec. 2018. Embedded Comput. Syst., Porto, Portugal, Tech. Rep. CISTER-
[42] A. S. Bharadwaj, R. Rego, and A. Chowdhury, ‘‘IoT based solid waste TR-180110, 2018. [Online]. Available: http://www.cister.isep.
management system: A conceptual approach with an architectural solution ipp.pt/docs/rpl_over_dsme__a_technical_report/1347/view.pdf
as a smart city application,’’ in Proc. IEEE Annu. India Conf. (INDICON), [64] H. Kurunathan, R. Severino, A. Koubaa, and E. Tovar, ‘‘Poster abstract:
Dec. 2016, pp. 1–6. An efficient approach to multisuperframe tuning for DSME networks,’’
[43] J. Rasouli, A. Motamedi, M. Baseri, and M. Parsa, ‘‘A reliable commu- in Proc. 17th ACM/IEEE Int. Conf. Inf. Process. Sensor Netw. (IPSN),
nication model based on IEEE802.15.4 for WSANs in smart grids,’’ Apr. 2018, pp. 162–163.
in Research Trends Challenges Smart Grids. Rijeka, Croatia: InTech, [65] C. Cano and D. Malone, ‘‘A learning approach to decentralised beacon
May 2019, doi: 10.5772/intechopen.84288. scheduling,’’ Ad Hoc Netw., vol. 49, pp. 58–69, Oct. 2016.
[44] D. De Guglielmo, S. Brienza, and G. Anastasi, ‘‘IEEE 802.15.4e: A sur- [66] P. Li, T. Vermeulen, H. Liy, and S. Pollin, ‘‘An adaptive channel selection
vey,’’ Comput. Commun., vol. 88, pp. 1–24, Aug. 2016. scheme for reliable TSCH-based communication,’’ in Proc. Int. Symp.
Wireless Commun. Syst. (ISWCS), Aug. 2015, pp. 511–515.
[45] A. Moravejosharieh and J. Lloret, ‘‘A survey of IEEE 802.15.4 effective
[67] T. Watteyne, A. Mehta, and K. Pister, ‘‘Reliability through frequency
system parameters for wireless body sensor networks,’’ Int. J. Commun.
diversity: Why channel hopping makes sense,’’ in Proc. Symp. Perform.
Syst., vol. 29, no. 7, pp. 1269–1292, Dec. 2015.
Eval. Wireless Ad Hoc, Sensor, Ubiquitous Netw., Oct. 2009, pp. 116–123.
[46] G. Patti and L. Lo Bello, ‘‘A priority-aware multichannel adaptive frame- [68] I. Juc, O. Alphand, R. Guizzetti, M. Favre, and A. Duda, ‘‘Energy con-
work for the IEEE 802.15.4e-LLDN,’’ IEEE Trans. Ind. Electron., vol. 63, sumption and performance of IEEE 802.15.4e TSCH and DSME,’’ in Proc.
no. 10, pp. 6360–6370, Oct. 2016. IEEE Wireless Commun. Netw. Conf., Apr. 2016, pp. 1–7.
[47] C. Ouanteur, D. Aïssani, L. Bouallouche-Medjkoune, M. Yazid, and [69] D. De Guglielmo, G. Anastasi, and A. Seghetti, ‘‘From IEEE 802.15. 4 to
H. Castel-Taleb, ‘‘Modeling and performance evaluation of the IEEE IEEE 802.15. 4e: A step towards the Internet of Things,’’ in Proc. Adv. Onto
802.15.4e LLDN mechanism designed for industrial applications in Internet Things Cham, Switzerland: Springer, Jan. 2014, pp. 135–152.
WSNs,’’ Wireless Netw., vol. 23, no. 5, pp. 1343–1358, Feb. 2016. [70] H. Rasouli, Y. S. Kavian, and H. F. Rashvand, ‘‘ADCA: Adaptive duty
[48] A. Elsts, X. Fafoutis, R. Piechocki, and I. Craddock, ‘‘Adaptive channel cycle algorithm for energy efficient IEEE 802.15.4 beacon-enabled wire-
selection in IEEE 802.15.4 TSCH networks,’’ in Proc. Global Internet less sensor networks,’’ IEEE Sensors J., vol. 14, no. 11, pp. 3893–3902,
Things Summit (GIoTS), Jun. 2017, pp. 1–6. Nov. 2014.
[49] D. Zorbas, V. Kotsiou, F. Théoleyre, G. Z. Papadopoulos, and [71] F. Kauer, M. Köstler, and V. Turau, ‘‘Reliable wireless multi-hop
C. Douligeris, ‘‘Lost: Localized blacklisting aware scheduling algorithm networks with decentralized slot management: An analysis of IEEE
for IEEE 802.15. 4-TSCH networks,’’ in Proc. Wireless Days (WD), 802.15.4 DSME,’’ 2018, arXiv:1806.10521. [Online]. Available:
May 2018, pp. 110–115. http://arxiv.org/abs/1806.10521

