Internet Protocol over Wireless Sensor Networks, from

JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
189
Internet Protocol over Wireless Sensor Networks,
from Myth to Reality
Paulo Alexandre Correia da Silva Neves
Instituto de Telecomunicações, Portugal
Department of Informatics, University of Beira Interior, Covilhã, Portugal
Superior School of Technology, Polytechnic Institute of Castelo Branco, Castelo Branco, Portugal
Email: pneves@est.ipcb.pt
Joel José Puga Coelho Rodrigues
Instituto de Telecomunicações, Portugal
Department of Informatics, University of Beira Interior, Covilhã, Portugal
Email: joeljr@ieee.org
Abstract— Internet Protocol (IP) is a standard network
layer protocol of the Internet architecture, allowing
communication among heterogeneous networks. For a given
network to be accessible from the Internet it must have a
router that complies with this protocol. Wireless sensor
networks have many smart sensing nodes with
computational, communication and sensing capabilities.
Such smart sensors cooperate to gather relevant data and
present it to the user. The connection of sensor networks
and the Internet has been realized using gateway or proxybased approaches. Historically, several routing protocols
were specifically created, discarding IP. However, recent
research, prototypes and even implementation tools show
that it is possible to combine the advantages of IP access
with sensor networks challenges, with a major contribution
from the 6LoWPAN Working Group. This paper presents
the advantages and challenges of IP on sensor networks,
surveys the state-of-art with some implementation examples,
and points further research topics in this area.
Index Terms— wireless sensor networks, Internet
connectivity on wireless sensor networks, Internet protocol,
ubiquitous networks
I. INTRODUCTION
Many smart sensing nodes that cooperate to sense the
environment may constitute a wireless sensor network
(WSN), providing sensing services to an ever-growing
application space. Each node has a wireless radio to
communicate, a processing unit and memory to process
tasks and an autonomous energy unit, a battery. WSN
applications started with initial military applications (a
common ground with many of the technologies we use
today), to many other application areas such as
environmental observation, health monitoring, structural
monitoring, habitat monitoring, smart classroom, and
tracking among others [1]. A growing interest is drawn to
enable ubiquitous applications using WSN, enabling
sensing services that can effectively used in behalf of the
Manuscript received Feb. 22, 2009; revised Sept. 1, 2009; accepted
Nov. 17, 2009.
© 2010 ACADEMY PUBLISHER
doi:10.4304/jcm.5.3.189-196
user, such as smart spaces and augmented reality.
WSN may have thousands of small smart sensors with
computational capability and memory, one or more
sensors and a limited power supply. The limited power
supply and computational power are the main constraints
at the smart sensor level, dictating the feasibility of a
given network protocol. Several applications do not allow
battery recharge, such as deployment in harsh
environments. Moreover, recharge circuitry uses board
space and provides for extra weight and cost. Typically,
the nodes of a WSN cooperate and drive information to a
special more powerful node: the sink node. Many
different sensors can be used such as temperature, light
intensity, accelerometer, and pressure, among others.
The main goal of a sensor network is to provide
sensing services to the user or other systems. The Internet
has 1 billion users worldwide, so it makes sense to
provide WSN services to this ever-growing community
[2]. The Internet of computers is becoming the Internet of
machines or “the tangible Internet”, a global network that
will not only connect computers, but all kinds of
processor-enabled machines, such as domestic
appliances, mobile phones, and hopefully WSNs.
The connection of WSN to the Internet has been
achieved using a proxy-based approach. In this approach
sensor nodes communicate through dedicated WSN
protocols, whereas the sink acts as a proxy to the Internet,
converting IP to/from the dedicated WSN protocol. This
approach allows existing networks to be connected with
minimal changes|.
Another approach is to use IP as the protocol for the
sensor network itself, avoiding the need to use a proxy.
However, there is a common belief that IP is not suited
for sensor node hardware limitations, mainly due to
energy considerations (header overhead), protocol
complexity and memory needs. We will prove with this
paper that not only IP is feasible for WSN, but also a
significant number of implementation attempts exist and
more are expected to surface as research continues.
The Internet Protocol (IP) [3] suite is the de-facto
routing protocol in the largest of all world networks - the
190
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
Internet. Though not perfect, it clearly served us well for
nearly 30 years in its version 4 specifications, so much
that IPv4 is commonly named IP. With the introduction
of version 6, the addressing space grew to a number close
to 2128 addresses, leading to 667x1023 addresses per
square meter of earth surface. This protocol will empower
the next generation of networks, where more
heterogeneity is expected on the Internet.
The motivation to connect sensor networks to the
Internet draws from remote access to data, making
ubiquitous computing realistic. Eventually WSN and
ubiquitous computing will somehow merge [4]. As an
example, an Internet-connected WSN is monitoring a
parking lot; a given application can consume data to
guide a user to the nearest available space. Moreover, it
can even reserve a parking space; provide expected
occupation time, among other interesting features, all
tightly coupled to provide transparent services to the user.
To the best of the authors’ knowledge, no other paper
surfaces the current approaches on the application of the
IP protocol over WSN. In [5] authors provide an
overview of the myths that drove away IP from WSN,
however, the paper focus is on presentation of a system
architecture, which is also present afterwards. As a result,
a great motivation is drawn to the writing of this paper.
The rest of the paper is organized as follows. Section II
elaborates on some background knowledge of the main
technologies that this paper focuses on: WSN, routing
protocols for WSN, IP version 4 and 6, and body sensor
networks. Section III presents the main challenges for
applying IP as the routing protocol for WSN. Section IV
elaborates on the current research and system
implementation. Section V concludes the paper and point
relevant topics for further research.
