Modelos emergentes para el desarrollo de aplicaciones móviles sociales: people as a service y social devices. Prueba de concepto.
Abstract
Nowadays mobile devices have gained a great importance in our society and they are present everywhere. These devices have a huge amount of data about their owners and context data of their surroundings, and many communication channels. Nevertheless with current models mobile devices cannot take advantage of this situation. For these reasons, new models for mobile applications have appeared recently. People-as-a-Service (PeaaS) allows interactions between mobile devices in their surroundings, using sociological profiles of their owners to create communication between interested devices. Social Devices is a recent model that implements the PeaaS concept, taking advantage the devices are almost always with the user it creates a new kind of interactions between nearby devices with the main target of improve social relations. For this purpose each device is able to provide services to nearby devices and consume them, keeping the user’s data in the device and sharing it only when it is authorized. This concept could be integrated with technologies like Internet-of-Things and Ubiquitous Computing, allowing devices to connect with nearby objects like elevators, traffic lights or your own vehicle. In particular, there is a new platform that implements both concepts, which is the study object of this work: OrchestratorJS. The platform is still under development and is not too stable yet, furthermore there is no documentation available. The main target of this work is to explore the platform and to create a basic documentation which could be used for anybody interested on the in, acting as a starting point for future works.
Full text
3
ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA INFORMÁTICA
GRADO EN INGENIERÍA DEL SOFTWARE
MODELOS EMERGENTES PARA EL DESARROLLO DE
APLICACIONES MÓVILES SOCIALES: PEOPLE AS A
SERVICE Y SOCIAL DEVICES. PRUEBA DE CONCEPTO.
EMERGING MODELS FOR THE DEVELOPMENT OF SOCIAL
MOBILE APPLICATIONS: PEOPLE AS A SERVICE, AND
SOCIAL DEVICES. A PROOF OF CONCEPT.
Realizado po
JOSÉ ANDRÉS CORDERO BENÍTEZ
Tu o izado po
CARLOS CANAL VELASCO
Depa amen o
LENGUAJES Y CIENCIAS DE LA COMPUTACIÓN
UNIVERSIDAD DE MÁLAGA
MÁLAGA, no iemb e 2014
Fecha de ensa:
El Sec e a io del T ibunal
5
Resumen: Ac ualmen e los disposi i os mó iles han ganado una g an impo ancia
en nues a sociedad y es án p esen es en cualquie luga . Es os disposi i os ienen una
g an can idad de in o mación sob e sus dueños y da os con ex uales de sus
al ededo es, y muchos canales de comunicación. No obs an e con los modelos ac uales
los disposi i os mó iles no pueden ob ene en aja de es a si uación.
Po es os mo i os, han comenzado a su gi ecien emen e nue os modelos de
aplicaciones mó iles. People-as-a-Se ice (PeaaS) pe mi e in e acciones en e
disposi i os mó iles ce canos, usando pe iles sociológicos de sus dueños pa a c ea
comunicaciones en e disposi i os in e esados. Social De ices es un modelo ecien e
que implemen a el concep o de PeaaS, ap o echando que los disposi i os es án casi
siemp e con el usua io c ea una nue a línea de in e acciones en e los disposi i os
ce canos con el obje i o de aumen a las elaciones sociales. Pa a ello cada disposi i o
es capaz de p o ee y consumi se icios de disposi i os ce canos, man eniendo la
in o mación del usua io en el p opio disposi i o y compa iéndola solamen e cuando
es é au o izado. Es e concep o ambién puede in eg a se con ecnologías como
In e ne -de-las-Cosas y la Compu ación Ubicua, pe mi iendo que los disposi i os se
conec en con o os obje os ce canos ales como ascenso es, semá o os o u p opio
ehículo.
En pa icula exis e una pla a o ma que implemen a ambos concep os, y que es el
obje o de es udio de es e abajo: O ches a o JS. La pla a o ma aún se encuen a en
desa ollo y no es demasiado es able, además de no con a con ninguna
documen ación. El obje i o de es e abajo es explo a la pla a o ma y c ea una
documen ación básica que pueda se u ilizada po cualquie a que es é in e esado en
ella, ac uando como pun o de pa ida pa a u u os abajos.
Palab as cla e: Social De ices, People-as-a-Se ice, Coope ación basada en
p oximidad, Compu ación Ubicua, O ches a o JS.
6
Abs ac : Nowadays mobile de ices ha e gained a g ea impo ance in ou socie y
and hey a e p esen e e ywhe e. These de ices ha e a huge amoun o da a abou
hei owne s and con ex da a o hei su oundings, and many communica ion channels.
Ne e heless wi h cu en models mobile de ices canno ake ad an age o his
si ua ion.
Fo hese easons, new models o mobile applica ions ha e appea ed ecen ly.
People-as-a-Se ice (PeaaS) allows in e ac ions be ween mobile de ices in hei
su oundings, using sociological p o iles o hei owne s o c ea e communica ion
be ween in e es ed de ices. Social De ices is a ecen model ha implemen s he
PeaaS concep , aking ad an age he de ices a e almos always wi h he use i c ea es
a new kind o in e ac ions be ween nea by de ices wi h he main a ge o imp o e social
ela ions. Fo his pu pose each de ice is able o p o ide se ices o nea by de ices and
consume hem, keeping he use ’s da a in he de ice and sha ing i only when i is
au ho ized. This concep could be in eg a ed wi h echnologies like In e ne -o -Things
and Ubiqui ous Compu ing, allowing de ices o connec wi h nea by objec s like
ele a o s, a ic ligh s o you own ehicle.
In pa icula , he e is a new pla o m ha implemen s bo h concep s, which is he
s udy objec o his wo k: O ches a o JS. The pla o m is s ill unde de elopmen and is
no oo s able ye , u he mo e he e is no documen a ion a ailable. The main a ge o
his wo k is o explo e he pla o m and o c ea e a basic documen a ion which could be
used o anybody in e es ed on he in, ac ing as a s a ing poin o u u e wo ks.
Keywo ds: Social De ices, People as a Se ice, P oximi y-based Coope a ion,
Ubiqui ous Compu ing, O ches a o JS.
7
Index
INDEX ............................................................................................................................................... 7
INTRODUCTION AND GOALS ........................................................................................................ 9
CHAPTER 1: PEOPLE-AS-A-SERVICE ........................................................................................ 13
CLOUD-ENABLED MOBILE DEVICES ................................................................................................ 13
PEOPLE-AS-A-SERVICE .................................................................................................................. 15
PEAAS ARCHITECTURE ................................................................................................................. 16
CURRENT IMPLEMENTATIONS ......................................................................................................... 18
THE INTERNET OF PEOPLE ............................................................................................................. 22
INTEGRATION WITH SOCIAL DEVICES AND INTERNET OF THINGS ...................................................... 26
CHAPTER 2: SOCIAL DEVICES ................................................................................................... 31
THE ACTION-ORIENTED PROGRAMMING MODEL ............................................................................. 32
REQUIREMENTS AND CHARACTERISTICS OF SOCIAL DEVICES .......................................................... 37
ARCHITECTURE ............................................................................................................................. 39
ORCHESTRATOR.JS ....................................................................................................................... 40
TOOLS .......................................................................................................................................... 41
GETTING THINGS WORK ................................................................................................................. 49
ANDROID CLIENT ........................................................................................................................... 52
CHANGES AND ADDITIONAL FEATURES ............................................................................................ 54
GOALS.......................................................................................................................................... 56
EXAMPLE: THE COFFEE BREAK MACHINE ....................................................................................... 57
CHAPTER 3: RELATED TECHNOLOGIES .................................................................................. 61
ANDROID PROGRAMMING .............................................................................................................. 61
MONGODB ................................................................................................................................... 67
NODE.JS ....................................................................................................................................... 70
CHAPTER 4: GETTING STARTED WITH SOCIAL DEVICES ..................................................... 75
DEPENDENCIES ............................................................................................................................. 75
INSTALLATION ............................................................................................................................... 76
GETTING STARTED ........................................................................................................................ 79
EXECUTING A FIRST ACTION ........................................................................................................... 82
STARTING THE DEVELOPMENT ........................................................................................................ 83
CHAPTER 5: APPLICATION HITHERE! ....................................................................................... 89
CHAPTER 6: APPLICATION EGREETINGS ................................................................................ 97
CHAPTER 7: APPLICATION FOLLOWME ................................................................................. 103
CHAPTER 8: APPLICATION HAVEYOUMET ............................................................................ 109
CHAPTER 9: APPLICATION TREASUREHUNT ........................................................................ 115
CONCLUSIONS AND FUTURE WORK ...................................................................................... 119
CONCLUSIONES Y TRABAJO FUTURO ................................................................................... 121
ANNEX 1: NPM MODULES ......................................................................................................... 123
REFERENCES ............................................................................................................................. 127
8
9
In oduc ion and goals
Nowadays sma de ices ha e an excellen po en ial and connec i i y, and can be
used o mul iple pu poses: en e aining, socializing o sha ing in o ma ion. We ha e
applica ions capable o do e e y hing wi h he help o web se ices and cloud
compu ing. Howe e , we ha e cu en ly some limi a ions.
These days e e y hing is some hing-as-a-se ice. We a e used o So wa e-as-a-
Se ice and Pla o m-as-a-Se ice ha ha e been well accep ed by he indus y. Now,
when he usage o mobile de ices ha e become cus oma y o he majo i y o
popula ion, p obably i is he momen o a new model. Mobile de ices a e becoming he
mos impo an piece in he communica ions, and hey con ain a lo o in o ma ion abou
he owne . We could use his ad an age o c ea e a new ype o communica ion, whe e
de ices in e ac wi h o he de ices as people would do.
The goal o his wo k is o lea n abou a new model, called People-as-a-Se ice
[18][19], whe e mobile de ices can deploy se ices o in e ac wi h o he de ices in hei
su oundings. In pa icula , his wo k s udies he pla o m called Social De ices,
de eloped as a esea ch wo k by he Uni e si y o Tampe e in conjunc ion wi h Nokia
[26]. This pla o m is cu en ly in de elopmen and he e is no documen a ion o u he
in o ma ion a ailable.
The main a ge o his wo k is o analyse he po en ial o he pla o m Social De ices,
o s udy all he capabili ies and c ea e a base documen a ion abou i ha could be used
o u u e wo ks. I ha e c ea ed some case s udies o s udy he po en ial and capabili ies
o he pla o m, which a e comple ely documen ed o p epa e he way o u u e p ojec s.
All he e o has been concen a ed in esea ching and unde s anding he pla o m,
documen ing e e y hing p ope ly o acili a e he wo k o u u e p ojec s using his
pla o m and echnology.
The pla o m is no inished and he e is a long way o in eg a e e e y hing, I ha e
been collabo a ing wi h Niko Mäki alo and his eam in he Uni e si y o Tampe e
imp o ing some aspec s o he pla o m and he And oid clien . The code is public o
e e yone and can be accessed in he nex link:
h ps://gi hub.com/nikkis
This wo k uses NodeJS
1
and Py hon
2
as basic echnologies o de elop new
capabili ies and unc ionali y in he pla o m. The se e se ice, called O ches a o JS is
cu en ly de eloped in NodeJS, and he capabili ies in Py hon. On he o he hand, he
clien is ocused in mobile de ices, so we use Ja a as main echnology o de elop he
clien unc ionali y o And oid de ices.
1
h p://nodejs.o g/
2
h ps://www.py hon.o g/
10
This documen is di ided in o h ee main pa s:
- The i s pa includes he chap e s om 1 o 3, which a e a b ie in oduc ion o
he echnologies used in his wo k: Cloud Compu ing
1
, specially he model
People-as-a-Se ice and he Ac ion-O ien ed P og amming Model [1], which a e
used as basic a chi ec u e o he Social De ices pla o m. The se e is
de eloped using Node.js [35], and connec ed o a MongoDB [32] Da abase.
And oid P og amming [4] is he echnology used in he clien .
- The second pa con ains all he ele an in o ma ion o unde s and he Social
De ices Pla o m and O ches a o JS. I explain in Chap e 4 how he pla o ms
beha es, hei a chi ec u e and he base o knowledge. Also I explain how o
p epa e and ins all he en i onmen o s a he de elopmen .
- The hi d pa includes he chap e s om 5 o 9, one chap e o each case s udy.
I ha e de eloped 5 small applica ions ha can be used as a p oo o concep o
see wha he pla o m can o e . Each chap e con ains all he documen a ion o
use he applica ion and de ails abou i s implemen a ion.
Social De ices has cu en ly wo unc ional pla o ms implemen ing he concep . In
chap e 4 we will see he i s implemen a ion, called Social De ices Pla o m (SDP) as
an in oduc ion o he new pla o m O ches a o .js (OJS) [25].
1
h p://en.wikipedia.o g/wiki/Cloud_compu ing
11
Mo i a ion
This p ojec is aligned wi h he s a egic objec i es o he esea ch g oups SCENIC,
om he Uni e si y o Málaga, and QUERCUS, om Ex emadu a on Mobile Cloud
Compu ing and ela ed wi h he esea ch p ojec s SeaClouds
1
and SOFIA
2
, in which is
cu en ly wo king he Uni e si y o Málaga as many o he uni e si ies ac oss Eu ope.
These esea ch g oups a e ocused in cloud echnologies and mobile compu ing.
Cloud Compu ing is a new concep and he e is a long u u e o i . I educes ime- o-
ma ke and p o ides on-demand scalabili y a a low cos o he business and a new
concep is being de eloped: People-as-a-Se ice.
PeaaS is a new echnology in de elopmen and he e a e only a ew pla o ms ha
in eg a es he concep . The pu pose o his p ojec is o c ea e a p oo o concep o wha
People-as-a-Se ice can b ing o us, ocused in he pla o m Social De ices, cu en ly in
de elopmen as a PhD, by Niko Mäki alo in he Uni e si y o Tampe e.
The a ge o his p ojec is o do he p ope in es iga ion in o he pla o m and
elabo a e a well know-how abou i , explo ing he possibili ies and p epa ing all he
documen a ion o u he esea ch and de elopmen s, as he e is no documen a ion
abou i . The pla o m is cu en ly in de elopmen and is no comple ely s able.
The concep o PeaaS will be also in e es ing o people, because his new
echnology will allow new ways o in e ac ing be ween people and de ices, p o iding
new capabili ies in how we use hem. Combined wi h In e ne o Things, people will be
able o in e ac wi h he su oundings, he de ices will know he con ex da a o he use ,
and change i s beha iou depending on i , making he wo k o he use wi hou o cing
him o send commands o he de ice.
Social De ices is a new pla o m ha can be used o his pu pose, i was in oduced
in 2011 as a join wo k by Nokia Resea ch Cen e , Aal o Uni esi y, and Tampe e
Uni e si y o echnology [26]. Sma phones ha e no only a lo o in o ma ion abou heis
owne s, bu also capabili ies ha enable hem o esemble humans ( hey can ansla e
ex in o speech, o example).
1
h p://www.seaclouds-p ojec .eu/
2
h p://so ia.lcc.uma.es/
12
19
Fo example, an ad e ise can use his pla o m o c ea e an applica ion ha sends
ad e ising o special a ge s, based in hei geog aphical posi ion, age, in e es s, social
s a us, e c.
