Ar c hitectural design of IoT -c loud computing integration platf orm Adhity a Bha wiyuga *1 , Dan y Primanita Kar tikasari 2 , Kasyful Amr on 3 , Oc ki Ba gus Pratama 4 , Moc h. Wildan Habibi 5 F aculty of Computer Science , Br a wija y a Univ ersity , K ota Malang, Indonesia Abstract An integ r ation betw een the Inter net of Things (IoT) and cloud computing can potentially le v er age the utilization of both sides . As the IoT based system is mostly composed b y the interconnection of per v asiv e and constr ained de vices , it can tak e a benefit of vir tually unlimited resources of cloud entity i.e stor age and com putation ser vices to store and process its sensed data. On the other hand, the cloud computing system ma y get benefit from IoT b y broadening its reach to real w or ld en vironment applicati ons . In order to incar na te this idea, a cloud softw are platf o r m is n eeded to pro vide an integ r ation la y er betw een the IoT and cloud computing taking into account the heterogenity of netw or k comm unication protocols as w ell as the secur ity and data management issues . In this study , an architect ur al design of IoT -cloud platf or m f or IoT and cloud computing integ r ation is presented. The proposed softw are platf or m can be decomposed into fiv e main components namely cloud-to-de vice interf ace , authentication, data management, and cloud-to-user interf ace component. In gener al, the cloud-to-de vice interf ace acts as a data tr ansmission endpoint betw een the whole cloud platf or m system and its IoT de vices counter par t. Bef ore a session of data tr ansmission estab lished, the comm unication inter f ace contact the authentication component to mak e sure that the corresponding IoT de vice is legitimate bef ore it allo w ed f or sending the sensor data to cloud en vironment. Notice that a v alid IoT de vice can be registered to the cloud system through w eb console component. The receiv ed sensor data are then collected in data stor age component. An y stored data can be fur ther analyz ed b y data processing component. User or an y de v eloped applications can then retr ie v e collected data, either r a w or processed data, through API data access and w eb console . Cop yright c 2019 Univer sitas Ahmad Dahlan. All rights reser ved. 1. Intr oduction The de v elopment of industr ial re v olution 4.0 ha v e led to an emerging research issue in Receiv ed J une 7, 2018; Re vised J an uar y 28, 2019; Accepted F ebr uar y 28, 2019 TE LK OMNIKA , V ol. 17, No . 3, J une 2019, pp . 1399 - 1408 ISSN : 1693-6930, ac cre dit ed Fi rs t Gr ad e b y K emen r is t ekdik t i , De cree No : 21 /E / K PT/20 18 DOI : 10.129 28/TE L K OMN I KA .v 17 i 3 . 11 78 6 1399 * Correspo nding au t h or , e-m ail: bh a wiyuga@ ub .ac.id 1 , da n y .jali n @ub .ac.id 2 , kasyful@u b .ac.id 3 , the area of Inter net of Things (IoT) . In IoT , the per v asiv e and ubiquitous de vic es equipped wit h micro controller , sensors and actuators are e xpec ted to be interc onnect ed t hrough a tr ansc eiv er module using v ar ious kind of com m unication protoc ols . Thos e kind of interc onnect ion lea ds to the abil ity of an IoT based system to be uniquely identified, perf or ming percept ion of surrounding   e nv ironme nt and e xchanging those pe rceiv ed data through v ar ious kind of com m unicat ion medium. As a result, an y promis ing sm ar t applicat ion s can be de v eloped on t op of that ecos ys tem su ch as: precision ag r icultur e , b uildi ng healt h monitor ing and sm ar t g r id [1] .   In gener al, IoT   b uilding b loc k com pose of six diff erent par ts . There are cons is t   of   ide ntific ation, sensi ng,   com m uni c a tion, com putat ion, ser vic es and sem antic s   par t [2].   Th e   co mput ation par t located on mic rocont roller att ached to IoT de vic e , has the main rol e to process   data acquired   fro m sens ing par t. It has basic ability to proces s sim ple comp utation taks such as analog to digital con v ersion locally .   Ho w e v er as a sy st e m r unning   in a longer   per iod,   the   data coll ected might e x- ceeds the amount st or age of local IoT   de vices to sa v e   and   proce ss   it   as requi red   to   be   analyz ed.   Ther ef ore ,   more   reliab le   st or age capacit y and po w erful com putatio nally entity is highly demanded. In   co ntr ast wit h IoT de vic es , a cloud computi ng entity off ers   a relat iv ely unlimit ed capabilit ies to st ore and proces s huge   data using vi r tual ization technology applied   on   it.   [3]   Not   only Cloud   com puting   reliability   of proces sing computa tion [4] b ut also   it max imizing the resourc e as   ment ion in [5].   K e yw or d s : cloud   computing,   inter ne t of thing s , softw are platf or m Evaluation Warning : The document was created with Spire.PDF for Python.