This section provides some insight on WSN,
considering their routing protocols, their application on
body sensor networks, and surfaces IP in both fourth and
sixth versions.
A. WSN
WSN first depicted in military applications, to track
enemy movement and survey battlefield areas. With the
promise of low node cost, powered by MEMS (micro
electromechanical systems) technology, WSN began its
application to broader areas like environmental
monitoring, building automation and monitoring,
underwater surveillance, up to healthcare monitoring [6].
The new application areas pose different challenges when
compared with military applications, such as patient
safety on healthcare applications or signal acquisition on
underwater applications, among others. However, some
challenges are horizontal, such as energy constraints,
wireless communication coverage and bandwidth, and
limited computing resources.
Another very interesting application of WSN is in
biofeedback. Several sensors are placed on the human
body directly or through wearable clothing. These sensors
monitor health parameters such as heartbeat, body
© 2010 ACADEMY PUBLISHER
Sensor 1
...
II. TECHNOLOGICAL BACKGROUND
temperature, Electrocardiogram (ECG) and may be
wirelessly connected to a sink node. The sink node is
responsible for capturing the reading of all sensors,
sending the data to an interface device, such as a mobile
phone or PDA [7]. This kind of network is commonly
known as a body sensor network (BSN), since its scope is
body-wide [8]. The connection of BSN to the Internet is
also mandatory for remote access.
The smart sensor nodes (also referred as motes) of a
WSN have one or more sensors, a processing unit with
RAM and program memory, a limited power supply, and
a wireless transceiver, as depicted in Figure 1. In terms of
energy, the majority is spent on wireless communication.
Sometimes a bit transmission uses 1000 times the power
of computing an instruction. Energy is of major
importance, with some designs featuring some sort of
energy harvesting, as in [9]. As the node operates, energy
is depleted; as nodes begin to fail the network coverage
shrinks, eventually rendering it useless [10]. As a result, a
great effort is drawn on employing power-efficient
routing protocols.
Among the challenges that this kind of network
presents, apart from the energy constraints, we find node
placement, node mobility, node resources (also
considering energy), and data aggregation. Node
placement clearly depends on the application. For a
healthcare application, node placement must be very
precise and difficult to control without human
intervention, while on a military application it may be
impossible to control, e.g. dropped from an aircraft. Node
mobility is not supported on many designs. However,
some applications like healthcare where a patient travels
along the hospital, or sensors placed on animals to study
their behavior, node mobility has to be considered.
Processing Unit
ADC
Wireless
Transceiver
FLASH
Sensor n
RAM
Battery
Figure 1. Typical block diagram of a wireless sensor network smart
node.
Node resources are scarce, and on a flat network, every
node plays every role: communication, sensing, and
computation. On a non-flat design a clustered or
hierarchical network and some nodes may be only
relaying information. Since a node must be cheap enough
for the overall network to be cost-efficient, hardware
resources must be bound. It is common to find 8 or 16-bit
microcontrollers, 4-10KB of RAM and 48-256KB of
Flash. Examples of such equipments can be found in
Crossbow Technology (http://xbow.com).
B. WSN Routing Protocols
Routing in WSN is of major importance, since nodes
need to communicate to each other for the information to
reach the sink node, constituting a multi-hop network.
Due to the previously outlined challenges that WSN
presents, research for dedicated routing protocols has
evolved.
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
Good surveys on routing protocols developed
specifically for WSN exist and may be found on [11, 12].
In [11], authors discuss system architecture and design
issues that influence the performance of a given protocol.
These are network dynamics, node deployment, energy
considerations, data delivery models, node capabilities
and data aggregation/fusion.
Routing protocols for WSN can be classified by type,
based on their main characteristics. Then, the following
three categories may be considered: data-centric
protocols, hierarchical routing protocols, and locationbased routing protocols [11]. Some other protocols are
based on network flow or quality of service (QoS)
modeling. The evaluated protocols feature energy
considerations, but still lack on QoS and real-time
applications.
The concept of data-centric protocols relies on a sink
node that queries sensor network areas for specific data
(data-centric). The query must clearly indicate the
required data, through e.g. attribute-based naming.
Among the protocols based on this approach are SPIN,
Directed Diffusion, Rumor Routing, Energy-aware
routing, and Gradient-based routing, among others.
Directed Diffusion clearly made a leap forward on this
type of protocols, with some others following the same
approach or similar concept.
Hierarchical protocols introduce multi-tier routing on
WSN. If a single tier is used on a large WSN, energy may
be depleted faster on the intermediate nodes. Hierarchical
routing takes advantage of network clustering, where
several sensing nodes elect a typically more powerful
node to communicate with the sink in their behalf.
Several cluster heads communicate with each other to
reach the sink, relieving nodes from multi-hop
communication, thus saving battery life. This kind of
protocols also can take advantage of data aggregation on
cluster heads, relieving the sink from lower-level data
aggregation, thus minimizing needed computing power.
Examples of hierarchical routing protocols are the
LEACH (Low-Energy Adaptative Clustering Hierarchy),
PEGASIS (Power-Efficient GAthering in Sensor
Information Systems), and TEEN (Threshold Efficient
sensor Network protocol). LEACH is one of the most
popular hierarchical routing algorithms for sensor
networks, while PEGASIS introduces improvements on
the LEACH protocol.