The nimBees a chi ec u e [Figu e 2] is a specializa ion o he one de ined o he
PeaaS model. Hence, he se e and mobile a chi ec u es a e designed as ollows:
- Ca alogue se e . Based on he concep o he ca alogue se e , a non-mobile
elemen wi h a ixed and well de ined se ice en ypoin whe e e e y new mobile
de ice added o he PeaaS ne wo k mus be connec ed, in o de o au hen ica e
and p o ide in o ma ion abou i sel , and o e ie e in o ma ion abou he es o
connec ed de ices. This ca alogue se e is called The Hi e in nimBees.
- Mobile Se ice En i onmen . They a e basically mobile applica ions ha mus
be ins alled on e e y mobile de ice ha wan s o be pa o a PeaaS ne wo k,
p o iding an execu ion pla o m in which PeasS-designed se ices can be
deployed and execu ed. This is di ided in ou laye s o abs ac ion:
o Se ice Manage : This laye cons i u es he co e o he mobile se ice
en i onmen . I p o ides he capabili ies equi ed o deploying, emo ing,
unning o s opping se ices in he se ice con aine , and a basic
communica ion in e ace o es ablish a link wi h he ca alogue se e . I
also p o ides an in e ace o he use o he de ice o con igu e he mobile
se ice en i onmen p ope ly. In nimBees he in e ence engine is also
included in his le el. I consis s on a componen ha moni o s use
ac i i y o ex ac beha iou al pa e ns ha en ich he pe sonal in o ma ion
p o ile. The in o ma ion gene a ed by his componen is used o de e mine
whe he he owne ma ches he use p o ile equi ed by he ad e ise o
a ce ain se ice execu ion.
o Se ice Con aine : This laye p o ides he execu ion en i onmen whe e
he PeaaS-based se ices a e alloca ed, simila o a i ualiza ion engine.
The li ecycle o he alloca ed se ices is managed by he se ice manage
laye . In nimBees his laye has been implemen ed as a sc ip ing engine
ha p o ides an execu ion en i onmen in which un he nimBees
se ices, which a e w i en using an own sc ip ing language.
o Pe mission Con ol Laye : The access pe missions o he esou ces and
se ices p o ided by he mobile de ice a e e i ied by his laye , he eby
a oiding access o ce ain esou ces by unau ho ized se ices. Use s can
block he access o ce ain esou ces o in o ma ion in e e y momen and
o any se ice.
o Resou ces access Laye : This laye abs ac s PeaaS-based se ices
om pla o m-speci ic implemen a ions o each esou ce and se ice
o e ed by mobile de ices, p o iding a s anda d in e ace o be used
homogeneously by he se ices. In nimBees he access o he de ice’s
20
esou ces is p o ided o he deployed se ices in o m o “sou ces o
in o ma ion”. Such sou ce can be accessed by he sc ip ing language wi h
he p ope syn ax.
In conclusion, nimBees is a new pla o m ha uses he powe o People-as-a-Se ice
concep o p o ide a new communica ion channel o he use s, which can be used o
ad e ising companies o communica e wi h speci ic g oups o use s in a new and mo e
powe ul segmen a ion capabili ies.
In oducing BeeFun
Mobile ma ke ing is a g ea niche o PeaaS, Google and Facebook a e clea
examples o how collec ing sociological in o ma ion om ad e ising a ge s has become
ex emely impo an o companies [10] [16], bu wi h hei model is di icul o gain use
sa is ac ion. They y o c ea e apps as ools o launching pe sonalized
ecommenda ions o de ice owne s, while digi al ad e ise s can ge segmen ed da a
abou he use s and show a ge ed and ele an ad e ising o hem.
Howe e , he sociological p o ile is always cen alized on company se e s, p o iding
no anspa ency o he use and wi hou a p i acy gua an ee abou his da a. He e is
whe e PeaaS p esen s an al e na i e applied o mobile ma ke ing ha could bene i bo h
consume s and ad e ise s.
BeeFun
1
is a mobile app de elope o And oid and iOS ha makes use o he
nimBees messaging in as uc u e. I was de eloped by Gloin and allow use s o in e ac
wi h new kinds o a ge ed messages based in c i e ia such as age, ci y o esidence,
sex, cu en loca ion, e c. I is a po en ial example o his ma ke and he capabili y o
p o ide apid eedback on i s bene i s.
Wha BeeFun o e s is he abili y o ad e ise s o anonymously que y ad e ising
a ge s’ sociological p o iles o e ie e collec i e in o ma ion abou hei p e e ences,
a eas o in e es , and so on. This comple ely segmen ed da a le s ad e ise s e ine hei
ma ke ing s a egies and po en ially inc ease hei campaigns’ e ec i eness.
Fu he mo e, he in e ope a ion possibili ies suppo ed by PeaaS, and consequen ly by
BeeFun, le ad e ise s send ou in e ac i e ads o he app’s use s on he basis o hei
p o iles.
In e ope a ion be ween ad e ise s and a ge s is ca ied ou h ough he se ices
ha ad e ise s can deploy in he de ices. BeeFun is a he momen a me ely a p oo o
concep and he applica ions and se ices ha e o be w i en in his p op ie a y sc ip ing
language, which con ains basic ins uc ions o ask he in e ence engine abou he use ’s
sociological p o iles, eques no i ica ions and isual con en o display on he use
in e ace, and e u n and execu ion esponse o ad e ise s.
1
h p://www.bee unapp.com/
21
All he se ices ha ad e ise s can deploy in mobile de ices a e main ained in an
ex e nal ca alogue o a Hi e sys em, which egis e s all mobile de ices connec ed o he
ne wo k and manages he deploymen and execu ion o se ices and applica ions in he
de ices.
The mos impo an hing o his sys em, and which di e en ia es he PeaaS model
om o he s (like Google o Facebook), is he acco dance wi h he ou h pilla . The
sociological p o iles a e always kep in hei de ices, wi h only some o he in o ma ion,
olun a ily gi en by he use s) sa ed in he hi e. This basic in o ma ion allows he
ad e ise s o in e ac wi h a ge s, because i would be di icul o ind he a ge s wi hou
a cen alized sys em.
The hi e sys em le s he sys em p eselec po en ial a ge s on he basis o his
in o ma ion. Then, he de ices’ in e ence engines s ill mus e alua e he a ge ’s
sui abili y.
Conclusions
Th ough his implemen a ion o PeaaS we can explo e new o ms o mobile
ma ke ing ha allow ad e ise s o each hei a ge s mo e p ecisely. I allows accessing
de ailed sociological p o iles o he a ge s and e ie ing collec i e in o ma ion abou
hem. Ta ge s a e able o p o ide hei opinions abou p oduc s and ads while p ese ing
hei p i acy, wha is an inno a ion in his ield. The echnology allows he gene a ion o
a iew o collec i e in e es s, which could be ele an o ma ke s udies.
By ocusing on people and hei mobile de ices as in o ma ion sou ces, we can
con inue o hink o a new ad anced compu ing models ha p o ide nume ous bene i s.
Mobile-cen ic compu ing a chi ec u es can p o e hei g ea capabili y o gene a e
sociological p o iles h ough applica ions such as BeeFun.
Applying his model in ields wi h a high social impac could be ex emely bene icial
o socie y. PeaaS could le o ganiza ions iden i y pa e ns ha could be con ibu ing o
heal h diso de s among la ge communi ies. They could use his in o ma ion o plan
s a egic ac ions ha could help inc ease indi iduals’ li e expec ancies. In go e nmen ,
PeaaS could allow e ie al o collec i e in o ma ion abou ci izens’ poli ical iews and
p e e ed cou ses o ac ion o enhance li ing s anda ds and p omo e democ acy.
Howe e , he success o PeaaS equi es a wide adop ion by mobile de ice use s,
hey need o ecei e some bene i s in exchange o pa icipa ing. PeaaS gi es use s
mo e con ol o e hei p i acy, bu pe haps hey could ecei e bene i s om he
companies using hei in o ma ion. Fo example, discoun s o people using he PeaaS
pla o m. Addi ionally, a hi d-pa y company ha uses he pla o m o each use s could
mone a ily compensa e hose use s o gi e hem i s p oduc s o se ices o ee o a
discoun ed p ice.
22
The In e ne o People
The e is g owing in e es in de eloping applica ions on he In e ne o Things. The
main objec i e o hese applica ions is o in eg a e echnology in o people’s e e yday li e
in o de o be a se ice o hem. Howe e he e a e many hings o imp o e in he
implemen a ion. Usually, he use needs o se pa ame e s wi hin he applica ion, and e-
con igu e hem when he con ex changes.
Nowadays In e ne is a ailable almos e e ywhe e, connec ing e e yday physical
objec s o he Ne is ge ing e e easie . One o he main goals behind such sma
in eg a ion o de ices is o simpli y people’s li es by ha ing echnology wo k o hem
seamlessly [6]. Fo ins ance, you can emo ely swi ch on you house’s ai -condi ioning
sys em using you sma phone o ge a com o able empe a u e when a i ing home.
Howe e , how IoT echnologies a e cu en ly in eg a ed wi h human s ill lea es much
oom o imp o emen [17]. We need o de elop sui able mechanisms o p ope ly adap
o people’s con ex and make he ecnology able o wo k o people. Cu en ly people
a e o ced o change hei con ex o i echnological equi emen s.
Technology should ake people’s con ex in o accoun , lea n om i , an ake
p oac i e s eps acco ding o hei si ua ion and excep a ions, a oiding use in e en ion
as much as possible. In his desi able scena io, he ai -condi ioning sys em will know
whe e we a e in ou way back home, ins ead o swi ching on a speci ic p econ igu ed
ime.
Enabling such scena ios equi es mo ing om he In e ne o Things o he In e ne
o People (IoP) [30]. In his in as uc u e, sma phones play a cen al ole, e lec ing
hei cu en use as he main in e ace allowing people o be connec ed o he In e ne .
Howe e , he sma phones ha e o mee a se ie o equi emen s and capabili ies [28]:
- The sma phone needs o be capable o lea ning abou i s owne and hei
con ex , cons uc ing a digi al p o ile.
- I needs o anspa en ly nego ia e and p opose in e ac ions wi h o he de ices on
he In e ne , eac ing o s imuli and handling ela ionships.
- I needs o be able o manage digi al p o iles and ac acco dingly, p o iding i s
owne ’s con ex as a se ice o o he s, and scanning o se ices ha migh be
o in e es o i s owne o o upda e he digi al p o ile.
The in as uc u e o suppo he IoP concep can be achie ed wi h a combina ion o
he Social De ices
1
and People-as-a-Se ice pla o ms. Social De ices enhances he
p oac i e capabili ies o sma phones o o ches a e hei in e ac ions wi h o he de ices
connec ed o he IoT. PeaaS p o ides sma phones wi h se ing capabili ies ha allow
people o o e se ices om hei de ices, including hei con ex and sociological
p o ile. The combina ion o hese wo con ibu es o cons uc ing an IoP by allowing
1
h p://socialde ices.gi hub.io/
23
people’s con ex in o ma ion o be included in he coo dina ion and in e ac ion
managemen o IoT de ices. The main bene i is he in eg a ion o people as i s -class
ci izens in he IoT uni e se, opening he way o he de elopmen o new kinds o human-
iendly se ices and applica ions.
As sma phones con ains a lo o in o ma ion abou hei owne s, and senso s o
e ie e con ex ual in o ma ion abou he su oundings o i and, consequen ly, abou he
owne , hey a e he pe ec de ice o be he cen al piece o his new concep [30].
Sma phones can educe he complexi y o use in e ac ion wi h IoT-linked de ices. They
can know he use ’s ci cums ances, as hey a e almos always wi h he use . The
con ex ual da a can be used o know in o ma ion abou he use and o lea n abou him
and o de ec how he pe son beha es when hey a e a ce ain loca ions o su ounded
by ce ain o he people.
One o he mos impo an aspec s o People-as-a-Se ice is he ac ha all he
in o ma ion abou he use is kep in he owne ’s de ice, and is secu ely p o ided as a
se ice o hi d pa ies di ec ly om he sma phone. The use can limi who can access
o ha in o ma ion and know who and when is ecei ing i . The use can con ol his
sociological p o ile and all he in o ma ion easily, jus modi ying o dele ing i om he
de ice.
PeaaS allows a a ie y o in o ma ion o be collec ed, such as he moods, ends,
social s a uses, and heal h habi s o a g oup o people, in o de o de ine hei digi al
p o ile.
Cu en IoT echnology needs people-cen ic enhancemen s. As se ou in (Mi anda
e al., 2014) [30], he goals o any pe asi e compu ing con ex should be:
- Be social: In e ac ions in which hings, de ices and people pa icipa e mus be
social. The IoP should allow o he e ogenei y by suppo ing he di e en ypes o
de ices ha people use, allowing hem o in e ac wi h people and o he de ices.
Use s need o be empowe ed o adjus hei p e e ences and policies abou when
and wi h whom hei de ices a e socializing.
- Be pe sonalized: In e ac ions be ween de ices mus be pe sonalized o he
sociological p o iles and con ex s o hei use s, p o iding a anspa en
mechanism o his cus omiza ion. Use s mus be able o adjus hei p e e ences
in o de o con ol how hei p o ile is used by o he IoP s akeholde s.
- Be p oac i e: The igge ing o in e ac ions mus be p oac i e, no manually
commanded by he use . The IoT should allow o de ice he e ogenei y o allow
hem all o in e ac mo e p oac i ely, wi hou he need o manage all connec ed
hings. Use s mus be empowe ed o adjus hei p e e ences o con ol how
p oac i e hei su ounding hings a e, his is done by p oac i i y policy. I is e y
impo an because can in oduce secu i y isks.
- Be p edic able: In e ac ions mus be p edic able and igge ed acco ding o a
p edic able con ex ha he use had p e iously iden i ied, and wi h a speci ic
24
beha iou de ined. Use s mus be able o iden i y and ag ha con ex , de ine he
expec ed beha iou and se he p i acy policies o sha ing hei in o ma ion. The
use has o know how an in e ac ion can be s opped immedia ely and how o
p e en a misbeha iou in u u e.
In e ne o People Middlewa e a chi ec u e
Figu e 3. A chi ec u al desc ip ion o he IoP middlewa e.
In his a chi ec u e, he cen al componen is he De ice Regis y, who main ains he
in o ma ion abou he di e en de ices managed by he sys em. The de ices a e
classi ied in o Companion and NonCompanion.
Companion de ices a e hose ha main ain con ex ual and sociological in o ma ion
abou hei owne s, while NonCompanions a e “ hings” which can egis e hei
capabili ies and addi ional in o ma ion abou hei s a e.
The Ac ion Reposi o y manages he middlewa e’s se o ac ions. Each ac ion de ines
how he de ices should in e ac wi h each o he and wi h people.
The Applica ion Reposi o y componen s o es he di e en applica ions de ined in he
IoT con ex and managed by he middlewa e.
25
The Applica ion Manage allows people o enable o disable he di e en applica ions
acco ding o whe he hey would like o be p esen and con ibu e wi h hei sociological
p o iles.
The applica ion and ac ion speci ica ions may include eques s o he di e en
Sociological P o ile componen s o he Companion de ices in ol ed, so ha he
speci ied beha iou can be adap ed o he use ’s con ex , pe sonali y o mood.
Sociological P o iles
The Sociological P o ile componen is a undamen al pa o e e y Companion
de ice. This componen main ains he in o ma ion ha has been ga he ed and in e ed
abou people and hei con ex s. This in o ma ion co e s people’s beha iou ,
p e e ences and also con ex ual da a such as loca ion. De ices can also p ocess
con ex ual in o ma ion such as a p oximi y g aph o su ounding people/de ices/ hings.