1400 ISSN: 1693-6930 Despite of its benefit, there e xists se v er al challenges ar ise from integ r ating the IoT to cloud computing including netw or k comm unication, secur ity and dat a management. [9] The first issue is netw or k comm unication which or iginates from the div ersity and v ar iability of a v ailab le netw or king protocols . T o name a f e w , there are TCP based protocols such as HTTP , MQTT and AMQP while on the other hand there e xists UDP based protocol such as CoAP [10] and [11]. The second issue is related to the ability of cloud system to confir m that its IoT de vice counter par t is a v alid par tner [12]. 2. Preliminar y Stud y on IoT -Cloud Platf orm TELK OMNIKA V ol. 17, No . 3, J une 2019 : 1399 - 1408 Aspects IoT Cloud Displacement P er v asiv e Centr aliz ed Reachability Limited Ubiquitous Components Real w or ld things Vir tual resources Computational Capabilities Limited Vir tually unlimited Stor age Limited or none Vir tually unlimited Role of the Inter net P oint of con v ergence Means f or deliv er ing ser vices Big data Source Means to manage T ab le 1. Complementar y aspects of IoT and cloud computing The int eg r ati on of IoT de vi ces and cl ou d computing en vi r onment has a ch ance to es cala te the le v er ag e of sy stem util ization on both si de s . On the othe r hand , IoT   sy stem ma y ge t be nefi t from tak ing stor age and computat ion re s our c e s pro vi de d b y the cl ou d en t ity . Fu r ther more , clo ud co mpu ting has more stab le natu re compar ed to IoT de vi ces in the IoT   s ys tem su ch tha t to en sure the a v ail a bil ity of IoT sensor data that has been col lected. Th e f or ma t ion of IoT -Clo ud area s rece iv e ser ious attention within the fiel d of IoT de v elopment as men tion in [6] .   In 20 10,   [7]   ha s proposed an e v ent- dr iv en sensor vi r tualiz ation appr oach f o r Inter ne t of Thin gs clo ud wh ich de monstr ate the de v elopment of IoT applic ation with re as on i ng ca pa bilit y us ing a g reen sc hoo l motorcycl e cas e study . In 2014 , [8] has pub li shed a util ity pa r ad igm f or IoT kn o wn as The se nsing cl ou d. Finally , aft er the v alidit y of IoT de vic e is confi r med and the data from sens or is receiv ed, ther e m ust be a mec hanism to eff ectiv ely st ore the data so that it can be proces sed and acc ess ed in the future . T o deal with issue s men ti on abo v e , an integ r ation la y er betw een the IoT and cloud co m puti n g is requi red. Thus in this study propose an archit ect u r al design of IoT -cloud sof tw are pla tf or m f or IoT and cloud com puting integ r ation. The proposed platf or m is com posed b y fiv e main co m ponent s nam ely cloud- to-de vice interf ace , authent icat io n, data management , and cloud   t o us er inter f ace com ponent. In gener al, the cloud   t o   d e vic e interf ace act s as a data tr ansm iss io n end point betw een the whole cloud platf or m sy st em and its IoT de vic es count er par t. Bef ore a sessi on of data tr ansmis sio n is est ab lish ed, the com m unicat ion interf ace contac ts authenticati on com ponent to mak e sure that the corres p o n ding IoT de vic e is legitim ate be- f or e it is g r anted to send the sensor data to cloud. Not ice that a v alid IoT de vic e can be regist ere d to the cl oud sys t e m throug h w eb console com ponent. The receiv ed sens or data are then collec ted in data stor age com ponent. An y st ored data can be fur ther analyz ed b y data proces sing co m ponent . User or an y de v eloped applicat io n s can then retr ie v e collec ted data, either r a w or pr ocess e d data, throug h API data access and w eb cons ole . The IoT com bined with Clo ud com puting func tionalit ies leads to a ne w conc ept kno wn as cl oud of things [13, 14]. The ter m IoT Cloud w as coined in [15] and [16] to e xplain the integ r ation of IoT and cloud com puting. Cloud comput ing and IoT basic ally de v eloped in diff erent w a y . F rom Bot ta’ s w or k, som e com plem entar y aspec t of cloud and IoT is illust r ated in T ab le 1 . I oT can tak e a ben efit of stor age and com putatio n resources pro vided b y cloud entit y .   