Towards the Internet of Relevant Things: the IEEE 802.15.4e Standard

Towards the
Internet of Relevant Things
The IEEE 802.15.4e Standard
Towards the
Internet of Relevant Things
The IEEE 802.15.4e Standard
PerLab
G. Anastasi, S. Brienza, D. De Guglielmo
Pervasive Computing & Networking Lab (PerLab)
Dept. of Information Engineering, University of Pisa
Website: www.iet.unipi.it/~anastasi/
ACM SAC 2016 | Pisa, Italy, April 4‐8, 2016

PerLab
Towards the Internet of Relevant Things 2
Overview
 Introduction
 IEEE 802.15.4 standard
 limitations
 IEEE 802.15.4e
 802.15.4e DSME mode
 literature survey
 open issues
 802.15.4e TSCH mode
 open issues
 IoT with 802.15.4e TSCH
 The 6TiSCH initiative
 Conclusions

PerLab
References
 D. De Guglielmo, S. Brienza, G. Anastasi, IEEE 802.15.4e: a Tutorial and
Survey, Computer Communications, available on line at:
http://www.sciencedirect.com/science/article/pii/S0140366416301980
 G. Anastasi, D. De Guglielmo, A Seghetti, From IEEE 802.15.4 to IEEE
802.15.4e: a Step towards the Internet of Things, Chapter 10 in Advances
onto the Internet of Things, pp. 135‐152, January 2014. Springer

PerLab
Internet of Things
“The next logical step in the technological revolution connecting people anytime,
anywhere is to connect inanimate objects. This is the vision underlying the
Internet of things: anytime, anywhere, by anyone and anything”
(ITU, Nov. 2005)
More than 26 billion devices will
be wirelessly connected to the
Internet of Things by 2020
 computers and communication devices
 cars, robots, machine tools
 persons, animals, and plants
 garments, food, drugs, etc.

PerLab
Wireless Sensor & Actuator Networks (WSANs)
“WSANs will behave as a digital skin for the IoT
providing a virtual layer through which any
computational system can interact with the physical
world”
C. Alcaraz, P. Najera, J. Lopez, R. Roman, Wireless Sensor Networks and the Internet of Things: Do we need a Complete
Integration? Proc. International workshop on the Security of The internet of Things (SecIoT 2010), Tokyo, Japan, 2010.

PerLab
Standards for WSANs
 IEEE 802.15.4/802.15.4e
 ZigBee
 Bluetooth
 WirelessHART
 ISA‐100.11a
 6LoWPAN
 IPv6 over Low power WPAN
 RPL
 Routing Protocol for Low power
and Lossy networks
 CoAP
 Constrained Application Protocol

IEEE 802.15.4 StandardIEEE 802.15.4 Standard
Reference technology for WSANs
Expected to be a major enabling technology
also for IoT

PerLab
IEEE 802.15.4 standard
 Standard for Personal Area Networks (PANs)
 low‐rate and low‐power
 PHY and MAC layers
 Main features
 transceiver management
 channel access
 PAN management

PerLab
802.15.4 Network Topologies
Ed Callaway, Paul Gorday, Lance Hester, Jose A. Gutierrez, Marco Naeve, Bob Heile, Home Networking with IEEE 802.15.4: Developing
Standard for Low‐Rate Wireless Personal Area Networks, IEEE Communications Magazine, August 2002.

PerLab
Channel frequencies
(20 Kbps) (40 Kbps)
(250 Kbps)
Europe USA
Ed Callaway, Paul Gorday, Lance Hester, Jose A. Gutierrez, Marco Naeve, Bob Heile, Home Networking with IEEE 802.15.4: Developing
Standard for Low‐Rate Wireless Personal Area Networks, IEEE Communications Magazine, August 2002.

PerLab
802.15.4 MAC protocols
 Two different channel access methods
 Beacon‐Enabled duty‐cycled mode
 Non‐Beacon Enabled mode (aka Beacon Disabled mode)
Non-beacon enabled Beacon enabled
Superframe Structure
Contention free
Reserved time slot
Contention based
Slotted CSMA-CA
Contention based
Unslotted CSMA-CA

PerLab
802.15.4 Beacon Enabled mode
GTS GTS
CAP CFP
SD = aBaseSuperFrameDuration*2SO
sym
BI = aBaseSuperFrameDuration*2BO
sym
Inactive
Active
Beacon Beacon

PerLab
At each trial the backoff‐
window size is doubled
Only a limited number of
attempts is permitted
(macMaxCSMABackoffs)
CSMA-CA: Beacon-Enabled mode
Yes
No
Yes
No
Yes
No
NB=0, CW=2, BE= BE=macMinBE

PerLab
At each trial the backoff‐
window size is doubled
Only a limited number of
attempts is permitted
(macMaxCSMABackoffs)
CSMA-CA: Non-Beacon Enabled mode
NB = 0,
BE = macMinBE
Delay for
random(2BE
- 1) unit
backoff periods
Perform CCA
Channel idle?
NB = NB+1,
BE = min(BE+1, aMaxBE)
NB>
macMaxCSMABackoffs
?
Failure Success
Un-slotted CSMA
Y
Y
N
N

PerLab
Acknowledgement Mechanism
 Optional mechanism
 Destination Side
 ACK sent upon successful reception of a data frame
 Sender side
 Retransmission if ACK not (correctly) received within the
timeout
 At each retransmission attempt the backoff window size is
re‐initialized
 Only a maximum number of retransmissions allowed
(macMaxFrameRetries)

PerLab
Limits of IEEE 802.15.4 MAC
 Reliability and scalability issues
 Unbounded latency
 Due to contention‐based CSMA‐CA algorithm
 No guaranteed bandwidth
 Unless GTS is used
 GTS only provides a limited service (7 slots)
 No built‐in frequency hopping technique
 Prone to failures due to interferences and multi‐path fading
G. Anastasi, D. De Guglielmo, A Seghetti, From IEEE 802.15.4 to IEEE 802.15.4e: a Step towards the Internet of Things,
Chapter 10 in Advances onto the Internet of Things (S. Gaglio, G. Lo Re, Editors), pp. 135‐152. January 2014. Springer

PerLab
Performance of CSMA-CA in BE mode
G. Anastasi, M. Conti, M. Di Francesco, The MAC Unreliability Problem in IEEE 802.15.4 Wireless Sensor Networks, Proceedings
ACM MSWIM 2009, Tenerife, Spain, October 26‐30, 2009
G. Anastasi, M. Conti, M. Di Francesco, A Comprehensive Analysis of the MAC Unreliability Problem in IEEE 802.15.4 Wireless
Sensor Networks, IEEE Transactions in Industrial Informatics, Vol. 7, N. 1, Feb 2011.

