Computer Networks - Georgia Institute of Technology

Computer Networks 52 (2008) 2260–2279

Contents lists available at ScienceDirect

Computer Networks

journal homepage: www.elsevier .com/locate /comnet

Nanonetworks: A new communication paradigm

Ian F. Akyildiz a,*, Fernando Brunetti b,1, Cristina Blázquez c,2

a Broadband Wireless Networking (BWN) Laboratory, School of Electrical and Computer Engineering, Georgia Institute of Technology,250 14th Street, Atlanta, GA 30332, USAb Bioengineering Group, Institute of Industrial Automation, Spanish Research Council (CSIC), Madrid, Spainc Departament d’Arquitectura de Computadors Universitat Politècnica de Catalunya, Barcelona, Spain

a r t i c l e i n f o

Article history:Received 1 February 2008Received in revised form 1 April 2008Accepted 15 May 2008Available online 6 April 2008

Keywords:NanonetworksNanocommunicationMolecular communication

1389-1286/$ - see front matter � 2008 Elsevier B.Vdoi:10.1016/j.comnet.2008.04.001

* Corresponding author. Tel.: +1 404 894 5141; faE-mail addresses: [email protected] (I.F. Akyild

es (F. Brunetti), [email protected] (C. Blázque1 This work was conducted during his stay at BWN2 This work was conducted during her stay at BW

a b s t r a c t

Nanotechnologies promise new solutions for several applications in biomedical, industrialand military fields. At nano-scale, a nano-machine can be considered as the most basic func-tional unit. Nano-machines are tiny components consisting of an arranged set of molecules,which are able to perform very simple tasks. Nanonetworks. i.e., the interconnection ofnano-machines are expected to expand the capabilities of single nano-machines by allow-ing them to cooperate and share information. Traditional communication technologies arenot suitable for nanonetworks mainly due to the size and power consumption of transceiv-ers, receivers and other components. The use of molecules, instead of electromagnetic oracoustic waves, to encode and transmit the information represents a new communicationparadigm that demands novel solutions such as molecular transceivers, channel modelsor protocols for nanonetworks. In this paper, first the state-of-the-art in nano-machines,including architectural aspects, expected features of future nano-machines, and currentdevelopments are presented for a better understanding of nanonetwork scenarios. More-over, nanonetworks features and components are explained and compared with traditionalcommunication networks. Also some interesting and important applications for nanonet-works are highlighted to motivate the communication needs between the nano-machines.Furthermore, nanonetworks for short-range communication based on calcium signalingand molecular motors as well as for long-range communication based on pheromones areexplained in detail. Finally, open research challenges, such as the development of networkcomponents, molecular communication theory, and the development of new architecturesand protocols, are presented which need to be solved in order to pave the way for the devel-opment and deployment of nanonetworks within the next couple of decades.

� 2008 Elsevier B.V. All rights reserved.

1. Introduction

The concepts in nanotechnology was first pointed outby the 1965 nobel laureate physicist Richard Feynman inhis famous speech entitled ‘‘There’s Plenty of Room at theBottom” in December 1959. The main focus of his speechwas about the field of miniaturization and how he be-

. All rights reserved.

x: +1 404 894 7883.iz), [email protected]).

Lab in 2007–2008.N Lab in 2007–2008.

lieved humans would create increasingly tinier and pow-erful devices in the future. The term ‘‘nanotechnology”was first defined by [59] 15 years later as: ‘‘Nanotechnol-ogy mainly consists of the processing of, separation, con-solidation, and deformation of materials by one atom orby one molecule”. In the 1980s, the basic idea of this def-inition was explored in much more depth by [26] whotook the Feynman concept of a billion tiny factories andadded the idea that they could replicate themselves, viacomputer control instead of control by a human operator.Activity surrounding nanotechnology began to slowly in-crease and the advancements really began to acceleratein the early 2000s.

mailto:[email protected]




http://www.sciencedirect.com/science/journal/13891286

http://www.elsevier.com/locate/comnet

I.F. Akyildiz et al. / Computer Networks 52 (2008) 2260–2279 2261

Nanotechnology enables the miniaturization and fabri-cation of devices in a scale ranging from 1 to 100 nanome-ters. At this scale, a nano-machine can be considered as themost basic functional unit. Nano-machines are tiny com-ponents consisting of an arranged set of molecules whichare able to perform very simple computation, sensingand/or actuation tasks [57]. Nano-machines can be furtherused as building blocks for the development of more com-plex systems such as nano-robots and computing devicessuch as nano-processors, nano-memory or nano-clocks.

Nano-machines can be interconnected to execute collab-orative tasks in a distributed manner. Resulting nanonet-works are envisaged to expand the capabilities andapplications of single nano-machines in the following ways:

� Nano-machines such as chemical sensors, nano-valves,nano-switches, or molecular elevators [4], cannot exe-cute complex tasks by themselves. The exchange ofinformation and commands between networked nano-machines will allow them to work in a cooperativeand synchronous manner to perform more complextasks such as in-body drug delivery or diseasetreatments.

� The workspace of a single nano-machine is extremelylimited. Nanonetworks will allow dense deploymentsof interconnected nano-machines. Thus, larger applica-tion scenarios will be enabled, such as monitoring andcontrol of chemical agents in ambient air.

� In some application scenarios, nano-machines will bedeployed over large areas, ranging from meters to kilo-meters. In these scenarios, the control of a specificnano-machine is extremely difficult due to its small size.Nanonetworks will enable the interaction with remotenano-machines by means of broadcasting and multihopcommunication mechanisms.

Communication between nano-machines can be real-ized through nanomechanical, acoustic, electromagneticand chemical or molecular communication means [31].

Nanomechanical communication is defined as the trans-mission of information through mechanical contactbetween the transmitter and the receiver. In acoustic com-munication, the transmitted message is encoded usingacoustic energy, i.e., pressure variations. Electromagneticcommunication is based on the modulation of electromag-netic waves to transmit information. Molecular communica-tion can be formally defined as the use of molecules asmessages between transmitters and receivers.

Molecular communication is the most promising ap-proach for nano-networking based on the followingadvantages:

� Due to the size and principles of traditional acoustictransducers and radiofrequency transceivers, their inte-gration at molecular or nano-scale is not feasible [31].By contrast, molecular transceivers are intrinsically con-ceived at nano-scale. These are nano-machines whichare able to emit and receive molecules.

� In nanomechanical communication, transmitters andreceivers need to be in direct contact. This is not arestriction for molecular communication over large

areas, where transmitters and receivers can be remotelylocated as long as the transmitted molecules reach theintended receiver.

In the recent literature, the term ‘‘nanonetworks” re-fers to electronic components and their interconnectionwithin a single chip on a nano-scale [12]. This conceptis also known as Network on Chip (NoC). This term isalso referred to as the network-like interconnection ofnanomaterials as well, e.g., carbon nanotubes arrays[38,15]. In this paper, we use the term ‘‘nanonetworks”strictly for the interconnection of nano-machines basedon molecular communication.

This paper follows the bio-inspired approach to explore,from a telecommunication point of view, the potential ofmolecular communication for nanonetworks. First, in Sec-tion 2, we present the nano-machines including the state-of-the-art in research and current approaches for theirdevelopment. In Section 4, we explain nanonetwork fea-tures and their advantages and disadvantages over tradi-tional communication networks. In Section 3, we explainpotential applications of nanonetworks. We explore exist-ing biological models and techniques for molecular com-munication for short and long-ranges in Sections 5–7,respectively. Finally, we outline open research issues andcontrast them with traditional communication networkchallenges in Section 8. We conclude the paper in Section9.

2. Overview of Nano-machines

A nano-machine is defined as ‘‘an artificial eutacticmechanical device that relies on nanometer-scale compo-nents” [24]. Also the term ‘‘molecular machine” is definedas ‘‘a mechanical device that performs a useful function usingcomponents of nanometer-scale and defined molecular struc-ture; includes both artificial nano-machines and naturallyoccurring devices found in biological systems”.

In general terms, we define a nano-machine as ‘‘a de-vice, consisting of nano-scale components, able to perform aspecific task at nano-level, such as communicating, comput-ing, data storing, sensing and/or actuation”. The tasks per-formed by one nano-machine are very simple andrestricted to its close environment due to its low complex-ity and small size.

There are three different approaches for the develop-ment of nano-machines as depicted in Fig. 1. In the top-down approach, nano-machines are developed by meansof downscaling current microelectronic and micro-elec-tro-mechanical technologies without atomic level control.In the bottom-up approach, the design of nano-machinesis realized from molecular components, which assemblethemselves chemically by principles of molecular recogni-tion arranging molecule by molecule. Recently, a third ap-proach called bio-hybrid is proposed for the developmentof nano-machines [63]. This approach is based on the useof existing biological nano-machines, such as molecularmotors, as components or models for the development ofnew nano-machines. In Fig. 1, different systems aremapped according to their origin, biological or man-made

Nature

Man-made

Scalem mm μm nm

Humans CellsInsects Cellorganells

Computers Micro-electonics Nano-electronics

MEMS NEMS

Robots Micro-robots Nano-robots NANOMACHINESMolecules

Top-downBottom-up

Bio-hybridBacteria

Fig. 1. Approaches for the development of nano-machines.

2262 I.F. Akyildiz et al. / Computer Networks 52 (2008) 2260–2279

systems, and to their size, ranging from nanometers to me-ters. In the future, nano-machines will be obtained follow-ing any of these three approaches. However, the existenceof successful biological nano-machines, which are highlyoptimized in terms of architecture, power consumptionand communication, motivate their use as models or build-ing blocks for new developments.

2.1. Development of nano-machines

2.1.1. Top-down approachRecently, newest manufacturing processes, such as the

45 nm lithographic process, have made the integration ofnano-scale electronic components in a single device possi-ble [39]. The top-down approach is focused on the develop-ment of nano-scale objects by downscaling currentexisting micro-scale level device components. To achievethis goal, advanced manufacturing techniques, such as elec-tron beam lithography [61] and micro-contact printing [42],are used. Resulting devices keep the architecture of pre-existing micro-scale components such as microelectronicdevices and micro-electro-mechanical systems (MEMS).

Nano-machines, such as nano-electromechanical sys-tems (NEMS) components, are being developed using thisapproach [19,34,45]. However, the fabrication and assem-bly of these nano-machines is still at an early stage. Sofar, only simple mechanical structures, such as nano-gears[66], can be created following this approach.

2.1.2. Bottom-up approachIn the bottom-up approach, nano-machines are devel-

oped using individual molecules as the building blocks.Recently, many nano-machines, such as moleculardifferential gears and pumps [51], have been theoreticallydesigned using a discrete number of molecules. Manufac-turing technologies able to assemble nano-machines mol-ecule by molecule do not exist, but once they do; nano-machines could be efficiently created by the precise andcontrolled arrangement of molecules. This process is calledmolecular manufacturing [24].

