A Study on Autonomous Coopertion etween Things in We of Things Jehk Yu, Hyunjoong Kng, Hyo-Chn Bng, MyungNm Be 2 Electronics nd Telecommunictions Reserch Institute, 38 Gjeongno, Yuseong-gu, Dejeon, 305-700, Kore {dzzng, knghj, ngs, mne}@etri.re.kr Astrct. Recently, the Internet of Things hs minly focused on estlishing connectivity in vriety of chllenging nd constrined network environments, nd the next logicl step will e focusing on the ppliction lyer on top network connectivity. We of Things plys criticl role in enling smrt things s eing first-clss citizens of the We. In this pper, we proposed we roker model on We of Thing tht cn perform collortion nd efficient opertion with vrious devices. This model is designed on the sis of coopertive gme theory so tht it my perform utonomous coopertion mong devices. In ddition, the model proposed ws so designed s to gurntee the mximum reliility y minimizing the informtion loss in mjor devices nd services. Keywords: Internet of Things, We of Things, Clssifiction of things, Device coopertion nd non-coopertion, We roker. Introduction The Internet of Things (IoT) is novel prdigm tht is rpidly gining ground in the scenrio of modern wireless telecommunictions. The sic ide of this concept is the pervsive presence round us of vriety of things or ojects such s Rdio- Frequency IDentifiction (RFID) tgs, sensors, ctutors, moile phones, etc. which, through unique ddressing schemes, re le to interct with ech other nd cooperte with their neighors to rech common gols [], [2]. With the generlity of the We, the Internet of Things is on the verge of experiencing mssive evolution. Indeed, strting from n Internet of nerly one hundred million computers, the We now turns to e n Internet of nerly 50 illions of things [3] presging trnsition from n Internet of Things to We of Things (). Some inititives re lredy offering things s We resources either coupling RFID reders to We sed infrstructure [3] or y directly inding We protocols onto physicl electronic crds prticipting in providing profusion of new, oject-enled, pplictions nd even the mens to msh them up together. Nevertheless, while pving the wy to novel pplictions through proves of concept, First uthor, 2 Corresponding uthor. SUComS 203, ASTL Vol. 26, pp. - 6, 203 SERSC 203
Proceedings, The 4th Interntionl Conference on Security-enriched Urn Computing nd Smrt Grid prcticl use of the We of Things () in lrge-scle deployment scheme still needs to e strengthened. We of Things plys criticl role in enling smrt things s eing first-clss citizens of the We. In this wy, we tools nd techniques cn e directly pplied to the rel world [4]. Gme Theory is discipline imed t modeling situtions in which decision mkers hve to mke specific ctions tht hve mutul, possily conflicting, nd consequences. Since then, the pplictions of gme re found in economics, politics, computer science, psychology nd sociology. In 950, John Nsh proved tht finite gmes lwys hve n equilirium point, t which ll plyers choose ctions which re est for them when given their opponents choices [5], [6]. In coopertive gme theory, the pyoffs of ech potentil group, or colition, cn e otined y the coopertion of its memers. On the other hnd, non-coopertive gme theory is concerned with the nlysis of the strtegic choices, the detils of the ordering nd the timing of plyers choices, which nd importnt to determine the outcome of gme. However, the relity is tht there is no efficient nd utonomous collortive technology etween things or services in we roker, which is core of. In this pper, we proposed we roker model on We of Thing tht cn perform collortion nd efficient opertion with vrious devices. This model is designed on the sis of coopertive gme theory so tht it my perform utonomous coopertion mong devices. In ddition, this model cn perform n efficient informtion opertion nd collortion etween devices through the coopertion etween unconstrined devices nd the coopertion etween constrined devices nd gtewys. It is expected tht the utonomous coopertion etween such devices will e ctively utilized in constructing environment. To stisfy this, gme theory of Nsh rgining solution (NBS) method y introducing system opertion nd performnce ws ttempting to mximize. In ddition, the model proposed ws so designed s to gurntee the mximum reliility y minimizing the informtion loss in mjor things or services. This pper is orgnized s follows. In Section 2, descries relted work nd the coopertive gme theory. In Section 3, the proposed coopertion model etween devise in We of Things is presented. Finlly, Section 4 discusses the conclusions nd future reserch directions. 2 Relted Work 2. The We of Things With the pulic interest of Internet of Things, the next logicl step will focus on the ppliction lyer on top of network connectivity, prticulrly We of Things. Tody, more nd more devices re uiquitously running round us. Trditionl communiction schemes mke use of heterogeneous protocols, softwre nd user interfces, mking hrd devices interction. Users would like to esily ccess pulic devices, whtever their implementtion choices [4]. In the pper, this glol devices interconnection is clled the Internet of Things. This does not refer to ny technology 2