PerLab
 Prone to failure due to interferences and multi‐path
fading

PerLab
Channel Hopping
 Subsequent packets are sent at a different frequency
 following a pseudo‐random hopping pattern
 If a transmission fails
 retransmission will happen on a different frequency
 Greater chance of successful transmission
 than if the retransmission happened on the same channel
 Since a different frequency means different effects of multi‐path fading
and interference
 Increased security
 Against Selective Jamming attack
 Channel Hopping used in
 Bluetooth
 WirelessHart

Key QuestionKey Question
How to do in real‐life
critical scenarios?

PerLab
Some Real-life Scenarios
 Smart Cities
 Smart Buildings
 Smart Homes
 Industrial Settings
 ....
Requirements
 M. Dohler, T. Watteyne, T. Winter, and D. Barthel, Routing Requirements for Urban Low‐Power
and Lossy Networks, IETF ROLL Std. RFC5548, May 2009.
 J. Martocci, P. De Mil, N. Riou, andW. Vermeylen, Building Automation Routing Requirements in
Low‐Power and Lossy Networks, IETF ROLL Std. RFC5867, June 2010.
 A. Brandt, J. Buron, G. Porcu, Home Automation Routing Requirements in Low‐Power and Lossy
Networks, IETF ROLL Std. RFC5826, April 2010
 K. Pister, P. Thubert, S. Dwars, and T. Phinney, Industrial Routing Requirements in Low‐Power
and Lossy Networks, IETF ROLL Std. RFC5673, Octobert 2009.

PerLab
Requirements
 Energy Efficiency
 Target battery lifetime: 5 years, or more
 Scalability
 Large network sizes
 Timeliness
 Alert applications, process monitoring, ...
 Reliability
 Wire‐like reliability may be required, e.g., 99.9% or better

PerLab
IEEE 802.15.4e
 IEEE 802.15 Task Group 4e
 chartered to define a MAC amendment to the existing
standard 802.15.4‐2006.
 The intent of this amendment was to enhance and add
functionalities to the 802.15.4‐2006 MAC
 better support the industrial markets
 increase robustness against external interference
 On February 6, 2012 the IEEE Standards Association
Board approved the IEEE 802.15.4e MAC Enhancement
Standard document for publication.
 http://www.ieee802.org/15/pub/TG4e.html

PerLab
Major Changes
 General functional improvements
 not tied to any specific application domain
 MAC Behavior Modes
 support of specific application domains
Remarks:
Many ideas borrowed from previous industrial
standards
 WirelessHART and ISA 100.11.a
 slotted access, shared and dedicated slots, multi‐channel
communication, and frequency hopping.

PerLab
General Functional Improvements
 Low Energy (LE)
 Information Elements (IE)
 Enhanced Beacons (EB)
 Multipurpose Frame
 MAC Performance Metrics
 Fast Association (FastA)

PerLab
MAC Behavior Modes
 Radio Frequency Identification Blink (BLINK)
 item and people identification, location, and tracking
 Asynchronous Multi‐Channel Adaptation (AMCA)
 application domains where large deployments are required (e.g.,  process
automation/control, infrastructure monitoring, etc.)
 Deterministic & Synchronous Multi‐channel Extension (DSME)
 industrial and commercial applications with stringent timeliness and
reliability requirements
 Low Latency Deterministic Network (LLDN)
 applications requiring very low latency requirement (e.g.,  factory
automation, robot control)
 Time Slotted Channel Hopping (TSCH)
 application domains such as process automation

IEEE 802.15.4e DSMEIEEE 802.15.4e DSME
Deterministic &
Synchronous
Multichannel Extension

PerLab
IEEE 802.15.4e DSME
 Target Application Domains
 Industrial applications
 Commercial applications
 Healthcare applications
 Provides
 Flexibility
 Robustness
 High reliability
 Deterministic latency
 Scalability
 Efficiency

PerLab
Main Features
 Multi‐channel, multi‐superframe
 Mesh extension to GTS
 Two channel diversity modes
 channel adaptation
 channel hopping
 Distributed beacon scheduling
 Distributed slot allocation
 Group acknowledgments
 Many topologies
 Star, cluster‐tree and peer‐to‐peer

PerLab
Multi superframe structure
When CAP reduction is enabled, only the first superframe in the multi‐
superframe has the CAP

PerLab
Enhanced Beacons (EBs)
 Sent by each coordinator at regular intervals
 At the beginning of superframes
 Contains parameters
 To regulate the multi‐superframe structure
 Indicates the enabled options
 Used for global time synchronization
 Each node associates with a coordinator (i.e., its time synchronization
parent) and tracks its EBs
 In order to join the network, a node performs a passive channel scan
over a given list of channels

PerLab
Beacon Scheduling
Each coordinator transmits its EBs during a
superframe
 Distributed scheduling algorithm
 Beacon schedule information shared through EBs
 A bitmap sequence indicates the usage of beacon slots by its one‐hop
neighbors
 A bit in the bitmap is set to one if the corresponding beacon slot is
already used for EB transmission
 A prospective coordinator searches for a vacant slot in the
bitmaps of all the received EBs
 Once the new device finds a vacant beacon slot, it uses the slot
as its own beacon slot

PerLab
Channel Diversity
 Channel Adaptation
 Two neighboring nodes can communicate using any of the free
available channels
 The channel is decided basing on the quality estimated by nodes
 The channel used in a DSME‐GTS does not change over time
 unless its quality degrades
 Channel Hopping
 Nodes change the communication channel at each DSME‐GTS
 The series of channels used is referred to as Hopping Sequence
 Each node has a Channel Offset
 an integer indicating the starting position in the Hopping Sequence
 A transmitting node has to switch to the channel used by the
receiver