Companion de ices can also no i y he egis y abou signi ican changes in hei
Sociological P o ile, so ha subsc ibed applica ions will be no i ied when an on-change
e en occu s, allowing hem o e- un he que ies o ge he eshly upda ed in o ma ion.
The componen also allows use s o de ine and adjus he sha ing schema o he
p o ile. Use s should de ine wha in o ma ion will be a ailable o which applica ions,
de ices o use s, and in which con ex he in o ma ion should be p o ided.
Finally, he Sociological P o ile p o ides an in e ace o de ice owne s o pe sonalize
he way in which he de ice beha es unde speci ic ci cums ances.
Conclusion
In e ne o Things is ha ing a success ul g ow h, howe e , people’s in e ac ion wi h
his kind o sys ems is s ill a om iendly. The e m In e ne o People has been used
be o e, bu i e e ed o adi ional Web sys ems designed only o humans o use. Now
he idea o his e m is o b ing he In e ne o Things close o people.
Mos o he p o ile’s cus omizable policies and p e e ences will be suppo ed by a
use - iendly wiza d. The componen will p og essi ely lea n he cus omiza ion du ing
he de ice’s use om he in e ac ions and con ex s in which i is in ol ed, and om he
owne ’s decisions abou how he de ice should beha e in hose in e ac ions.
To cons uc mo e human-like, p edic able, p oac i e and social in e ac ions be ween
people and “ hings”, he IoP middlewa e o e s a Web-based IDE and ools
1
, which we
will discuss in nex chap e .
1
h p://o ches a o js.o g
26
In eg a ion wi h Social De ices and In e ne o Things
People-as-a-Se ice and Social De ices combined p o ides new in e ac ions and
capabili ies o use s when hey a e co-loca ed, bu IoT can be also in eg a ed wi h hem,
p o iding new capabili ies and communica ions wi h e e yday objec s nea he use . In
some yea s 50 billion o de ices could be connec ed [11].
Some IoT- ela ed scena ios ha e been al eady implemen ed and in oduced using
Social De ices, o example The Social Co e Machine [25]. This co ee machine asks
o he use , based on his con ex , i he wan s o ha e a co ee, and hen he machine
in i ed o he people o join he co ee b eak.
Vázquez and López-de-Ipina [50] ha e been wo king on his subjec oo, hey
in oduced he concep o seman ic de ice as a sys em “able o spon aneously disco e ,
exchange and sha e con ex in o ma ion wi h o he ellow seman ic de ices as well as
augmen his con ex in o ma ion ia easoning in o de o be e unde s and he
si ua ion and pe o m he app op ia e eac i e esponse.” [48][49].
They ha e de eloped ou p o o ypes o wha hey call Social De ices, in eg a ed
wi h The In e ne o Things o make some objec s capable o in e ac wi h use s. They
eu ilize he concep o Seman ic Gadge , which was in oduced by Lassila and Adle as
de ices capable o pe o ming “disco e y and u iliza ion o se ices wi hou human
guidance o in e en ion” [24].
This concep emphasizes e en mo e he bene i s o he communica ion wi h o he
de ices in o de o de e mine he app opia e beha iou . In e ne is he enabling
echnology in his case, since he knowledge p o ided by local de ices can be
complemen ed wi h in o ma ion ob ained om he In e ne (which is p o ided by o he
de ices, se ices o humans).
Flexeo’s Senso Map: a wi eless senso ne wo k
This is he i s p o o ype by Vazquez and Lopez-de-Ipina. The goal o his p ojec
was designing a sys em o emo ely moni o ing wi eless senso ne wo ks h ough he
In e ne , and in eg a ing all he da a in o business p ocesses in o de o execu e da a-
mining ope a ions and de e mine co ela ions among da a. Scena ios ela ed o “hea lh
a home” and emo e moni o ing o indus ial acili ies we e designed in o de o es he
esul ing sys em.
As an addi ional ou come, a mash-up o senso collec ed da a and Google Maps was
de eloped, called Senso Map. This subsys em illus a ed he po en ial applica ions o
publishing up- o-da e senso da a on he In e ne in o de o dynamically in eg a e
di e en sou ces o in o ma ion o p o iding highe le el se ices.
The nex [Figu e 4], om [50], shows a senso ized chai wi h wo p essu e senso s
(a he sea and a he back) connec ed o a wi eless senso ne wo k node in o de o
p o ide in o ma ion abou i s use o local and emo e objec s. [Figu e 5] con ains a
27
sc eensho o he Senso Map applica ion showing he s a e o he chai (i someone is
sa on and/o leaned backwa ds).
Figu e 4. P essu e senso s in he chai .
Figu e 5. Flexeo’s Senso Map showing he s a e o a chai .
28
RealWidge s: om he desk op o he eal wo ld
Wi h RealWidge s hey wan ed o embody he unc ional capabili ies o hese digi al
en i ies in o eal wo ld iny wi eless displays, in o de o ha e small “windows” deployed
e e ywhe e opened o he in o ma ion om he In e ne .
The RealWidge is o med by an OLED display, wi h high esolu ion and con as
while small ene gy consump ion, in eg a ed wi h a C ossbowMo e2 wi eless senso
ne wo k node. A compu e ac ed as a ga eway be ween he In e ne and he wi eless
senso ne wo k, unning a RealWidge Managemen Applica ion ha connec ed o he
equi ed si es on he Web, downloaded he in o ma ion, analyzed i and inally sen he
con en o he app op ia e widge as con igu ed by he use . [Figu e 6] illus a es a
RealWidge showing in o ma ion abou he liquid le el in a emo e chemical con aine
ha was moni o ed by a wi eless senso ne wo k o e he In e ne .
Figu e 6. RealWidged ale ing abou he liquid le el in a ank.
Two bu ons we e p o ided o in e ac ing wi h he RealWidge , basically o
managing he ene gy by pos poning he in o ma ion o a la e ime, o immedia ely
disca ding i .
35
Se ice
Desc ip ion
Ac ion egis y
A egis y o all possible ac ions and in e aces ha a e a ailable.
De ice egis y
A egis y o all de ices and hei ins alled capabili ies. This egis y may
also con ain addi ional dynamic in o ma ion abou he de ices, such as
de ice p oximi ies and s a es, as well as con ain in o ma ion abou use s
and de ice owne s.
T igge egis y
A egis y ha consis s o a igge ing able, which holds in o ma ion abou
possible igge s and co esponding ac ions; and o an ac i e igge s able,
which holds in o ma ion abou igge ins ances ha ha e been obse ed,
a e s ill ali e, and ha e no ye been success ully scheduled.
Scheduling
Managemen o he scheduling ask.
Execu ion
Managemen o he ac ion execu ion, ha is, synch oniza ion and
coo dina ion o he de ices.
Table 3. The Se ices in he Ac ion-O ien ed P og amming Model In as uc u e.
Figu e 10. Ac o s, a e ac s and esponsibili ies in he ac ion-o ien ed p og amming model
in as uc u e.
36
The e a e se e al ac o s and hey in e ac wi h he a e ac s in di e en ways [Figu e
10].
- The in as uc u e p o ide deploys and adminis e s he Ac ion-O ien ed
in as uc u e.
- The capabili y de elope is esponsible o implemen ing he capabili ies o he
de ices, acco dingly o he de ined in e aces.
- The ac ion de elope uses he in e aces o execu e he necessa y capabili ies in
each de ice. The de elope doesn’ know he iden i ies o cha ac e is ics o he
de ices, and how he capabili ies a e implemen ed.
- The use ins alls he desi ed capabili ies o he de ice.
- The de ice could be conside ed as an ac o oo, i will egis e i sel o he
in as uc u e a un ime, and cons an ly epo s dynamic in o ma ion abou i s
s a e.
The o de o he illus a ed e en s can be di e en , because he ac ion de elope can
euse exis ing de ice capabili ies ins ead o w i ing code o new de ice capabili ies o
suppo mul iple mobile pla o ms.
The h ee basic s eps o c ea e an Ac ion-O ien ed p og am would be he ollowing:
- De ine o implemen he in e aces o he necessa y capabili ies o he de ice.
- De ine he ac ions o be execu ed, he p og am i sel , and speci y he needed
p econdi ions o s a ing he ac ion.
- De ine he igge s ha ini ia e he ac ions.
37
Requi emen s and cha ac e is ics o Social De ices
Ou de ices ha e a lo o in o ma ion abou us, maybe ou mobile phone knows
whe e we wan o d i e when we en e in he ca , and i can communica e ou des iny
au oma ically o ou GPS sys em. Also a he o ice, employees’ de ices can synch onize
each use calenda o s ablish he day and ime o a mee ing, and hen communica e he
esul o each employee. The concep can be aken beyond, o example, ou de ice
knows we ha e a mee ing in he hi d loo , so i could send ha in o ma ion o he
ele a o when we en e in i .
The key o Social De ices is o u ilize esou ces and capabili ies o mul iple physically
co-loca ed de ices and he p oximi y o he de ices in a use obse able manne . Some
o hese de ices can be s a iona y, like an ele a o o he sc een in a mee ing oom. This
echnology would i well in conjunc ion o o he echnologies like pe asi e compu ing
1
[7] o In e ne o Things [17].
Howe e , Social De ices equi es some de ices capabili ies o be able o wo k in
each de ice. The de ice mus be a compu a ionally capable de ice and be able o
in e ac wi h o he de ices and use s in hei p oximi y, o example ia Blue oo h. This
equi emen is easy o accomplish o de ices like mobile phones, lap ops o compu e s,
bu is di icul o o he kind o de ices like ele a o s, GPS sys ems, ca s, e c.
The capable de ices need o ha e iden i ies, so he pla o m would be able o keep
ack o he de ice iden i ies. De ices collabo a e in a way ha is o en pe sonal o he
use and ela ed o use p e e ences. The e is a cen alized egis y whe e de ices need
o egis e hei iden i y and capabili ies. The pla o m allows o egis e mul iple de ices
o each use and selec which o hem is he p incipal, he mos used.
Ano he equi emen is he need o know which o he de ices a e nea by he de ice.
Social De ices in e ac ions wo k in a co-loca ed way, and equi e ha use s and de ices
a e nea each o he . Thus, he p oximi y o de ices needs o be disco e ed and s o ed in
he Social De ices Pla o m, who main ains p oximi y in o ma ion in a o m o a
ma hema ical g aph, whe e nodes deno e de ices and edges deno e he dis ance o he
de ices. This in o ma ion is acqui ed by he de ices and epo ed o he cen alized
se e , who upda es he p oximi y g aph.
The e is also a need o know wha di e en collabo a i e ac ions a e a ailable
be ween de ices. The pla o m manages ac ion desc ip ions cen ally. As we ha e seen,
an ac ion has a body and a p econdi ion. The body desc ibes wha he ac ion will do,
and he p econdi ion is a se o equi emen o be comply he de ice o be able o
execu e he ac ion. Each de ice has a se o enabled capabili ies ha can be used by
he unning ac ions, ha ing a equi ed capabili y could be a p econdi ion o execu e a
de e mined ac ion. Also we may need o know which kinds o de ices can pa icipa e in
an ac ion, wha will be also cap u ed in he ac ion p econdi ion.
1
Pe asi e compu ing, also called ubiqui ous compu ing, is a concep whe e compu ing is made
o appea e e ywhe e and anywhe e.
38
Capabili ies a e a way o manage pe missions in Social De ices. The use can
enable o disable di e en capabili ies. Fo example, he e is a capabili y able o speak
aloud by syn hesizing ex , i he use doesn’ wan he de ice can speak, he can disable
he capabili y. Howe e he use won’ be able o execu e ac ions ha equi e ha
capabili y.
The condi ion when ac ions a e execu ed needs o be decided au oma ically a
un ime, bu he p econdi ions a e no only abou s a ic capabili ies o de ices, bu hey
may cons ain dynamic p ope ies changing on he de ice. Dynamic condi ions ha
a ec whe he a de ice can pa icipa e in an ac ion a e moni o ed ia de ice s a es. The
de ices con inuously epo hei s a es o he cen al se e , so i has all he necessa y
in o ma ion o decide i he ac ion can ake place o no . S a es can be based on gene al
p e e ences o he use as, o example, i he de ice is in silen mode.
All he condi ions ha e o be e alua ed wi hou use in e ac ion, bu con inuous
sea ching o de ices and e alua ing p econdi ions is compu a ionally ha d. He e is
whe e igge s ake place o ma k he need o ac ion execu ion. Ce ain ac ions a e
a emp ed o be scheduled when a ce ain igge is ecei ed. A igge can be an upda e
in he s a e o he de ice, a p oximi y change o an ex e nal e en . The de ices only
need o iden i y he condi ion o aise a igge , which will be ecei ed by he cen al
se e , who decides whe he and which ac ion mus be ini ia ed.
Also when execu ing an ac ion, one has o decide which conc e e de ices will
pa icipa e in i . The main ask is o ela e he po en ially a ailable de ices o he
p econdi ions s a ed in he ac ion desc ip ion and o decide he oles ha will be
assigned. Some imes he same se o de ices can pa icipa e in he ac ion wi h di e en
oles, o example in a dialog applica ion.
Finally, he ac ion needs o be execu ed, bu each de ice has a di e en so wa e
and ha dwa e ea u es. The ac ion coo dina ion o o ches a ion equi es de ining a
con ol low o each de ice, ega dless he di e en so wa e o ha dwa e speci ica ions.
In he SDP he e is a gene ic sys em whe e a ious coope a ion possibili ies a e
a ailable de ining a pla o m in o which coo dina ion is buil in.
These equi emen s and cha ac e is ics a e necessa y p ima ily o use and p i a e
de ices, because hey mus allow he owne o change hei p e e ences and sa e he
p i acy o he use . Howe e , o public Social De ices, like an ele a o , he e is no need
o e alua e p econdi ions o capabili ies, cause is a simple de ice which only will accep
eques s o nea by de ices and ac acco dingly.
39
A chi ec u e
This is he basic a chi ec u e o Social De ices, he de ices a e egis e ed in he
Con olle componen in he se e , and he use can ins all he in e aces and
capabili ies he wan s. Once connec ed, he de ice will look o nea by de ices and
epo hem o he se e , wha will igge some e en s. The se e hen e alua es i
he e is an ac ion a ailable o be execu ed o his de ice, and o ches a o execu es i .
Figu e 11. The basic a chi ec u e o Social De ices, and connec ions done s ep by s ep.
40
O ches a o .js
We ha e seen a small desc ip ion abou Social De ices Pla o m, who implemen s
he concep o Social De ices. Howe e we ha e seen i as an in oduc ion o he new
pla o m, called O ches a o .js and implemen ed by he same g oup o de elope s. This
new pla o m is he ocus o his wo k, and i is mo e powe ul han SDP.
O ches a o .js (OJS) is a comple ely new middlewa e ha suppo s Social De ices
concep . I is implemen ed as a single cloud se ice ha o e s a comple e se o ools o
implemen and p o o ype Social De ices applica ions. I is specially designed o
implemen in e ac ions be ween de ices and people, and o obse ing si ua ions when
hese in e ac ions should p oac i ely ge igge ed.
Social De ices applica ions di e om adi ional mobile apps, bu O ches a o .js
can be easily hos ed on any egula compu e and be used o quickly de eloping mul i-
de ice applica ions. The main ask o O ches a o .js is o main ain pe sis en connec ion
o de ices, s o e applica ions and con ex ual in o ma ion o de ices, and o e means o
igge ing applica ions.