Molecular manufacturing could be developed from cur-rent technologies in couple decades if adequate resourcesare invested. Current development of nano-machines usingthis bottom-up approach, such as molecular switches [5]

and molecular shuttles [6], are based on self-assemblymolecular properties [7].

2.1.3. Bio-hybrid approachSeveral biological structures found in living organisms

can be considered as nano-machines. Most of these bio-logical nano-machines can be found in cells. Biologicalnano-machines in cells include: nano-biosensors, nano-actuators, biological data storing components, tools andcontrol units [25]. Expected features of future nano-machines are already present in a living cell, which canbe defined as a self-replicating collection of nano-machines[63]. Several biological nano-machines are interconnectedin order to perform more complex tasks such as celldivision. The resulting nanonetwork is based on molecularsignaling. This communication technique is also used forinter-cell communication allowing multiple cells to coop-erate to achieve a common objective such as the controlof hormonal activities or immune system responses inhumans.

The bio-hybrid approach proposes the use of these bio-logical nano-machines as models to develop new nano-machines or to use them as building blocks integratingthem into more complex systems such as nano-robots. Fol-lowing this approach, the use of a biological nano-motor topower a nano-device has been reported in [56]. Anotherexample in line with this approach is the use of bacteriaas controlled propulsion mechanisms for the transport ofmicro-scale objects [10].

2.2. Expected features of nano-machines

The main constraint in the development of nano-ma-chines is the lack of tools which are able to handle andassemble molecular structures in a precise way. However,given the rapid advances in manufacturing technologies,efficient fabrication of more complex nano-machines willbe possible in the near future. They are expected to includemost of the functionalities of existing devices at micro-scale. In addition, nano-machines will present novel fea-tures enabled by molecular properties of the materials thatcan be exploited at nano-scale. The most important and ex-pected features of future nano-machines can be describedas follows:

Processing and ControlUnit

Storage Unit

Sensors Actuators

Power Unit

Energy Scavenger

Transceiver LocationSystem

1

23

4 5

67

Microrobot node


� Nano-machines will be intrinsically self-contained. Thismeans that each nano-machine will contain a set ofinstructions or code to realize the intended task. Theseinstructions or sequence of operations can be embeddedin the molecular structure of nano-machines, or can beread from another molecular structure in which theinstruction set is stored.

� Self-assembly is defined as the process in which severaldisordered elements form an organized structure with-out external intervention, as a result of local interactionsbetween them. At nano-level, self-assembly is naturallydriven by molecular affinities between two different ele-ments. Self-assembly will leverage the development ofnano-machines and will allow them to interact withexternal molecules in an autonomous way.

� Self-replication is defined as the process in which adevice makes a copy of itself using external elements.This potential process will enable the creation of largenumber of nano-machines to realize macroscopic tasksin an inexpensive way [43]. Similar to the first feature,self-replication implies that the nano-machine containsthe instructions to create a copy of itself.

� Locomotion is the ability to move from one place toanother. Nano-machines are aimed to accomplish spe-cific tasks, which are usually described by a spatial-tem-poral actuation. This means that a nano-machine shouldbe located in the right place at the right time to accom-plish the task. However, no single nano-machine is ableto move towards a previously identified target. Morecomplex systems could use embedded nano-sensorsand nano-propellers to detect and follow specific tracesof the target. Locomotion will enable the use of nano-machines in applications where mobile actors areneeded, e.g., nano-robots for disease treatments [18].

� Communication between nano-machines is needed toallow them to realize more complex tasks in a coopera-tive manner. At this level, as explained in Section 1, themost promising technique is based on molecular com-munication. Further advances in nano-sensors andnano-actuators are expected to enable the integrationof molecular transceivers into nano-machines.

Gap Junctions

Nucleus

Receptors

Flagellum

Mitochondrion

ADN

Vacuoles

1

2

3

4

5

67

Nanonetwork node(Cell inspired)

Fig. 2. Functional architecture mapping between nano-machines of amicro or nano-robot, and nano-machines found in cells.

2.3. Nano-machine architecture

A nano-machine could consist of one or more compo-nents, resulting in different levels of complexity, whichcould be from simple molecular switches to nano-robots[18]. The most complete nano-machines will include thefollowing architecture components:

(1) Control unit. It is aimed at executing the instructionsto perform the intended tasks. To achieve this goal, itcan control all the other components of the nano-machine. The control unit could include a storageunit, in which the information of the nano-machineis saved.

(2) Communication unit. It consists of a transceiver ableto transmit and receive messages at nano-level,e.g., molecules.

(3) Reproduction unit. The function of this unit is to fab-ricate each component of the nano-machine using

external elements, and then assemble them to repli-cate the nano-machine. This unit is provided with allthe instructions needed to realize this task.

(4) Power unit. This unit is aimed at powering all thecomponents of the nano-machine. The unit will beable to scavenge energy from external sources suchas light, temperature and store it for a later distribu-tion and consumption.

(5) Sensor and actuators. Similar to the communicationunit, these components act as an interface betweenthe environment and the nano-machine. Several sen-sors and/or actuators can be included in a nano-machine, e.g., temperature sensors, chemical sensors,clamps, pumps, motor or locomotion mechanisms.

Currently such complex nano-machines cannot be built.However, there exist systems found in the nature, such asliving cells, with similar architectures. According to thebio-hybrid approach these biological models, i.e., the cells,can be used to learn and understand the principles govern-ing the operation of nano-machines and their interactions.This knowledge is expected to contribute to the develop-ment of new bio-inspired nano-machines and systems forspecific purposes. In Fig. 2, we show a component mapping


between a generic architecture of a nano-machine and aliving cell, including its biological nano-machines. Similarto the architecture of a nano-machine, a cell contains thefollowing components:

(1) Control unit. The nucleus can be considered as thecontrol unit of the cell. It contains all the instruc-tions to realize the intended cell functions.

(2) Communication. The gap junctions and hormonal andpheromonal receptors, located on the cell mem-brane, act as molecular transceivers for inter-cellcommunication.

(3) Reproduction. Several nano-machines are involved inthe reproduction process of the cell such as the cen-trosome and some molecular motors. The code ofthe nano-machine is stored in molecular sequences,which are duplicated before the cell division. Eachresulting cell will contain a copy of the originalDNA sequence.

(4) Power unit. Cells can include different nano-machines for power generation. One of them is themitochondrion that generates most of the chemicalsubstances, which are used as energy in many cellu-lar processes. Another interesting nano-machine isthe chloroplast, which converts sunlight into chemi-cal fuel.

(5) Sensors and actuators. Cells can include several sen-sors and actuators such as the Transient ReceptorPotential channels for tastes and the flagellum ofthe bacteria for locomotion. The chloroplast of theplants can also be considered as an actuator sinceit transforms water to oxygen that is later releasedto the environment.

A bio-hybrid approach cannot only be used to developnovel nano-machines, but also to understand their interac-tions in larger systems such as cells. These interactions,exclusively enabled by molecular communication tech-niques, are essential since this is the only way to exploretheir capabilities to achieve more complex tasks in a coop-erative manner.

In this paper, we expand the bio-hybrid approach be-yond the models for novel nano-machines in order to studyand develop molecular communication techniques fortheir interconnection.

3. Nanonetworks applications

The potential applications of nanonetworks are unlim-ited. We classify them in four groups: biomedical, environ-mental, industrial and military applications. However,since nanotechnologies have a key role in the manufactur-ing process of several devices, nanonetworks could be usedextensively in many other fields such as consumer elec-tronics, life style and home appliances among others.

3.1. Biomedical applications

The most direct applications of nano-machines andnanonetworks are in the biomedical field. Biological mod-

els inspire and encourage the use of nanotechnologies tointeract with organs and tissues. The advantages providedby nanonetworks are clearly in terms of size, biocompati-bility and biostability, enabled by the control of systemcomponents at molecular level. These are some of theenvisaged applications:

� Immune system support. The immune system is com-posed by several nano-machines that protect organismagainst diseases. These nano-machines are a collectionof nano, micro and macro systems, including sensorsand actuators, acting in a coordinated way to identifyand control foreign and pathogen elements. Nano-machines can be used to help the detection and elimina-tion procedures. They could realize tasks of localizationand response to malicious agents and cells, such as can-cer cells [20,32], resulting in a less aggressive and inva-sive treatments compared to the existing ones.

� Bio-hybrid implants. These are aimed at supporting orreplacing components such as organs, nervous tracksor lost tissues [24,30]. Nanonetworks can providefriendly interfaces between the implant and the envi-ronment. Restoration of central nervous system tracksis a possible application of bio-hybrid implants.

� Drug delivery systems. These are another specific type ofimplants. For instance, these systems could help to com-pensate metabolism diseases such as diabetes. In thissense, nano-sensors and smart glucose reservoirs or pro-ducers can work in a cooperative manner to support reg-ulating mechanisms [29]. Drug delivery systems couldalso help to mitigate the effects of neurodegenerativediseases by delivering neurotransmitters or specificdrugs [64].

� Health monitoring. Oxygen and cholesterol level, hor-monal disorders, and early diagnosis are some examplesof possible applications that can take advantage of in-body nano-sensor networks [22,30]. The informationretrieved by these systems must be accessible by rele-vant actors. Thus, nanonetworks must provide theproper level of connectivity to deliver the sensedinformation.

� Genetic engineering. Manipulation and modification ofnano-structures such as molecular sequences and genescan be achieved by nano-machines. The use of nanonet-works will allow expanding the potential applications ingenetic engineering.

3.2. Industrial and consumer goods applications

Nanonetworks will be used not only the intra-body butalso in industries. Nanonetworks can help with the devel-opment of new materials, manufacturing processes andquality control procedures. More specifically, these appli-cations have already been proposed:

� Food and water quality control. Similar to health monitor-ing applications, food and fluids quality control can takeadvantage of nanonetworks. Nano-sensor networks canhelp detecting small bacteria and toxic components thatcan affect to the product quality and cannot be detected


using traditional sensing technologies [3]. Advancedself-powered nano-sensor networks can be used to detectvery small amount of chemical or biological agentsinstalled in water supplies across the country.

� Functionalized materials and fabrics. Nanonetworks canbe included in advanced fabrics and materials to getnew and improved functionalities. Antimicrobial andstain-repeller textiles are being developed using nano-functionalized materials [60]. For instance, nano-actua-tors can help to improve the airflow in smart fabrics.These nano-actuators can communicate to nano-sensorsto control the proper reaction based on the externalconditions.

3.3. Military applications

Nanotechnologies can also have several applications inthe military field. While in the applications pointed out be-fore, the range covered by nanonetworks is short, in mili-tary field the deployment range of nanonetworks can bewidely variable depending on the application. Battlefieldmonitoring and actuation demand a dense deployment ofnanonetworks over large areas, while systems aimed atmonitoring soldier performance are deployed in smallerareas, i.e., human body. Among the military applications,we can mention:

� Nuclear, biological and chemical (NBC) defenses. This is aclassic application for large area deployment. Nanonet-works can be deployed over the battlefield or targetedareas to detect aggressive chemical and biologicalagents and coordinate the defensive response [55]. Theoverall system can be compounded of nano-sensorsand nano-actuators, which would detect and controlthe hostile agents. Nano-sensor networks could also bedeployed into cargo containers to detect the unautho-rized entrance of chemical, biological or radiologicalmaterials.