PerLab
Channel Adaptation

PerLab
Channel Hopping
Channel number C at the given DSME‐GTS Slot ID i in SDIndex j, shall
be determined as follows:

PerLab
ALLOCATION COUNTER TABLE (ACT)
 A table containing an entry for each Guaranteed TimeSlot
allocated to the node
 Each entry stores the following information:
 Superframe ID
 Slot ID
 Channel ID
 Direction
 Address
Data Structures (1/2)

PerLab
SLOT ALLOCATION BITMAP (SAB)
 A bitmap used to store which GTSs in the
multi‐superframe are allocated to the
node and its neighbors
 The SAB contains a bit for each possible
pair [ch, ts], indicating whether the
channel ch, during timeslot ts, is used (1)
or not (0).
Data Structures (2/2)

PerLab
DSME-GTSs
 A DSME‐GTS is used for communication between a specific
source device and a specific destination device.
 A DSME‐GTS is allocated by the destination device,
basing on
 requirements specified by the source node in the request
 current slot availability
 DSME‐GTSs are allocated on a
FCFS basis
 Three‐step allocation
(during CAP)

PerLab
 B sends a REQUEST in unicast to C
 The REQUEST indicates the available slots for allocation
A B C D
• In this example, B wants to allocate a DSME‐GTS with C
DSME-GTSs: Allocation (1/3)
 After transmission, node B waits for a REPLY from node C
 If the REPLY frame is not received in time, the allocation is
assumed to be failed (i.e., a Timeout occured)

PerLab
A B C D
 Once the REQUEST has been received, C search in its
bitmap for free GTSs in common with B
 If they are found, C updates its own data structures
 C sends a REPLY in Broadcast, indicating the newly
allocated GTSs

PerLab
A B C D
 Upon receiving the REPLY, B updates its data structures
 B broadcasts a NOTIFY message to inform its neighbors
 When receiving a REPLY/NOTIFY, A and D check if they
already use the just allocated GTS:
• If not, they update their SAB
• Otherwise, they send a error message to the sender

Literature Review

PerLab
Performance Evaluation
 G. Alderisi, G. Patti, O. Mirabella, L. Lo Bello, Simulative assessments of the IEEE
802.15.4e DSME and TSCH in realistic process automation scenarios, In Proceedings
of the 13th IEEE International Conference on Industrial Informatics (INDIN),
Cambridge, United Kingdom, 22‐24 July 2015.
 J. Lee, W.C. Jeong, Performance analysis of IEEE 802.15.4e DSME MAC protocol under
WLAN interference, In Proceedings of the IEEE International Conference on ICT
Convergence (ICTC), Jeju Island, South Korea, 15‐17 October 2012.
 W.C. Jeong, J. Lee, Performance evaluation of IEEE 802.15. 4e DSME MAC protocol for
wireless sensor networks, In Proceedings of the first IEEE Workshop on Enabling
Technologies for Smartphone and Internet of Things (ETSIoT), Seoul, South Korea, 18
June 2012.
 T. Paso, J. Haapola, J. Iinatti, Feasibility study of IEEE 802.15.4e DSME utilizing IR‐
UWB and S‐Aloha, In Proceedings of the 24th IEEE International Symposium on
Personal Indoor and Mobile Radio Communications (PIMRC), London, United Kingdom,
8‐11 September 2013.

PerLab
802.15.4e DSME vs. 802.15.4 BE
 DSME always offers a minimum throughput
 only limited by the number of available DSME‐GTSs
 In 802.15.4 BE the delivery ratio decreases when the # of nodes
increases
 DSME guarantees a deterministic delay
 it is unpredictable in 802.15.4 CSMA‐CA (or S‐Aloha)
 DSME shows a significant lower consumption
 as collisions cannot happen and CCA are not needed
 DSME is much more resistant to
external interferences,
 thanks to its channel diversity
mechanisms

PerLab
802.15.4e DSME vs. TSCH
 Both are robust against channel noise
 assuring high reliability
 TSCH offers lower delays
 when the number of nodes is limited
 the DSME multi‐superframe structure does not adapt to the network size
 TSCH needs larger timeslots
 to accommodate ACKs
 DSME can benefit from the Group ACK option
 DSME exhibits lower delays
 when the number of nodes grows
 more DSME‐GTSs in the multi‐superframe are used.
G. Alderisi, G. Patti, O. Mirabella, L. Lo Bello, Simulative assessments of the IEEE 802.15.4e DSME and TSCH in realistic process
automation scenarios, Proc. IEEE International Conference on Industrial Informatics (INDIN 2015), Cambridge, UK, July 2015

PerLab
General Enhancements (1/2)
 ELPIDA (Enhancements for Low‐Power Instrumentation DSME Applications)
 Aims at making DSME more energy efficient in cluster‐tree
networks
 Limits the time spent by leaf nodes radio in RX state (e.g.
during CAP)
 Introduces three main changes to DSME:
 CAP Wake‐up
 DSME‐GTS Wake‐up
 Beacon Look‐up
 ELPIDA decreases the energy consumption of end devices
 up to a factor 7, if compared with standard DSME
S. Capone, R. Brama, F. Ricciato, G. Boggia, and A. Malvasi, Modeling and Simulation of Energy Efficient Enhancements for IEEE
802.15.4e DSME, Proceedings of Wireless Telecommunications Symposium (WTS), Washington, DC, USA, 9‐11 April 2014.

PerLab
General Enhancements (2/2)
 EFastA (Enhanced Fast Association)
 Addresses the association phase
 Many collisions may occur with the association procedure
described in the standard
 Many Beacon Intervals can be necessary to associate all the nodes
to the network
 In EFastA, nodes do not sent association requests
immediately after the EB reception
 Nodes wait for a randomly selected superframe inside the
multi‐superframe
 Simulations show that EFastA reduces association time
 up to about 90%, with respect to FastA
X. Liu, X. Li, S. Su, Z. Fan, G. Wang, Enhanced Fast Association for 802.15.4e‐2012 DSME MAC Protocol, Proceedings of the
2nd International Conference on Computer Science and Electronics Engineering (ICCSEE), Hangzhou, China, 22‐23 March 2013.