Fur ther more , a more sta b le nature of cloud com puting sys tem com pared to its IoT de vic es count er par t ma y giv e more as sur anc e on the a v ailab ility of sens or data. On the other hand, the cloud com puting sy st em ma y get ben efit from IoT b y broade ning its reac h to real w or ld en vironment. Theref ore , more per v asiv e se r vi ce s can be de v eloped on top if this integ r ation. Fur ther more , the cloud can poss ib ly tak e a ro le as mi ddle w are to br idge betw een the user or applicat ion to the IoT de vic es . Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 1693-6930 1401 3. Resear c h Method The research methods conducted with se v er al steps . F or the first step is do liter ature study on the area of IoT , cloud computing and the issues related to the integ r ation of both entities presented in Section 2. Ne xt step is mak e a design of the system architecture of proposed IoT cloud platf or m which presented in Section 4. Based on the design, the ne xt step is to implement each component composing the sys t em in Section 5. After finish the implementation step , the system design is tested in ter m of its functionality and perf or mance . T esting result is presented in Section 6. Finally , the last par t is conclusion of the w or k mention in the conclusion in Section 7. 4. System Design Figure 1 illustr ates a gener al en vironment of the IoT based system using three actors namely IoT sensing d e vice , IoT gate w a y de vice and IoT cloud platf or m. Notice that, to simplify the ter minology IoT de vice is used to state both IoT sensing and gate w a y de vices . Figure 1. Gener al en vironment of IoT based system Architectur al design of IoT -cloud computing integ r ation platf or m... (Adhity a Bha wiyuga) The cloud should then meet the requiremen t of IoT applications . It has to be ab le to ser v e diff erent type of IoT application. All subject related to IoT application ha v e to be off ered the ser vices pro vided b y the IoT cloud. An IoT -Cloud system typically has three components . The first component is natur al en vironment where the sensors , actuators and connecting de vices located. Basically sensors as electronic de vices that inter act wit h the p h ys i c a l w or ld, gener ate data from their surrounding en vironment. The ne xt component is cloud computing system as par t f or stor ing, processing and analyzing gener ated data from the en vironment. IoT middle w are collects data from sensors , then tr ansf er it to Cloud en vironment. Vir tual machines r unning on the ph ysical ser v er in the cloud will handle the data to be store in a stor age system and do computation function. And the last component of proposed system is user applications . In user applications , user can do some actions such as visualizing result data from the sensor and controlling the en vironment remotely . Not only IoT sensing de vice has a ro le to perceiv e contempor ar y circumst ances of surrounding en vironment through in v olv ement of sensor s , b ut also has function to giv e a specific re spon s e to the en vironment b y adding actuator par t in IoT de vice . T h e collected sensor data will be tr ansmitted to the IoT ga t e w a y thr ough a wireless connection such as Wifi [17], Bluetooth Lo w   Energy   [ 18] ,   zigbee   [19]   o r   low   power   wide   area   network   (LPWAN)   [20] .   O nce   the   data   is   received,   IoT   gateway   device   relays   the   data   to   IoT   cloud   platf orm   thro u gh   a   cellular   or   backbone   network   connection.   In   this   sense,   the   IoT   gateway   should   have   capability   to   communicate   with   both   local   IoT   sensing   device   and   glo bal   IoT   cloud   platf orm.   