Another routing approach in WSN is location routing.
This kind of routing takes advantage of the knowledge of
smart sensor location to send queries to specific regions.
The location information can also be used to minimize
the impact of communication on energy consumption,
taking shorter paths requiring less transmission power.
Most of these protocols come from ad hoc networks, but
may be well applicable to sensor networks. MECN
(Minimum Energy Communication Network), GAF
(Geographic Adaptive Fidelity), and GEAR (Geographic
and Energy-Aware Routing) are examples of location
protocols.
None of the identified protocols were developed with
Internet connectivity as a main goal. As a result a proxy-
© 2010 ACADEMY PUBLISHER
191
based scenario is the only solution for Internet
connectivity, as opposed to smart sensor node stack based
IP implementation.
C. Surfacing IP version 4 and IP version 6
Since IP over WSN is the focus of this writing, we
surface both IPv4 and IPv6 protocols in this section. IPv4
was developed in the early seventies to ease
communication between restrict and closed number of
researchers and academics in the United States [13]. Later
RFC 791 presented the Internet Protocol, a protocol for
wired computer networks that is able to connect different
networks by the means of a compatible router. The age of
the Internet was just beginning, and back then no one
predicted the number of users Internet has today.
The version 4 of the IP protocol survived until now
with a great success, in spite of its limitations. However,
the introduction of several mobile Internet access devices
pushes IPv4 addressing space boundaries. The Internet
Engineering Task Force (IETF) began working on the
successor of IPv4 around the early 1990’s. In 1993, the
IETF started the Internet protocol next generation (IPng)
for proposals investigation and recommendations for the
future IPv6.
The use of network address translators allows the reuse
of IPv4 addressing space inside a given network, a
solution adopted by many Internet service providers (ISP)
for household utilization. Such amend allows IPv4 to be
used even today, even with billions of Internet-enabled
devices. However, we all eventually adopt IPv6.
D. Body Sensor Networks
A very interesting application area of WSN is
healthcare promotion and biofeedback. In this area,
WSNs are known as wireless body sensor network
(WBSN), since the objective is to place sensors on the
human body to study health parameters.
Much research is focused on the development of
biosensors [14, 15], medical systems [16, 17] and
development of suitable interfaces [18-20]. This presents
a very challenging and interesting application area that
can also benefit from the integration of IP.
The raw sensor data is transferred to a more powerful
node, the sink. Data must be processed, stored and
presented to the medical staff and/or patients. The
feasibility of such networks depend on their ability to
operate for weeks on a 24/7 basis without intervention,
operation under extreme temperatures and provide lower
cost than current solutions [21]. This kind of networks
plays well with remote monitoring, thus also drawing
great benefit from Internet connectivity.
III. IP OVER WSN
This section provides an overview with the motivation
and challenges for the implementing the IP protocol on
WSN. The use of IP on WSN is more than a mere
academic research thrust; it provides significant
advantages and tackles the need for the augmentation of
the Internet to provide more ubiquitous services.
However, such advantages come with a price, namely on
192
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
the addressing of some new challenges that a protocol
created for wired networks presents.
A. Background
Two basic architectures emerged when connecting
WSN to the Internet, a proxy-based solution or
integration of IP at the smart sensor level (also referred as
IP stack). As may be seen in Figure 2, the sensor network
has another routing protocol different from IP, and the
sink acts as a protocol-mapping device, which connects
the network to the Internet.
Non-IP
Sensor
Network
Dedicated
protocols
Internet
IP
SINK/PROXY
Figure 2. Connecting WSN to the Internet using a proxy-based solution.
This transformation can be performed at the
application level [22]. For example, the sink node may
query the sensor network using dedicated protocols, store
data locally, while performing data aggregation. This
approach may be shown in Figure 3, scenario (a). When
an Internet host requires data, it communicates via IP
protocol with the sink that sends data from the local
database to the requiring host.
Dedicated Protocols
Sensor
Network
Database
Application Layer
Transformation
TCP/IP
Internet
TCP/IP
TCP/IP
Internet
TCP/IP
SINK NODE
(a)
Dedicated Protocols
Sensor
Network
Network
Transformation Layer
SINK NODE
(b)
Figure 3. Different approaches for proxy-based WSB connection to the
Internet: (a) with proxy database, and (b) without memory.
Another proxy-based solution may be applied at the
network level, where a sink performs protocol
transformation without the use of a local database,
depicted by the scenario (b) of Figure 3. When an Internet
host sends a query to the sink, the sink queries the sensor
network. While this approach may lead to freshness of
data, it may also lead to delays since the database is not
present to provide some buffer effect.
Another solution is to integrate an IP stack on the
smart sensor and directly use IP as routing protocol inside
the network, as Figure 4 depicts. This is the scenario we
are interested on in this paper. This scenario presents
specific challenges as above mentioned in Section 3.2.
Another motivation comes from the fact that IP protocol
presents several advantages that mainly draw from the
decades that is being used with success on computer
networking, with various mechanisms and protocols
already developed, validated and operationally deployed
[22]. Moreover, tools for network management,
commissioning, configuring and debugging developed for
IP-enabled networks could also be used [23]. IP brings an
© 2010 ACADEMY PUBLISHER
open standard to WSNs, and presents a very attractive
learning curve (availability of bibliography published on
the subject), almost transparent Internet integration,
network maintenance and scalability.
Figure 4. Connecting WSN to the Internet at the smart sensor level.