PerLab
Beacon Scheduling (1/3)
 The DSME Beacon Scheduling algorithm presents a number
of limitations
 long waiting time for network construction
 storage wastage
 variable and growing length of EB frame size
 potential beacon collision issues

PerLab
 E‐DSME (Enhanced DSME)
 Exploits a distributed permission notification mechanism
 A node waits for a Permission Notification to complete the
allocation
 Only the coordinator that has transmitted the latest beacon can
send permission notifications
 CAP is divided into Allocation Contention Periods (ACP) and
Permission Notification Periods (PNP)
 Simulation results show that E‐DSME allows
 to achieve very high success ratio (near to 100%)
 to avoid beacon collisions even in complex networks
K. Hwang, S. Nam, Analysis and Enhancement of IEEE 802.15.4e DSME Beacon Scheduling Model, Journal of Applied
Mathematics.

PerLab
 DFBS (Distributed Fast Beacon Scheduling)
 many problems are caused by the use of a bitmap to
represent slot allocation data
 The bitmap is replaced by a single integer, namely RINSD
(Representative Indicator of Neighbor Superframe Duration)
 It represents the index of the superframe to be used to send EBs
 Each node maintains its own RINSD
 To join the network, a node performs an Active Association
 Simulation results show that
 allows to solve the highlighted beacon collision problems
 reduces the network construction time (up to 75% with 25 nodes)
W. Lee, K. Hwang, Y.A. Jeon, S.S. Choi, Distributed Fast Beacon Scheduling for Mesh Networks, Proceedings of the 8th IEEE
International Conference on Mobile Adhoc and Sensor Systems (MASS), Valencia, Spain, 17‐22 October 2011.

Open Issues

PerLab
Open Issues (1/2)
 A complete implementation of DSME is still missing
 Current works do not consider mesh topologies
 Many aspects are left to implementers
 How to choose the DSME‐GTSs to be allocated
 How to accommodate traffic in GTSs
 How to select the communication channel when using
channel adaptation
 The beacon scheduling algorithm is presented without any
concrete outline and implementation details
 …

PerLab
Open Issues (2/2)
 More attention should be devoted to the energy topic
 DSME conceals a number of security issues
 The distributed DSME‐GTS allocation mechanism assumes
that all nodes are trustworthy
 Feasibility of the integration with the upper layers of
the network stack should be investigated
 with reference to the Internet of Things protocols,
such as 6LoWPAN, RPL, and CoAP.

IEEE 802.15.4e TSCHIEEE 802.15.4e TSCH
Time Slotted
Channel Hopping

PerLab
Time Slotted Channel Hopping (TSCH)
 Time slotted access, multi‐channel communication and
channel hopping
 Particularly suitable for multi‐hop mesh networks
 Time‐slotted access
 Predictable and bounded latency
 Guaranteed bandwidth
 Multi‐channel communication
 More nodes can communicate at the same time (i.e., same slot) using
different channels (identified by different channel offsets)
 increased network capacity
 Channel hopping
 mitigates the effects of interferences and multipath fading
 improves reliability
 No CAP in TSCH

PerLab
Periodic Slotframe

PerLab
L→B
I→D
E→C
G→C
H→D
C→A
B→A
D→A
0 1 3
timeslot
ChannelOffset 0
1
2
3
4
2
TSCH Link
 Link = Pairwise assignment of a directed communication
between devices in a specific slot, with a given channel offset
 t slot in the slotframe
 chOf channel offset
Dedicated links
Shared link

PerLab
Dedicated vs. Shared Links
 Dedicated links
 Deterministic traffic
 Periodic transmissions
 One transmitter – One
receiver
 Direct access
 Shared links
 Sporadic, unpredictable
traffic
 Discovery and routing
messages
 Multiple
transmitters/receivers
 CSMA‐CA based access

PerLab
TSCH CSMA-CA for Shared Links

PerLab
TSCH CSMA-CA vs. 802.15.4 CSMA-CA
 Backoff mechanism
 is activated only after the node has experienced a collision (to
avoid repeated collisions)
 Backoff unit duration
 TSCH: a shared slot
 802.15.4: 320 sec
 Clear Channel Assessment (CCA)
 CCAs are not used to prevent collisions among nodes, but to avoid
transmitting a packet if a strong external interference is detected
 Packet dropping
 a packet is dropped only if it reaches the maximum number of
retransmissions
D. De Guglielmo, B. Al Nahas, S. Duquennoy, T. Voigt, G. Anastasi, Analysis and Experimental Evaluation of IEEE 802.15.4e TSCH
CSMA‐CA Algorithm, IEEE Transactions on Vehicular Technology, to appear.

PerLab
TSCH Link
 Link=Pairwise assignment of a directed communication
between devices in a specific slot, with a given channel offset
 t slot in the slotframe
 chOf channel offset
L→B
I→D
E→C
G→C
H→D
C→A
B→A
D→A
0 1 3
timeslot
ChannelOffset
0
1
2
3
4
2

PerLab
Frequency Translation - Channel Hopping
The channel offset of a Link = (t, chOf ) is translated in
an operating frequency f
 ASN: total # of slots elapsed since the network was
deployed
 nch: number of used physical channels
 F is implemented as a look‐up‐table containing the set
of available channels

PerLab
t
Slotframe 1 2
ASN 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
channelOffset
0 3
slotOffset 0 1 2 3
TSCH: Multi-channel + Frequency Hopping
4
Channel (f) 5 9
0 1 2 3
13
0 1 2 3
0
1
15
1
0 1 2 3
5
0 1 2 3
9
0 1 2 3
13
0 1 2 3
Slotframe size and nch should be relatively prime
Each link rotates through the nch available channels over nch slotframes.