Finally,   the   sensor   data   are   received   by   IoT   cloud   platf orm   and   stored   to   its   datab ase   systemf or   a   m o re   complex   data   processing   as   well   asfurther   data   access   by   users   or   developed   applicatio ns.   On   reverse   direction,   the   IoT   cloud   may   receive   a   specific   comm and   from   user   thr o ugh   a   predetermined   application   programmin g   inter face   (API)   and   send   it   to   IoT   device   through   the   assistance   of   IoT   gateway. Evaluation Warning : The document was created with Spire.PDF for Python.
1402 ISSN: 1693-6930 Figure 2. Building b loc k of proposed IoT cloud platf or m 5. Component Implementation In this section, the implementation of each component in proposed IoT cloud platf or m is e xplained. 5.1.1. Restful HTTP Interface TELK OMNIKA V ol. 17, No . 3, J une 2019 : 1399 - 1408 5.1. Cloud- to-D e vice Inter face As   stat ed   in   section   4.,   the   clou d-t o-device   interface   component   has   a   vit al   role   for   pro vi ding a comm un icat i o n en dp oint bet w een an IoT de vice and the whole cloud platf or m syste m. In orde r to pro vide such fu nctiona l ity a messaging protocol named Restful HTTP [21 ] is uti li z ed. T hi s proto col is chose n du e to their wide ado ption in the area of machine-to-machine (M2M) com m un ic a tion . Fi gure 3 illustr ates the design of REST ful HT TP in our proposed IoT cloud platf or m. R ESTf ul HTTP interf ace is designed using three main act ors . Nam ely I oT gate w a y/ de vic e act ing as a HT TP client, IoT Gate w a y’ s corres po nding HT TP ser v er locat ed in cloud platf or m and the data base .   Fi rst,   the   IoT   devices   send   the   acquired   sens or   data   toget her   wit h   its   aut hen ti cation   toke nin   form   of   HT TP   reques t.   Upon   recept i o n ,   the   HT TP   serv er   authent icates   th e   dev ice   identit y   by   com paring   t oken   c redential   w ith   one   in   aut henticatio n   dat abase.   Onc e   it   is   cl a r ifi ed,   t he   H T TP   ser ver   t hen   st ores   rec eived   s ensor   dat a   t o   t he   dat abase   as   s torage   component   and   send   back   a   r esponse   indicating   t hat   t he   operatio n   is   s ucc essf ully   perf or m ed.   I n   detail ,   t he   r equest   and   r esponse   f ormat   is   pres ented   in   T abl e  2   w hile   t he   dat a   f ormat   of   reques t   m e ssage   is   pr e sent ed   in   s ecti on   5 . 1. 2. In   t his   st udy   is   f ocus ed on   the   desig n and implem e n tation of IoT cl oud   pl atf or m   par t spes ifical ly . The b uilding b loc k of proposed   IoT clo ud platf or m can be ill ustr at ed   in   Fi gure 2   Th ere   e xist s fiv e   ma in   c omponent s   in   the sy st em. The y are cl oud-to- d e vice   inter f ace , authentic ation,   d ata management ,   and   cl oud   to-us er interf ace co m ponent.   In   gener al,   the cloud -to-de vic e   inter f ace act s as a data tr ansm iss ion end- point betw een   th e   whole   cl oud platf or m   sy st em   and   its   IoT gate w a y de vic es count er par t.   Bef ore a ses sion   of   data   tr ansmission   est ab lished,   the com m unicat ion interf ace cont act s the authentic ation   com ponent   to   ensur e   that   the corresponding IoT gate w a y is legitim ate bef ore bei ng   g r anted   to   se nd   ac quire d   sens or   data   to cloud en vironm ent.   Not ice that a v alid IoT gate w a y   can   be   regis t ered   to   the   cloud   sy st em through w eb cons ole com ponent.   The receiv ed se nsor data t hat has arr iv ed   in   cloud   sy st em   is then   co llected in data st or age com ponent. An y stor ed   data   can   be fur ther analyz ed b y data proces sing com ponent. Us er can retr ie v e the co llected data throu gh API data access and w eb cons ole. Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 1693-6930 1403 Figure 3. Design of RESTful HTTP interf ace T ab le 2. Request and response f or mat f or sensor data tr ansmission. 