A Study on Autonomous Coopertion etween Things in We of Things nor ny network structure, ut only to the ide of interconnecting ojects s well s we interconnect computers with the Internet. Use of the We s the pltform hosting nd exposing connected ojects, cn e explined y multiple technologicl nd usiness enefits, few of which including deployment, high vilility nd verstility, use of stndrdized communiction protocols nd the ecosystem creted thnks to We 2.0 prdigm. A thing ecomes Internet-enled if it is ssocited with networking cpility, which uniquely identifies it on the Internet (refer to Fig. ()). Tody, devices such s sensors, street lights, electric meters, nd ccess crds re lredy networked nd ccessed on the Internet; even IP-connected pcemkers re used to monitor the helth of ptients [7]. A thing ecomes We-enled when it is ugmented with We server (refer to Fig. ()) so tht it cn expose its functionl nd non-functionl cpilities on the We through HTTP. Reserchers hve lredy successfully emedded tiny We servers on resource-constrined things, mking We-enled things relity. Definitely, there is scope for REST in the re of We services, dvnces in REST sed We service rchitectures is propgting the strction of physicl things s services on the We. This trend gives rise to the possiilities of wrpping things in the physicl world s We services (refer to Fig. (c)). THING THING THING WEB SERCIVE SOAP/ REST INTERNET CONNECTION INTERNET CONNECTION WEB SERVER HTTP INTERNET CONNECTION WEB SERVER HTTP () Internet Enled Thing () We Enled Thing Fig.. Connecting things on the We (c) Thing s We Service 2.2 A Survey of Gme Theory Gme Theory is discipline imed t modeling situtions in which decision mkers hve to mke specific ctions tht hve mutul, possily conflicting, nd consequences. John Von Neumnn nd O. Morgenstern hve developed clssicl theory of two-person zero-sum gmes in their ook [5] Theory of Gmes nd Economic Behvior. Since then, the pplictions of Gme re found in economics, politics, computer science, psychology nd sociology. In 950, John Nsh proved tht finite gmes lwys hve n equilirium point, t which ll plyers choose ctions which re est for them when given their opponents choices [6]. In the 950s nd 960s, Gme Theory ws extended nd developed theoreticlly nd pplied to prolems of wr nd politics. Since the 970s, it hs creted revolution in economic theory. Moreover, it hs found pplictions in sociology nd psychology, nd found links with evolution nd iology. At the end of the 990s, Gme Theory hs een used in the design of uctions for llocting rights to the use of the electromgnetic spectrum nds to the moile telecommunictions industry. 3
Proceedings, The 4th Interntionl Conference on Security-enriched Urn Computing nd Smrt Grid The rgining prolem of the coopertive gme theory cn e descried s follows [6], [8]. Let k =,2,..., K e the set of plyers, nd S e closed nd K convex suset of R to represent the set of fesile pyoff lloctions tht the plyers cn get if they ll work together. Let U e the miniml pyoff k -person i rgining prolem. Given the cooperte things, we cn prove tht our prolem is two-person rgining prolem. First, we use S to represent the set of fesile pyoff lloctions for user i, i.e., S _ i { ( ) } i = U i U i = U i p i, m, n, m 0, n 0, m + n. () Then, the set of fesile pyoff lloctions tht the two plyers cn get when they work together is S { U = ( U, U ) U S, U }. (2) = S 2 2 2 From the definition of the rgining prolem, we know tht S should e close 2 nd convex suset of R. Since it is ovious tht S is closed, we only need to prove tht S is convex, which mens for ny 0 θ, if U = ( U, U ) S nd 2 U ( U, U ), θ U + ( - θ ) U S. By simple derivtion, we cn get = S 2 θ U LW θ U M f γ α β + (, ) + ( - ) = [ ( )( - - ) ( ) ] (3) 2 p where α = θ m + ( - θ ) m, nd β = θ n - ( - θ ) n. Becuse m, m, n, n 0, m + n, +, Thus, ( ) θ U m n nd 0 θ, it is esy to derive - θ U = + S α 0, β 0, + β. f γ AF α (4). We cn lso prove y the sme method tht θ U + ( - θ ) U = S. Therefore, θ U + ( - θ ) U S, nd S is convex. This wy, 2 2 2 we hve proved tht the gme etween users in our system is indeed two-person rgining prolem. β 3 The utonomous coopertion model etween things in In coopertive gme theory, when nlyzing the k-person rgining prolem, the coopertive solution should stisfy four xioms i.e., fesiility, preto efficiency, symmetry, invrince to liner trnsformtion, nd independence of irrelevnt lterntives. In this pper, we proposed to mximize system-wide performnce using Nsh rgining solution (NBS). Nsh showed tht there exists unique solution f stisfying the xioms of the ove, nd it tkes the following form * * ( ) ( ) ( )( ). u, v = f S, d = mx u - d v - d 2 ( u, v ) S (5) 4