VOLUME 8, 2020 41949

 N. Choudhury et al.: Performance-to-Cost Analysis of IEEE 802.15.4 MAC With 802.15.4e MAC Modes

[72] T. Watteyne, M.-R. Palattella, and L. A. Grieco, ‘‘Using IEEE 802.15.4e MITHUN MUKHERJEE (Member, IEEE)
time-slotted channel hopping (TSCH) in the Internet of Things (IoT): received the B.E. degree in electronics and
Problem statement,’’ Internet Requests Comments, RFC Editor, RFC, communication engineering from the Univer-
INRIA, France, Tech. Rep. 7554, May 2015. [Online]. Available: sity Institute of Technology, Burdwan University,
https://tools.ietf.org/html/rfc7554 Bardhaman, India, in 2007, the M.E. degree in
[73] S. Gandham, M. Dawande, and R. Prakash, ‘‘Link scheduling in wireless information and communication engineering from
sensor networks: Distributed edge-coloring revisited,’’ J. Parallel Distrib. the Indian Institute of Science and Technology,
Comput., vol. 68, no. 8, pp. 1122–1134, Aug. 2008.
Shibpur, India, in 2009, and the Ph.D. degree in
[74] OMNeT++ Network Simulator. Accessed: Dec. 25, 2019. [Online]. Avail-
electrical engineering from IIT Patna, Patna, India,
able: https://omnetpp.org/download/
[75] Simulator for TSCH-Based Networks. Accessed: Dec. 25, 2019. [Online]. in 2015. He is currently an Assistant Professor
Available: https://bitbucket.org/6tisch/simulator/src/master/ with the Guangdong Provincial Key Laboratory of Petrochemical Equip-
[76] Opendsme: Open Source Portable Implementation of IEEE ment Fault Diagnosis, Guangdong University of Petrochemical Technol-
802.15.4 DSME. Accessed: Dec. 25, 2019. [Online]. Available: ogy, Maoming, China. He has (co)authored more than 80 publications
https://github.com/openDSME/inet-dsme/ in peer-reviewed international transactions/journals and conferences. His
[77] Y. Xiao, X. Shen, B. Sun, and L. Cai, ‘‘Security and privacy in RFID current research interests include wireless communications, fog computing,
and applications in telemedicine,’’ IEEE Commun. Mag., vol. 44, no. 4, and ultrareliable low-latency communications. He was a recipient of the
pp. 64–72, Apr. 2006. 2016 EAI International Wireless Internet Conference, the 2017 International
[78] D. Williams. The Strategic Implications of Wal-Mart’s RFID Man- Conference on Recent Advances on Signal Processing, Telecommunications
date. Accessed: Aug. 2019. [Online]. Available: https://github.com/ and Computing, the 2018 IEEE SYSTEMS JOURNAL, and the 2018 IEEE
Nikumani/synchronization International Conference on Advanced Networks and Telecommunications
[79] Y. Li, Y.-Q. Song, R. Schott, Z. Wang, and Y. Sun, ‘‘Impact of link unreli- Systems (ANTS) Best Paper Award. He has been an Associate Editor of
ability and asymmetry on the quality of connectivity in large-scale sensor IEEE ACCESS and a Guest Editor of the IEEE INTERNET OF THINGS JOURNAL,
networks,’’ Sensors, vol. 8, no. 10, pp. 6674–6691, Oct. 2008.
the IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, ACM/Springer Mobile
[80] S. Williams, L. Parker, and A. Howard, ‘‘Terrain reconstruction of glacial
Networks and Applications, and Sensors.
surfaces: Robotic surveying techniques,’’ IEEE Robot. Autom. Mag.,
vol. 19, no. 4, pp. 59–71, Dec. 2012.