5.1.2. Data Format A compact data f or mat is required t o represent v ar ious kind of sensor data coming from 5.2. A uthentication The authentication component pla ys an impor tant roles to ensure the v alidity of IoT de Architectur al design of IoT -cloud computing integ r ation platf or m... (Adhity a Bha wiyuga) f ”results”:”Sensordatas has successfully added. g Request : POST /sensordatas/ HTTP/1.1 Host: Cloud IP:P or t Content-T ype: application/json A uthor ization: JWT tok en Sensor data in JSON f or mat. Detail in section 5.1.2. Response HTTP/1.1 201 Created di f f eren t IoT sens i ng de vic e s . Theref ore , this st udy is using a k e y   v alue data in the f or m of J a v a s cr ip t Obje c t Notation (J S O N). Th e JS O N off ers a relativ ely lig ht w eigh t data-interchange f o r ma t whi c h is easil y reada b le bot h b y human and mac hine   [22] . The com plete f or mat used b y IoT d e vices to send it s sens or dat a has tw o main k e y-v alue . The firs t k e y v alue is ”label k e y in dica t in g the id enti er of an IoT de vice register e d in cloud platf or m. The second k e y v alue is ”sen sors” k e y in dicati ng al l a v ai lab le se n s or data. Ins i de the ”sensors” par t there e xists another k e y-v a lue pai r s , nam ely ”v al ue” k e y indi c ating the list of sensor v alue with its corres pondin g time s ta mp an d a sen- sor la bel to classify those list. By using this data f or mat , an IoT gate w a y de vice can perf or m a da ta ag g regat ion of sensor data com ing fr om diff erent att ached IoT se nsin g de vic es in a cer tai n ti me windo w .A com pact data f or mat is required to represent v ar io us kind of sens or dat a comin g fro m diff erent IoT sensing de vices . The r e f ore , in this st udy , a k e y-v al ue da ta in the f or m of J a v ascr ipt Object Notation (JS ON) is used. The JS O N off ers a rela tiv el y li ght w ei gh t da ta- i nterc h ange f or mat which is easily readab le both b y human and machi ne . vi ces conne cte d to the clou d pla t f o r m. In order to do tha t, the cl o ud platf or m softw are shou l d ha v e t h e abil it y to pe rf or m a n a uthe ntica t ion process . Ho w e v er , instead of put ting the alre ady sensiti v e user name-p ass w o r d cre de n t ial on the de vi ce , the pr opo sed system is using the tok en ba s e d in f or m of JSON W eb T ok en (JWT). The JWT is chosen si n ce it ha s an e xpir ation me c ha ni sm wh i ch is saf er f r om a wireles s tam pe r ing an d sniffing tha n tha t of the user nam e- pass w or d. Fi g ur e 4 il lustra te s   th e   flo w   of   au the ntic ati o n   for   bo th   RE S Tful   H T TP.   Since   the   RE S Tful   H T TP   is   a   stateless   protoco l ,   the refo re,   for   ea ch   se ns or   da ta   tr ansmissi on ,   it   shou l d   alwa y s   conta i n s   the   token   creden ti al   loc ated   i n   th e   HT T P   h e ad er   pa r t. Evaluation Warning : The document was created with Spire.PDF for Python.
1404 ISSN: 1693-6930 5.3. Data Mana g ement The sensor data receiv ed b y cloud platf or m should then be stored on a database man- agement system (DBMS) f or fur ther processing and access . There e xists se v er al char acter istics of IoT sensor data that should be considered in selecting the r ight DBMS . First, the v ar iety of data can be e xpanded as additional sensor attached on IoT de vices . Fur ther more , a sensor data can be t r ansmitted in the f or m of streaming r anging from daily per iod to a near real time f ashion which ma y leads to a huge v olume of sensor data stor age . Theref ore , DBMS with a less str ict schema r ules combined with f ast wr iting perf or mance is suitab le f or IoT data stor age pur poses . T aking into account af orementioned requirement, in this study , a NoSQL DBMS called MongoDB is utiliz ed. In contr ast with SQL based DBMS , the NoSQL off ers a schema-less f eature in which doesn’t require to specify a predefined str ucture of the tab le . It means , the tab le pro per ty can be elastically defined dur ing the inser tion of ne w data. Fur ther more , MongoDB can be hor iz ontally scaled thanks to its sharding and replication capabilities . F or using MongoDB , the system need to define a big picture of data stor age in ter m of document. Figure 5 illustr ates the design of proposed data stor age using MongoDB . Figure 5. Design of proposed data stor age using mongoDB document In the design, MongoDB is consists with f our documents: Users , Nodes , Sensors and Sensor data. The User Document is utiliz ed to store user inf or mation which is used mainly f or TELK OMNIKA V ol. 17, No . 