A chi ec u e
O ches a o .js is based on Node.js
1
and he communica ion is based on Socke .IO
2
P o ocol. This o e s highly e icien inpu /ou pu ope a ions be ween O ches a o .js
se e and i s de ices. The se e is o med by he ollowing componen s:
- The O ches a o Co e c ea es new ac ion ins ances and makes su e ha he
ac ions wo k p ope ly. I main ains he connec ions o de ices and coo dina es
he de ices by handling incoming and ou going e en s based on he logic de ined
in ac ions.
- The Applica ion Manage s o es and execu es he applica ions, allowing use s
o s a /s op hem, and s o es use - ela ed se ings ha may be equi ed by he
applica ion. The applica ions a e like daemons o child p ocesses ha un
cons an ly and moni o changes wi h help o obse e s. When he app de ec s
changes, i uses some scheduling algo i hm de ined by he de elope o
de e mine whe he some in e ac ion should ake place.
- The De ice Regis y main ains con ex ual s a e in o ma ion abou each de ice,
which is s o ed o MongoDB
3
da abase. Mo eo e , he egis y main ains he
p oximi y g aph o he de ices, which is o med based on Blue oo h signal
s eng h alues (RSSI). This allows igge ing applica ions o a se o de ices
ha a e close o each o he .
1
Node.js is a se e -side Ja aSc ip pla o m buil on Ch ome’s Ja aSc ip un ime.
2
Socke .IO is a new non-s anda dized p o ocol o elying e en s be ween de ice and se e .
3
MongoDB is an open-sou ce documen Non-SQL da abase.
41
- The Ac ion Reposi o y con ains he ac ion de ini ions s o ed by he de elope s.
Basically hese de ini ions a e Ja aSc ip iles which each con ains p econdi ion,
e o handling, and he desc ip ion o join beha iou o de ices and in e ac ions
be ween use s by u ilizing de ice capabili ies.
- The Capabili y Reposi o y con ains desc ip ions o each capabili y, like Ja a’s
in e aces. The unc ionali y is no w i en he e, his only de ines wha me hods
he capabili y mus implemen , and wha pa ame e each me hod akes. The
de elope s mus implemen he capabili ies like hey wish and using wha e e
language, as long as hey ollow he capabili y desc ip ion.
Tools
O ches a o JS o e s a web-based use in e ace named Web Console. Use s can
manage hei p o ile and hei de ices, and o s a and s op he applica ions om his
console. I can be used di ec ly om mobile phone’s b owse and also be embedded
inside mobile apps wi h WebView UI componen s.
Figu e 12. Web Console iew whe e he use can add, modi y o emo e his de ices.
F om his iew, use s can add, emo e and manage hei de ices. In [Figu e 12] we
can see 2 de ices egis e ed o my accoun . Each de ice has an iden i ie , composed by
he use name and he de ice name.
We can see all he egis e ed de ices in he De ice Moni o iew [Figu e 13]. In his
iew we ha e all he in o ma ion a ailable by he pla o m o all he de ices. Fo
example, we can see he Blue oo h Add ess, UUID, Owne , Type o de ice, enabled
capabili ies, and he con ex da a.
Only egis e ed de ices can be used in he pla o m, so when a new use wan s no
use he applica ion, he has o egis e he de ice using he Web Console, and enable all
he desi ed capabili ies.
De ices can be modi ied and emo ed, so i he use doesn’ wan o con inue using
he pla o m, he can emo e his egis e ed de ices. All he in o ma ion abou ha de ice
will be o ally emo ed om he pla o m, and won’ be possible o use i anymo e.
42
Figu e 13. De ice Moni o wi h my egis e ed de ices.
All he enabled capabili ies appea in he iew, hey a e codi ied as a colou ed
squa e. We can enable he desi ed capabili ies in he se e -side om he same iew
whe e we egis e new de ices.
Figu e 14. Regis e new De ice iew. F om his iew we can add all he in o ma ion abou he de ice,
and enable he capabili ies we wan o use.
43
The Web Console also o e s a Web-based IDE o implemen ing he applica ions
wi hou ins all any so wa e. De elope s can de ine capabili ies, ac ions and obse e s
di ec ly om hei b owse s, and igge hem manually o a p ede ined se o de ices.
We can see, c ea e and modi y ac ions om he Ac ions
De ini ions menu. We will
see a iew like [Figu e 15]. All he de ined ac ions will appea on he igh menu, also
c ea e a new Ac ion is possible wi h he i s op ion.
On he le we ha e a big edi o whe e we can in oduce he body o he ac ion. This
edi o is e y powe ul, i will colou s he syn ax while we a e coding and show he line
numbe s. This IDE allows us o p og am ac ions en i ely wi hou he need o an ex e nal
so wa e. We can sa e he ac ion di ec ly o he se e p essing he Push bu on.
Figu e 15. Example o he Ac ions De ini ion iew, wi h he Apocalymbics ac ion loaded in he edi o .
We ha e wo ields below he edi o which allows us o execu e he ac ion in a
de ice. Pa ame e s can be de ices o simple da a like S ings, In ege s, e c.
Figu e 16. We can selec de ices o execu e he ac ion di ec ly in he d opdown lis .
44
Capabili ies a e implemen ed wi h he na i e pla o m speci ic ools, like wi h Eclipse
(And oid), howe e , he O ches a o JS pla o m o e s a ool o de ining he capabili y
desc ip ions. This iew [Figu e 17] is p e y simila o he Ac ions De ini ion iew, wi h all
he capabili ies on he igh menu and he edi o on he le .
Figu e 17. Capabili ies in e ace example, in his case wi h he Playe Capabili y loaded.
In his in e ace we can de ine all he unc ionali y o his capabili y. We can add
commen s o explain hem. All o he unc ions de ined in he in e ace mus be
implemen ed in he de ice [Figu e 18].
Figu e 18. Implemen a ion example o he unc ion se Volume o he Playe Capabili y in And oid.
O ches a o JS is a ailable o And oid, iOS, .NET Gadgee e and Py hon using
hese capabili ies. We only ha e o implemen he capabili ies in each de ice in hei
na i e language, and he Pla o m will call he unc ions.
51
I we need o ead in o ma ion om a websi e we can use he class eques . In oking
ac ions is done wi h he h p eques class and calling he REST API o he pla o m.
Figu e 26. Calling an ac ion om an applica ion. In oking an ac ion is achie ed sending he
co esponding POST message o he REST API.
We can also c ea e and manage Fibe s, which will allow us o c ea e pe iodic
ac ions, poll websi es, and check i any use has eques in o ma ion, e c. A Fibe is
basically as a Th ead in any con en ional p og amming language.
In he same way, we can use any unc ion and lib a y in NodeJS. The possibili ies
a e endless: access a da abase, c ea e socke s, connec o a web se ice, e c.
Figu e 27. A basic Fibe p in ing a message each 10 seconds while he applica ion is unning.
52
And oid Clien
Now we unde s and how he se e wo ks, le ’s see how he And oid clien connec s
o he se e , manages capabili ies and execu es he ac ions sen by he pla o m.
The main class in And oid is he O ches a o JSAc i i y.ja a. This class con ains he
basic unc ionali y o he clien : connec ion o he se e , managemen o capabili ies,
in oca ion o ac ions, se ings and con igu a ions. All he classes mus be inside he
main package com.ojs.
The main class will execu e when we
open he applica ion, i has a minimalized
iew wi h wo small bu ons, o connec and
disconnec . The i s ime we execu e he
applica ion i p obably won’ connec o ou
se e because i needs o be con igu ed.
In chap e 4 we will see how o ins all and
con igu e e e y hing.
When he de ice connec s o he
pla o m a small message appea s, i he
connec ion is los i will be also no iced wi h
a small message a he bo om o he
sc een. The de ice will y o load all he
enabled capabili ies.
Capabili ies a e di e en ia ed h ough
packages, each package means a di e en
capabili y.
All he capabili ies mus be in a package called com.ojs.capabili ies.myCapabili y and
hey mus implemen he public me hod ini Capabili y( Con ex applica ionCon ex ).
Figu e 28. ini Capabili y me hod om he hashTagCapabili y.
53
No e: The name o he package and he capabili y class a e e y impo an . The clien
will y o loca e he capabili y in a package wi h he same name as he capabili y in he
se e , bu wi h he i s le e in lowe case. Also he name o he class ha implemen s
he capabili y mus ha e exac ly he same name as in he se e ’s capabili y in e ace.
As we can see in [Figu e 28], he capabili y
name is HashTagCapabili y, which is he name
in he se e ’s capabili y in e ace. The name o
he class implemen ing ha capabili y mus be he same, and he name he package is
hashTagCapabili y, wi h he same le e in lowe case and inside he package
com.ojs.capabili ies.
The capabili y mus also implemen s he me hod ini Capabili y as in he igu e. In one
o hese s eps ail, he capabili y will be au oma ically disabled om he clien
1
.
Once his is accomplished, he capabili y will be able o load in he de ice and i is
eady o execu e he ac ions sen by he se e . The capabili y should implemen all he
unc ions de ined by he se e ’s capabili y in e ace. I one o he unc ions is no
implemen ed, he de ice will send an excep ion o he se e when an ac ion ies o
igge he me hod.
Usually a capabili y will c ea e a new Ac i i y in he de ice, his is no oo much
di e en as c ea ing a no mal ac i i y in a usual And oid applica ion. Me hods in
capabili ies can ha e a gumen s o ecei e in o ma ion om he se e , in his example I
ha e he me hod showNo i ica ion, which ecei es a i le and a message. Bo h
a gumen s a e sen o he new ac i i y h ough an In en .
Figu e 29. Implemen a ion code in Ja a o he me hod showNo i ica ion o No i ica ionCapabili y.
Then, we can ead he In en in he c ea ed ac i i y and use he da a p ope ly.
1
This is one o he ea u es I ha e implemen ed.
54
Figu e 30. No i ica ionCapabili y in e ace deployed in he se e . The capabili y ins alled in he clien
should implemen all o hese unc ions.
Changes and addi ional ea u es
To accomplish he a ge o his p ojec and be able o de elop he i e es
applica ions I ha e o make some changes and de elop new ea u es in he pla o m. I
will explain he e wha hese changes a e and how hey wo k:
- T igge ing obse e s when he a i ing da a is he same: In he o iginal
pla o m when con ex da a a i es o he pla o m, i didn’ igge he obse e
i he con en was exac ly he same as he pla o m al eady had. I ha e
changed his beha iou because some imes I need o igge he obse e .
Fo example, in some applica ions he se e asks he use o in oduce da a,
like he phone numbe . When he use in oduces his da a he applica ion can
s a , howe e , i he use in oduced he same phone numbe as he exis ing
one in he pla o m (wha is e y p obable), he applica ion didn’ igge he
ac ion.
- Changed he RSSI Blue oo h alues: I ha e
c ea ed 5 di e en anges o de e mine he
dis ance o a de ice based in he RSSI Blue oo h
alue. I changed his alues because i would be
be e i we ha e a be e app oxima ion o he
dis ance. Also he o iginal alues had nega i e
alues (and my de ices send a alue be ween 0
and 255), and some imes he p oximi y g aph
didn’ d aw co ec ly because alues we e oo
much high.
- Added suppo o Blue oo h p oximi y in he And oid clien : o iginally when I
s a ed my p ojec he And oid clien had no Blue oo h suppo . I ha e
implemen ed i in my e sion. La e , a Blue oo h se ice was in eg a ed in he
main pla o m. Howe e , bo h se ices can coexis in he applica ion. I had
implemen ed he Blue oo h suppo as a capabili y, wha he use can enable
o disable om he de ice. Se e applica ions call his capabili y when an
upda e o he Blue oo h p oximi y is needed.
55
- Au oma ically disable unimplemen ed capabili ies in he And oid clien : i a
new capabili y in e ace is deployed in he se e , p obably ou ins alled
And oid applica ion has no suppo o i . I ha e changed he beha iou o how
capabili ies a e ac i a ed, so he use won’ be able o enable capabili ies no
implemen ed in he clien . Unins alled capabili ies can’ be selec ed in he
menu [Figu e 31].
- Handle disconnec ion om he se e
o inish capabili ies: in my
implemen a ion I use some
capabili ies as a se ice, his se ices
can, o example, cons an ly upda e
Blue oo h p oximi y o GPS posi ion. I
ha e added a igge when he
connec ion wi h he se e is los , o
allow inish some ope a ions ha
ha e no much sense o be in
execu ion once he clien is
disconnec ed.
- Imp o ed debug messaged and
clean-up: some imes he console
was di icul o ead because some
debug messages appea eally as . I
ha e done small imp o emen s in
his aspec o be able o ead he
debug console in an easie way. I
ha e also added some e o
messages o he And oid clien .
Figu e 31. No ins alled applica ions can’ be enabled.
- Bug ixes: some small ixes and imp o emen s in he pla o m and he clien ,
o a oid c ashes and andom disconnec ions o he use s.
56
Goals
The main goal o OJS is o suppo Social De ices concep , his means ha he
middlewa e is a ge ed o acili a e and inc ease in e ac ions be ween people and
de ices when hey mee ace- o- ace. The objec i e is o coo dina e de ices whe e e
hey a e in he p oximi y o each o he .
Se ice composi ions may be di e en as ano he pla o m, he main sepa a ion is
ha he de ices can independen ly socialize, s a ing o esemble humans, and ge ing
simila cha ac e is ics.
Apa he main goal, he e is a se o ou goals he sys em should mee [8]
(B ons ed, Hansen, & Ings up, Se ice composi ion issues in pe asi e compu ing,
2010):
Con ex Awa eness
The sys em should be able o espond con ex changes by p o iding in o ma ion o
execu ing a command. OJS is specially designed o compose he se ices in an ad hoc
manne as he in e ac ions (ac ions) a e e y momen a y by hei na u e. Fu he mo e,
OJS can epo and s o e he e ogeneous and spa se con ex da a. As we ha e seen, he
pla o m suppo s he implemen a ion o obse e s, which can be used o epo ing and
ecei ing no i ica ions o eac hese changes in he con ex da a. The de elope s can
de ine hei own logic o do ha , and decide how o in o m he use .
Managing Con ingencies
The second goal is o manage con ingencies. OJS ac ions a e Ja aSc ip modules
ha con ain he in e ac ion de ini ion (body), bu also con ain h ee op ional me hods:
p econdi ion, wa ning and e o . I he e o me hod has been de ined i ge s called when
an excep ion is h own in he ac ion body, o in de ice-side code. OJS au oma ically
elays excep ions om de ice o he se e .
- The de elope can use he e o me hod o ace wha happened, and eac
acco dingly, o example in o ming he use . I he e o me hod is no de ined,
he de aul me hod kills he ac ion p ocess.
- The p econdi ion is used o making su e ha he ac ion can ake place, which
helps o p e en he con ingencies al eady be o ehand. Fo example, we can
de ine as a p econdi ion ha a ce ain de ice mus be connec ed, o he
con ex da a has a special alue. So he ac ion won’ be execu ed un il hese
p econdi ions a e no sa is ied.
- The wa ning me hod wo ks simila ly han e o me hod, bu i is used by he
de ice-side de elope o manually in o m he ac ion abou some mino e o
cases. The de aul handle does no kill he ac ion.