PerLab
Literature Review
A lot of attention by the research community
 Researches mainly focused on:
 Performance Evaluation of TSCH mode
 Network formation
 Link scheduling
 Network synchronization
 IETF 6TiSCH Initiative
 Integration of TSCH with IoT protocols
 Use of IPv6 on top of 802.15.4e TSCH

Performance
Evaluation

PerLab
Performance Evaluation of TSCH Mode
 S. Zats, R. Su, T. Watteyne, K. Pister, Scalability of Time Synchronized
Wireless Sensor Networking, Proceedings of Annual Conference on IEEE
Industrial Electronics Society (IECON 2011), November 7‐10, 2011.
 S. Chen, T. Sun, J. Yuan, X. Geng, C. Li, U. Sana, M. Abdullah Alnuem,
Performance Analysis of IEEE 802.15.4e Time Slotted Channel Hopping for
Low‐Rate Wireless Networks, KSII Transactions on Internet and Information
Systems, Vol. 7, N. 1, 2013.
 D. De Guglielmo, G. Anastasi, and A. Seghetti, From IEEE 802.15.4 to IEEE
802.15.4e: A Step Towards the Internet of Things, Chapter in Advances onto
the Internet of Things, pp. 135–152, 2014. Springer.
 X. Vilajosana, Q. Wang, F. Chraim, T. Watteyne, C. Tengfei, K. Pister, A
Realistic Energy Consumption Model for TSCH Networks, IEEE Sensors
Journal, 14 (2), 482‐489, Feb. 2014.
 T. Watteyne, J. Weiss, L. Doherty, J. Simon, Industrial IEEE802.15.4e
Networks: Performance and Trade‐offs, Proceedings of IEEE International
Conference on Communications (ICC 2015), June 8‐12, 2015.
 ....

PerLab
802.15.4e TSCH vs. 802.15.4 CSMA-CA
 Comparison
 802.15.4 CSMA‐CA in BE and NBE mode
 802.15.4e in TSCH mode
 Star Network
 Periodic Reporting
 1 packet every T (BI in BE mode)
 TSCH: single channel used
 Performance Metrics
 Delivery Ratio
 Average Latency
 Energy per Delivered Packet
G. Anastasi, D. De Guglielmo, A Seghetti, From IEEE 802.15.4 to IEEE 802.15.4e: a Step towards the Internet of Things, Chapter
10 in Advances onto the Internet of Things (S. Gaglio, G. Lo Re, Editors), pp. 135‐152. January 2014. Springer

PerLab
802.15.4e TSCH vs. 802.15.4 CSMA-CA
Simulation Results
G. Anastasi, D. De Guglielmo, A Seghetti, From IEEE 802.15.4 to IEEE 802.15.4e: a Step towards the Internet of Things, Chapter
10 in Advances onto the Internet of Things (S. Gaglio, G. Lo Re, Editors), pp. 135‐152. January 2014. Springer

PerLab
Scalability
 Network composed of 10,000 nodes
 deployed in an area of 0.1 km2
 E.g., oil refinery where miles of piping are equipped with 100‐1000
temperature, pressure, level and corrosion sensors, deployed in a
relatively small geographical area
 50 Access Points
 i.e., special nodes that collect data generated by sensor nodes
 Multi‐hop communication towards the AP
– routes are constructed using the upstream algorithm of RPL
– with a metric which introduces load‐balancing between APs
 Periodic data generation
 1 packet every 10s
S. Zats, R. Su, T. Watteyne, K. Pister, Scalability of Time Synchronized Wireless Sensor Networking, Proceedings of Annual
Conference on IEEE Industrial Electronics Society (IECON 2011), November 7‐10, 2011.

PerLab
Scalability
 Results
 delivery ratio above 99.9%
 end‐to‐end latency of 2.25 s
 network lifetime of 8.4 years
 if 2200mAh AA batteries are used
 Conclusions
 By tiling 100 of these networks, the same results hold for a
network with
 1 million nodes
 deployed in 10 km2
 using 5,000 APs.
S. Zats, R. Su, T. Watteyne, K. Pister, Scalability of Time Synchronized Wireless Sensor Networking, Proceedings of Annual
Conference on IEEE Industrial Electronics Society (IECON 2011), November 7‐10, 2011.

PerLab
Suitability to Real-life Applications
Goal
 exploring the capabilities and performance of a
TSCH network under different conditions
 trade‐off between throughput, latency, reliability, and
power consumption for different scenarios
Use Cases
 Smart City
 Smart Building
 Industrial Setting
T. Watteyne, J. Weiss, L. Doherty, J. Simon, Industrial IEEE802.15.4e Networks: Performance and Trade‐offs, Proceedings of IEEE
International Conference on Communications (ICC 2015), June 8‐12, 2015.

PerLab
Smart City
 10‐100 thousands node with moderate density
 Low data rate (data reporting period often in the order of hours)
 Delay tolerant (a fraction of the reporting period)
 99% reliability
 Target battery lifetime: 10 years (or more)
Smart Building
 Large networks, but broken in application‐specific networks
 Moderate data rate (e.g., 1 packet every 1‐30 s)
 Maximum delay less than 1s for security or fire applications
 Higher reliability than in urban networks, e.g., 99.9% reliability

PerLab
Industrial Setting
 Process monitoring and control applications
 Harsh environment
 Network size typically smaller than in urban/building environments, but
with larger radius (i.e., # of hops)
 Low‐to‐moderate data rate
 1 data packet every 1‐60 s
 Delay tolerant for some monitoring applications, stringent requirements
for control applications
 99.9% (or better) reliability
 Larger batteries can be used because of lower cost sensitivity

PerLab
 Performance Estimator for TSCH networks
 Advanced model providing latency, power consumption and
throughput achieved by nodes
 assumes that topology, quality of wireless links, and traffic
demands are known
 Validation through experimental measurements
 based on SmartMesh IP

PerLab
manager/gateway
 Topology
 Multi‐hop mesh topology
 Up to 100 nodes per manager/gateway
 In a real environment many such
networks can be tiled to form a
large network
 Common Parameters
 Packet size: 90 bytes