B. Challenges
Several reasons supported the idea that IP cannot be
used directly at the smart sensor level, reserving the
routing for dedicated protocols. In this subsection we
provide a vision of the major challenges with some
discussion.
Header overhead. IP adds significant amount of data
on the header block of the packet, introducing undesirable
overhead. Since the majority of energy is spent on
wireless communication, this may be a very limiting
factor for the use of IP on the smart sensor node. The
minimum IPv4 header has 20 bytes plus the payload.
Extensions can be used that further enlarge the size of the
header.
IPv6 uses a different approach, where a fixed 40 bytes
header (double the one in IPv4) is used. The header size
increase is mostly due to 128-bit addresses instead of 32bit addresses of IPv4, even though IPv6 header is
optimized leaving some IPv4 header bits behind. As a
result header overhead may increase. To tackle this
challenge, header compression must be used. The
compression can be applied to the addresses (by using
link-local addresses for instance) or even applying the
compression mechanisms defined by the 6LoWPAN
specification.
Addressing Scheme. IP addressing scheme relies on
the knowledge of the source address and a destination
address, and both must be unique inside a given network.
While IPv4 can use dynamic host configuration protocol
(DHCP) for address attribution, it contributes for more
protocol overhead; while IPv6 provides an intrinsic
mechanism for stateless auto-configuration. In IPv6 both
anycast and multicast addresses also make it possible to
address a group of nodes with a single address. Anycast
delivers a packet to the nearest interface of the identified
group alone, while multicast delivers to all the network
interfaces of the identified group.
While
traditional
computer
networks
hosts
communicate to transfer data among applications, WSNs
are intended to provide sensing services. In many
applications, it is desirable to know the sensed data, not
the address of the smart sensor(s) that produced such
data. As a result, the most important asset is the data
produced by the smart sensor, and typically a data-centric
approach is preferable to address-centric. However, when
considering body sensor networks, an address-centric
approach may be preferable, since typically no redundant
sensors are used, mainly because of patient discomfort
and setup time. As a result, this special kind of WSN may
benefit from a unique addressing scheme.
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
Limited Bandwidth. Small smart sensors have limited
wireless bandwidth; 250kbps is very common in IEEE
802.15.4 implementations [24]. The more bits must be
transmitted, the longer the data transmission will take,
and longer the latency for medium access. With limited
bandwidth one wants to waste as minimum as possible in
overhead bits, let it be for header, error control, or others.
IP may pose significant challenges, since it was
developed for wired connections. To tackle this challenge
header compression mechanisms
Limited Energy. One of the distinct factors of WSNs
is the limited energy of the nodes, since nodes must be
small and cost-effective. Wireless communications on the
nodes consume the maximum amount of energy,
involving both transmission and reception [25]. In some
cases the energy cost of one bit transmission corresponds
to 1000 processor instructions or more. In many scenarios
it is not viable to provide battery replacement or even
recharge. Then, when a node looses power it dies. When
a given number of nodes in a network die, the network
ceases to provide sensing services, rendering it useless.
This challenge is tackled with a conjunction of several
mechanisms. The first is header compression. With
transmission of fewer bits, energy wasted on transmission
of a single packet is minimized. The second mechanism
is stateless auto-configuration of IPv6. This mechanism
allows the association of an IPv6 link-local address to an
interface, which may be enough for a given network.
Link-local addresses in IPv6 start with FE80.
Implementation
Challenges.
One
of
the
implementation challenges is the internetworking
between layer 2 protocols and IP. Since a specification
for Ethernet exists and most computer local area
networks use this protocol, the problem is solved. A
growing interest is given to IEEE 802.15.4 standard and
mechanisms must be implemented to internetwork with
IP. Other implementation challenges are related to
development of suitable security mechanisms, and
specification of ad-hoc networking and autoconfiguration for ad-hoc deployment. In this specific
challenge the 6LoWPAN specification provides an
overlay network layer that can transmit IPv6 data over
IEEE 802.15.4 frames. 6LoWPAN is presented with
some detail in section IV.A. Also the work from Adam
Dunkels addresses some implementation challenges with
the first IP stack for 8-bit microcontrollers [26].
Transport Protocol. IP protocol does not guarantee
reliability in packet transmission, employing a best-effort
approach. When one considers IP on its usefulness for
global Internet connection, one considers transmission
control protocol (TCP) as the transport protocol to
achieve reliable packet transmission. However, TCP is
not energy-aware and requires acknowledgment packets
to be sent towards the transmitting host; which wastes
valuable bandwidth and energy resources.
Another alternative is the use of the user datagram
protocol (UDP). This protocol is sometimes used on non
mission-critical sensor networks, avoiding the
acknowledgement mechanism of TCP. A discussion on
some variations of TCP can be found on [27].
© 2010 ACADEMY PUBLISHER
193
IPv4 or IPv6. IPv4 currently still manages to satisfy
the great majority of computer communication needs
across the Internet, mainly due to several mechanisms
like network address translation (NAT). However, IP
addresses are becoming short, so IPv6 rises as a solution.
Moreover, features inside IPv6 provide functionality only
found on IPv4 plus one or more added mechanisms.
The IPv6 protocol may even aggravate the expected
overhead of IP for WSN. However, a detailed study
proves that overhead increases by a very small amount
[28]. The paper presents both simulation and prototype
results, pointing several advantages of using IPv6 on such
hardware constrained devices, like auto-configuration and
stateless mechanisms, the growing adoption of IPv6 and
increase in addressing space.