57
Le e aging De ice He e ogenei y
Acco ding o he hi d goal, he sys em should dis ibu e esponsibili ies in he
composi ion on he basis o capabili ies o di e en ypes o de ices. One o he main
concep s o he Ac ion-O ien ed P og amming Language a e capabili ies which a e used
o desc ibing wha a de ice can do. This makes i s aigh o wa d o he app de elope
o pick app op ia e de ices o he in e ac ions and se hei oles.
The con ex ual in o ma ion s o ed in egis y, as well as he in o ma ion in social
media and o he pa y se ices, can be accessed wi hin app and in e ac ions. I is
possible o de ine algo i hms o always selec he de ice ha bes i s o each pu pose.
Mo eo e , he cons an ly ex ending suppo o mobile and embedded de ices allows
aking ad an age o many o he e e yday physical objec s o OJS, which also suppo s
he goal as i helps o build cus om de ices o some speci ic pu poses.
Empowe ing Use s
The ou h goal demands ha a pe asi e app should change pu pose and
unc ionali y o e hei li e ime, and also ha he au oma ic solu ions o his do no
always wo k, so he middlewa e should o e means o he use o ecompose he
se ice.
The a chi ec u e o OJS and he concep s o AcOP model allow use s o con ol
in e ac ions in se e al ways. The use can now easily con ol wha applica ion he wan s
o un and wha in e ac ions his phone can ake pa o, by enabling o disabling de ice
capabili ies.
I he use ne e wan s he phone o alk, he can easily disable ha capabili y. I he
de ice is aking pa in an undesi ed in e ac ion, he can easily s op he applica ion.
Example: The Co ee B eak Machine
As we in oduced be o e, he e is an example o Co ee Machine de eloped using he
Social De ices concep , and in eg a ed in he O ches a o .js Pla o m [25]. He e we will
see a small abs ac o how his p o o ype was de eloped. I hink his is a e y
in e es ing p o o ype because i illus a es how he pla o m can be in eg a ed o apply
he In e ne o Things and he In e ne o People models.
In his implemen a ion, Bob is a wo ke ha akes he bus e e y mo ning o his o ice.
His mobile phone de ec s when he is in he bus and ib a es o ask i he wan s o ha e a
co ee when he a i es o he o ice.
The phone sends da a om senso s o De ice Regis y o he O ches a o JS
Pla o m, o example wi h GPS loca ion o nea by WIFI ne wo ks, o know he use is
lea ing home and way o he o ice.
58
Figu e 32. Pa ial code o he Co ee Machine P o o ype.
The use ecei es a no i ica ion asking i he wan s o ha e he co ee, he can always
decline ha ing a co ee, o jus igno e he no i ica ion and i will anish a e a sho
pe iod. The con ex ual da a can be s o ed in any o ma , and hen be p ocessed by he
app. Maybe i ’s necessa y o ansla e he GPS coo dina es o add ess, o calcula e he
dis ance and speed o know when Bob will a i e he o ice.
The De ice Regis y allows s o ing con ex ual da a in any o ma and publishing i
h ough pub/sub API. Co ee B eak app has wo obse e s ha ge no i ied when use ’s
loca ion changes, and when co ee is eady [Figu e 33]. In lines 13 o 18 he applica ion
checks i he use hasn’ ye been asked abou ha ing co ee.
When all he condi ions a e me , he applica ion calls he REST API using he
h p eques lib a y o igge he MakeCo ee ac ion. The second obse e [line 28]
checks when he co eeReady con ex da a a i es om he Co ee Machine, which
indica es ha he co ee is eady. The applica ion will hen igge he In i eFo Co ee
ac ion o in i e Bob and all his colleagues a he o ice o ha e a co ee oge he in he
co ee oom.
59
Figu e 33. O ches a o JS and all he componen s used in he Co ee Machine implemen a ion.
The applica ion was de eloped using he Web IDE and i is a small example o wha
can be done wi h he pla o m. This p o o ype was de eloped using he pla o m
F amewo k o .NET Gadgee e . This amewo k allows o make he echnology
a ailable o di e en kinds o ha dwa e de ices, which can help o in eg a e wi h IoT.
60
67
Ac i i y S a es
An ac i i y can be in one o he di e en s a es du ing he cou se o i s execu ion
wi hin an applica ion:
- Ac i e / Running: The ac i i y is a he op o he Ac i i y S ack and is he
o eg ound ask isible on he de ice sc een, wi h he ocus and in e ac ing wi h
he use . This would be he las ac i i y o be e mina ed in case he de ice is low
o esou ces.
- Paused: The ac i i y is isible o he use , bu does no cu en ly ha e ocus, o
example because he e is a dialog o ano he small ac i i y on op asking o an
ac ion om he use . These paused ac i i ies a e hold in memo y, and can be
quickly es o ed when gains he ocus again.
- S opped: The ac i i y is no isible o he use and i is e ained in memo y like in
paused s a e. Howe e ac i i ies wi h his s a e ha e a highe isk o be
e mina ed in low memo y si ua ions.
- Killed: The ac i i y has been e mina ed by he un ime sys em o ee up
memo y. The ac i i y is no longe p esen in he Ac i i y S ack and has o be
es a ed.
Usually, when he un ime sys em is going o change he s a e o an ac i i y, i sends
a no i ica ion be o e he change. This allows he ac i i y o eac acco dingly. Fo
example, we can sa e all he wo k be o e being killed.
MongoDB
MongoDB [31] is a c oss-pla o m Documen ed-O ien ed Da abase
1
. I is classi ied as
a NoSQL da abase and he da a is sa ed in a s uc u e o JSON-like documen s wi h
dynamic schemas, making he in eg a ion o da a in ce ain ypes o applica ions easie
and as e .
I was de eloped by he so wa e company 10gen (now MongoDB Inc.) in Oc obe
2007 as a componen o a planned Pla o m-as-a-Se ice p oduc . In 2009 i was shi ed
o an open sou ce de elopmen model.
MongoDB ocuses on lexibili y, powe , speed and ease o use [32]:
- Flexibili y: MongoDB s o es da a in JSON documen s, which is se ialized o
BSON (Bina y JSON). JSON p o ides a ich da a model ha seamlessly maps o
na i e p og amming language ypes, and he dynamic schema makes i easie o
e ol e you da a model han wi h a sys em wi h en o ced schemas such as a
Rela ional Da abase.
1
A Documen -O ien ed Da abase is a kind o NoSQL da abase designed o s o ing, e ie ing,
and managing documen -o ien ed in o ma ion, also known as semi-s uc u ed da a [13].
68
- Powe : MongoDB p o ides a lo o he ea u es o a adi ional Rela ional
Da abase Managemen Sys em such as seconda y indexes, dynamic que ies,
so ing, ich upda es, upse s (upda e i exis s, inse i i doesn’ ), and easy
agg ega ion. This gi es he wide unc ionali y om a RDBMS, wi h he lexibili y
and scaling capabili y ha he non- ela ional model allows.
- Speed and scaling: By keeping ela ed da a oge he in documen s, que ies can
be much as e han in a ela ional da abase whe e ela ed da a is sepa a es in o
mul iple ables and hen needs o be joined la e . MongoDB also makes i easy o
scale ou you da abase. Au o-sha ding
1
allows o scale you clus e linea ly by
adding mo e machines. I is possible o inc ease capaci y wi hou any down ime,
which is e y impo an on he web when load can inc ease suddenly and
b inging down he websi e o ex ended main enance can cos you business
la ge amoun s o e enue.
- Ease o use: MongoDB wo ks ha d o be e y easy o ins all, con igu e, main ain,
and use. To his end, MongoDB p o ides ew con igu a ion op ions, and ins ead
ies o au oma ically do he “ igh hing” whene e possible. This means ha
MongoDB wo ks igh ou o he box, and you can di e igh in o de eloping you
applica ion, ins ead o spending a lo o ime ine- uning obscu e da abase
con igu a ions.
Some impo an ea u es ha MongoDB suppo s a e:
- Replica ion: MongoDB p o ides high a ailabili y wi h eplica se s. A eplica se
consis s o wo o mo e copies o he da a. Each eplica se membe may ac in
he ole o p ima y o seconda y eplica a any ime. The p ima y eplica pe o ms
all w i es and eads by de aul . Seconda y eplicas main ain a copy o he da a on
he p ima y using buil -in eplica ion. When a p ima y eplica ails, he eplica se
au oma ically conduc s an elec ion p ocess o de e mine which seconda y should
become he p ima y. Seconda ies can also pe o m ead ope a ions, bu he da a
is e en ually consis en by de aul .
- Agg ega ion: MapReduce can be used o ba ch p ocessing o da a and
agg ega ion ope a ions. The agg ega ion amewo k enables use s o ob ain he
kind o esul s o which he SQL GROUP BY clause is used.
- Se e -side Ja aSc ip execu ion: Ja aSc ip can be used in que ies,
agg ega ion unc ions (such as MapReduce), and sen di ec ly o he da abase o
be execu ed.
- Capped collec ions: MongoDB suppo s ixed-size collec ions called capped
collec ions. This ype o collec ion main ains inse ion o de and, once he
speci ied size has been eached, beha es like a ci cula queue.
1
h p://en.wikipedia.o g/wiki/Sha d_(da abase_a chi ec u e)
69
How i wo ks
MongoDB can be managed in Linux om a command-line e minal. To connec o a
unning MongoDB se e we simply execu e he command:
mongo [--hos localhos ] [--po 21017]
Pa ame e s hos and po a e op ional, i hey a e no speci ied i will au oma ically
connec o a da abase lis ening on po 21017 on he localhos in e ace.
To show a lis o he da abases exis ing in he se e we can execu e he command:
show dbs
We can swi ch o a da abase using he command use and he name o he da abase.
No e ha i he da abase doesn’ exis i will c ea e i au oma ically. Tha ’s why MongoDB
is comple ely lexible DBMS:
use mydb
db
The las command db will show he cu en selec ed da abase. Now you a e using
he new c ea ed da abase called mydb. This new da abase is emp y a his momen , and
i won’ be shown in he lis o da abases e u ned by he show dbs un il any da a is
s o ed on i .
Le ’s inse some da a in JSON (called documen s). C ea ing a new documen is
done in he same way as we would do in Ja aSc ip :
Now we can inse his new documen s in o he da abase. We will inse hem in a
collec ion called myTes :
The show collec ions command will show all he collec ions exis ing in he da abase.
The ind command will ou pu all he documen s exis ing in he speci ied collec ion.
70
We can speci y a gumen s wi h he alue o he keys, so i will e u n only he
documen s ma ching he speci ied sea ch:
Upda ing and dele ing da a is no e y di e en o complica ed. MongoDB p o ides
also mechanisms o i e a e h ough e u ned da a om que ies, called Cu so s.
An impo an example o whe e MongoDB is e y use ul is combined wi h In e ne o
Things [33]. In e ne o Things needs o s o e in o ma ion om huge amoun s o senso s
and objec s, such amoun o da a is oo much o a ela ional da abase.
Bosch has buil i s In e ne o Things sui e on MongoDB, b inging he powe o big
da a o a new ange o Indus ial In e ne applica ions including manu ac u ing,
au omo i e, e ail, ene gy and many o he s.
In e ne o Things is a big challenge o da abases, each new gene a ion o ‘ hings’
comes wi h new senso s, which c ea e new da a and new unc ionali y equi emen s.
Rela ional da abases make i ha d o inco po a e new da a and i e a e on you da a
model. MongoDB can manage da a o any s uc u e, no ma e how o en i changes.
You can ship new unc ionali y wi hou edesigning he da abase, allowing build apps
ha could ne e build be o e.
Also, ela ional da abases a e no designed o handle da a om 40 billion de ices,
while MongoDB is buil o scale ou on commodi y ha dwa e.
Finally, MongoDB can analyse da a o any s uc u e di ec ly wi hin he da abase and
in eal ime. In ela ional da abases you p obably need o p ocess da a and p epa e o
ou pu in he co ec o ma .
Node.js
Node.js [29] is a way o unning Ja aSc ip on he se e , i allows o c ea e Web
applica ions using Ja aSc ip en i ely. Node.js is an in e ace o he V8 Ja aSc ip
un ime, he as Ja aSc ip in e p e e implemen ed by Google ha uns in he Ch ome
b owse .
Con en ional se e s, like Apache, spawns a new h ead o each connec ion o he
se e , bu his doesn’ scale well and i ’s e y easy o quickly lose pe o mance o he
si e goes down. Wi h Node.js his is no he case, you will need o handle he co e se e
a chi ec u e you sel . This can be o e whelming when you s a ou wi h Node.js, bu
he e a e modules o make his easie , and esul will be a high pe o mance web
applica ion.
Ja aSc ip is an e en -based language, so any hing ha happens on he se e
igge s a non-blocking e en . Each new connec ion i es an e en , ecei ed da a om a
o m i es an e en , eques ed da a om he da abase i es an e en . This means a
71
Node.js websi e will ne e lock up and can suppo ens o housands o concu en
use s [9].
In Node.js you don’ ha e a main code ha uns cons an ly in he backg ound, you
don’ ha e o check i some hing has inished o con inue, you will ne e wai o he
esponse o a da abase eques . In Node.js you speci y unc ions ha should be un o
comple ion o some hing else, and ha unc ions will execu e when hey a e igge ed by
he speci ied e en .
Figu e 38. A basic Hello Wo ld p og am in Node.js.
[Figu e 38] shows a Hello Wo ld p og am ha simply p in s he Hello Wo ld! message
when a new connec ion is c ea ed in he po 8080. This basic example can help o
unde s and how asynch onous class wo k.
Node.js is based in an e en d i en model, wha is e en -d i en p og amming? I liked
his desc ip ion om a P ince on Uni e si y Page [15] (E en -d i en p og amming,
2014):
“In compu e p og amming, e en -d i en p og amming o e en -based
p og amming is a p og amming pa adigm in which he low o he p og am is
de e mined by e en s—i.e., senso ou pu s o use ac ions (mouse clicks, key
p esses) o messages om o he p og ams o h eads.
E en -d i en p og amming can also be de ined as an applica ion
a chi ec u e echnique in which he applica ion has a main loop which is clea ly
di ided down o wo sec ions: he i s is e en selec ion (o e en de ec ion), and
he second is e en handling. In embedded sys ems he same may be achie ed
using in e up s ins ead o a cons an ly unning main loop; in ha case he
o me po ion o he a chi ec u e esides comple ely in ha dwa e.
E en -d i en p og ams can be w i en in any language, al hough he ask is
easie in languages ha p o ide high-le el abs ac ions, such as closu es. Some
in eg a ed de elopmen en i onmen s p o ide code gene a ion assis an s ha
au oma e he mos epe i i e asks equi ed o e en handling.”
72
To do a login sys em, we simply call he da abase que y when he use sends he
o m da a, Node.js sends he que y o he da abase, bu i won’ wai o he esponse.
When he esponse om he da abase a i es, we will ha e a di e en unc ion igge ed
when he da a a i es, which will send he esul back o he use . This allows he se e
o execu e ins uc ions ( o ano he use , o example) ins ead o was ing CPU cycles
wai ing o an IO esponse [52].
E en -d i en p og amming is mo e complex han usual p og amming, howe e
Node.js pla o m p o ides abs ac ions and modules ha make i easie o c ea e e en -
d i en applica ions.
Node.js is single h eaded, i can use only 1 CPU a any gi en ime, bu i p o ides
unc ionali y o ha e mul iple p ocesses lis en on he same po . Each indi idual p ocess
can use a single CPU so you can scale ou a mul i-CPU se e by using mul iple
p ocesses. You can also ake ad an age o scaling ac oss mul iple hos by using a
e e se p oxy, load balancing web se e [20].