JAIME LLORET (Senior Member, IEEE) received

the M.Sc. degree in physics, in 1997, the M.Sc.
NIKUMANI CHOUDHURY (Student Member, degree in electronic engineering, in 2003, and
IEEE) received the B.Tech. degree in informa- the Ph.D. degree in telecommunication engi-
tion technology from Assam University, in 2012, neering, in 2006. He is currently an Associate
and the M.Tech. degree in information technology Professor with the Polytechnic University of
from Gauhati University, in 2014. He is currently Valencia. He was the Director of the University
pursuing the Ph.D. degree with the Computer Sci- Master ‘‘Digital Post Production’’ from 2012 to
ence and Engineering Department, IIIT Guwahati, 2016. He is also the Chair of the Integrated Man-
India. He was a Project Fellow with IIIT Guwahati. agement Coastal Research Institute (IGIC) and the
His research interests include low-power wireless Head of the ‘‘Active and collaborative techniques and use of technologic
sensor networks and the Internet of Things. resources in the education (EITACURTE)’’ Innovation Group. He is also
the Director of the University Diploma ‘‘Redes y Comunicaciones de Orde-
nadores.’’ He has authored 22 book chapters and more than 480 research
articles published in national and international conferences, and international
journals (more than 220 with ISI Thomson JCR). He has been the co-editor
of 40 conference proceedings and a guest editor of several international
books and journals. Since 2016, he has been the Spanish Researcher with
highest H-index in the Telecommunications journal list according to the
Clarivate Analytics Ranking. He is a Senior Member of ACM and a Fellow
RAKESH MATAM (Member, IEEE) received the of IARIA. He was the Vice-Chair of the Europe/Africa Region of Cogni-
B.Tech. degree in computer science and engineer- tive Networks Technical Committee, IEEE Communications Society, from
ing from Jawaharlal Nehru Technological Univer- 2010 to 2012, and the Internet Technical Committee, IEEE Communications
sity Hyderabad, the M.Tech. degree from Kakatiya Society and Internet Society, from 2011 to 2013. He was the Chair of
University, Warangal, India, and the Ph.D. degree the Internet Technical Committee, IEEE Communications Society, and the
in computer science from IIT Patna, in 2014. Internet Society, from 2013 to 2015. He has been the Chair of the Internet
In 2014, he joined the Department of Computer Technical Committee, IEEE Communications Society and Internet Society,
Science, IIIT Guwahati, as an Assistant Profes- from 2013 to 2015. He is also the Chair of the Working Group of the Standard
sor. He is currently a member of the Design and IEEE 1907.1. He is also the Editor-in-Chief of Ad Hoc and Sensor Wireless
Innovation Center, IIIT Guwahati, and a Principal Networks (with ISI Thomson Impact Factor), the international journal Net-
Investigator of a research project sponsored by the Government of India. works Protocols and Algorithms, the International Journal of Multimedia
His research interests include wireless networks, network security, and cloud Communications, and Network Protocols and Algorithms.
computing.

41950 VOLUME 8, 2020