IV. CURRENT APPROACHES
This section reviews the most relevant efforts to bring
IP to WSNs. The first part presents an overview of the
first approaches, while the second part presents and
discusses some 6LoWPAN implementations. Finally,
section C presents some real implementations on the
field.
A. Breaking the walls
Breaking the walls presents initial work on IP over
WSN, that progressively put IP on the WSN space. The
first breakthrough was the introduction of a full TCP/IP
stack on very limited hardware, through the work of
Adam Dunkels [26]. In this work two implementations
are presented, one for very limited 8-bit architectures
(uIP) and other with more functionalities (lwIP), conform
to a subset of RFC 1122. While lwIP provides a fullscale, but simplified, implementation of IPv4, ICMP,
TCP and UDP, uIP only can handle one interface, does
not implement UDP, focusing on IP, ICMP and TCP
protocols.
Adam Dunkels et al. present possible approaches for
the header overhead problem, by applying header
compression techniques. The use of an application
overlay network implements a data-centric routing with
address distribution based on sensor location. The
Distributed TCP caching mechanism enables lower
energy consumption, each node is able to cache a single
TCP segment, enabling single-hop retransmissions [29].
Another work on header compression presents a
layered approach that dissociates the network from the
transport compression, enabling compression on different
links, domains and even networks [30]. Performance
evaluation is drawn for a tree-shaped sensor network with
one sink node, and shows energy consumption gains from
the proposed header compression architecture.
In [5] some challenges are identified and the
architecture of IPSense, a system that allows IPv6 over
WSNs. IPSense features flexible addressing, enhanced
mobility, and a clustering mechanism with sensor routers.
Sensor routers are responsible for communication
management with the sink node, aggregating several
sensor nodes, and are also faced as gateway points for
194
other networks to communicate, thus alleviating the very
constrained sensor nodes.
A TCP/IP implementation is described by Xiaohua
Luo et al., but is based on a proxy approach [31]. The
base station converts external TCP/IP requests to an
Active Message, while sensor nodes implement a
protocol called SIP – Sensor Internet Protocol. This
approach lightens the computational requirements on the
sensor nodes, but the sensor network itself does not use
TCP/IP. The SIP protocol assumes that motes never need
to communicate with external hosts, so only respond to
queries. The base station is responsible for receiving
TCP/IP requests from Ethernet, 802.11b wireless network
or Bluetooth, translating into an Active Message.
B. 6LoWPAN
A task force named 6LoWPAN Working Group (WG)
from the Internet Engineering Task Force (IETF) is
working on a standard protocol definition: 6LoWPAN
[23]. The main goal is to enable IPv6 packets over low
power wireless networks, with emphasis on the IEEE
802.15.4 standard, supporting small/pico sensor network
nodes. The group aim is to define an encoding
mechanism that is layer 2 and 3 agnostic.
Initial implementations of 6LoWPAN show that a
mere 32KB Flash ROM is needed. Moreover, the WG has
managed to ditch DHCP and NAT by using Zero-Conf
and Neighbor Discovery capabilities of IPv6. Also,
stacked headers are used to minimize header overhead,
through header compression.
The working group successfully addresses the header
overhead problem of IPv6, removing the need for
configuration servers (namely, DHCP and NAT), use of
EUI-64 and 16-bit unique addresses within the personal
area network (after an association event) [32]. However,
in order to use 16 bit addresses, a PAN coordinator must
dole the address in an association event, limiting the
validity to the lifetime of the association.
Zach Shelby considers IP-enabled WSNs as the Wi-Fi
of the embedded world [33], namely IP over IEEE
802.15.4. The paper refers the 6LoWPAN initiative as a
means to achieve the desired functionality, namely with
the use of NanoSensors™.
According to [34] management tools must take into
account the special characteristics of WSNs. The LNMP
network management protocol provides network
management with simple network management protocol
(SNMP) support. Coordinator nodes capture the state of
sensor nodes and relay data to the gateway, which filters
state data from the list of reporting coordinators. SNMP
protocol is supported on the external networks, and the
gateway acts as a proxy between SNMP and the local
management framework.
Even with 6LoWPAN communication between
different networks can result in relatively high overhead
due to the bits needed for addressing hosts. The 6GLAD
[35] architecture proposes a twice-NAT and reverse
network address translation mechanisms. The twice-NAT
features both source and destination IP address
modifications, and by means of reverse NAT mapping on
the WSN gateway, mapping IPv6 addresses to node’s
© 2010 ACADEMY PUBLISHER
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
short addresses inside the network, allowing the use of
less bits for addressing.
A similar approach for efficient address utilization is
the dual addressing scheme (DAS) for IPv6 over IEEE
802.15.4 networks [36]. DAS maps a global IPv6 node
address to a link local address to save energy and
resources. Also the gateway maps IPv6 global addresses
into link local lightweight addresses to reach a specific
node inside the network. The proposed scheme is
validated through simulation that shows a significant
reduction in overhead.
C. Some Implementations
The first attempts to integrate a sensor network in the
Internet were proxy-based. The above-mentioned SIP
protocol is based on a proxy scheme [31]. The proxy
deals with the communication between external hosts and
wireless sensors through an Active Message mechanism.
The proxy is transparent in terms of TCP/IP operation,
since it does not require any changes on the wireless
sensors or the Internet hosts. Such approach is still used
on ZigBee Bridge for instance [37].
One example of smart sensor stack comes from a
sensor network for intrusion monitoring featuring an IPbased WSN with the ESB platform from FU Berlin [38].