Figu e 39. How a Node.js applica ion scales on a pe p ocesso basis and ac oss se e s.
Node.js can also ake a huge ad an age o JSON, Ja aSc ip Objec No a ion, which
acili a es da a ansmission be ween he se e and he applica ion wi hou he need o
XML. JSON is ligh weigh , human- eadable and can be pa sed di ec ly in Ja aSc ip .
Node.js is also a good choice o de elop REST ul se ices e y e icien ly and can
handle housands o eques s concu en ly, whe e PHP would jus collapse.
The majo downside o Node.js is ha i is s ill in he ea ly s ages o de elopmen ,
meaning ha some ea u es a e likely o change as de elopmen p og esses. In ac , he
documen a ion includes a s abili y index, which shows how isky use o each ea u e is
cu en ly [Figu e 40].
73
Figu e 40. S abili y index.
In conclusion, Node.js scales much be e han o he pla o ms: consumes less
memo y and is as e . The de elopmen can be complica ed o people no amilia ized
wi h he e en -d i en p og amming model and wi h Ja aSc ip , howe e is a good
al e na i e o de elop small and middle web applica ions wi h housands o connec ed
use s as could be, o example, a REST ul se ice.
In eg a ion wi h MongoDB
Node.js can be easily in eg a ed wi h MongoDB using he package MongoJS, which
le s you access MongoDB using an API ha is ex emely simila o MongoDB’s
Ja aSc ip shell.
The package can be ins alled om NPM simply using he command:
npm ins all mongojs
Le ’s see some examples. Fi s s ep would be he connec ion om Node.js o
MongoDB:
a da abaseU l = "mydb";
a collec ions = ["use s", "messages"];
a db = equi e("mongojs").connec (da abaseU l, collec ions);
This example will c ea e a new connec ion called db connec ed o he da abase
mydb using he de aul hos name and po (localhos and 21017). We can speci y
di e en alues as:
a da abaseU l = "use name:passwo [email protected]/mydb";
74
The pa ame e collec ions can be used o speci y he se o collec ions ou applica on
uses. This pa ame e is op ional, bu I ecommend o se i because i will allow us o
emula e a MongoDB Ja aSc ip clien like API wi hin ou Node.js applica ion.
Now, o ind a documen he syn ax is e y simila o he MongoDB console:
db.use s. ind({sex: " emale"}, unc ion(e , use s) {
i ( e || !use s) console.log("No emale use s ound");
else use s. o Each( unc ion( emaleUse ) {
console.log( emaleUse );
} );
});
This will e u n all he emale use s. This que y is no like usual que ies in o he
di e en p og amming languages. We a e no wai ing un il he esul o he que y, we a e
sending he que y o he DB and speci ying he unc ion ha will be execu ed when he
da abase e u ns he esul .
Sa ing a new documen in o he MongoDB da abase is also e y simila :
db.use s.sa e({email: “use @email.com”, passwo d: “ es MongoDB”, sex:
“male”}, unc ion(e , sa ed) {
i ( e || !sa ed ) console.log(“Use no sa ed”);
else console.log(“Use sa ed”);
});
Finally, le ’s see how o upda e a documen . In his case we a e changing he
passwo d o he use [email p o ec ed].
db.use s.upda e({email: "use @email.com"}, {$se : {passwo d:
" es Upda e"}}, unc ion(e , upda ed) {
i ( e || !upda ed ) console.log("Use no upda ed");
else console.log("Use upda ed");
});
As you ha e p obably ealized, all he unc ions has a pa ame e e , which will allow
us o de e mine i an e o occu ed, and a pa ame e wi h he esul o he ope a ion. In
he i s que y, he pa ame e use s will include he lis o esul s.
75
Chap e 4
Ge ing s a ed wi h Social De ices
In his chap e we will see how o ins all and con igu e he O ches a o JS Pla o m
and all he needed so wa e o ge he pla o m p epa ed o use and de elop. This guide
explains how o ins all he pla o m and all he equi ed so wa e in Ubun u. I ha e used
a Vi ual Machine o ins all e e y hing, bu i is alid o any e sion and compu e .
No e: I ’s e y impo an o keep he same e sion o some o he dependencies,
especially he npm modules, as he newe e sions ha e changed some names and
unc ions and hey a e no backwa d compa ible.
Dependencies
My es se e uns on Ubun u 14.04 LTS, howe e should be no p oblem in un
ano he Linux Dis ibu ion o e sion. Howe e his u o ial and all he dependencies
included in he CD-ROM a e speci ic o his e sion, and could no be alid in any o he
e sion o dis ibu ion.
- Ubun u 14.04 LTS.
- NodeJS 0.10.25.
- Npm 1.3.10.
- MongoDB Shell e sion 2.4.9.
- O ches a o JS Pla o m om my Gi hub eposi o y
1
.
- And oid Clien o O ches a o JS, which can be downloaded om my eposi o y
2
.
To es he applica ions we would need a leas wo And oid de ices, wi h And oid 4.0
o highe , and Blue oo h capabili ies.
No e: Due he Pla o m is s ill unde de elopmen , some applica ions could s op wo king
i speci ica ions change in he Gi hub eposi o y. I you wan o es hese applica ions
p ope ly I encou age you use my eposi o y ins ead he o icial.
1
h ps://gi hub.com/Jacb667/O ches a o JS.gi
2
h ps://gi hub.com/Jacb667/O ches a o JSAnd oid.gi
76
Ins alla ion
Ins all Node.js
P obably you can ins all i om he eposi o y o you Linux Dis ibu ion, bu i i ’s no
possible he e you ha e one eposi o y you migh use. To add he eposi o y o Ubun u
14.04 you can un a e minal and w i e hese commands:
sudo add-ap - eposi o y ppa:ch is-lea/node.js
sudo ap -ge upda e
Once he eposi o y is success ully added, we can ins all Node.js:
sudo ap -ge ins all py hon-so wa e-p ope ies py hon g++ make nodejs
I you ha e any p oblem o ins all Node.js o you wan o ins all i in a di e en
pla o m you can download and ins all om he Node.js websi e:
h p://nodejs.o g/download/
Ins all MongoDB
As in Node.js, p obably you can ins all i au oma ically om you dis ibu ion’s
eposi o y, i no , you can add his eposi o y o Ubun u and ins all i :
sudo ap -key ad --keyse e hkp://keyse e .ubun u.com:80 -- ec 7F0CEB10
echo ‘deb h p://downloads-dis o.mongodb.o g/ epo/ubun u-ups a dis
10gen’ | sudo ee /e c/ap /sou ces.lis .d/mongodb.lis
sudo ap -ge upda e
sudo ap -ge ins all mongodb-10gen
I you ha e any p oblem you can download and ins all i om he MongoDB websi e:
h p://www.mongodb.o g/downloads
Check i e e y hing is co ec
We can check i all he dependencies was co ec ly ins alled execu ing hese simple
commands in he e minal:
Figu e 41. Te minal sc een in Ubun u showing he e sion o all he ins alled
dependencies.
83
When hese capabili ies a e enabled, we can go o he menu Ac ions
De ini ions
and selec he ac ion called DialogTes . This ac ion [Figu e 47] will send a basic dialog o
he de ice, asking o answe YES o NO. The ques ion will be also said loudly by he
de ice, using he Tex -To-Speech Capabili y called TalkingCapabili y.
The ac ion will hen sleep o one second con inuously while he use answe s he
ques ion, once he use has answe ed, he ac ion lea es he loop and eads he
esponse, which will be ou pu o he se e console. Then, i calls he TalkingCapabili y
again, which will say he chosen op ion o he use .
To execu e he ac ion we can go down o
he a gumen s ield and selec ou de ice in
he lis . This pa ame e will be sen o he
ac ion as he d1 pa ame e o he unc ion.
Then, p essing he bu on igge will send
he ac ion o he de ice.
The de ice will say he ques ion using
he TalkingCapabili y and show he ques ion
and he op ions o choose.
To unde s and how capabili ies wo k, you can go now o he se ings iew in you
de ice and disable he TalkingCapabili y. I you execu e he same ac ion now, i will
show he ques ion and ask he use o an op ion, bu he de ice won’ use he Tex -To-
Speech ea u e.
S a ing he de elopmen
In his sec ion I will explain he mos impo an echnical de ails ha e e y de elope
should know abou he pla o m. I hink his in o ma ion can be a g ea s a ing poin o
any de elope in e es ed in c ea e new applica ions o he O ches a o JS Pla o m.
No e: A he momen o his wo k he e is no u he in o ma ion abou how o de elop
applica ions o he pla o m, all o his in o ma ion is he esul o my in es iga ion, and
could a y in u u e e sions.
Capabili ies
Capabili ies a e he main piece o de elop when s a ing o implemen a new
unc ionali y, as hey a e he only piece ha is implemen ed in he clien de ice, and
hus, he only way o c ea e new ea u es.
We will use he Web IDE o c ea e a new capabili y. To open he IDE, selec he
menu Capabili ies in he main menu o he O ches a o JS main page.
84
[Figu e 48] shows he basic in e ace o a new capabili y called Capabili yName wi h
a unc ion unc ionName.
All he code inside he in e ace should be inside he JSON objec module.expo s.
Inside he objec we can add unc ions in he same way as we c ea e JSON Objec s, in
he o ma unc ionName : unc ion (pa ame e s…).
Figu e 48. Basic code o a new unc ionali y.
We can add as much unc ions as we wan , all o hese unc ions mus be
implemen ed in he clien o a oid p oblems and un ime e o s.
In his example I ha e c ea ed a unc ion called unc ionName wi h 3 pa ame e s:
pa ame e A, pa ame e B, and pa ame e C. The e is no need o speci y he ype o hese
pa ame e s, he clien will map hem au oma ically h ough he name.
Le ’s see how his capabili y would be implemen ed in an And oid clien . Fi s s ep
will be c ea e he igh package o s o e he new
capabili y. This package mus be inside he
package com.ojs.capabili ies and he name o he package mus be capabili yName,
wi h he i s le e in lowe case.
Inside he package we ha e o c ea e he main class o he capabili y. The class has
o be named exac ly he same as he capabili y: Capabili yName.
Figu e 49. Implemen a ion o he me hod unc ionName in Capabili yName.
Pa ame e s o he me hod mus be named as we de ined in he in e ace, bu hey
can be wha e e ype, including a JSONObjec . Also we can speci y a e u n ype. The
e u ned alue will be aken by he se e when execu ing he ac ion, bu he se e
won’ wai o i .
85
We can speci y any hing inside he body o he unc ion. Fo example c ea ing a new
And oid Ac i i y, calling a web se ice, asking an inpu om he use , e c.
The me hod ini Capabili y is manda o y, we ha e o speci y i in he main class o all
he capabili ies. This me hod will be execu ed when he clien connec s o he se e and
i allows o adjus any ini ializa ion se ings he capabili y could need.
[Figu e 50] shows how o send con ex da a o he se e . The con ex da a mus be
a JSONObjec , which can include any kind o da a (JSONA ays, S ings, In ege s,
Booleans, Doubles, Floa s, e c.). This da a needs a name o di e en ia e i in he se e -
side.
Figu e 50. Sending con ex da a o he pla o m.
Ac ions
Le ’s c ea e an ac ion which uses he new capabili y. To c ea e an ac ion om he
Web IDE we can go o he menu Ac ions
De ini ions.
Ac ions ha e o be decla ed also inside he module.expo s objec , and hey a e
eally a new JSON Objec . The body o he ac ion mus be always a unc ion decla ed
wi h he name body.
Addi ional unc ions can be decla ed, o example, an excep ion handle , which will
be execu ed when he de ice h ows an excep ion.
[Figu e 51] shows a small example ac ion which calls he me hod unc ionName om
ou capabili y. I included he lib a y misc.js, which has a sleep me hod ha will be use ul
as he e u n alue om ou capabili y me hod won’ block he ac ion wai ing o he
esul . We need o poll he me hod un il he alue a i es.
I ha e implemen ed an excep ionHandle in his example, his me hod is op ional and
by de aul i will p in he excep ion message in console and s op he ac ion. We can
speci y he e any beha iou o ea he excep ion.
86
Figu e 51. Ac ion example using ou new capabili y Capabili yName.
The body unc ion has a pa ame e called de ice, his will be he de ice o execu e
he ac ion. This pa ame e can be se in he a gumen s lis when igge ing he ac ion.
Applica ions
Finally, we will implemen a simple applica ion which will call ou new Ac ion. Simply
we ha e o go o he Apps menu and selec he bu on NEW APP in he g id. This will
open us he Web IDE wi h an example code o how o implemen basic hings.
An applica ion will basically moni o con ex da a ecei ed om he de ices and
igge ac ions based on ha da a. We can also c ea e h eads (Fibe s) o execu e
pe iodic asks, like poll a websi e o se ice.
[Figu e 53] shows he i s pa o a basic applica ion. The i s we should do is o
include all he equi ed modules om Node.js:
Then, he JSON Objec modules.expo s is decla ed, in he same way o ac ions and
capabili ies, we ha e o de ine he name o he objec and he alue. An applica ion wil
ha e basically wo objec s:
87
- Se ings: his objec will include a lis o pa ame e s equi ed o he applica ion.
This pa ame e s will be eques ed o he use when unning he applica ion
[Figu e 52].
- Logic: his objec con ains
a unc ion wi h he body o
he applica ion.
The se ings de ined can ha e
a de aul alue speci ied.
Figu e 52. When he use uns he applica ion, i will be
asked o in oduce all he pa ame e s de ined in he se ings
objec .
In he example in [Figu e 53] I ha e de ined an obse e [Line 16] which is eading
he online con ex da a. This da a is au oma ically sen o ue when a new de ice is
connec ed o he pla o m.
I he de ice is connec ed (i he con ex Value is se o ue), and he iden i y o he
connec ed de ice is he iden i y he use has speci ied in he pa ame e de iceId, he
p og am will con inue, o he wise i will e u n wi hou execu ing any ins uc ion.
Then, he applica ion will call he REST API using he h p eques module, execu ing
he ac ion Ac ionName on he de ice con ained in he a iable de iceIden i y.
Figu e 53. Example applica ion (pa 1).
88
[Figu e 54] shows he second pa o he code. This pa only c ea es a h ead (Fibe )
ha simply p in s he wo d ick in console e e y 10 seconds.
Figu e 54. Simple example o a Fibe o he applica ion (pa 2).
No e: Always emembe o sa e you code be o e execu ing i . To sa e you code
simply p ess he bu on ‘Push’. I you change some hing in you code and you p ess he
bu on ‘S a ’ o ‘T igge ’ ( o ac ions), i will execu e he code sa ed in he se e , no he
code you a e edi ing in you b owse .
89
Chap e 5
Applica ion hiThe e!
In his chap e I will in oduce he i s s udy applica ion. The applica ion hiThe e! is
he i s o i e applica ions o p o e he po en ial o he pla o m. Also I will explain how
he en i e applica ion, i s ac ions, and i s capabili ies a e de eloped, o help o
unde s and he implemen a ion p ocess.
The goal o his applica ion is o ead he lis o con ac s om he de ice, and
au oma ically send a no i ica ion o he use i one o his con ac s is in he su oundings.
This is achie ed using a Blue oo h de ec ion o nea by de ices.
Each use will be asked o in oduce his mobile phone numbe , his phone numbe
will be mapped o he use ’s iden i y in he pla o m and, consequen ly, o his Blue oo h
MAC add ess, which will be used o de ec he p oximi y.