A Study on Autonomous Coopertion etween Things in We of Things The solution to the following mximiztion prolem is then given y N α i u,..., u = f S, d = mx ( u - d ). (6) N i i ( u, v ) S i = * * ( ) ( ) In this pper, we proposed we roker model on tht cn perform collortion nd efficient opertion with vrious devices. This model is designed on the sis of NBS coopertive gme theory so tht it my perform utonomous coopertion mong devices. In ddition, this model cn perform n efficient informtion opertion nd collortion etween devices through the coopertion etween unconstrined devices nd the coopertion etween constrined devices nd gtewys. In this pper, we proposed devices coopertion system in We roker is composed of four modules: Applictions Service Lyer RESTful Interfce Service Profile Mngement Service Mngement Service Profile Repository Service Profile Service Control Service Composition Device Lyer Device Mngement Device ID Mngement Device Control Device Coopertion sed on NBS Device # Device #2 Device #3 Device #N Constrined Things Unconstrined Things Fig. 2. The utonomous coopertion model etween things in ) Service Profile Mngement module stores the profile informtion of ll services provided through roker. Minly the type of service, service provider informtion, such s the service nme is stored. Service control lock nd service composition lock serch or through the Service is utilized when comined. 2) Service Mngement module runs the ctul service, serch, registrtion nd deletion is performed. In ddition, the ppliction to request the services of roker in the to get the first module. In ddition, service composition lock y two or more existing services to provide new services to crete nd perform the role. 3) Device ID Mngement module stores the physicl device IDs s well s the gent IDs elonging to those devices. It lso stores su-network types nd service loction informtion. 5
Proceedings, The 4th Interntionl Conference on Security-enriched Urn Computing nd Smrt Grid 4) As function to mnge the resources of ctul devices, Device Mngement module contins the privilege informtion of control, registrtion, deletion, ccess nd execution for the devices. It lso includes device coopertion lock using NBS of gme theory in order to perform utonomous coopertion etween devices. 4 Conclusion nd future work In this pper, we proposed we roker model on We of Thing tht cn perform collortion nd efficient opertion with vrious devices. This model is designed on the sis of coopertive gme theory so tht it my perform utonomous coopertion mong devices. In ddition, this model cn perform n efficient informtion opertion nd collortion etween devices through the coopertion etween unconstrined devices nd the coopertion etween constrined devices nd gtewys. It is expected tht the utonomous coopertion etween such devices will e ctively utilized in constructing environment. In this pper, in order to stisfy, gme theory of Nsh rgining solution (NBS) method y introducing system opertion nd performnce ws ttempting to mximize. For the future work, we re plnning to develop prototype system sed on the proposed mechnism for things nd pplictions coopertion in We roker. Acknowledgments. This work ws supported y Electronics nd Telecommunictions Reserch Institute (ETRI) grnt funded y the Kore government [3ZC30, Development of USN/ Convergence Pltform for Internet of Relity Service Provision]. References. Atzori, L., Ier, A., Morito, G.: The Internet of Things: A survey. J Computer Networks. 54, 5, 2787--2805 (200) 2. Chenzhou, Y., Shen, W.J., Sun, W., Hu, G., Sun, X.Z.: Active Linked Dt for Humn Centric Semntic We of Things. In: 202 IEEE Interntionl Conference on Green Computing nd Communictions, pp. 9--6. (202) 3. Christophe, B.: Semntic Profiles to Model the "We of Things". In: 20 Seventh Interntionl Conference on Semntics, Knowledge nd Grid, pp. 5--58. (20) 4. Duquennoy, S., Grimud, G., Vndewlle, J. J.: The We of Things: Interconnecting Devices with High Usility nd Performnce. In: Interntionl Conference on Emedded Softwre nd Systems, pp. 323--330. (2009) 5. Neumn, J.V., Morgenstern, O.: Theory of Gmes nd Economic Behvior, Princeton University Press, (947) 6. Slly, H.A., Rfie, M.: A survey of Gme Theory using Evolutionry Algorithms. In: 200 Interntionl Symposium in Informtion Technology, pp. 39--325. (200) 7. Mthew, S.S., Atif, Y., Sheng, Q.Z., Mmr, Z.: We of Things: Description, Discovery nd Integrtion. In: 20 Interntionl Conference on Internet of Things nd 4th Interntionl Conference on Cyer, Physicl nd Socil Computing, pp. 9--5. (20) 8. Rsmusen, E.: Gmes nd Informtion. Blckwell, Oxford, U.K. (989) 6