In this work authors use the ContikiOS with the uIP
stack, providing TCP/IP support. Addresses are
distributed based on node location on a grid, which
require location information at the sensor level.
An implementation of a TCP/IP stack for Tiny OS
based on a code base from HP Labs, featuring IEEE
802.15.4, a port of TCP/UDP/IP uIP stack, Telnet and
HTTP (dynamic web pages support) servers. The
implementation also features a lightweight version of the
protocols SIP, DHCP, NTP and an IMAP-like message
service. It also features a Linux-based IEEE 802.15.4
access point through a Telos mote plugged into the
computer [39].
ContikiOS, currently on version 2.2.3 presents the
uIPv6 stack with 6LoWPAN implementation [40], the
evolution of above-mentioned uIP. This version was
awarded with the IPv6 Ready silver seal, featuring IPv6
over IEEE 802.15.4 through 6LoWPAN. The stack
features IPv6, TCP, UDP, ICMPv6, and neighborhood
discovery (ND).
TinyOS (http://tinyos.net), currently on version 2.1,
also has a 6LoWPAN implementation (b6loWPAN) with
support for stateless auto-configuration, multi-hop
routing, and fragmentation for 1280bytes MTU. The
implementation features tools like ping, nc6 and tracert6.
V. CONCLUSIONS AND FURTHER RESEARCH TOPICS
In this paper we surfaced the state-of-the-art on the
efforts to bring IP over WSN. Motivation is clearly
focused on their connection to the Internet in the most
transparent possible way, allowing realistic ubiquitous
computing applications. IP over WSN is more than a
myth; it is a reality. Efforts are aimed at the sensor level,
namely the 6LoWPAN group and software resources like
TinyOS and ContikiOS.
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
A sensor network presents many challenges that
ultimately result in node failure, leading to dynamic
routing of information inside the network. As in
traditional IP networks, dynamic routing may be needed.
One research topic that must be considered is the search
for an optimal dynamic routing protocol such as routing
information protocol (RIP), open shortest path first
(OSPF), or other to route information.
IP over WSNs is a reality, but the vision of Internet
connectivity we have on traditional computing systems,
with the required ease of use and auto-configuration is yet
beyond reach. We believe a Plug-and-Play like approach
for WSN is a very interesting research topic for network
deployment.
The development of adequate applications for remote
network management taking into account the specificities
of the sensor network are yet to be achieved. Such
software tools must run on several hardware platforms
(personal computer, smart phone, or even dedicated
devices), provide accurate and timely information about
the network, and also provide a convenient applicationprogramming interface (API) for remote data acquisition.
With IP, a dynamic routing protocol and a convenient
set of tools, WSN present an invaluable resource for the
vision of ubiquitous computing.
195
[8]
[9]
[10]
[11]
[12]
[13]
[14]
AKNOWLEDGEMENT
Part of this work has been supported by Instituto de
Telecomunicações, Next Generation Networks and
Applications Group, Covilhã, Portugal, and by the
Euro-NF Network of Excellence from the Seventh
Framework Program of EU.
[15]
[16]
REFERENCES
[1] I. Khemapech, I. Duncan, and A. Miller, "A Survey of
Wireless Sensor Networks Technology", in 6th Annual
Postgraduate Symposium on the Convergence of
Telecommunications, Networking and Broadcasting,
Liverpool, UK, June, 2005.
[2] T. Kim, Y. Han, S. Cheon, I. Kim, J. Oh, and J. Ryou,
"Network Traffic
Analysis System Based on Data
Engineering Methodology using System Entity Structure",
in International Conference on Convergence Information
Technology, Gyeongju, South Korea, 2007, pp. 1752-1757.
[3] R. Braden, Requirements for Internet Hosts Communication Layers, Internet Engineering Task Force,
RFC 1122, October 1989.
[4] J. A. Stankovic, "When Sensor and Actuator Networks
Cover the World", ETRI Journal, vol. 30, pp. 627-633, No.
5, 2008.
[5] T. Camilo, J. S. Silva, and F. Boavida, "Some Notes and
Proposals on the use of IP-based Approaches in Wireless
Sensor
Networks",
Ubiquitous
Computing
and
Communication Journal, vol. Special Issue on Ubiquitous
Sensor Networks, 2007.
[6] K. Römer and F. Mattern, "The Design Space of Wireless
Sensor Networks", IEEE Wireless Communications
Magazine, vol. 11, pp. 54-61, December 2004.
[7] L. Zhong, M. Sinclair, and R. Bittner, "A Phone-centered
Body Sensor Network Platform Cost, Energy Efficiency &
User Interface", in International Workshop on Wearable
© 2010 ACADEMY PUBLISHER
[17]
[18]
[19]
[20]
[21]
[22]
and Implantable Body Sensor Networks (BSN 2006),
Cambridge, MA, USA, 2006.
S.-L. Chen, H.-Y. Lee, C.-A. Chen, C.-C. Lin, and C.-H.
Luo, "A Wireless Body Sensor Network System for
Healthcare Monitoring Application", in IEEE Biomedical
Circuits and Systems Conference (BIOCAS 2007),
Montréal, Qc, Canada, 2007, pp. 243-246.
Y. Ammar, A. Buhrig, M. Marzencki, B. Charlot, S.
Basrour, K. Matou, and M. Renaudin, "Wireless Sensor
Network Node with Asynchronous Architecture and
Vibration Harvesting Micro Power Generator", in Joint
Conference on Smart Objects and Ambient Intelligence:
Innovative Context-aware Services: Usages and
Technologies, Grenoble, France, 2005, pp. 287-292.