PerLab
Parameters
 Nodes: 100
 Network Radius: 8 hops
 Similar # of nodes at each
hop
 Data report period: 3600 s
 No peer‐to‐peer traffic
 Ploss=20%
 Slotframe=1024
 10000 mAh battery
Performance
 Lifetime: 51 years
 Larger than 10‐year requirement
 Latency
 1‐hop nodes: 1.50 s
 Deep nodes: 12.0 s
Smart City

PerLab
Parameters
 Nodes: 100
 40, 30, 20, 10 nodes per hop
 Peer‐to‐peer traffic and
downstream control
 Ploss=30%
 Slotframe=256
 2000 mAh battery (a pair of AA batteries)
Performance
 Lifetime: 5.3 years
 Larger than 5‐year requirement
 Latency
Smart Building

PerLab
Parameters
 Nodes: 100
 5 nodes for each hop
 Ploss=30%
 Slotframe=256
 10000 mAh battery
Performance
 Latency
Industrial Setting

TSCH
Network Formation

PerLab
Network Formation
 Based on Enhanced Beacons (EBs)
 Regularly emitted by the PAN Coordinator and other nodes
 EBs are special frames containing
 Synchronization information
 allows new devices to synchronize to the network
 Channel hopping information
 allows new devices to learn the channel hopping sequence
 Timeslot information
 describes when to expect a frame transmission and when to send an
acknowledgment
 Initial link and slotframe information
 allows new devices to know:
– when to listen for transmissions from the advertising device
– when to transmit to the advertising device
The EB advertising policy is not part of the 802.15.4e standard

PerLab
Network Formation
 A joining device starts listening for possible EBs
 on a certain frequency
 Upon receiving an EB
 The MAC layer notifies the higher layer
 The higher layer initializes the slotframe and links
 Using information in the received EB message
 and switches the device into TSCH mode
 At this point the device is connected to the network
 Then, the device allocates communication resources
 (i.e., slotframes and links)
 and starts advertising, on its turn

PerLab
Network Formation
Coordinator Node
Joining Node
Radio ON using frequency f2
EB EB EBEB
f0 f3f1 f2
Legenda

PerLab
When to transmit EBs?
Slotframe
0 1 2 3 10
channelOffset
0
1
2
15

PerLab
Goals
 Minimum Joining time
 Devices must keep the radio ON during the joining
phase
 EBs should be sent frequently
 Frequent EB transmissions
 Reduce communication resources
 Increase energy consumption at network nodes

PerLab
When to transmit EBs?
Slotframe
0 1 2 3 10
channelOffset
0
1
2
15
Random Vertical
E.Vogli, G.Ribezzo, L.A.Grieco, and G.Boggia, Fast Join and Synchronization Schema in the IEEE 802.15.4e MAC, Proc. IEEE
Wireless Communications and Networking Conference Workshops (WCNCW 2015), 2015.
Slotframe
0 1 2 3 10
channelOffset
0
1
2
15
Random Horizontal

PerLab
Model-based Approach
Model of the Joining Phase
 Based on DTMC
 Analytical expression of joining time
 Avg. joining time depends on the EB transmission strategy
D. De Guglielmo, S. Brienza, G. Anastasi, A Model‐based Beacon Scheduling Algorithm for IEEE 802.15.4e TSCH Networks,
Proc. IEEE Int’l Symposium on a World of Wireless, Mobile, and Multimedia Networks (WoWMoM 2016), Coimbra, Portugal,
June 21‐24, 2016.

PerLab
Model of the Joining Phase
 Based on DTMC
 Analytical expression of joining time
 Avg. joining time depends on the EB transmission strategy
∑ 1
, , 1
2
, 1 ∑ ,
1
≡
June 21‐24, 2016.

PerLab
Optimization Problem
min

∈ ∀

 … are the unknowns
‒ distances, in terms of timeslots, between successive EBs
 T: slotframe duration
June 21‐24, 2016.

PerLab
 Model‐based Sheduling Algorithm
 Given the total number of EBs to allocate
 provides the EB allocation strategy that minimizes the
joining time
Slotframe
0 1 2 3 10
channelOffset
0
1
2
15

PerLab
Performance Comparison
 Model‐based Scheduling (MBS)
 Random (RD)
 Extended Random Vertical (ERV)
 Extended Random Horizontal (ERH)
Nb (# of EBs advertised per superframe) equal for all the
considered options
Performance Index: Avg. Network Formation Time
average time (in # of timeslots) from when the PAN coordinator transmits the
first EB to when the last node in the group joins the network

PerLab
Simulation Results
Scenario
 nodes willing to set up a network
 Random deployment over an area
of 80 80 m2
 Ploss = 0%
D. De Guglielmo, S. Brienza, G. Anastasi, A Model‐based Beacon Scheduling Algorithm for IEEE 802.15.4e TSCH Networks, Proc. IEEE
Int’l Symposium on a World of Wireless, Mobile, and Multimedia Networks (WoWMoM 2016), Coimbra, Portugal, June 21‐24, 2016.

TSCH
Link Scheduling

PerLab
Link Scheduling
 Assignment of links to nodes for data transmissions
 Duplex‐conflict free
 A node cannot receive simultaneously from many senders
 A node cannot transmit and receive at the same time
 Interference‐conflict free
 neighboring nodes should not transmit in the same timeslot and
with the same channel offset
 Multi‐channel comm. makes link scheduling easier
 Optimal schedule is hard to find
 large networks with multi‐hop topology
 dynamic networks (mobility, power management, …)
IEEE 802.15.4e does NOT specify how
to derive an appropriate link schedule

PerLab
TDMA Scheduling
 TDMA scheduling algorithms typically address
single‐channel networks
 recently, multi‐channel TDMA scheduling solutions have
been proposed
 Most existing multi‐channel scheduling schemes are
not suitable for TSCH networks
 do not allow per‐packet channel hopping
 do not address resource‐constrained nodes
 e.g., they are not memory efficient
 do not consider the spatial reuse of channels