� Nano-functionalized equipments. These applications aresimilar to those found in consumer goods field, but itis focused on military equipment. Advanced camouflageas well as army uniforms can take advantage of nanon-etworks. For instance, equipments can be manufacturedusing advanced materials containing nanonetworks thatself-regulate the temperature underneath soldiersclothes [27] and even detect whether the soldier hasbeen injured.

3.4. Environmental applications

Since nanoneworks are inspired in biological systemsfound in nature, they can also be applied in environmentalfields achieving several goals that could not be solved withcurrent technologies. Some environmental applicationsare:

� Biodegradation. There is an existing and growing prob-lem with garbage handling around the world. In thisline, nanonetworks could help with the biodegradation

process in the garbage dumps. Nanonetworks can helpthe biodegradation process by sensing and tagging dif-ferent materials that can be later located and processedby smart nano-actuators.

� Animals and biodiversity control. Nanonetworks can bealso used in natural environments to control severalspecies. For instance, as seen in nature, nanonetworksusing pheromones as messages can trigger certainbehaviors on animals. As a result it is possible to interactwith those animals and also to control their presence inparticular areas.

� Air pollution control. Similar to the quality control appli-cations, air can be monitored using nanonetworks.Moreover, nano-filters can be developed to improvethe air quality by removing harmful substances con-tained in it [35].

4. Communication among nano-machines

Among all of the expected features of future nano-ma-chines, the communication capabilities are also veryimportant. This is the only feature that enables them towork in a synchronous, supervised and cooperative man-ner to pursuit a common objective.

Nano-machines communication can include the twofollowing bidirectional scenarios:

(1) Communication between a nano-machine and a lar-ger system such as electronic micro-devices, and

(2) Communication between two or more nano-machines.

Different communication technologies, such as electro-magnetic, acoustic, nanomechanical or molecular; havebeen proposed for each scenario in [31].

Communication based on electromagnetic waves is themost common technique to interconnect microelectronicdevices. These waves can propagate with minimal lossesalong wires or through air. However, given the size ofnano-machines, wiring a large quantity of them is unfeasi-ble. As an alternative, wireless solutions could be used. Toestablish a bidirectional wireless communication, a radio-frequency transceiver should be integrated in the nano-machine. Nano-scale antennas could be developed for veryhigh frequency communication. However, due to the sizeand current complexity of the transceivers, they still can-not be easily integrated into nano-machines. In addition,if the integration were possible, the output power of thenano-transceiver would not be enough to establish a bidi-rectional communication channel [31]. As a consequence,electromagnetic communication could be used to transmitinformation from a micro-device to a nano-machine, butnot from nano-machines to micro-devices, nor amongnano-machines.

At nano-level, acoustic communication is mainly basedon the transmission of ultrasonic waves. Similar to thecommunication based on electromagnetic waves, theacoustic communication implies the integration of ultra-sonic transducers in the nano-machines. These transducersshould be capable to sense the rapid variations of pressureproduced by ultrasonic waves and to emit acoustic signals


accordingly. Again, the size of these transducers representsthe major barrier in their integration in the nano-machines.

In nanomechanical communication, the information istransmitted through hard junctions between linked de-vices at nano-level. The main drawback of this type of com-munication is that a physical contact between thetransmitter and the receiver is required. Moreover, thiscoupling should be precise enough to guarantee that thedesired mechanical transceivers are aligned correctly. Thiscommunication technique is not suitable in many applica-tion scenarios where nano-machines are deployed overlarge areas without any direct contact between them. Inaddition, without precise navigation systems in nano-ma-chines, their positioning for a correct nanomechanicalcommunication is a major barrier.

Molecular communication is defined as the transmissionand reception of information encoded in molecules. Molec-ular communication is a new and interdisciplinary fieldthat spans nano, bio and communication technologies[46]. Unlike previous communication techniques, the inte-gration process of molecular transceivers in nano-ma-chines is more feasible due to the size and naturaldomain of molecular transceivers, i.e., nano-scale frame-work. These transceivers are nano-machines which areable to react to specific molecules, and to release othersas a response to an internal command.

Molecular communication can be used to interconnectmultiple nano-machines, resulting in nanonetworks as de-fined in Section 1. Nanonetworks expand the capabilities ofsingle nano-machines in the following terms:

� More complex objectives can be achieved if multiplenano-machines cooperate. Nanonetworks enable thiscooperation by providing mechanisms to exchangeinformation between different nano-machines such asmolecular motors or nano-switches.

� Single nano-machines can only perform tasks at nano-level, and therefore their workspace is very limited.However, if a large number of nano-machines are inter-connected, they can pursuit macro-scale objectives, andwork over larger areas, such as treatment of cancertumors or air pollution monitoring.

� If multiple nano-machines are deployed over largeareas, the interaction with a specific nano-machine isextremely difficult due to its small size. This interactionincludes procedures such as nano-machines activation/deactivation, configuration of parameters, data acquisi-tion or actuation commands. Nanonetworks willenable this interaction by providing the infrastructureand mechanism to broadcast the information overthese areas. In addition, two nano-machines couldinteract indirectly by using other nano-machines asrepeaters.

Besides expanding capabilities of single nano-machines,nanonetworks represent a potential solution for someapplications where available communication networksand micro-devices are not suitable. Compared to currentcommunication network technologies, nanonetworks havethe following advantages:

� The reduced size of nano-machines and the resultingnanonetwork components can be an advantage in manyapplications where the dimension of the involved sys-tems is critic. For instance, in the biomedical field,nano-machines can be used for intra-body applicationsallowing nanonetworks to lead to nano-invasive andmore selective treatments.

� Biocompatibility is defined as the quality of a device tooperate accordingly in biological environments withoutaffecting them negatively. In some biomedical applica-tions, many electronic devices have to cope with hostileenvironments as well as many organisms rejectimplants and drugs. Nanotechnologies can be used toenhance the compatibility between nano-machinesand natural organs or tissues by means of more friendlymaterials and interfaces. For instance, bio-hybrid nano-machines compound by biological elements can interactwith natural processes without any side effect. In addi-tion, nano-machines and molecular messages may alsobe programmed to deactivate after completing thenanonetwork task preventing removal procedures.Nanotechnologies, allow the control of materials atmolecular level. Using these materials, we can designnanonetworks nodes according to specific environmen-tal conditions improving the biocompatibility of thesystem.

� Chemical reactions are highly efficient in terms of energyconsumption [47]. These reactions will power the nanon-etworks nodes and processes. Chemical reactions canalso represent complex computation and decision pro-cesses, which in traditional communication could meanmultiple operations.

4.1. Nanonetworks versus Traditional communicationnetworks

Nanonetworks are not a simple extension of traditionalcommunication networks at the nano-scale. They are acomplete new communication paradigm, in which mostof the communication processes are inspired by biologicalsystems found in nature. Nanonetworks have the followingdifferences with traditional communication networks:

� In nanonetworks, the message is encoded using molecules;while in traditional communication networks, the infor-mation is encoded in electromagnetic, acoustic or opti-cal signals. Two different and complementary codingtechniques can be considered to represent the informa-tion in nanonetworks. The first one uses temporalsequences to encode the information, such as the tem-poral concentration of specific molecules in the med-ium. According to the level of the concentration, i.e.,the number of molecules per volume, the receptordecodes the received message. For instance, this tech-nique is used by the Central Nervous System to propa-gate the neural impulses. This technique can beconsidered similar to those used in traditional networkswhere time-varying sequences transport the informa-tion. The second technique, hereinafter called molecularencoding, uses internal parameters of the molecules to


encode the information such as the chemical structure,relative positioning of molecular elements or polariza-tion [46]. In this case, the receiver must be able to detectthese specific molecules to decode the information. Thistechnique is similar to the use of encrypted packets incommunication networks, in which only the intendedreceiver is capable to read the information. In nature,molecular encoding is used in pheromonal communica-tion, where only members of the transmitter specie candecode the transmitted message.

� The propagation speed of signals used in traditional com-munication networks, such as electromagnetic or acous-tic waves, is much faster than the propagation ofmolecular messages. In nanonetworks, the information,i.e., molecules, has to be physically transported fromthe transmitter to the receiver. In addition, moleculescan be subject to random diffusion processes and envi-ronmental conditions, such as temperature, which canaffect the propagation of the molecular messages.

� In traditional communication networks, noise isdescribed as an undesired signal overlapped with thesignals transporting the information. In nanonetworks,according to the coding techniques, two different typesof noise can affect the messages. First, as occurs in tradi-tional communication systems, noise can be overlappedwith molecular signals such as concentration level ofmolecules. This means that another source emits thesame molecules used to encode the message, thereforethey affect the concentration sensed by the receiver. Innanonetworks, noise can also be understood as an unde-sired reaction occurring between information moleculesand other molecules present in the environment. Thesereactions can modify the structure of the informationmolecules and therefore the receiver would not be ableto detect the transmitted message.

� Text, voice and video are usually transmitted overtraditional communication networks. By contrast, innanonetworks, since the message is a molecule, thetransmitted information is more related to phenomena,chemical states and processes [11].

� In nanonetworks, most of the processes are chemicallydriven resulting in low power consumption. In tradi-tional communication networks the communicationprocesses consume electrical power that is obtainedfrom batteries or from external sources such as electro-magnetic induction.

We summarize nanonetworks and traditional commu-nication network features in Table 1 [37]. Most of the exist-

Table 1Main differences between traditional communication networks and nanonetwork

Communication Traditional

Communication carrier Electromagnetic wavesSignal type Electronic and optical (ElectromPropagation speed Light (3� 108 m/s)Medium conditions Wired: Almost immune

Wireless: Affect communicationNoise Electromagnetic fields and signaEncoded information Voice, text and videoOther features High energy consumption

ing communication networks knowledge is not suitable fornanonetworks due to their particular features. Nanonet-works require innovative networking solutions accordingto the characteristics of the network components and themolecular communication processes.

4.2. Nanonetwork components

The first models of nanonetworks are based on thoseused in Information and Communication Technologies(ICT) for telecommunication networks. In Fig. 3, we showthe general concepts of nanonetworks versus existing tele-communication systems. Nanonetworks components arefunctionally similar to those found in traditional networks.

In nanonetworks, we can identify five different compo-nents: the transmitter node, the receiver node, the mes-sages, the carrier, and the medium. Each of thesecomponents affects the overall communication process,which includes the following steps as Fig. 3b depicts:

(1) The transmitter encodes the message ontomolecules.

(2) The transmitter inserts the message into the med-ium by releasing the molecules to the environmentor attaching them to molecular carriers.

(3) The message propagates from the transmitter to thereceiver.