3, J une 2019 : 1399 - 1408 Figure 4. A uthentication mechanism Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 1693-6930 1405 authentication pur poses . A user ma y ha v e se v er al IoT de vice nodes under his/her super vision while a node can possib le ha v e more than one sensor . At last, there e xists Sensordata’ s document to store the receiv ed sensor data. It has relation with document nodes and sensors to giv e status about the data or iginated inf or mation. 5.4. Cloud-to-User Interface A cloud-to-user interf ace is de v eloped to pro vide an access to stored sensor data f or both user and other application using RESTful w ebser vice based application prog r amming interf ace . The specification of API f or request f or mat are presented in T ab le ?? . In addition, a w eb console to pro vide a g r aphical user interf ace (GUI) f or both managing allo w ed IoT de vices and accessing the sensor data is de v eloped. T ab le 3. API Spesification request f or mat f or sensor data access Request Descr iption POST /auth/ A uthenticate the user using JWT tok en GET /sensordatas/ Get all sensor data from all nodes GET /sensordatas/node/ node-id Get all sensor data a node with specific Get all sensor data a node with specific node-id and sensor-id 6. Result and Anal ysis In this section is presenting the result and analysis of the testing perf or med to the pro- posed IoT cloud platf or m in ter m of functional and perf or mance testing. All components are deplo y ed on a vir tual pr iv ate ser v er (VPS) with specifications 1.6 GHz Single Co re CPU with 1 GB RAM, 30 GB SSD dr iv e and pub lic IP address r unning Ub untu 14. 04 as ser v er oper ating sys- tem. W e use Apache v ersion 2.4.7 [23] as w ebser v er and Django F r ame w or k [24] f or de v eloping RESTful HTTP ser vice as w ell as JWT authenticationser v er . F or helping our w or k in database management system, MongoDB [25] v ersion 3.4.14 is chosen. 6.1. Functional T esting T o perf or m the functional testing, a set of hardw are prototype is de v eloped to r un IoT sensing and gate w a y de vice function. IoT de vice prototype is b uilt usin g Arduino Nano with A T - Mega 328 microprocessor . It also equipped with a 433 MHz RFM95 LoRa comm unication module . The IoT de vice prototy e is connected to sensor de vices to measure soil moisture , air temper ature- humidity and r ain le v el. The IoT de vices gather data from surrounding and tr ansmit the acquired data to IoT gate w a y . IoT gate w a y prototype is b uilt using Raspberr y Pi B ser ies with 1.2 GHz ARM Processor , 1 GB RAM and 8 GB MicroSD card. As comm unication interf ace , IoT gate w a y prototype is engaged with 433 MHz RFM95 LoRa comm unication module and IEEE 802.11 b/g/n module . Once t he IoT gate w a y de vice receiv es data from its conn ected sensing de vices , IoT gate- w a y de vice will send the data to cloud platf or m b y using RESTful HTTP protocol. First, W e tested the authentication mechanism b y sending a sensor data request to cloud system using both v alid and in v alid tok en. Figure 6a and 6b sho w the HTTP request and response with v alid and in v alid tok en, respectiv ely . F rom the result, w e obser v e that the authentication component successfully perf or m a v alidation b y giving diff erent responses to the both v alid and in v alid tok ens . The sensor data is then stored b y management component on a MongoDB instance . Figure 7 and 8 sho w the stored sensor data and its visualization in w eb console , respectiv ely . 6.2. P erf ormance T esting This study perf or med a test to measure the perf or mance of proposed IoT cloud platf or m in ter m of: response time of sensor data reception and throughput of RESTful HTTP ser v er . Architectur al design of IoT -cloud computing integ r ation platf or m... (Adhity a Bha wiyuga) no de -id GET   /sensordatas/ n o de / no de - id / sensor/ senso r -id Evaluation Warning : The document was created with Spire.PDF for Python.