PerLab
TSCH Scheduling
 Centralized Scheduling
 Distributed Scheduling

PerLab
TSCH Scheduling
Centralized Scheduling
 Link schedule computed and distributed by a
special node
 Coordinator (Manager) node
 Based on information received by all the nodes of the network
 network topology
 traffic pattern
 Re‐computed dynamically
 Whenever a change in the operating conditions occurs
 Not appealing for
 dynamic networks (mobile nodes, power management)
 large‐scale networks

PerLab
TSCH Scheduling
Distributed Scheduling
 No central entity
 Link schedule computed autonomously by each node
 based on local, partial information exchanged with its
neighbors.
 The overall schedule is typically non optimal
 Limited Overhead
 Suitable for energy‐constrained nodes

PerLab
TSCH Centralized Scheduling
 TASA (Traffic Aware Scheduling Algorithm) [1] [2]
 Tree network topology
 Converge‐cast communication model
 The coordinator has a single radio interface
 Heterogeneous traffic conditions
 MODESA (Multi‐channel Optimized Delay Slot Assignment) [3]
 The coordinator has multiple radio interfaces
 Homogeneous traffic conditions (heterogeneous conditions in [4])
[1] M.R. Palattella, N. Accettura, M. Dohler, L.A. Grieco, G. Boggia, Traffic Aware Scheduling Algorithm for reliable low‐power multi‐hop IEEE
802.15.4e networks, Proc. IEEE Int’l Symposium on Personal Indoor and Mobile Radio Communications (PIMRC 2012), Sept. 12, 2012
[2] M.R. Palattella, N. Accettura, L.A. Grieco, G. Boggia, M. Dohler, T. Engel. On Optimal Scheduling in Duty‐Cycled Industrial IoT Applications
Using IEEE802.15.4e TSCH, IEEE Sensors Journal, Vol. 13, N. 10, pp. 3655‐3666, October 2013
[3] R. Soua, P. Minet, E. Livolant, MODESA: An optimized multichannel slot assignment for raw data convergecast in wireless sensor
networks, Proc. IEEE International Performance Computing and Communications Conference (IPCCC 2012), Dec. 1‐3, 2012
[4] RSoua, E. Livolant, P. Minet, MUSIKA: A multichannel multi‐sink data gathering algorithm in wireless sensor networks, Proc.
International Wireless Communications and Mobile Computing Conference (IWCMC 2013), July 1‐5, 2013

PerLab
TSCH Centralized Scheduling --TASA
 Every node regularly updates the Manager with
 the list of other nodes it can hear
 the amount of data it generates
 The Manager
 draws the connectivity graph …
 Based on information received from nodes
 assigns slots to different links in the graph …
 Conflict‐free schedule taking an iterative approach
 Based on a combination of matching and vertex coloring techniques
 informs each node about links it is involved in
 If the connectivity graph changes
 the manager updates its schedule and informs affected nodes
Palattella, M. R., Accettura, N., Dohler, M., Grieco, L. A., & Boggia, G., Traffic Aware Scheduling Algorithm for reliable low‐power
multi‐hop IEEE 802.15. 4e networks, Proc. IEEE International Symposium on Personal Indoor and Mobile Radio Communications
(PIMRC 2012), pp. 327‐332, 2012

PerLab
TSCH Centralized Scheduling
 TASA (Traffic Aware Scheduling Algorithm) [1] [2]
 The coordinator has a single radio interface
 Heterogeneous traffic conditions
 MODESA (Multi‐channel Optimized DElay Slot Assignment) [3]
 The coordinator has multiple radio interfaces
 Homogeneous traffic conditions (heterogeneous conditions in [4])
[1] M.R. Palattella, N. Accettura, M. Dohler, L.A. Grieco, G. Boggia, Traffic Aware Scheduling Algorithm for reliable low‐power multi‐hop IEEE
802.15.4e networks, Proc. IEEE Int’l Symposium on Personal Indoor and Mobile Radio Communications (PIMRC 2012), Sept. 12, 2012
[2] M.R. Palattella, N. Accettura, L.A. Grieco, G. Boggia, M. Dohler, T. Engel. On Optimal Scheduling in Duty‐Cycled Industrial IoT Applications
Using IEEE802.15.4e TSCH, IEEE Sensors Journal, Vol. 13, N. 10, pp. 3655‐3666, October 2013
[3] R. Soua, P. Minet, E. Livolant, MODESA: An optimized multichannel slot assignment for raw data convergecast in wireless sensor
networks, Proc. IEEE International Performance Computing and Communications Conference (IPCCC 2012), Dec. 1‐3, 2012
[4] RSoua, E. Livolant, P. Minet, MUSIKA: A multichannel multi‐sink data gathering algorithm in wireless sensor networks, Proc.
International Wireless Communications and Mobile Computing Conference (IWCMC 2013), July 1‐5, 2013

PerLab
TSCH Centralized Scheduling --MODESA
 Basic idea
 selecting one node at each iteration …
 … and choosing a link to accommodate one of its required
transmissions
 The execution terminates when the transmissions of all
nodes in the network have been accommodated
 Node selection
 based on dynamic priority
 depending on the number of packets still to transmit
R. Soua, P. Minet, E. Livolant, MODESA: An optimized multichannel slot assignment for raw data convergecast in wireless
sensor networks, Proc. IEEE International Performance Computing and Communications Conference (IPCCC 2012), Dec. 1‐3, 2012

PerLab
MODESA Algorithm
Iterates over the set of nodes with data to transmit
 and having an available interface with their parent in that slot
 Sorted according to their priorities
 Picks the node with the highest priority
 Schedules its first transmission
 on that timeslot
 on the first channel offset
 Then, selects another node
 If the node is in conflict with a previously scheduled node its transmission
is allocated on a different channel offset
 otherwise, the same channel offset is used
 Continues until transmissions of all nodes have been scheduled

PerLab
 MODESA has been shown to be optimal
 For linear and balanced‐tree topologies
 Performance is close to that of an optimal algorithm
 Also for random network topologies
 Impact of # of communication channels
 Drastically reduce the schedule length
 Similar trend when increasing the # of interfaces at the coordinator
 Guidelines
 It is useless to equip the coordinator with a number of interfaces greater
than the number of its children
 It is pointless to use a number of channels higher than the number of
coordinator interfaces.