(4) The receiver detects the message.(5) The receiver decodes the molecular message into

useful information such as reaction, data storing,actuation commands, etc.

4.2.1. Transmitters and receiversThe transmitter role is played by a nano-machine such

as a modified living cell, a biomedical implant, or a nano-robot. In most of traditional communication networks,the nodes encode the information by modulating electro-magnetic signals. In nanonetworks the nodes encode theinformation by modifying molecules by means of chemicalreactions. Nano-machines also play the receiver role. Thereceiver should be able to extract the message from themedium. Since nano-machines are very simple, and themessage is basically a molecule, handling this messagecould not be a trivial task. The message contains usefulinformation that is going to be used by the receiver. Atmolecular level, this can mean to forward the message,to store it, or to react to it. Nanonetwork nodes transmitand receive molecules. These molecules are a finite source.

s enabled by molecular communication [37]

Molecular

Moleculesagnetic) Chemical

Extremely lowAffect communication

ls Particles and molecules in mediumPhenomena, chemical states or processesLow energy consumption

TX

RX

Electromagneticwaves

Propagation speed (Light)

RX

TXNano-machine

RXNano-machine

Encoder

Decoder

RXNano-machine

Matter(Molecules)

Low propagation speed

(Random process)

Sensitive toobstacles

Noise

a

b

Fig. 3. Overview of (a) traditional communication systems and (b) nanonetworks: energy transmission vs. molecular transmission.


This has two consequences. First, the transmitter nano-ma-chines should be able to obtain, from an external source,raw molecules and store them for a later use as messages.Second, the receiver nano-machines should be able to buf-fer a limited number of molecules, and also include releasemechanisms to empty this molecular buffer to be ready forthe next messages.

4.2.2. MessageIn traditional computer networks, the information is

represented using a binary system and the transmittedmessage is usually a set of bits. In molecular communica-tion, the message is a molecule. This molecular messagewill have three important characteristics [31]. First, it willpresent a predefined external structure that will allow aneasy recognition at the receiver. Second, it will be inactive.This means that molecular messages will not be prone toreact to other molecules in the medium. Finally, molecularmessages should easily be eliminated without any side ef-fect once they are decoded at the receiver nano-machine.In [37], some molecular features such as polarity, motion,magnetization, and structures, are suggested for informa-tion coding. A potential candidate for molecular messagingis the partially fluorinated polyethylene proposed in [24].

This molecule can be compounded by a sequence of hydro-gen and flour atoms, resulting in a binary sequence able tostore digital data [9].

4.2.3. CarrierIn classical communication paradigm, a carrier is used to

transport the message. The carrier is used because of its sig-nal feature advantages, especially in terms of propagation.In traditional computer networks, the carriers also allowcreating multiple communication channels. In nanonet-works, the carriers are particular molecules which are ableto transport chemosignals or molecular structures contain-ing the information. The aim of the molecular carriers is toenhance the propagation features of single informationmolecules, to create more reliable communication by pro-tecting molecules from external noise sources, and finallyto enable the creation of multiple independent channelsusing the same medium. The use of molecules as informa-tion carriers in molecular communication was observed inbiological systems. According to these observations, twopotential types of carriers have been identified: the molecu-lar motors and the calcium ions [50]. Molecular motors, e.g.,kinesin, dynein and myosin, are proteins that can generatemovements using chemical energy. These protein motors


can transport a data packet, i.e., molecule, from the trans-mitter to the receiver as described in [49]. The second typeof carrier consists of calcium ions (Ca2+). The transmitter canmodulate the concentration of these ions in amplitude andfrequency to encode the information [17].

4.2.4. MediumWireless communication networks have been devel-

oped for almost every type of media found in nature suchas airborne, waterborne, and underground. Many wirelesstransmission technologies have also been applied success-fully to communicate with implantable devices. The propa-gation of typical network signals, such as acoustic, opticalor electromagnetic, can be affected by the medium condi-tions. In molecular communication, the medium can bewet or dry, e.g., in-body or environmental monitoring nan-onetworks, and the propagation is even more dependent onthe medium conditions. The speed of the medium, which isfaster than the propagation speed of the molecules, can af-fect the communication between nano-machines. In addi-tion, physical obstacles on the signal pathway can impedethe propagation of the molecules, which can go through so-lid objects. These radical changes oblige the review of manycommon communication concepts such radiation, range, aswell as the development of new propagation models.

5. Short-range communication using molecular motors

First nanonetworks models are inspired by molecularcommunication schemes observed in biological systems.Biological nanonetworks are used for intra-cell, inter-celland intra-specie communication. Intra-cell and inter-cellcommunication are referred as short-range techniquesdue to the size of living cells and their internal compo-nents. Thus, in the framework of nanonetworks, we classifyas short-range the communication process that takes placein a range from nm to few mm. Complementarily, we clas-sify as long-range communication, the interconnection oftwo nano-machines in a range of mm to km as occurs inpheromonal communication.

Most of the intra-cell communications are based onmolecular motors. Molecular motors, e.g., dynein, are pro-teins or protein complexes that transform chemical energyinto mechanical work at a nano-scale [16]. Their use asinformation shuttles or communication carriers for nano-machines within a short-range has been widely proposed[21,46,54]. These molecular motors can be found ineukaryotic cells in living organisms.

Inside the cells, molecular motors are aimed at trans-porting essential particles among organelles during differ-ent stages of the cell life cycle. They travel or move alongmolecular rails called microtubules. These microtubulesare widely deployed setting a complete railway networkfor intra-cell substance transportation. The microtubulesgo from the centrosome, i.e., cell organelle, outwards thecell membrane resembling a star topology in communica-tion networks.

Specific molecular motors able to carry other moleculesuse these intracellular molecular rails. Depending on thecargo, molecular motors move at a speed up to 400 mm/

day. The ability to move molecules makes molecular mo-tors a feasible way to transport information packets, i.e.,molecules, from the transmitter to the receiver nano-ma-chine. As an additional advantage, molecular motors andmicrotubules can be used as a bio-hybrid communicationinterface between man-made nano-machines and biologi-cal structures. For instance, nano-machines could use thiscommunication interface to interact with cell organellesto achieve specific tasks, such as drug delivery.

The deployment of the microtubules, i.e., network infra-structure, to support the movement of the molecular mo-tors is a key issue. Two different solutions could be usedfor the deployment of the network infrastructure. The firstone is based on natural mechanisms found in cells, inwhich microtubules can be developed by means of molec-ular self-assembly [62]. The second solution is based on thedevelopment of lithographic tracks of molecular motors tointerconnect nano-machines. These tracks, which are madeof molecular motors attached to a surface, can move struc-tures similar to the microtubules from one point to another[36]. In this solution, the information molecule is not at-tached to the molecular motor but to the microtubule ora structure with similar characteristics [48].

The concept behind these two solutions is the same. Inboth solutions, the movement created by molecular mo-tors is used to transport the information molecules. Fur-ther description of this molecular communicationtechnique will consider the scenario according to the firstsolution. The scenario is assumed to include system com-ponents such as pre-deployed microtubules betweennano-machines and traveling molecular motors. Thesecomponents, shown in Fig. 4, should present specific fea-tures to support all of the communication processes be-tween the nano-machines.

5.1. Communication features

Molecular communication enabled by molecular mo-tors takes place in aqueous medium. The environmentshould include the necessary components at biologicallyappropriate conditions [28], such as temperature, humid-ity, medium viscosity and pH. Due to the organic andchemical nature of the involved nano-machines and infor-mation packets, the nanonetwork is highly sensitive tothese conditions and the communication process can benegatively affected by sudden variations of the environ-mental conditions.

In molecular communication based on molecular mo-tors, the communication process includes one transmitterand only one receiver. When the communication processoccurs, the information molecule is altered and the en-coded information is lost. Therefore the molecule informa-tion restricts the communication process to one receiver.The outcome is similar to a point-to-point physical com-munication link in traditional communication networks.To implement more complex communication schemesincluding multihoping techniques, the receiver nano-ma-chines should be able to amplify, decode and redirect theinformation.

On the transmitter side, the information moleculesare loaded on molecular motors, which transport the

TX

RX

Decoder

Encoder

De de

Molecular rail

Molecular motor

Fig. 4. System components in molecular motors communication systems.


information along the microtubules to the receiver.The packets can be encapsulated in vesicles. A vesicle is afluid or an air-filled cavity that can store or digest cellularproducts [1]. The objective if this encapsulation of theinformation is twofold [48]. First, it allows enhancingthe compatibility between the information molecule andthe molecular motor, enabling the use of diverse typesof molecules as information packets. Second, theencapsulation protects the information molecules avoidingthem to react with antagonistic receptors present in themedium.

The network infrastructure should be deployed priorto the beginning of the communication process. Severalmicrotubules could be deployed to interconnect onenode, i.e., nano-machine, with many others. Dependingon the nanonetwork infrastructure, a transmitter nano-machine will be able to use unicasting or multicastingmechanisms. To implement the first mechanism, thetransmitter should be able to select a specific molecularrail to transmit the information molecule to the intendedreceiver. To achieve a multicast transmission, the trans-mitter could release several molecules, containing eachone the same data, which are self-assembled onto molec-ular motors traveling on different molecular rails andhence different destinations. The propagation of molecu-lar motors along a microtubule is unidirectional. Thepolarity of the microtubule indicates the movementdirection of specific molecular motors, e.g., kinesin movestowards the (+) end of the microtubule, and dynein to-wards the (�) end. Thus, bidirectional communicationlinks can be obtained using different molecular motorsor a pair of opposite microtubules between two nano-machines.

5.2. Communication process using molecular motors

Molecular motors are used to carry vesicles containinginformation molecules that are transmitted from the trans-mitter to the receiver. To facilitate the reception, the trans-mitter uses protein tags that bind to specific receptors onreceiver nano-machines. Once the infrastructure is de-ployed between the transmitter and receiver nano-ma-chines, the process includes the following tasks [46]:

� Encoding. This task involves the generation of the infor-mation molecules. When an external stimulus is appliedto transmitter nano-machines, they generate theseinformation molecules. The encoding process consistsof selecting the right molecules that represent the infor-mation or the reaction to be invoked at the receiver.Thus, it is possible to control the reaction at the receiverby selecting the proper stimulus applied to the transmit-ter nano-machine.

� Transmission. This is a key task in this scenario and fur-ther research is needed to establish the process to attachthe information packet to carrier molecules in an accu-rate way. The feasibility of this process is based onligand-receptor binding process [41]. A ligand molecule,usually a protein, can bind to another larger moleculereferred as receptor according to the affinity betweenthem. In chemical terms, affinity is defined as the trendof dissimilar elements to form chemical compoundsbased on their electronic properties. The transmitternano-machine should guarantee the high affinitybetween the information molecules and the molecularmotors. If this affinity exists, the information moleculecan bind to the molecular motors to start the propaga-tion process. Encapsulation techniques can be used toensure the affinity between the information moleculeand the molecular motors. Vesicles are biological cap-sules with high affinity with molecular motors, there-fore, they can be used as standard interfaces betweendifferent information molecules and molecular motors.Other biological process such as cell fission, reproduc-tion or pore formation could also be used to releaseinformation packet to the medium [48].