T. Hailun, "Maximizing Network Lifetime in Energyconstrained Wireless Sensor Network", in 2006 Int. Conf.
on Communications and Mobile Computing, Vancouver,
British Columbia, Canada, 2006, pp. 1091-1096.
K. Akkaya and M. Younis, "A Survey on Routing
Protocols for Wireless Sensor Networks", Elsevier Ad Hoc
Network Journal, vol. 3/3, pp. 325-349, 2005.
J. N. Al-Karaki and A. E. Kamal, "Routing Techniques in
Wireless Sensor Networks: A Survey", IEEE Wireless
Communications, vol. 11, pp. 6-28, No. 6, December 2004.
S. Hagen, IPv6 Essentials: O'Reilly, 2006.
C. H. Chan, C. C. Y. Poon, R. C. S. Womg, and Y. T.
Zhang, "A Hybrid Body Sensor Network for Continuous
and Long-term Measurement of Arterial Blood Pressure",
in 4th IEEE/EMBS International Summer School and
Symposium on Medical Devices and Biosensors,
Cambridge, UK, 2007, pp. 121-123.
P. Cao, S. Jia, X. Wang, and J. Zhou, "Wearable and
Wireless Multi-Electrophysiological System", in 3rd
IEEE/EMBS International Summer School on Medical
Devices and Biosensors, Cambridge, MA, USA,
September, 2006, pp. 83-85.
U. Varshney, "Pervasive Healthcare and Wireless Health
Monitoring", Mobile Networks and Applications, vol. 12,
pp. 113-127, No. 2-3, March 2007.
CodeBlue: Wireless Sensors for Medical Care, URL:
http://fiji.eecs.harvard.edu/CodeBlue,
Accessed
in
February 2008.
O. Pereira, P. Neves, and J. Rodrigues, "Mobile Solution
for Three-tier Biofeedback Data Acquisition and
Processing", in IEEE Global Communications Conference
(IEEE GLOBECOM 2008), New Orleans, LA, USA,
November 30 - December 4, 2008, pp. 1-5.
T. Gao, C. Pesto, L. Selavo, Y. Chen, J. Ko, J. Lim, A.
Terzis, A. Watt, J. Deng, B.-r. Chen, K. Lorincz, and M.
Welsh, "Wireless Medical Sensor Networks in Emergency
Response: Implementation and Pilot Results", in IEEE
Conference on Technologies for Homeland Security,
Waltham, MA, 2008, pp. 187-192.
D. Yun, J. Kang, J.-e. Kim, and D. Kim, "A Body Sensor
Network Platform with Two-Level Communications", in
IEEE International Symposium on Consumer Electronics
(ISCE 2007), Irving, TX, USA, 2007, pp. 1-6.
P. Kulkarni and Y. Öztürk, "Requirements and Design
Spaces of Mobile Medical Care", ACM SIGMOBILE
Mobile Computing and Communications Review, vol. 11,
pp. 12-30, No. 3, July 2007 2007.
K. Mayer and W. Fritsche, "IP-enabled Wireless Sensor
Networks and Their Integration Into the Internet", in First
International Conference on Integrated ad-hoc and Sensor
Networks, Nice, France, 2006.
196
[23] G. Mulligan and L. W. Group, "The 6LoWPAN
Architecture", in 4th Workshop on Embedded Networked
Sensor, Cork, Ireland, 2007, pp. 78-82.
[24] "Part 15.4: Wirless Medium Access Control (MAC) and
Physical Layer (PHY) Specifications for Low-Rate
Wireless Personal Area Networks (WPANs) - IEEE Std
802.15.4-2006": IEEE Computer Society, 2006.
[25] I. F. Akyldiz, W. Su, Y. Sankarasubramaniam, and E.
Cayirci, "Wireless Sensor Networks: a Survey", Computer
Networks (Elsevier), vol. 38, pp. 393-422, No. 4, 2002.
[26] A. Dunkels, "Full TCP/IP for 8-bit Microarchitectures", in
First Int. Conf. on Mobile Systems, Applications and Services,
San Francisco, CA, USA, 5-8 May, 2003, pp. 85-98.
[27] V. Tsaoussidis and I. Matta, "Open Issues on TCP for
Mobile Computing", Wireless Communications and
Mobile Computing, vol. 2, pp. 3-20, No. 1, 2002.
[28] J. S. Silva, R. Ruivo, T. Camilo, and G. Pereira, "IP in
Wireless Sensor Networks - Issues and Lessons Learnt", in
Third International Conference on Communication
Systems, Software and Middleware (COSMWARE 2008)
Bangalore, India: IEEE Communication Society, 2008.
[29] A. Dunkels, J. Alonso, and T. Voigt, "Making TCP/IP
Viable for Wireless Sensor Networks", in First European
Workshop on Wireless Sensor Networks (EWSN 2004),
Berlin, Germany, 2004.
[30] C. Westphal, "Layered IP Header Compression for IPenabled Sensor Networks", in IEEE Conference on
Communications (ICC '06), Instanbul, Trukey, June 2006,
2006, pp. 3542-3547.
[31] X. Luo, K. Zheng, Y. Pan, and Z. Wu, "A TCP/IP
Implementation for Wireless Sensor Networks", in IEEE
International Conference on Systems, Man and
Cybernetics, Hague, Netherlands, 2004, pp. 6081-6086.