1406 ISSN: 1693-6930 (a) HTTP request-response with v alid tok en k e y (b) HTTP request-response with v alid tok en k e y Figure 6. A uthentication mechanism testing using v alid and in v alid tok en k e y Figure 7. Sensor data stored in mongoDB . 6.2.1. P erf ormance of RESTful HTTP Interface F or measur ing the perf or mance of RESTful HTTP interf ace , this study utiliz es JMeter tool to sim ulate concurrent sensor dat a request from 50, 100, 150 IoT de vices . Figure 9 (a) and (b) sho ws the result of measurement in ter m of response time (in seconds) and the throughput of HTTP ser v er (in requests per second), respectiv ely . F rom the result can be obser v e that there is a conflicting trend betw een the response time and the throughput of sensor data reception. As the n umber of de vice increases , the response time tends to increase while the throughput decrease . This can be happened since the HTTP ser v er should handle and maintain more HTTP request a s the n umber of de vices increases . Ho w e v er , from the result has sho wn that at w orst the a v er age response is about 3 seconds while the ser v er can handle the 29 .23 requests/second f or 250 de vices concurrent connections which are still acceptab le consider ing the specification of the ser v er machine . TELK OMNIKA V ol. 17, No . 3, J une 2019 : 1399 - 1408 Figure 8. Data Visualization in W eb Console . Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA ISSN: 1693-6930 1407 Ref erences [9] D ´ ıaz, Man uel, Cr istian Mar t ´ ın, and Bar tolom ´ Architectur al design of IoT -cloud computing integ r ation platf or m... (Adhity a Bha wiyuga) Fi gure 9. (a) Resp onse time and (b) Throughput of IoT Cloud Plat f or m Sensor Dat a Rec eption f or V ar iou s Am ount of Conc urrent Connec t io ns (a) Res po ns e Ti me                                                                                                                                                     (b) Thro ug h pu t 7. Conc lusion The architectur al design of IoT -cloud platf or m f or IoT and cloud computing integ r ation has be en prese nted. Th e prop ose d so ftw ar e pla t f or m can be de compo sed into fiv e main compo ne nts : cloud -t o- de vic e interf ac e , au then tication, da ta man age ment, an d clou d- to- user inter f ace comp onent. In ge ne r al, the clou d-to-de vice inter f ace acts as a da ta tr an smission en dp oint be tw ee n the who le clou d platf or m syst e m an d its IoT de vices coun ter pa r t. Be f or e a session of da ta tr an smis- sion es t a b lish , the comm un ication inter f ace contact the au then t ica t ion comp onent to mak e su r e that the corre s po ndin g IoT de vice is leg itimate be f or e it is g r an ted to sen d the se nso r da ta to clou d. Not i ce that a v ali d IoT de vice can be re gister ed to the cloud sys tem through w eb co ns ole co mpone nt. The re c e iv ed sensor da ta ar e then colle ct ed in da ta stor ag e co mpone nt. An y stored  data  can  be  further  analyzed  by  data  processing  component.  User  or  any  developed  applications   can   then   retrieve   collected   data,   either   raw   or   processed   data,   through   API   data  access   and   web   console.   From   functional   testing   result   shown   that   the   proposed   system   has  been  able  to  provide  the  communication,  security  and  storage  functionalities.  Furthermore,  the  performance   result   shows   that   there   exists   an   impact   of   the   increase   of   concurrent   device  connections  on  the  delay  and  throughput  of  sensor data reception from IoT devices to the cloud system.   [8] Dist ef ano , S , et al. A Uti lit y pa r ad i g m f or IoT : The Sensing Clou d. P er v asiv e Mobile Co mputing . 2 01 5; 20 : 12 7 -144 . [1] Le e , In, an d K y oo chu n Le e . The Inter ne t of Thing s (IoT ): Applic a ti o ns , in v estmen ts , an d challeng es f or en t er pr ises . Bus i n ess Hor iz on s . 20 15 ; 58(4): 43 1-440 . [2] Al -Fuqaha , Al a , et al. Inter ne t of thing s: A sur v e y on en ab li n g tech nologies , pro tocols , an d ap pli catio ns . IE EE Com m un i c ation s Sur v e ys & T utor ials . 