� Propagation. This part of the process refers to the move-ment of the information molecules through the medium.Microtubules or molecular rails interconnecting nano-machines restrict the movement to themselves. As aresult, molecules move in the direction of these railsinstead of diffusing or moving randomly through themedium. In this scenario, the propagation is similar topropagation of electrical signals through conductingmaterials. The difference lies on the propagation speed,which will depend on the information molecule, themolecular motors and the molecular rail. The natural


domain for molecular motors is intra-cell communica-tion. However, new advances in biomolecular sciencecan enable inter-cell and bio-hybrid scenarios, asreported in [36].

� Reception. It is the process in which molecular motorscontaining information molecules arrive at the receivernano-machine. The information molecules are detachedfrom the molecular motors. This is achieved thanks tothe affinity between the protein tags or the molecularencapsulation and receptors at the receiver. A receivercan have several reception properties in order to detachdifferent information molecules from the molecularrails. Similar to the transmission, the extraction ofinformation from a vesicle can also be done throughdifferent cellular processes such as fusion or pore forma-tion [58].

� Decoding. The receiver nano-machine invokes thedesired reaction according to the received informationmolecule. These biological reactions depend on theinformation molecule sent by the transmitter nano-machine. A nano-machine can be equipped with severalreceptors. Each receptor would be able to react to a spe-cific information molecule. The decoding can be highlyassociated to these specific receptors. For instance, eachreceptor can be mapped to a specific action of thereceiver nano-machine. Another option is to decodethe information molecules inside the nano-machines.In this case the nano-machine could have one receptorand the different messages should use a common inter-face such as vesicles. As we can see, molecular motorsjust provide a means to transport the information fromone point to another. Nano-machines will have toinclude all the mechanisms to support traditional net-work tasks such as message coding/decoding, multi-hopping and routing.

6. Short-range communication using calcium signaling

Inter-cell communication based on calcium signaling isone of the most well-known molecular communicationtechniques. It is responsible for many coordinated cellulartasks such as fertilization, contraction or secretion [11].Calcium signaling is used in two different deployment sce-narios. It can be used to exchange information among cellsphysically located one next to each or among cells de-ployed separately without any physical contact. Dependingon the scenario, the propagation of the chemosignals in-cludes different messengers such as ATP substance, cal-cium ions (Ca2+) or inositol 1,4,5-triphosphate (IP3)among others.

The description of the sequential propagation of thechemosignals driven by different messengers is known asthe chemosignal pathway. One messenger transports theinformation until certain point of the pathway where theinformation is transferred to another messenger. The infor-mation transfer from one messenger to another continuesuntil the information reaches its destination. This multi-messenger propagation scheme presents two advantages.First, the signal can be amplified at different levels of thechemosignal pathway. The amplification is usually donewhen the information is transferred among messengers.

Second, we could use a set of messengers, i.e., chemicalagents, to obtain different signal transduction and decod-ing results. The combination of different messengers couldlead to different results and could be used to adapt thepropagation parameters to a specific environment. The li-gand-receptor binding principle is one of the most impor-tant processes participating in the information transferamong messengers. This process consists in the boundingof two molecules resulting in a local reaction, which in turncan trigger other processes.

If cells are deployed separately in the chemosignalpathway, as it is shown in Fig. 5b, the information mole-cule, i.e., the ligand, binds to a molecular receptor locatedon the external membrane of the cell. This binding gener-ates an inner cell signal, which can be decoded by cell com-ponents. In this case, the ligand is considered as the firstmessenger while the inner signal molecules are consideredas second messengers of the chemosignal pathway. Firstmessengers are molecules that transport the informationoutside the cell. Complementarily, second messengerstransport the information inside the cells. Membranereceptors convert the first messenger signals into secondmessenger signals by chemical reactions.

When cells are located next to each other, as it is shownin Fig. 5a, they can be connected through gap junctions.These biological gates allow different molecules and ionsto pass freely between cells [1]. In this deployment sce-nario, the signal travels along cells using second messen-gers, such IP3 and without the intervention of firstmessengers. The IP3 is a messenger that provokes the re-lease of calcium ions by cell organelles. Thus, the diffusionof IP3 through gap junctions propagates a Ca2+ wave alongthe interconnected cells.

Several cellular components participate in the flux reg-ulation of Ca2+ that makes signal propagation possible. Thestorage of Ca2+ inside the cell is enabled by certain proteinsof the cytoplasm and organelles such as the endoplasmaticreticulum that can act as Ca2+ buffers or reservoirs. Extra-cellular mechanisms, such as ion channel located on thecell membrane, are also involved in this process in orderto maintain this finite resource.

Calcium signaling provides a mean to interconnectneighboring cells. Unlike communication based on molec-ular motors and molecular rails, as explained in Section 5,calcium signaling provides more flexible transmissionschemes. Using calcium signaling, all the surroundingnano-machines can receive a message sent by a uniquetransmitter. This communication technique is similar tobroadcast networks in which all the transceivers locatedin the transmission range of the transmitter can sensethe electromagnetic signal.


Calcium signaling can be used for short-range commu-nication among several nano-machines [50]. Similar tonatural models, nano-machines should be near each other.Similar to the natural models, the propagation of theinformation can be performed in two differentschemes depending on the deployment of the nano-machines:

TX

RX

GapJunctio

CalciumSignalin

Encoder

Decoder

TX

RX

DiffusionProcessCalcium

Signaling

Encoder

Decoder

Fig. 5. Signal propagation in calcium signaling communication systems by (a) gap junctions signal forwarding and (b) by diffusion.


� Direct access. If nano-machines are physically con-nected, Ca2+ signals travel from one nano-machine tothe next one through the gates, as shown in Fig. 5a.These gates should work similarly to the gap junctionsallowing the flux of ions and molecules from onenano-machine to the interior of another nano-machine.

� Indirect access. If the nano-machines are not in directcontact, the transmitter nano-machine should be ableto release the information molecules to the mediumas first messenger in the chemosignal pathway. Infor-mation molecules will move through the medium fol-lowing a diffusion process. They will generate acalcium signal inside the receiver nano-machine.Transmitters can encode the information varying theconcentration of first messengers as shown inFig. 5b. Biological systems encode the information onfrequency and amplitude of concentration changes,usually referred to as Ca2+ oscillation and Ca2+ spikes[11].

These two propagation schemes enable the formation ofnetworks supporting multicast or broadcasting transmis-sion mechanisms. The overall communication systemcould work as follows: transmitter and receiver nano-ma-chines are connected to each other through a signaling net-work consisting of interconnected nodes, which propagatethe information using Ca2+ signals.

Due to the close interaction between nano-machines,the propagation of the information is not affected by thenatural speed of the medium. However, the channel isnot expected to be noiseless. Ca2+ can bind easily tocharged particles that can be found in the networkmedium. A nanonetwork could use several different mes-sengers to design the chemosignal pathway according tothe working conditions. Nanonetworks should also beaware of potential communication process running in theenvironment. One of the potential drawbacks on using cal-cium signaling is the interference that can be caused inbiological process such as muscle contractions orneurosignals.


6.2. Communication process within calcium signaling

For a better description of the communication process,we use two different network scenarios as shown inFig. 5a and b, respectively. The direct access scenario con-tains one transmitter nano-machine and many other cellsor similar nano-machines that are able to propagate thesignal, i.e., the message. These components are physicallylocated one next to each other and they are interconnectedthrough functional gates or gap junctions. The receivernano-machine, to which the message is addressed, can beany of these nano-machines members of the network.The indirect access scenario includes nano-machines thatare deployed separately and where the signals propagatealong the medium where the nano-machine are deployed.

In these two scenarios, the communication processbased on calcium signaling includes the following steps:

� Encoding. This task refers to the generation of the infor-mation molecules. If neighboring nano-machines areused for the propagation of the signal as occurs in directaccess scenario, the transmitter nano-machine encodesthe information using Ca2+. The information is preciselyencoded in amplitude and frequency of the functiondescribing the concentration of Ca2+ signal. Theseencoding methods are known as amplitude modulation(AM) and frequency modulation (FM). In the indirectaccess case, the transmitter should encode the informa-tion in the molecule to be used as a first messenger. Thefirst messengers could also be encoded using AM andFM techniques. An external stimulus can be used to startthe generation encoding process. For instance, it hasbeen reported that some mechanical stimulus appliedto cells provoke the generation of IP3 substance insidethe cell. The presence of IP3 unleashes the release ofCa2+.

� Transmission. This task involves the signaling initiation.In direct access scheme, transmitter nano-machinesstimulate neighboring cells and consequently the signal-ing process starts. The signaling generates the initiationof propagation of Ca2+ waves. IP3, which was generatedpreviously in the transmitter nano-machine, starts flow-ing into neighboring cells through the gates or gap junc-tions. Thus, neighboring nano-machines would releasemore Ca2+ driven by the presence on IP3 substance. Inthe indirect access case, transmitters may initiate sig-naling by releasing substances to the environment. Sim-ilar processes to cell fission or pore formation could beused by nano-machines to release the information mol-ecules to the medium.

� Signal propagation. When nano-machines use directaccess, IP3 transmitted to neighboring cells or nano-machines induces the release of Ca2+ from the IP3-sensi-tive Ca2+ stores. The diffusion of the IP3 substance con-tinues to new neighboring nano-machines inducingagain the release of Ca2+ on these nano-machines. Thischain reaction provokes an increase of the Ca2+ concen-tration and as a result, the Ca2+ wave propagates acrossthe networked nodes affected by IP3. This propagationcan be controlled varying the permeability of the gatesor gap junctions. When nano-machines use indirect

access, the propagation of first messengers containingthe information can be described by using diffusion orBrownian motion models. When these information mol-ecules bind to the receptors of the receiver nano-machines, they are translated into Ca2+ internal signals.The indirect access is much more sensitive to the med-ium conditions than the direct access. Transmitternano-machines should consider the medium conditionssuch as wind, flow, temperature and noise to ensure thatthe information molecules will arrive to the intendedreceiver.

� Reception. When using direct reception, receiver nano-machine establishes gap junctions with the neighboringcells and perceives the Ca2+ concentration from inside ofthese cells. Once the message is received, receiver nano-machine can stop the IP3 propagation by closing thegates or gap junctions connecting with other nano-machines. In the case of indirect reception, the receptorplays the most import role. They have to translate theinformation molecule into internal Ca2+ signals. Anano-machine could be equipped with several receptorscapable to detect different information molecules. Thiscan be used to establish several parallel communicationchannels among different nano-machines.