PerLab
TSCH Distributed Scheduling
Distributed
Scheduling
Non‐Energy
Constrained Nodes
Energy‐constrained
Nodes
Single Sinks Multiple Sinks
D. De Guglielmo, S. Brienza, G. Anastasi, IEEE 802.15.4e: a Survey, Computer Communications, Accepted with minor revision.

Summary &
Open Issues

PerLab
Summary
TSCH combines
 Time‐slotted access
 multi‐channel communication
 frequency hopping
 Energy efficiency
 Reliability
 Guaranteed bandwidth
 Limited and predictable delay
 Scalability (high aggregate throughput)
 Support for multi‐hop mesh communication

PerLab
Open Issues
 Communication Model
 Most of the previous studies have considered a convergecast
scenario
 Node‐to‐node communication needs to be investigated
 In the perspective of the Internet of Things
 Lightweight scheduling and routing solutions for quickly
establishing node‐to‐node paths
 Network Formation
 Currently available solutions may be inefficient for
 Mobile Objects  Frequent joins and leaves
 Nodes operating on a duty cycle

PerLab
Open Issues
 Security
 TDMA networks are typically prone to Selective Jamming
(SJ) attacks
 An attacker can easily identify slots assigned to victim node
 and destroy all communications by that node
 Frequency Hopping and Secure Beacons should make TSCH
networks more secure
 How robust is a TSCH network against SJ attacks?

PerLab
Open Issues
 TSCH and IoT integration
 IoT protocols rely on a flexible and dynamic paradigm
 E.g., the IPV6 Routing Protocol for Low‐Power and Lossy Networks
(RPL)
 They assume that smart objects
 are always active
 can be added and removed dynamically
Combining a dynamic and flexible routing mechanism
(e.g., RPL) with 802.15.4e TSCH introduces a number of
challenges

TSCH and
the Internet of Things

PerLab
The 6TiSCH Initiative
 Working Group set up by IETF
 Name: IPv6 over the TSCH mode of IEEE 802.15.4e
 Acronym: 6TiSCH
 Goal
 defining mechanisms to combine the high reliability and
low‐energy consumption of IEEE 802.15.4e TSCH with
the ease of interoperability and integration offered by
the IP protocol
https://datatracker.ietf.org/wg/6tisch/

PerLab
6TiSCH: Problem
TSCH does not define
 policies to build and maintain the communication schedule
 mechanisms to match the schedule to the multi‐hop paths
maintained by RPL
 mechanisms to adapt the resources allocated between
neighbor nodes to the data traffic flows
 techniques to allow differentiated treatment of packets
 data packets generated at the application layer
 signaling messages needed by 6LoWPAN and RPL to discover
neighbors and react to topology changes

PerLab
6TiSCH Reference Architecture

PerLab
6TiSCH Protocol Stack
New sub‐layer for integrating
higher IETF layers with IEEE
802.15.4e TSCH
Under definition by the 6TiSCH WG
Protocol stack for
deterministic IPv6‐enabled
wireless mesh networks

PerLab
6Top Sub-layer
 Controls the TSCH schedule
 Through a Management Entity (ME)
 Adds/removes links (cells)
 Collects connectivity information
 useful for upper layers (e.g., RPL)
 Monitors the performance of links
(cells)
 reschedules them if performance is not
as expected
 Both centralized and distributed
scheduling supported

PerLab
6TiSCH Scheduling
 Minimal Scheduling
 used during network bootstrap or
 when a better schedule is not available
 Centralized Scheduling
 Based on a central entity
 Path Computation Element (PCE)
 Distributed Scheduling
 distributed multi‐hop scheduling protocols
 neighbor‐to‐neighbor scheduling negotiation

PerLab
Minimal scheduling
 Static scheduling
 Pre‐configured
 Learnt by the node at joining time
 Minimal 6Tisch configuration
X. Vilajosana and K. Pister, “Minimal 6TiSCH Configuration, draft‐ietf‐6tisch‐
minimal‐00”, IETF, Fremont, CA, USA, 2013

PerLab
Centralized Scheduling
Path Computation Element (PCE)
 Collects
 network state information
 traffic requirements from all nodes
 Builds
 Communication schedule
 Making sure that QoS requirements of all flows are met
 Installs
 The schedule on the network
6TiSCH will define the protocol to be used for message exchange between PCE
and network nodes

PerLab
Distributed Scheduling
 Nodes agree on a common schedule
 distributed multi‐hop scheduling protocol
 neighbor‐to‐neighbor scheduling negotiation
 Reservation protocol
 to transport QoS requirements along a certain path
 Negotiation phase
 at each hop, 6top starts a negotiation with the next hop
 which and how many cells allocate to satisfy the QoS requirements
of the path
6TiSCH WG is currently identifying protocols to be used for QoS requirements.
Strategies for cells allocation are also under definition

IEEE 802.15.4eIEEE 802.15.4e
Summary and
Conclusions

PerLab
Summary
 IEEE 802.1.5.4 standard
 limitations
 IEEE 802.15.4e
 open issues
 open issues
 The 6TiSCH initiative

PerLab
Conclusions
 Only partially investigated
 A lot of open issues
 A lot of work has been done
 Many open issues still remains
 Network Formation
 Security
 Still a completely open issue

PerLab
References
 D. De Guglielmo, S. Brienza, G. Anastasi, IEEE 802.15.4e: a Tutorial and
Survey, Computer Communications, available on line at:
http://www.sciencedirect.com/science/article/pii/S0140366416301980
 G. Anastasi, D. De Guglielmo, A Seghetti, From IEEE 802.15.4 to IEEE
802.15.4e: a Step towards the Internet of Things, Chapter 10 in Advances
onto the Internet of Things, pp. 135‐152, January 2014. Springer

Towards the Internet of Relevant Things: the IEEE 802.15.4e Standard

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to Towards the Internet of Relevant Things: the IEEE 802.15.4e Standard (20)

Recently uploaded (20)

Towards the Internet of Relevant Things: the IEEE 802.15.4e Standard