20 15;1 7(4): 23 47-2 376. [3] Bot ta, Ales sio , et al. Integ r ation of clou d comput i n g an d inter ne t of thing s: a sur v e y . Fut ure Gen er ation Com puter Sys t e ms . 20 16 ; 56 : 68 4-7 00. [4] K ob us i ns ka , A, et al. T o w ard s incre as ing rea l ibili ty of cl o uds en vi ro nme nt with RESTful w eb ser vi ce s . Fut ure Gen er ation Com pu t er Sys t e ms . 20 18; 87 : 50 2-51 3. [5] Bi s w as , A. R , et al. IoT an d clou d Con v erg ence: Opp or tun i t i e s an d challeng es . IEE E W or ld F or um on Inter net of Thing s (WF -IO T) (20 14 ): 37 5-376 . [6] Vi llar i, M, et al. Le v er ag i n g the Inter ne t of Thing s Int eg r ation of Sensors an d cl o ud comput i n g. Int er na - ti o nal Jou r na l Distr ib uted Sen sor Netw or k . 20 16; 12 : 97 642 87. [7] Al am, S , et al. SenaaS: An event-driven sensor virtualization approach for Internet of Things cloud . in 2010 IEEE International Conference on Networked Embedded System for Enterprose Application . 20 16: 1-6. e Ru bio . State - o f-the -ar t, challeng es , an d op en is su es in the inte g r ation of Inter ne t of things an d clou d comput i n g. Jour na l of Ne tw or k an d Co mpute r App li cat i o ns . 20 16; 67: 99 -11 7. Evaluation Warning : The document was created with Spire.PDF for Python.
1408 ISSN: 1693-6930 TELK OMNIKA V ol. 17, No . 3, J une 2019 : 1399 - 1408     Jou r na l of the K ore a instit u te of elec troni c comm un i cat i on sci e nces . 20 15; 10 (3): 41 9-424 .     Expres s . 20 16 ; 2(3 ): 10 0-1 02.     54(1 2): :32– 39 .     Mag azine . 20 15; 53 (9): 72 -79 .          40 th Inter na ti on al Con v en t ion on Inf or mation an d Co mm un ication T echn olog y , El ectro nics an d                         Mi croelectron ics (MIPR O) . 20 16: 50 1-50 5.         [ 12 ]     Bha wiy uga, Adh i ty a, Ma he nd r a Data , an d Andr i W ard a. Architectural design of token based      authentica tion  of  MQTT  protocol  in  constrained  IoT  device.   Te lecommunicat i on   Systems   Services               [ 20]     Bard yn , Je an -P au l , et al. IoT: The era of LPW AN is st a rt ing no w. Euro pe an Solid-Sta t e Circuits                             Conference ,   ESSCIRC   Co nfere nc e   2 016 :   42 nd .   IEEE,   2 016 . [11] Al -Fuqa ha, Al a, et al. T o w ard be tter hor iz on tal integ r ation amon g IoT ser vi ces . IEE E Co mm un ications [ 10 ]   Hedi,   I,  et  al.  IoT  ne twork  pro toc ols  com p aris o n   for   the  purpo se  of  I oT  const rained   networks.  i n   2017     and  Application s (TSSA), 2017  11 th Inte rnat i ona l   Con f eren c e  on.  I E EE,  201 7.      [13] M. Aaza m an d E. N. Huh , Fog Com puti n g an d Smart Ga t ew ay Based Co mm u nic a t ion f or Clou d of     thi ngs-a su r v e y , Digit . Com m un . Netw . . 20 17 .       Things, in 20 14 Inte rna tiona l Con fere nce on Fu ture I nterne t of Th ings an d Clou d. 20 14 : 46 4– 47 0. [14] A. F ar ah za di, P . Sham s , J . Rez az ad eh , an d R. F ar ah ba khsh, Mi dd l e w are tech no logies f or cl ou d of [15] L. Hou et al., Inter ne t of Thing s Clou d: Arc hitectu re an d Implemen tat i o n, IEE E Co mm un . Mag. , 20 16 ;         mig r atio n in IoT clou d, Future Gen er . Comput . Sy st.,2 01 8; 85 : 17 3– 18 3. [17] Ahmed, N., H. Rah ma n, an d Md I. Huss a i n . A compa r is on of 80 2.1 1 ah an d 80 2.1 5. 4 f or IoT . IC T [18] Ryu, Dae -Hyu n. De v elop me nt of BLE senso r modu le ba sed on op en source f or IoT ap plications . The [21] K er ane n, A, et al. REST ful Des ign f or Int er ne t of Thing s Sy ste ms . (20 15). [22] Br a y , Tim. The ja v as cr ipt ob j e ct no t a t i o n (js o n) da ta interch ange f or mat. (20 17). [23] Apache W eb Ser v er . The Apa che HTTP Ser v er Project . (20 18). [24] Django . Th e w eb fr am e w or k f or pe rf ection i sts with de adlines . (20 18). [25] Mon goDB , C . R. U . D . Introd uction-M ongo DB Man ua l 3.4 . (20 17). [19] Han, Dae -Ma n, an d J ae -Hyu n Lim. Sm ar t ho me en erg y mana ge ment sys tem using IEE E 80 2.1 5. 4         and zigb ee . IE EE T r an sa ction s on Con sumer El ectro nics . 20 10 ; 56 (3). [16] G. J . L. P au lr aj, S . A. J . F r an cis , J . D . P ete r , an d I. J . Jeba du r ai, Re source-a w are vi r tua l machine Evaluation Warning : The document was created with Spire.PDF for Python.