[32] G. Montenegro, N. Kushalnagar, J. Hui, and D. Culler,
Transmission of IPv6 Packets over IEEE 802.15.4
Networks, IETF, RFC 4944, September 2007.
[33] Z. Shelby, "IP-based Wireless Embedded and Sensor
Networks: The WiFi of the Embedded World":
http://www.sensinode.com/pdfs/sensinode-wp1why_ip_wsn.pdf, 2008.
[34] H. Mukhatar, K. Kang-Myo, S. A. Chaudhry, A. H. Akbar,
K. Ki-Hyung, and S.-W. Yoo, "LNMP - Management
Architecture for IPv6 Based Low-power Wireless Personal
Area Networks (6LoWPAN)", in IEEE Network
Operations and Management Symposium (NOMS 2008),
Salvador, Bahia, Brazil, 2008, pp. 417-424.
[35] A. Zimmermann, J. S. Silva, J. Sobral, and F. Boavida,
"6GLAD: IPv6 Global to Link-layer ADdress Translation for
6LoWPAN Overhead Reducing", in Next Generation Internet
Networks (NGI 2008), Krakow, Poland, 2008, pp. 209-214.
[36] S. Yang, S. Park, E. J. Lee, J. H. Ryu, B.-S. Kim, and H. S.
Kim, "Dual Addressing Scheme in IPv6 over IEE 802.15.4
Wireless Sensor Networks", ETRI Journal, vol. 30, pp.
674-684, No. 5, 2008.
[37] M. Sveda and R. Trchalik, "ZigBee-to-Internet
Interconnection Architectures", in Second International
Conference on Systems (ICONS '07), Sainte-Luce,
Martinique, France, April 22-28, 2007.
[38] A. Dunkels, T. Voigt, N. Bergman, and M. Jonsson, "The
Design and Implementation of an IP-based Sensor
Network for Intrusion Monitoring", in Swedish National
Computer Networking Workshop, Karlstad, Sweden, 2004.
[39] TCP/IP
over
802.15.4
stack,
URL:
http://mail.millennium.berkeley.edu/pipermail/tinyosdevel/2005-August/000767.html, Accessed in July 2008.
[40] M. Durvy, J. Abeillé, P. Wetterwald, C. O'Flynn, B.
Leverett, E. Gnoske, M. Vidales, G. Mulligan, N. Tsiftes,
© 2010 ACADEMY PUBLISHER
JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 3, MARCH 2010
N. Finne, and A. Dunkels, "Making Sensor Networks IPv6
Ready", in Proceedings of the 6th ACM Conference on
Networked Embedded Sensor Systems (ACM SenSys 2008),
Raleigh, North Carolina, USA, 2008.
Paulo Alexandre Neves is a PhD
student on Informatics Engineering at
the University of Beira Interior under
supervision
by
professor
Joel
Rodrigues. He received his 5-year B.S.
degree (licentiate) in 1998 in
Electronics and Telecommunications
Engineering from University of Aveiro,
2008; and MsC degree in Electronics
and Telecommunications Engineering
from University of Aveiro, Portugal in 2001. He also teaches in
the Informatics Engineering Department at the Superior School
of Technology of the Polytechnic Institute of Castelo Branco,
Portugal. He is a PhD student member of the Institute of
Telecommunications, Portugal. His current research areas are
WSN, integration of the Internet Protocol on WSN, and WSN
management. He authors or co-authors more than 15
international conference papers, participates on several
Technical Program Committees, and also has two accepted
journal publications.
Joel José P. C. Rodrigues is a
Professor at the Department of
Informatics of the University of Beira
Interior,
Covilhã,
Portugal,
and
researcher
at
the
Instituto
de
Telecomunicações,
Portugal.
He
received a PhD degree in Informatics
Engineering, MSc degree from the
University of Beira Interior, Portugal,
and a 5-year B.S. degree (licentiate) in Informatics Engineering
from University of Coimbra, Portugal. His main research
interests include sensor networks, high-speed networks, and
mobile and ubiquitous computing. He is the Editor-in-Chief of
the International Journal on E-Health and Medical
Communications. He is or was the general Chair of the MAN
2009 and 2010 (in conjunction with IEEE ICC 2009 and 2010),
N&G 2010 (with IEEE AINA 2010), Chair of the
Communications Software, Services and Multimedia
Applications Symposium at IEEE Globecom 2010, Chair of the
Symposium on Ad-Hoc and Sensor Networks of the SoftCom
Conference and chaired many other technical committees. Hi is
or was member of many international program committees
(IEEE ICC, IEEE Globecom, IEEE WCNC, IEEE CCNC, IEEE
ISCC, IEEE ICCCN, ICTTA, SoftCOM, etc.) and several
editorial review boards (IEEE Communications Magazine,
Journal of Communications Software and Systems,
International Journal of Communications Systems, International
Journal of Business Data Communications and Networking,
etc.), and he has served as a guest editor for a number of
journals including the Journal of Communications Software and
System. He chaired many technical sessions and gave tutorials
at major international conferences. He has authored or coauthored over 90 papers in refereed international journals and
conferences, a book and a patent pending. He is a licensed
Professional Engineer and he is member of the ACM
SIGCOMM, a member of the Internet Society, IARIA Fellow,
and a Senior Member of the IEEE Computer Society, IEEE
Communications Society and IEEE Education Society, and a
member of several IEEE Technical Committees related with his
research areas.
Download PDF
Similar pages