[Figu e 55] shows how he applica ion a high le el. Now I will explain i s de ails in
deep, as ac ions, capabili ies and con ex da a.
Figu e 55. Ac i i y diag am wi h he beha iou o he applica ion hiThe e!
90
Implemen a ion
I will explain now he mos impo an hings abou he implemen a ion. This
applica ion will use he ac ions: PhoneDialog, Ge Con ac Lis , SendNo i ica ion and
Upda eP oximi y.
The PhoneDialog ac ion will call he me hod askPhoneNumbe om he
phoneNumbe Capabili y.
In he clien his me hod simply s a s a new And oid’s
Ac i i y, which will show a dialog o he use o inpu he
mobile phone numbe . The pla o m sends wo pa ame e s:
a message o show o he use , and a imeou , which
indica es he numbe o seconds o au oma ically close he
Ac i i y i he use doesn’ in oduce he phone numbe .
The Ac i i y will send he phone numbe in oduced by
he use in an asynch onous message, as a con ex da a.
The code in [Figu e 56] shows he me hod o send he
con ex da a o he pla o m.
Figu e 56. F agmen o code ha sends he phoneNumbe in oduced by he use as con ex da a o
he O ches a o JS Pla o m.
91
// De ices wi hou a alid phone numbe canno pa icipa e in his app
pubsub.on( 'phoneNumbe ', unc ion( con ex Value, de iceIden i y )
{
console.log(de iceIden i y + ": " + con ex Value);
phoneMap[de iceIden i y] = con ex Value;
} );
This is he code o he obse e lis ening o he con ex da a called phoneNumbe ,
he applica ion will ecei e he phone numbe in oduced by he use and will s o e i in
an objec indexed wi h he de ice iden i y.
A he end o he applica ion he e is a Fibe ha cons an ly execu es upda ing he
p oximi y o all connec ed de ices. This will i e a e h ough all he connec ed de ices
(de ices which we ha e he phone numbe in he map) and eques hei Blue oo h
p oximi y using he Upda eP oximi y ac ion.
// Upda e Blue oo h isibili y o all connec ed de ices
Fibe ( unc ion()
{
while ( ue )
{
console.log( 'blue oo h ick' );
// I e a e h ough all he de ices and upda e he p oximi y.
Objec .keys(phoneMap). o Each( unc ion(key)
{
console.log( " eques p oximi y on " + key );
a pa ams = {
ac ionName: 'Upda eP oximi y',
pa ame e s: [ 'de ice:' + key ]
};
h p eques ( {
u i: '/api/1/ac ionins ance',
me hod: "POST",
o m: pa ams
}, unc ion( e o , esponse, body ) {} );
});
ools.sleep( 30 );
}
} ). un();
92
Blue oo hCapabili y is as I added Blue oo h suppo o he O ches a o JS Pla o m.
One o he las s e sions o he pla o m has added Blue oo h suppo oo, bu i is
implemen ed as a se ice ha uns cons an ly on he de ice. My capabili y is compa ible
wi h his me hod, bu my capabili y allows he de elope o ask o a Blue oo h upda e
only when i is needed.
This capabili y has some ini ializa ion code in he And oid de ice, ha code is
execu ed when he use connec s o he pla o m. The mos impo an piece o code in
his capabili y is he Blue oo h disco e y:
Figu e 57. Piece o code ha shows how he Blue oo h disco e y is implemen ed.
This Recei e is execu ed o each disco e ed de ice in And oid. Wha I do is s o ing
all he disco e ed de ices in he mDe icesMap, which s o es he Blue oo h MAC
add ess o he de ice and he RSSI, o be used as an app oxima ion o he dis ance
be ween bo h de ices.
The capabili y has a imeou con igu ed by he de elope , a e ha imeou he clien
will send all he disco e ed de ices. In he ac ion Upda eP oximi y I con igu ed his alue
as 10 seconds. Tha means ha he de ice will send all he disco e ed de ices a e 10
seconds, bu i will con inue looking o de ices and when i inishes will send all o hem
again.
The deli e y is done in a simila way as he phone numbe , each de ice is a JSON
A ay wi h he Blue oo h MAC add ess and he RSSI. All he de ices a e s o ed in
ano he JSON A ay, and inally e e y hing is w apped in o a JSON Objec [Figu e 58].
99
I he de ice is disconnec ing om he pla o m, he applica ion will dele e all he
s o ed da a ( he objec s sendedMap and connec edDe ices).
When he use c ea es a new isi ing ca d, i is sen o he pla o m as con ex da a
called isi ingCa d. This da a con ains all he ields and he image in Base64 o ma :
p i a e oid sendVisi ingCa d()
{
S ing name = name.ge Tex (). oS ing();
S ing add ess = add ess.ge Tex (). oS ing();
S ing phone = phone.ge Tex (). oS ing();
S ing desc ip ion = desc ip ion.ge Tex (). oS ing();
S ing email = email.ge Tex (). oS ing();
by e[] b = {};
Bi map bm = Bi mapFac o y.decodeFile(selec edImagePa h);
i (bm == null)
bm = ((Bi mapD awable) image.ge D awable()).ge Bi map();
By eA ayOu pu S eam baos = new By eA ayOu pu S eam();
bm.comp ess(Bi map.Comp essFo ma .JPEG, 50, baos);
b = baos. oBy eA ay();
S ing image = Base64.encodeToS ing(b, Base64.DEFAULT);
y
{
// Objec o be sen
JSONObjec sendDa a = new JSONObjec ();
// Visi ing ca d
JSONObjec isi ingCa d = new JSONObjec ();
isi ingCa d.pu ("name", name);
isi ingCa d.pu ("add ess", add ess);
isi ingCa d.pu ("phone", phone);
isi ingCa d.pu ("email", email);
isi ingCa d.pu ("desc ip ion", desc ip ion);
isi ingCa d.pu ("image", image);
sendDa a.pu (" isi ingCa d", isi ingCa d);
O ches a o JsAc i i y.ojsCon ex Da a(sendDa a);
}
ca ch (JSONExcep ion ex)
{
ex.p in S ackT ace();
Toas .makeTex ( his, "E o !", Toas .LENGTH_SHORT).show();
}
Toas .makeTex ( his, "Visi ing Ca d c ea ed succes ully!",
Toas .LENGTH_SHORT).show();
}
Then, he se e will ecei e his JSON Objec and s o e he isi ing ca d in he
connec edDe ices map, wi h he de ice iden i y as key. This da a is s o ed in he se e
o acili a e he sha e o he in o ma ion be ween de ices wi hou ha ing o eques he
100
in o ma ion om he use each ime. Howe e he da a is emo ed om he se e when
he use disconnec s.
Finally, he las new pa in his code is he p oximi yDe ices. This code wo ks in a
simila way o he code in he las applica ion, howe e in his case he isi ing ca d will
be sen o e e y use (connec ed o he pla o m) in p oximi y. Also I ha e added a
con igu able imeou in se ings, se by de aul o 10 minu es. This means ha he isi ing
ca d will be sen maximum once each 10 minu es.
// Reques da a om p oximi y
pubsub.on( 'p oximi yDe ices', unc ion ( con ex Value, de iceIden i y )
{
i (de iceIden i y in connec edDe ices)
{
// My isi ingCa d o send
a isi ingCa d = connec edDe ices[de iceIden i y];
o ( a i = 0; i < con ex Value.leng h;i++)
{
a de ice = con ex Value[i][0];
i (de ice in connec edDe ices)
{
a now = new Da e().ge Time();
i (sendedMap[de iceIden i y] != null &&
sendedMap[de iceIden i y][de ice] != null &&
sendedMap[de iceIden i y][de ice] > now)
{
console.log("I ha e al eady sen o his de ice!!");
}
else
{
// Don' send o his de ice in a leas 10 minu es
sendedMap[de iceIden i y][de ice] = new Da e().ge Time() +
se ings. imeou ;
a pa ame e s = [ 'de ice:' + de ice, isi ingCa d.name,
isi ingCa d.phone, isi ingCa d.add ess,
isi ingCa d.email, isi ingCa d.desc ip ion,
isi ingCa d.image ]
execu eAc ionDe ice( 'ShowVisi ingCa d', pa ame e s );
No e ha he code has been simpli ied, I ha e emo ed logs and debug in o ma ion
no needed in his example.
101
Ac ions can ecei e only p imi i e da a
(S ings, In ege s, Doubles, Floa s, and
Booleans) as pa ame e s, hey don’ allow
JSON Objec s as pa ame e s, which is he
eason I ha e o send all he ields in he
isi ing ca d one by one ins ead o sending
he en i e objec .
When all he equi emen s a e me he
ac ion ShowVisi ingCa d is execu ed in he
a ge de ice, showing my isi ing ca d o he
o he de ice.
Each de ice will send he isi ing ca d
based on his own p oximi y. This means ha
when wo de ices mee , maybe hey don’
in e change hei isi ing ca ds in he same
momen . Some de ices can ake mo e ime
o disco e nea by de ices.
[Figu e 61] shows he code om
ShowVisi ingCa dAc i i y which akes all he
a gumen s, including he image, and shows
he isi ing ca d o he use .
Figu e 61. Example o code o ead all he a gumen s and show he isi ing ca d o he use . The
image is sen as a S ing codi ied in Base64. This s ing is decoded and showed in an ImageView.
102
Conclusions
This applica ion can be e y use ul o businessmen who ha e a lo o mee ings in a
day. Thei mobile de ices can send he isi ing ca d au oma ically o e e ybody in hei
su oundings. This can be a cheap way o do publici y.
Fo example, he applica ion could un in a small de ice a he en ance o a shop,
and i will send he isi ing ca d o he people who pass in on o he shop. This sys em
could be used o o e discoun coupons o clien s o o make small ad e s.
O cou se, some people can hink his may be dis u bing, howe e hey can disable
he capabili y in any momen and hey will s op ecei ing isi ing ca ds.
103
Chap e 7
Applica ion ollowMe
FollowMe is he hi d applica ion. This basic applica ion asks he use o en e hei
p e e ences in a hash ag way. Then, he applica ion will ge all he use ’s hash ags and
ma ch hem based on he de ices in you p oximi y.
Wi h his applica ion you can easily ind people wi h you in e es s in you
su oundings. When he de ices ind one o mo e common hash ags, hey ask he use s
i hey wan o sha e hei da a wi h he o he use . I bo h use s answe “yes”, hei
social ne wo k da a will be sha ed each o he .
Implemen a ion
When a new de ice connec s o he applica ion, i au oma ically sends wo ac ions.
The i s , Reques hashTags, is a iew o eques he use o in oduce his in e es s. The
second ac ion, Reques SocialDa a, will eques he social ne wo k da a o he use .
Figu e 62. Fi s pa o he ollowMe code applica ion wi h he login obse e .
This da a will be sen by he And oid clien , and ecei e by wo di e en obse e s:
// Obse e o hashTags
pubsub.on( 'hashTags', unc ion( con ex Value, de iceIden i y )
{
de iceHashTags[de iceIden i y] = con ex Value;
console.log("hashTags " + de iceHashTags[de iceIden i y]);
} );
104
// Obse e o socialDa a
pubsub.on( 'socialDa a', unc ion( con ex Value, de iceIden i y )
{
con ac De ails[de iceIden i y] = con ex Value;
console.log("socialDa a " + con ac De ails[de iceIden i y]);
} );
The use has no limi in he numbe o
hash ags he can use. To add a new hash ag he
has o simply p ess he bu on New ag. To
emo e a hash ag he can swipe i om he lis .
Used hash ags a e also sa ed in he And oid
clien , so he use doesn’ ha e o speci y hem
when he disconnec s and connec s again. They
will be he e and he use can send hem back o
he se e o edi hem be o e deli e ing.
Finally, his applica ion also uses he
Upda eP oximi y ac ion, which upda es he
p oximi y o all connec ed de ices each 20
seconds.
This is he code o e ie e he sa ed social
ne wo ks om he And oid de ice:
public oid eques De ails() h ows Excep ion
{
Log.d(O ches a o JsAc i i y.TAG, " eques ing Social Ne wo k da a!");
y
{
JSONObjec sendDa a = new JSONObjec ();
JSONA ay socialDa a = new JSONA ay();
Accoun [] accoun s = Accoun Manage .ge (applica ionCon ex _).ge Accoun s();
o (Accoun accoun : accoun s)
{
i (accoun . ype.equalsIgno eCase("com. wi e .and oid.au h.login"))
{
JSONObjec cuen a = new JSONObjec ();
cuen a.pu (" wi e ", accoun .name);
socialDa a.pu (cuen a);
}
else i (accoun . ype.equalsIgno eCase("com. acebook.au h.login"))
{
JSONObjec cuen a = new JSONObjec ();
cuen a.pu (" acebook", accoun .name);
socialDa a.pu (cuen a);
}
}
sendDa a.pu ("socialDa a", socialDa a);
O ches a o JsAc i i y.ojsCon ex Da a(sendDa a);
105
This code will ead you social ne wo k da a om Twi e and Facebook. Suppo o
mo e accoun s can be p o ided, and an op ion o le he use which o hem he p e e s
o sha e o no .
[Figu e 63] shows he main implemen a ion o he p oximi yDe ices obse e . This is
he main code o he applica ion. This obse e will i e a e all he nea by de ices
epo ed by you de ice, compa ing hei hash ags wi h you s o ind po en ial ma ches.
I has a cool down ime like he las applica ion o a oid sending he no i ica ion
se e al imes o he same use .
Figu e 63. Main implemen a ion o he p oximi yDe ices obse e .
106
Howe e , his applica ion doesn’ use you p oximi y da a o send you in o ma ion
(like he eG ee ings app). In his case, he i s de ice who inds he po en ial hash ag
ma ch will s a he ac ion. The ime will be se in bo h de ices, o a oid o call he ac ion
again when he second de ice’s p oximi y senso inds he cu en de ice.
Once he applica ion has ound all he common hash ags, i will call he ac ion
SendReques In o ma ionDialog wi h 5 pa ame e s: he de ice iden i y I ha e ound, my
own de ice iden i y, he lis o common hash ags, he JSON Objec con e ed o S ing
(S ingi ied) wi h he con ac de ails o he i s de ice, and he con ac de ails o mysel .
[Figu e 64] shows he implemen a ion code o he ac ion. This ac ion will send a
dialog o bo h de ices asking i hey wan o sha e hei con ac de ails wi h each o he .
This is done using he capabili y DialogCapabili y, one o he o iginal capabili ies om
he O ches a o JS Pla o m. This capabili y will c ea e a small dialog wi h he op ions
YES and NO, wi h a imeou o 60 seconds.
I bo h use s answe YES o he sha ing o con ac de ails, bo h de ices will sha e he
in o ma ion each o he and he ac ion will inish.
Figu e 64. Implemen a ion o he ac ion SendReques In o ma ionDialog.
As we ha e seen in p e ious examples, eques ing da a di ec ly om a capabili y is
no synch onized, wha means ha he ac ion won’ wai o he esponse. Tha means
107
ha he ac ion has o poll i cons an ly un il i
ecei es he esponse. This is achie ed
using he sleep me hod om he misc.js
lib a y, which will sleep he unning ac ion o
1 second.
Conclusions
This applica ion can be use ul o mee
new people wi h same in e es s. P obably
e e y day when you a e going o he wo k,
shopping, o home, you pass nea people
wi h you same in e es s.