� Decoding. The receiver nano-machine reacts to the inter-nal Ca2+ concentration. The concentration depends onthe influx of IP3 and Ca2+ that arrives to the receiver oron the bindings occurred between information mole-cules and nano-machines receptors. The Ca2+ signalcan be encoded in amplitude and frequency enablingthe activation of different processes. The basic differencebetween the receiver and those nano-machines or cellsparticipating in the signal propagation is that the recei-ver can decode the message, while other nano-machinesjust receive the signal and forward it. Since calcium sig-naling can be considered as a broadcast scheme, multi-ple receivers can decode the same message.

7. Long-range communication using pheromones

Long-range communication is referred to as the com-munication process in which the distance between trans-mitter and receiver nano-machines ranges frommillimeters up to kilometers. Control and communicationof nano-machines in long-range communication can beuseful in many applications, such as the military field orenvironmental monitoring.

Nanonetworks in long-range communication scenarioare also inspired in biological systems found in nature.Ants, butterflies, bees and many mammals use molecularmessages, i.e., pheromones, to communicate with mem-bers of the same species. Pheromones can be defined asmolecular compounds containing information that canonly be decoded by specific receivers and can invoke cer-tain reactions in them. Pheromonal communication is usedto build complex networks including micro and macro-scale mobile actors. For instance, in ants colony, the com-munication between members is based on pheromonesand a whole set of social behavior primitives is encodedin these nano-messages. Another example is the long-dis-tance communication between butterflies using phero-


mones where molecular messages can reach the range of afew kilometers [65].


Communication using pheromones is similar to theshort-range techniques described in Section 5 and 6. Thecommunication is based on the release of molecules thatcan be detected by a receiver. In long-range nanonetworksthe channel cannot be modeled as a deterministic neither aphysical link between transmitter and receiver. Thus, con-cepts like diffusion, flows and molecules concentration inthe medium play a key role in the communication process.

Regarding the medium, molecular communication usingpheromones can be used in dry [14] and wet [40] environ-ments. Once the molecules are released to the medium,they can be affected by several factors, such as antagonistagents, medium flow and temperature, and dispersion. Allthose factors can be considered as sources of noise, similarto those found in traditional communication channels, andthey can compromise the transmission reliability.

A key feature in long-range communication using pher-omones is the coding system. While in traditional commu-nication networks the message is encoded using a binarysystem, the information in long-range nanonetworks is en-coded on the molecule itself. Since messages consist of mol-ecules, there is a huge quantity of possible combinationsthat can be used to transmit data. Moreover, messagescan be compounded by several different molecules allow-ing even more combinations to encode the information.

The reception of the transmitted molecules is realizedby molecular receptors located on the receiver, as depictedin Fig. 6. This phenomenon is based on the ligand-receptorbinding process as occurs within calcium signaling trig-gered by first messengers. A ligand is a molecule that inter-acts with a protein, by specifically binding to this one. Inmolecular communication using pheromones, the receptorproteins can be considered as the receiver nano-machineantenna or transducer, which transforms energy containedin the message into a reaction at the receiver.

Nanonetworks based on pheromonal communicationare good examples of scalable molecular communicationsystems. In biological communication systems based onpheromones, the information is encoded on molecules atnano-scale, although the transmitter and receiver nano-

TXNano-machine

Encoder

Pheromones(the message

Scale

mic

rona

no

Micro-system

Fig. 6. Conceptual diagram of a pheromonal communication. Biological mod

machines, i.e., animals, can be considered as macro-sys-tems. However, the communication is still based onnano-transceivers and nano-messages and therefore is inline with the definition of nanonetworks. In these biologi-cal systems, complex neural networks and hormonal sys-tems act as the interface among these macro entities, e.g.,brain, and the nano communication system.

7.2. Long-range pheromonal communication process

The communication process includes five tasks similarto those found in short-range communication techniques.Currently, there are still no artificial nano-machines capa-ble of executing the tasks described in this section but theyare expected to be available in the near future. The objec-tive of the following description is to identify the existingbiological communication processes and to map them intothose communication processes found in traditional com-munication networks. This knowledge mapping can helpto understand the biological communication mechanismsand to develop solutions for future nano-machines.

� Encoding. This process includes the selection of the spe-cific pheromones or appropriate molecules to transmitthe information and produce the reaction at theintended receiver. On the biological systems found innature, animals release specific pheromones to invokecertain reactions at the receivers. For instance, thefemale Douglas-fir beetle releases pheromones to attractthe male to the host tree. When the male detects thesepheromones, it transmits an acoustic signal to stop thesexual pheromones production and also releases Doug-las-fir anti-odors to avoid the detection of the host byother males. Another example of coding through mole-cules can be found in colonies of bees. Honeybees useup to 15 known glands in their molecular communica-tion process. The released message is not a single com-ponent but rather a complex molecular structurecompound by different chemicals. Different pheromonescan be used as sexual, alarm, marking, attraction oraggregation and orientation messages.

� Transmission. This process consists of releasing theselected pheromones during the encoding process tothe medium. Since pheromones can be released tothe medium through physiological fluids, molecular

RXNano-machine

)

Receptors

Micro-system

els provide a useful example of molecular communication scalability.


messages can be in liquids or gases. The transmission ofinformation is defined as a voluntary action of the trans-mitter. It is important to emphasize the voluntary char-acter of the transmission, otherwise it is not consideredas a communication message. For instance, odors cancontain information and can trigger specific behaviorsat the receivers, but usually they are not transmittedvoluntarily and therefore they cannot be considered asan example for nanonetwork communication.

� Propagation. The propagation of pheromones, from thetransmitter to the receiver nano-machine, can be mod-eled using diffusion processes where each molecule issubjected to Brownian motion. The diffusion of mole-cules or propagation is very sensitive to environmentalconditions such as temperature, viscosity and pressure.Antagonist molecules present in the medium can alsoaffect the propagation negatively by modifying the infor-mation molecules before they reach the destination. Inbiology, the reception of pheromonal messages is stud-ied by measuring the concentration level of pheromonesnear the intended receiver [13]. These communicationmodels based on concentration level and diffusion mod-els, can help to determine some parameters, such as thesensitivity needed in the receptor to detect the molecularmessages. The model can also help to develop channel-multiplexing mechanisms and to assess the channelcapacity according to the propagation characteristics.

� Reception. Once the information message reaches thereceiver, it can use receptor proteins to detach the mol-ecule from the carrier. These receptors can be located onthe physical surface of the receiver. Receptors are pro-teins with a high molecular affinity to pheromonal mes-sages, so that they can bind to them. The ligand-receptorbinding process also depends on the environmental con-ditions. Special structures can be included in nano-machines for this task, e.g., the vomeronasal organ thatacts as the receiver for pheromonal communication inmost mammals [14]. This biological organ is a blind-ended mucus-filled tube, located in the nasal septum[23]. The vomeronasal sensory neurons (VSNs) trans-form these received molecules into bioelectric signalsthat are transmitted to the brain and later decoded intoa specific reaction.

� Decoding. This process is referred to the interpretation ofthe information transmitted or the reaction invoked bythe received message. For instance, in some insects thereception organs includes many different receptors thatcan be sensitive to different molecules or messages. Themolecular receptor of the fruit fly is its antennae and themaxillary palps that include 1300 olfactory receptorneurons (ORNs). These ORNs, which are connected tothe brain, can decode 40 different odors. The decodingsystem is embedded in the reception organs and canbe expressed as spatial-temporal activation patterns ofthe receptors.

8. Research challenges in nanonetworks

The interest in nanotechnologies is growing while moreapplications are being proposed. The potential impact of

these technologies leverages the continuous developmentof new tools, such as simulators [18], scanning probemicroscopes [44], or lithography machines able to patternnano-metric structures.

Over the last years, novel molecular manufacturingtechniques are enabling the development of more ad-vanced nano-machines. Using nanonetworks, these nano-machines can be interconnected to realize more complextasks in a coordinated and complementary manner. Firstdevelopments in nanonetworks are inspired by biologicalsystems based on molecular communication.

Recently, two molecular communication schemes havebeen proposed based on natural models. These schemesare based on inter and intra-cell molecular communicationtechniques [57]. Additionally, in this paper we propose athird communication scheme based on pheromones. Thesethree communication schemes are examples for short orlong-range molecular communication observed in nature.They are proposed as basic models for the developmentof future nanonetworks.

Nanonetworks are not only related to the molecularcommunication techniques, but also to nano-machines,which are able to communicate at this level. The nanonet-works development roadmap includes several stages, inwhich an interdisciplinary scientific approach is neededto address all the posted research challenges.

8.1. Development of nano-machines, testbeds and simulationtools

One of the first phases in the nanonetworks roadmap isaimed at developing nano-machines including moleculartransceivers. Latest efforts on nanotechnologies enabledthe creation of functional nano-structures that introduceswitching, memory, and light-emitting behaviors [8].These are still very simple but they will lead to more com-plex nano-machines, which will be capable of performingspecific communication processes. Future nano-machinesare expected to operate autonomously, and also to includeself-assembly, self-replication, locomotion and communi-cation capabilities as described in Section 4.

Despite the current existence of simulation tools formolecular assembly, and biological and genetic systems,there is none for nanonetworks up to now. Simulationtools should allow the use of different nanonetwork topol-ogies and molecular communication schemes such as cal-cium signaling or networks based on molecular motors. Itshould also include the medium parameters that affectthe propagation of the information molecules, and allowthe selection of different carriers to transport theinformation.

Simple nano-machines or molecular transducers can beconsidered for the development of first nanonetworkstestbeds for molecular communication solutions. This isthe case of robotic micro-noses, which could be used tomeasure concentration level of certain molecules in theenvironment [33], and therefore act as molecular receiv-ers. Using molecular releasers as transmitters, and sys-tems like the robotic noses, developments in signalmodulation, channel models, and medium access couldbe tested.


8.2. Theoretical approach: basis for a new communicationparadigm

A typical communication process includes the followingphases:

� The encoding phase in which the transmitter forms theinformation molecule.

� The transmission or release of these molecules to theenvironment.

� The propagation of the information molecules troughthe medium.

� The reception of the information molecules.� The decoding of the received information molecules.

First nanonetworks are based on many biological com-munication processes. However, the communication tasksoccurring among these biological components are stillnot completely understood.

When talking about nanonetworks, the first questionsthat arise are regarding the information encoding and laterdecoding. Based on the three bio-inspired communicationschemes presented in Sections 5–7, we can identify twodifferent methods to encode the information in molecularcommunication. The first method encodes the informationin the information molecule itself, e.g., structure or polar-ity. Depending on the molecular structure the message willbind only to specific receptors, resulting in a channel de-fined by these information molecules and proper receptors.Thus, the communication system could use different mol-ecules and receptors to encode the information as spa-tial-temporal activation patterns of the receptors, e.g.,decoding of odors in fruit flies. The second method encodesthe information in the fluctuations of the concentration ofions, similar to amplitude and frequency modulationmethods used in traditional telecommunication systems.How nano-machines decode these different fluctuationsis still uncertain although the activation of some cellularprocesses according to these variations has already beenproven [11].

The transmission and reception of information mole-cules is more related to the nano-machines, more specifi-cally with the molecular transceiver. There are manyopen questions regarding the transmission and receptionsuch as (i) how to acquire new molecules and modify themto encode the information, (ii) how to manage the receivedmolecules, and (iii) how to control the binding processes torelease the information molecules to the medium. In [52],models for flux and concentration detectors are presented.Such models demonstrate the dependence of the bindingprocess on the medium conditions. This kind of modelscan help to understand the interaction between the recep-tors and the information signal, and to assess some trans-ceivers features such as the molecule release rate or thesensitivity of the receptor.

The propagation of the communication signals in nan-onetworks is totally different than in classical communica-tion networks. Nanonetworks use molecules instead ofelectromagnetic or acoustic waves. The propagation ofthese molecules, which is subject to different mediumparameters such as viscosity, temperature or pH, can be

described using a particle diffusion model. Uncontrolledmedium conditions such as rain, obstacles, wind or tide,can also affect the communication pathway. Prior to thedevelopment of communication mechanisms or protocols,it is important to develop, analyze and understand thechannel model and the molecular signal propagation fea-tures. Brownian motion models were successfully used todescribe the free movement of particles in fluids [53].Using similar models, the communication based ondiffusion of pheromones in ant colonies has been describedin [13]. These models can help to develop the channel andpropagation models for nanonetworks. Transmitter and re-ceiver binding features can be added to these propagationmodels to assess the molecular channel capacity [2].Similar models, but including molecular propagation prin-ciples, can be used to explore more complex encodingtechniques for molecular transmissions, such as AM andFM. If molecular motors are used as carriers, the propaga-tion is restricted to the nanonetwork infrastructure andtherefore the communication process can be consideredto be more deterministic. The signal propagation speedusing this physical channel as well as the reliability ofthe transmission using molecular rails still need to bemeasured.

8.3. Knowledge transfer: architectures and protocols

Once the basic nanonetwork components are built, thetransmission is controlled and the propagation is under-stood, advanced networking knowledge can be applied todesign and realize more complex nanonetworks. Thedevelopment of the nanonetworks can be done using a lay-ered architecture including medium access protocols, rout-ing schemes and application interfaces.

A medium access control (MAC) protocol is needed todefine and apply mechanisms to ensure a fair use oftransmission channel. In biological molecular communi-cations, most of the channel access schemes are basedon code division. In nanonetworks, many molecular cod-ing schemes could be used to transmit the informationwhile these molecules do not interfere with each otherand do not affect the medium, e.g., changing the viscos-ity. An example of biological channel access based on dif-ferent codes can be found in pheromonal communication.In this example, each communication channel uses differ-ent information molecules that can only be decoded byintended receivers, resulting in multiple communicationchannels over the same medium, one per species. An-other medium access technique used by biological sys-tems is based on the modulation in frequency (FM) andamplitude (AM) of the molecular signal. According tothese last techniques, the communication channelsamong nano-machines can be established using differentfrequencies or signal amplitudes. Each frequency andamplitude channel can be decoded by specific nano-machines.

For more complex networks, an addressing scheme isneeded. The packet should include the information aboutits origin and destination to allow bidirectional and multi-hop communications. How to include this information inan information molecule is still an open issue. However,


according to natural models of nanonetworks, the informa-tion is usually encoded and broadcast without including aspecific address of the transmitter nano-machine. In bio-logical nanonetworks it is assumed that only the autho-rized transmitters can have access to the channel andsend the encoded message. When this message arrives atthe receptor, it reacts according to the molecular informa-tion ignoring its origin. Thus, these nanonetworks can beconsidered as data centric networks. This behavior can beobserved in pheromonal communication, or in living cellnetworks using calcium signaling. In these examples, thereactions at the receiver are triggered by the molecular sig-nal itself, which does not contain any specific data aboutthe origin. This principle is also the basis for drugs deliveryand the use of supplementary hormones in healthcare. Bycontrast, the destination address is crucial in data centricnetworks. Nano-machines participating in the propagationpathway should be able to route a packet towards the rightreceptor based on the destination address. Addressing androuting schemes are not trivial, and solutions will be devel-oped according to the level of complexity of nano-ma-chines participating in the propagation of the molecularsignals. In natural models encoded signals are broadcastin a medium where the intended receivers are located. Thissimple mechanism is used to reach the intended receiversof the network.

Nanonetworks are developed for a specific purpose orapplication. The objectives are described in a main control-ler or are distributed in the internal code of the participat-ing nano-machines. All the interfacing aspects between thecontroller of a nano-machine, in which the applicationtasks are described, and the nanonetwork must be ex-plored. Standard interfaces should be defined includingprotocols and primitives to start a transmission usingnano-machines and translate the received molecules intoinformation relevant to the application. These interfacesshould be available for macro-systems participating inthe nanonetworks, such as monitoring computer networksin biomedical or military fields.

Communication reliability is another key issue in nan-onetworks. Some applications demand reliable mecha-nisms to monitor or interrupt selective transmissions oron-going processes. Some molecular communicationschemes are subject to random processes; as a result, thereception of a transmitted message by the right receivercannot be guaranteed. The communication reliabilityshould be investigated by considering random communi-cation features and particular infrastructures in molecularcommunication.

9. Conclusions

The development of nanotechnologies will continueand will have a great impact in almost every field. Theuse and control of these technologies will be a majoradvantage in economics, homeland security, sustainablegrowth and healthcare. Intrinsic technological burdens willlimit the use of more advanced and smart materials, sen-

sors, actuators and devices at nano-scale, if they are notable to communicate to cooperate to perform more com-plex tasks. This need for a communication network willbe more plausible with the increased complexity of devel-oped nano-devices.

Molecular communication seems to provide efficientmechanisms for networking of nano-machines. It repre-sents a complete new communication paradigm in whichthe information is encoded into molecules. In this paperthe term ‘‘nanonetwork” was defined as the interconnec-tion of multiple nano-machines using molecular communi-cation. Nanonetworks demand innovative solutions tocreate reliable molecular communication channels amongnano-machines. First developments are bio-inspired byexisting biological nanonetworks. At nano-level, manycomponents and communication process has been studiedfrom a biological or chemical point of view.

Despite being a novel communication paradigm that re-quires an interdisciplinary approach, information and com-munication technologies (ICT) are called to be a keycontributor for the evolution of the nanonetworks. Net-work architectures, channel models, nano-machines andtransceivers architectures, medium access control androuting protocols are some of the contributions that are ex-pected from the ICT field.

Acknowledgement

The authors would like to thank Eylem Ekici, Mehmet C.Vuran, Linda Jiang Xie, Tommaso Melodia, and DarioPompili for their valuable comments.

References

[1] B. Alberts, Molecular Biology of the Cell, 4th Bk& Cdr edition, Ed.Garland Science, 2002.

[2] G. Alfano, D. Miorandi, On information transmission amongnanomachines, in: Proceedings of the First International Conferenceon Nano-Networks (NanoNet’06), September 2006.

[3] J.W. Aylott, Optical nanosensors – an enabling technology forintracelular measurements, Analyst 128 (2003) 309–312.

[4] J.D. Badjic, V. Balzani, A. Credi, S. Silvi, J.F. Stoddar, A molecularelevator, Science 303 (March) (2004) 1845–1849.

[5] R. Ballardini, V. Balzani, A. Credi, M. Gandolfi, M. Venturi, Artificialmolecular-level machines: which energy to make them work?,Accounts of Chemical Research 34 (May) (2001) 445–455

[6] V. Balzani, M. Gómez-López, J.F. Stoddart, Molecular machines,Accounts of Chemical Research 31 (May) (1998) 405–414.

[7] V. Balzani, A. Credi, S. Silvi, M. Venturi, Artificial nanomachines basedon interlocked molecular species: recent advances, Chemical SocietyReviews 35 (2006) 1135–1149.

[8] J. Batteas, C. Chidsey, C. Kagan, T. Seideman, Building electronicfunctions into nanoscale molecular architectures, National ScienceFoundation, Technical Report, June 2007.

[9] C. Bauschlicher, A. Ricca, R. Merkle, Chemical storage of data,Nanotechnology 8 (March) (1997) 1–5.

[10] B. Behkam, M. Sitti, Bacterial flagella-based propulsion and on/offmotion control of microscale objects, Applied Physics Letters 90(January) (2007).

[11] M.J. Berridge, The AM and FM of calcium signalling, Nature 386(April) (1997) 759–780.

[12] T. Bjerregaard, S. Mahadevan, A survey of research and practices ofnetwork-on-chip, ACM Computing Surveys 38 (2006) 1–50.

[13] W.H. Bossert, E. Wilson, The analysis of olfactory communicationamong animals, Journal of Theoretical Biology 5 (November) (1963)443–469.


[14] P.A. Brennan, E.B. Keverne, Something in the air? New insights intomammalian pheromones, Current Biology 14 (January) (2004) R81–R84.

[15] S.F. Bush, Y. Li, Nano-communications: A new field? An explorationinto a carbon nanotube communication network, General Electric,Technical Report, February 2006.

[16] C. Bustamante, Y. Chelma, N. Forde, D. Izhaky, Mechanical processesin biochemistry, Annual Review of Biochemistry 73 (July) (2004)705–748.

[17] E. Carafoli, Calcium signaling: a tale for all seasons, Proceedings ofthe National Academy of Sciences of the United States of America 99(3) (2002) 1115–1122.

[18] A. Cavalcanti, B. Shirinzadeh, R.A. Freitas, T. Hogg, Nanorobotarchitecture for medical target identification, Nanotechnology 19(1) (2008) 1–12.

[19] C.-Y. Chang, The highlights in the nano world, Proceedings of theIEEE 91 (November) (2003) 1756–1764.

[20] C. Chen, Y. Haik, J. Chatterjee, Development of nanotechnology forbiomedical applications, in: Proceedings of the EmergingInformation Technology Conference, August 2005.

[21] J. Dennis, J. Howard, V. Vogel, Molecular shuttles: directed motion ofmicrotubules along nanoscale kinesin tracks, Nanotechnology 10(September) (1999) 232–236.

[22] T. Donaldson, 24th century medicine, Cryonics (December) (1988)16–34.

[23] K. Doving, D. Trotier, Structure and function of the vomeronasalorgan, Journal of Experimental Biology 21 (November) (1998) 2913–2925.

[24] E. Drexler, Nanosystems: Molecular Machinery, Manufacturing, andComputation, John Wiley and Sons Inc., 1992.

[25] E. Drexler, C. Peterson, G. Pergami, Unbounding the future: Thenanotechnology revolution, Foresight Institute, Technical Report,1991.

[26] E. Drexler, Molecular engineering: Assemblers and future spacehardware, American Astronautical Society: AAS-86-415, 1986.

[27] M. Endo, T. Hayashi, Y.A. Kim, H. Muramatsu, Development andapplications of carbon nanotubes, Japanese Journal of AppliedPhysics 46 (June) (2006) 4883–4892.