This sys em can be e y use ul o help
people in he s ee . I someone is looking o
some hing he can con igu e an ale in he
de ice, o example, he is looking o a oom
o en in he neighbo hood. P obably you
a e en ing a oom, so you can con igu e he
ad e in you de ice and when you a e
nea by in he s ee Social De ices will do he es o join you.
108
115
Chap e 9
Applica ion easu eHun
The las applica ion is a small easu e hun ing game whe e he use s can play and
compe e each o he . This applica ion also combines GPS loca ion wi h Blue oo h
p oximi y o know whe e you a e and wha do you ha e in you su oundings.
Each easu e can ha e some in o ma ion like a name, a desc ip ion o hin , he GPS
loca ion, a sco e, e c. The e is also a sco eboa d wi h he sco es o he playe s.
Implemen a ion
T easu es a e s a ic de ices, hey canno mo e and hey can be implemen ed in any
de ice wi h Blue oo h capabili ies. They won’ need o execu e he clien nei he In e ne
connec ion. Fo example, a easu e could be a small A duino o simila de ice wi h
Blue oo h capabili ies.
They a e se in he i s lines o he applica ion:
This applica ion makes use o he ollowing capabili ies:
- Reques Playe P o ile: This is e y simila o he used in he las applica ion, i
eques s da a om he use , like he use name and he a a a ha will be shown
o ano he playe s. This ac ion will call he capabili y loca eCapabili y.
- Reques Loca eT easu e: Ve y simila o he Reques Loca ePe son o he las
applica ion, bu his will show he lis o all nea by easu es, so we can pick one
o s a he hun ing.
- SendPlaye Lis : as he name says, his will be he ac ion esponsible o sending
he sco eboa d o all he playe s. The sco eboa d is sen when a new use
connec s o he applica ion, and each ime a use inds a easu e.
- Se Blue oo hT easu e: simila o Se Blue oo hTa ge , his ac ion will send special
da a like posi ion and Blue oo h MAC add ess o he a ge easu e.
- SendNo i ica ion: basic no i ica ion o in o m he use when some hing was
w ong. Fo example, when he has al eady ound a easu e.
116
Figu e 72. Example o iews. The i s shows all he nea by easu es, he use should selec one o
hem o s a he loca ion. Each easu e shows he dis ance o he cu en posi ion o he use . The
second igu e is an example o he sco eboa d.
The code o he obse e loca eT easu e is simila o he loca ePe son in he las
applica ion, howe e now we ha e o check i he use has al eady ound he selec ed
easu e:
117
Finally, he MapT easu eCapabili yAc i i y will do he same as
MapLoca eCapabili yAc i i y, i will check pe iodically he loca ion o he easu e, gi ing
hin s o he o ien a ion and dis ance. When he easu e is inside he Blue oo h ange,
he sc een will change i s colo acco ding o he dis ance.
When you a e close enough he easu e (abou 2-3 me e s), you will ge you sco e,
and he possibili y o selec ano he easu e o hun . The sco eboa d will be sen o all
he playe s o know you ha e ound a new easu e.
Figu e 73. Code o he obse e oundT easu e.
[Figu e 73] shows he code o he oundT easu e obse e . This obse e will gi e
you he co esponding poin s (5 in his case, bu hey can be con igu ed). Then, i will
ask you o selec a new easu e o hun , and send he
sco eboa d o all he connec e playe s.
You will see a small anima ion [see igu e a igh ]
gi ing you he sco e. This easu e will be au oma ically
disabled o you, so you won’ ge you poin s again.
118
119
Conclusions and u u e wo k
Social De ices is aimed o inc ease, acili a e and en ich social in e ac ions be ween
people in a ious kinds o co-loca ed and ace- o- ace si ua ions. In his wo k I ha e
p esen ed he la es pla o m implemen ing his concep : O ches a o JS. This pla o m
allows o c ea e and manage all he de ices, capabili ies, ac ions and apps h ough he
websi e wi hou he need o ex e nal ools.
Fu he mo e we ha e seen how he concep can be in eg a ed wi h such
echnologies like In e ne o Things and pe asi e compu ing, which a e no he main
a ge o he pla o m, bu has a big po en ial wi h hem.
To acili a e he de elopmen , he pla o m is based in he Ac ion-O ien ed
P og amming Model, which o e s abs ac ions o de elope s ha help hem o s uc u e
pe asi e applica ions. The de elopmen o applica ions is e y lexible and modula
because he main piece a e Ac ions, a small piece o code which can c ea e a small
in e ac ion be ween some de ices. Ac ions a e eusable and easily in eg a ed in any
applica ion.
In his wo k I ha e p esen ed se e al p o o ypes o hi ds using he People-as-a-
Se ice concep o ela ed wi h i . To inish his wo k, I ha e de eloped i e s udy cases
which show he po en ial o he pla o m. These applica ions a e only small examples,
wi h an impo an didac ical objec i e, because a he momen o his wo k he e was no
documen a ion abou how o s a he de elopmen in he pla o m.
In he u u e, he pla o m could ha e new ea u es and capabili ies, ecen ly he
Blue oo h suppo has been added, and p obably GPS suppo will be a ea u e soon.
P obably in he ideogames ield his kind o pla o ms can ha e an impo an u u e,
he e a e wo example ideogames whe e use s can s a he game and in i e people in
hei su oundings o s a playing. Wi h he di e en amewo ks and he REST API, he
echnology can also be ex ended o many di e en ypes o de ices.
Fu he mo e he new Blue oo h 4.0 Low Ene gy can be a pe ec solu ion o a oid he
cu en Blue oo h p oblems, which make he de ices o was e huge amoun s o ene gy
when inding o communica ing wi h nea by de ices. Wi hou Blue oo h Low Ene gy
won’ be possible he deploymen o a echnology like Social De ices.
120
121
Conclusiones y abajo u u o
Social De ices es á o ien ado a aumen a , acili a y en iquece las in e acciones
sociales en e las pe sonas de di e en es o mas coope a i as en si uaciones ca a a
ca a. En es e abajo p esen o la úl ima pla a o ma que implemen a es e concep o:
O ches a o JS. Es a pla a o ma pe mi e c ea y ges iona odos los disposi i os,
capacidades, acciones y aplicaciones desde un si io web sin necesidad de he amien as
ex e nas.
Además hemos is o como el concep o puede se in eg ado con ecnologías como
In e ne de las Cosas y compu ación ubicua, que no son el p incipal obje i o de la
pla a o ma, pe o iene un g an po encial en ellas.
Pa a acili a el desa ollo, la pla a o ma es á basada en el Modelo de P og amación
O ien ado a Acciones, que o ece abs acciones que ayudan a es uc u a aplicaciones
ubicuas. El desa ollo de aplicaciones es muy lexible y modula ya que la pieza
p incipal son las Acciones, un pequeño agmen o de código que capaz de c ea una
pequeña in e acción en e a ios disposi i os. Las acciones son eu ilizables y
ácilmen e in eg ables en cualquie aplicación.
En es e abajo he p esen ado a ios p o o ipos de e ce os u ilizando el concep o de
People-as-a-Se ice o elacionados con él. Pa a e mina es e abajo, he desa ollado
cinco casos de es udio que mues an el po encial de la pla a o ma. Es as aplicaciones
son sólo pequeños ejemplos, con un impo an e obje i o didác ico, ya que en el
momen o de es e abajo no exis e documen ación sob e cómo empeza a desa olla
pa a es a pla a o ma.
En un u u o, la pla a o ma pod ía ene nue as ca ac e ís icas y capacidades,
ecien emen e se ha añadido sopo e Blue oo h, y p obablemen e el sopo e GPS se á
p on o una nue a uncionalidad. P obablemen e en el campo de los ideojuegos es e
ipo de pla a o mas puedan ene un impo an e u u o, exis en dos ideojuegos de
ejemplo donde los usua ios pueden comenza a juga e in i a a pe sonas a su
al ededo pa a juga jun os. Con los di e en es ma cos de abajo y la API REST, la
ecnología puede ambién ex ende se a muchos ipos di e en es de disposi i os.
Además la nue a e sión de Blue oo h 4.0 de baja ene gía puede se la solución
pe ec a pa a e i a los p oblemas del Blue oo h ac ual, que hace que los disposi i os
gas en g andes can idades de ene gía cuando buscan o se comunican con los
disposi i os ce canos. Sin Blue oo h de baja ene gía no se á posible el despliegue de
una ecnología como Social De ices.
122
123
Annex 1: npm modules
Due o he amoun o p oblems can appea because he e a e always new upda es o
he NodeJS lib a ies, I ha e done a dump o all my ins alled modules in npm. The idea is
o keep exac ly he same e sion o all he modules he e appea , o a oid any p oblem.
Fo example, he Socke s.IO lib a y is one o he mos p oblema ic lib a ies, because
om e sion 1.0 i doesn’ o e backwa d compa ibili y and he O ches a o JS Pla o m
won’ wo k as expec ed.
He e is he se o commands o ins all npm modules o he igh e sion. A he
momen hese commands should be enough, bu in he u u e mo e hings could a y.
Howe e he way o ins all old e sions would be he same:
npm ins all colo [email protected]
npm ins all c [email p o ec ed]
npm ins all ej[email p o ec ed].5
npm ins all exp [email protected]
npm ins all ibe [email protected]
npm ins all o e e @0.10.10
npm ins all o e e -moni [email p o ec ed].3
npm ins all [email protected]
npm ins all [email protected]
npm ins all mu2exp [email protected]
npm ins all no[email p o ec ed]
npm ins all no[email p o ec ed].1
npm ins all passpo @0.2.0
npm ins all passpo -loca[email p o ec ed]
npm ins all eques @2.34.0
npm ins all socke .i[email p o ec ed].16
npm ins all socke .io-clien[email p o ec ed]6
npm ins all wins [email protected]
To ins all he equi ed e sion we only ha e o speci y he name o he module and
he desi ed e sion, sepa a ed wi h ‘@’.
In he nex page is he en i e dump o npm ins alled modules.
124
├── colo [email protected]
├── c yp [email protected]
├── [email p o ec ed]
├─┬ exp [email protected]
│ ├── bu e -[email p o ec ed].1
│ ├── command[email p o ec ed]
│ ├─┬ connec @2.7.11
│ │ ├── by [email protected]
│ │ ├── cooki[email p o ec ed]
│ │ ├── o mi[email p o ec ed].14
│ │ ├── [email protected]
│ │ ├── [email protected]
│ │ └─┬ [email protected]
│ │ └── mim[email p o ec ed]
│ ├── [email protected]
│ ├── cookie-signa u [email protected]
│ ├── [email protected]
│ ├── [email protected]
│ ├── me [email protected]
│ ├── mkdi [email protected]
│ ├── ange-p[email p o ec ed].4
│ └─┬ [email protected]
│ └── [email protected]
├── ibe [email protected]
├─┬ o e e @0.10.10
│ ├─┬ cli @0.1.8
│ │ ├── [email protected]
│ │ └─┬ wins [email p o ec ed]
│ │ ├── asy[email p o ec ed]
│ │ ├── cyc[email p o ec ed]
│ │ ├── pkg[email p o ec ed].3
│ │ ├── eq[email p o ec ed].203
│ │ └── s ack- [email protected]
│ ├── colo [email protected]
│ ├─┬ la i o[email p o ec ed]
│ │ ├─┬ b oad[email p o ec ed]
│ │ │ ├── e en emi e [email p o ec ed]
│ │ │ └─┬ win[email p o ec ed].2
│ │ │ ├── a[email p o ec ed].22
│ │ │ ├── c[email p o ec ed].2
│ │ │ ├── e[email p o ec ed]
│ │ │ ├── pkgin [email protected]
│ │ │ ├── eques @2.9.203
│ │ │ └── s ack- ac[email p o ec ed]
│ │ ├── di ec[email p o ec ed]0
│ │ ├─┬ op im[email p o ec ed]
│ │ │ └── wo [email p o ec ed].2
│ │ └─┬ p omp[email p o ec ed]
│ │ ├─┬ ea[email p o ec ed]
│ │ │ └── mu e-s ea[email p o ec ed]
│ │ ├── e alida o @0.1.5
│ │ └─┬ win[email p o ec ed].2
│ │ ├── a[email p o ec ed].22
│ │ ├── c[email p o ec ed].2
│ │ ├── e[email p o ec ed]
│ │ ├── pkgin [email protected]
│ │ ├── eques @2.9.203
│ │ └── s ack- ac[email p o ec ed]
│ ├─┬ o e e -moni o @1.2.3
│ │ ├─┬ b oad[email p o ec ed]
│ │ │ ├── e en emi e [email p o ec ed]
│ │ │ ├─┬ nco[email p o ec ed]
│ │ │ │ ├── a[email p o ec ed].9
│ │ │ │ ├── i[email p o ec ed]
│ │ │ │ └─┬ op imis @0.6.0
│ │ │ │ ├── minimis @0.0.5
│ │ │ │ └── wo dw [email protected]
│ │ │ ├─┬ u i[email p o ec ed]
│ │ │ │ ├── a[email p o ec ed].9
│ │ │ │ ├── [email protected]
│ │ │ │ ├── [email protected]
│ │ │ │ ├── m[email p o ec ed].5
│ │ │ │ ├── n[email p o ec ed]
│ │ │ │ └── [email p o ec ed].5
│ │ │ └─┬ win[email p o ec ed].2
│ │ │ ├── a[email p o ec ed].9
│ │ │ ├── c[email p o ec ed].2
│ │ │ ├── e[email p o ec ed]
│ │ │ ├─┬ eques @2.16.6
│ │ │ │ ├── [email protected]
│ │ │ │ ├── cookie-j[email p o ec ed]
│ │ │ │ ├── o e e -agen @0.2.0
│ │ │ │ ├─┬ o m-da [email p o ec ed]
│ │ │ │ │ └─┬ combined[email p o ec ed].4
│ │ │ │ │ └── delayed-s [email protected]
│ │ │ │ ├─┬ [email p o ec ed]0.2
│ │ │ │ │ ├── [email protected]
│ │ │ │ │ ├── c yp i[email p o ec ed].3
│ │ │ │ │ ├── [email protected]
│ │ │ │ │ └── sn [email protected]
│ │ │ │ ├── json-s [email p o ec ed]
│ │ │ │ ├── [email p o ec ed].11
│ │ │ │ ├── node-uui[email p o ec ed]
│ │ │ │ ├── oau h-si[email p o ec ed]
│ │ │ │ ├── [email p o ec ed]
│ │ │ │ └── unnel-a[email p o ec ed]
│ │ │ └── s ack- ac[email p o ec ed]
│ │ ├─┬ minim[email p o ec ed].12
│ │ │ ├── l [email protected]
│ │ │ └── sig[email p o ec ed].0
│ │ ├─┬ ps- [email p o ec ed]
│ │ │ └─┬ e en -s [email protected]
│ │ │ └─┬ op imis @0.2.8
│ │ │ └── wo dw [email protected]
│ │ └── wa [email protected]
│ ├─┬ ncon @0.6.7
│ │ ├── [email protected]
│ │ ├── [email protected]
│ │ ├─┬ op im[email p o ec ed]
│ │ │ └── wo [email p o ec ed].2
│ │ └── pkgin[email p o ec ed]
│ ├─┬ nssocke[email p o ec ed]
│ │ ├── e en emi e [email protected]
│ │ └── [email protected]
│ ├─┬ op imis[email p o ec ed]
│ │ └── wo dw[email p o ec ed]
│ ├── pkgin [email protected]
│ ├── imespa[email p o ec ed]