[28] A. Enomoto, M. Moore, T. Nakano, R. Egashira, T. Suda, A. Kayasuga,H. Kojima, H. Sakibara, K. Oiw, A molecular communication systemusing a network of cytoskeletal filaments, in: Proceedings of the2006 NSTI Nanotechnology Conference, May 2006.

[29] R.A. Freitas, Pharmacytes: An ideal vehicle for targeted drug delivery,Journal of Nanoscience and Nanotechnology 6 (September) (2006)2769–2775.

[30] R. Freitas, What is nanomedicine?, Nanomedicine: Nanotechnology,Biology and Medicine 1 (November) (2004) 2–9

[31] R.A. Freitas, Nanomedicine, Volume I: Basic Capabilities. LandesBiosience, 1999.

[32] R.A. Freitas, Nanotechnology, nanomedicine and nanosurgery,International Journal of Surgery 3 (November) (2005) 243–246.

[33] J. Gardner, P. Bartlett, Electronic Noses: Principles and Applications,Oxford University Press, 1999.

[34] H. Goldstein, The race to the bottom, IEEE Spectrum (March) (2005)32–39.

[35] J. Han, J. Fu, R.B. Schoch, Molecular sieving using nanofilters:past, present and future, Lab on a Chip 8 (January) (2008)23–33.

[36] Y. Hiratsuka, T. Tada, K. Oiwa, T. Kanayama, T.Q.P. Uyeda, Controllingthe direction of kinesin-driven microtubule movements alongmicrolithographic tracks, Biophysical Journal 81 (September)(2001) 1555–1561.

[37] S. Hiyama, Y. Moritani, T. Suda, R. Egashira, A. Enomoto, M. Moore, T.Nakano, Molecular communication, in: Proceedings of the NSTINanotechnology Conference, October 2005.

[38] Y. Homma, Y. Kobayashi, T. Ogino, T. Yamashita, Growth ofsuspended carbon nanotube networks on 100-nm-scalesilicon pillars, Applied Physics Letters 81 (September) (2002)2261–2263.

[39] Intel, ‘‘Technology@intel magazine,” October 2006.[40] P. Karlson, A. Butenandt, Pheromones (ectohormones) in insects,

Annual Review of Entomology 4 (April) (1959) 49–58.[41] I. Koltz, D. Hunston, Mathematical models for ligand-receptor

binding. real sites, ghost sites, Journal of Biological Chemistry 259(16) (1984) 10060–10062. August.

[42] H.H. Lee, E. Menard, J.A.R.N.G. Tassi, G.B. Blanchet, Large areamicrocontact printing presses for plastic electronics, MaterialsResearch Society Bulletin 846 (2005) 731–736.

[43] R. Merkle, Self-replicating systems and molecular manufacturing,Journal of the British Interplanetary Society 45 (1992) 407–413.

[44] E. Meyer, H.J. Hug, R. Bennewitz, in: E. Meyer (Ed.), Scanning ProbeMicroscopy: The Lab on a Tip, Springer, 2004.

[45] M. Meyyappan, J. Li, J. Li, A. Cassell, Nanotechnology: An overviewand integration with MEMS, in: Proceedings of the 19th IEEEInternational Conference on Micro Electro Mechanical Systems(MEMS’06), January 2006, pp. 1–3.

[46] M. Moore, A. Enomoto, T. Nakano, R. Egashira, T. Suda, A. Kayasuga,H. Kojima, H. Sakakibara, K. Oiwa, A design of a molecularcommunication system for nanomachines using molecular motors,in Proceedings of the Fourth Annual IEEE International Conferenceon Pervasive Computing and Communications (PerCom’06), March,2006.

[47] M. Moore, A. Enomoto, T. Nakano, Y. Okaie, T. Suda, Interfacing withnanomachines through molecular communication, in: Proceedingsof the IEEE International Conference on Systems, Man andCybernetics, October 2007, pp. 18–23.

[48] Y. Moritani, S. Hiyama, T. Suda, Molecular communication amongnanomachines using vesicles, in: Proceedings of NSTINanotechnology Conference, May 2006.

[49] Y. Moritani, S. Hiyama, T. Suda, Molecular communication for healthcare applications, in: Proceedings of the Fourth Annual IEEEInternational Conference on Pervasive Computing andCommunications (PerCom’06), March 2006.

[50] T. Nakano, T. Suda, M. Moore, R. Egashira, A. Enomoto, K. Arima,Molecular communication for nanomachines using intercellularcalcium signaling, in: Proceedings of the Fifth IEEE Conference onNanotechnology, June 2005, pp. 478–481.

[51] C. Peterson, Taking technology to the molecular level, IEEE Computer33 (January) (2000) 46–53.

[52] J.-P. Rospars, V. Krivan, P. Lánky, Perireceptor and receptor events inolfaction. Comparison of concentration and flux detectors: amodeling study, Chemical Senses 25 (June) (2000) 293–311.

[53] D. Selmeczi, S. Mosler, P. Hagedorn, N. Lasen, H. Flyvbjerg, Cellmotility as persistent random motion: theories from experiments,Biophysical Journal 89 (June) (2005) 912–931.

[54] V. Serreli, C. Lee, E.R. Kay, D.A. Leigh, A molecular informationratchet, Nature 445 (February) (2007) 523–527.

[55] R. Smalley, M.S. Dresselhaus, G. Dresselhaus, P. Avouris, in: M.S.Dresselhaus (Ed.), Carbon Nanotubes: Synthesis, Structure,Properties and Applications, Springer, 2001.

[56] R.K. Soong, G.D. Bachand, H.P. Neves, A.G. Olkhovets, H.G. Craighead,C.D. Montemagno, Powering an inorganic nanodevice with abiomolecular motor, Science 24 (November) (2000) 1555–1558.

[57] T. Suda, M. Moore, T. Nakano, R. Egashira, A. Enomoto, Exploratoryresearch on molecular communication between nanomachines, in:Proceedings of Genetic and Evolutionary Computation Conference(GECCO’05), June 2005.

[58] T. Tanaka, R. Sano, Y. Yamashita, M. Yamazaki, Shape changes andvesicle fission of giant unilamellar vesicles of liquid-ordered phasemembrane induced by lysophosphatidylcholine, Langmuir 20(August) (2004) 9526–9534.

[59] N. Taniguchi, On the basic concept of nano-technology, in:Proceeding of the International Conference on ProductionEngineering, 1974.

[60] D. Tessier, I. Radu, M. Filteau, Antimicrobial fabrics coated withnano-sized silver salt crystals, NSTI Nanotech 1 (May) (2005) 762–764.

[61] A. Tseng, K. Chen, C. Chen, K. Ma, Electron beam lithographyin nanoscale fabrication: recent development, IEEE Transactionson Electronics Packaging Manufacturing 26 (April) (2003) 141–149.

[62] U. Tulu, C. Fagerstrom, N. Ferenz, P. Wadswort,Molecular requirements for kinetochore-associated microtubuleformation in mammalian cells, Current Biology 16 (March) (2006)536–541.

[63] G.M. Whitesides, The once and future nanomachine, ScientificAmerican (September) (2001) 78–83.

[64] B. Wowk, Cell repair technology, Cryonics, vol. July 1988, pp. 21–30,May.

[65] T.D. Wyatt, Pheromones and animal behaviour: communication bysmell and taste, in: T.D. Wyatt (Ed.), Cambridge University Press,2003. September.

[66] Y.J. Yun, C.S. Ah, S. Kim, W.S. Yun, B.C. Park, D.H. Ha, Manipulation offreestanding Au nanogears using an atomic force microscope,Nanotechnology 18 (November) (2007) 505304. 5 pp.

Networks 52 (2008) 2260–2279 2279

Ian F. Akyildiz (M’86-SM’89-F’96) receivedthe B.S., M.S., and Ph.D. degrees in ComputerEngineering from the University of Erlangen-Nuernberg, Germany, in 1978, 1981 and 1984,respectively.

Currently, he is the Ken Byers DistinguishedChair Professor with the School of Electricaland Computer Engineering, Georgia Instituteof Technology, Atlanta, and Director ofBroadband Wireless Networking Laboratory.He is an Editor-in-Chief of ComputerNetworks Journal (Elsevier) as well as thefounding Editor-in-Chief of the AdHoc Net-

work Journal (Elsevier) and the Physical Communication Journal (Else-vier). His current research interests are nanonetworks, cognitive radio

I.F. Akyildiz et al. / Computer

networks, wireless mesh networks and next generation sensor networks.He received the ‘‘Don Federico Santa Maria Medal” for his services to theUniversidad of Federico Santa Maria, in 1986. From 1989 to 1998, heserved as a National Lecturer for ACM and received the ACM OutstandingDistinguished Lecturer Award in 1994. He received the 1997 IEEE LeonardG. Abraham Prize Award (IEEE Communications Society) for his paperentitled ‘‘Multimedia Group Synchronization Protocols for IntegratedServices Architectures” published in the IEEE Journal of Selected Areas inCommunications (JSAC) in January 1996. He received the 2002 IEEE HarryM. Goode Memorial Award (IEEE Computer Society) with the citation ‘‘forsignificant and pioneering contributions to advanced architectures andprotocols for gíreles and satellite networking”. He received the 2003 IEEEBest Tutorial Award (IEEE Communication Society) for his paper entitled‘‘A Survey on Sensor Networks,” published in IEEE CommunicationsMagazine, in August 2002. He also received the 2003 ACM SigmobileOutstanding Contribution Award with the citation ‘‘for pioneering con-tributions in the area of mobility and resource management for gı́relescommunication networks”. He received the 2004 Georgia Tech FacultyResearch Author Award for his ‘‘outstanding record of publications ofpapers between 1999 and 2003”. He also received the 2005 Distinguished

Faculty Achievement Award from School of ECE, Georgia Tech. He has beena Fellow of the Association for Computing Machinery (ACM) since 1996.

Fernando Brunetti was born in Asunción,Paraguay. In 2004, he received his ElectricalEngineering diploma from the UniversidadCatólica de Asunción, in Paraguay. In 2005, hereceived a second M.S. degree from the Schoolof Informatics of the Universidad Politécnicade Madrid (UPM), Spain. Currently, he isaffiliated to the Bioengineering Group of theIndustrial Automation Institute, SpanishResearch Council (CSIC) and is a Ph.D. candi-date in the School of TelecommunicationsEngineering of the UPM. He has been a visit-ing researcher at the Broadband Wireless

Networking Laboratory, Georgia Institute of Technology, Atlanta, fromOctober 2007 to March 2008. His current research interests are nanon-
etworks, body area networks, and inertial sensing.
Cristina Blázquez was born in Barcelona(Spain) in 1983. She is pursuing her M.S.degree in Telecommunication Engineeringfrom the ETSETB-Telecom Barcelona, Univer-sitat Politècnica de Catalunya (UPC), Spain.From August 2007 she is working at theBroadband Wireless Networking Laboratory,School of Electrical and Computer Engineeringin Georgia Institute of Technology, Atlanta.

Documents

Computer Networks - Georgia Institute of Technology