UBICC_Aitenbichler_477

8/7/2019 UBICC_Aitenbichler_477

1/7

SHAPING THE FUTURE INTERNET

Erwin Aitenbichler, Alexander Behring, Dirk Bradler, Melanie Hartmann,Leonardo Martucci, Max Mhlhuser, Sebastian Ries, Dirk Schnelle-Walka,

Daniel Schreiber, Jrgen Steimle, Thorsten StrufeTelecooperation Lab, Technische Universitt Darmstadt, Germany

[email protected]

ABSTRACTThe Internet of Things (IoT) and the Internet of Services (IoS) are two well-knownexemplars of the emerging Internet variants. These variants will be tightlyinterwoven yet specific with respect to the supporting technologies needed. Thepresent paper discusses the five variants identified as essential by the authors: IoT,IoS, Internet-of-Humans, Internet-of-Crowds, and Internet-of-Clouds. For eachvariant, a non-comprehensive set of research challenges is cited and related to thestate of the art and to ongoing projects of the lab .

Keywords: future internet, research challenges, Internet of Services, Internet of Things, cloud computing, social networks, HCI, security, privacy, trust.

1 INTRODUCTION

Academic research about and public andcommercial interest in the Future Internet isobviously split into two camps. The first camp sees apressing need for the (r)evolution of the IP andrelated network protocols plus correspondingconcepts in order to satisfy perceived future trafficcharacteristics and quality demands. Others look atfuture challenges more from an applicationperspective and see an urgent need to providesupport for several interwoven special variants of the Future Internet - each of which is likely toeventually outgrow the entire Internet as it is today.

In the eyes of the second camp, the conceptsdeveloped by the first camp shall provide the bitpipe for the concepts they work on. The authors of the present paper consider themselves members of the second camp. The paper will discuss fiveInternet variants which the authors identified asessential.

After a brief introduction of these variants below,the remainder of the paper will strictly follow thesequence of these five variants. For each variant (andpaper section), major research challenges will beidentified and the state of the art and relevantprojects will be cited. Due to the large extent of current research, the treatment can by no means becomprehensive. Nevertheless, interested readersshould be able to get an interesting insight into theissues of relations among (second-camp) FutureInternet research. As depicted in Fig. 1, the followingfive Future Internet variants shall be considered.

1.1 Internet-of-ThingsThis oldest variant relates to non-traditional

entities connected to the Internet (neither desktopcomputers nor servers nor humans using them): interms of ubiquitous computing, Internet-of-Thingselements can be separated into things worn(mobile devices, biosensors, wearables) and thingsencountered (smart labels, networked sensors, andembedded devices). While smart labels andparticularly RFIDs and corresponding infrastructuresattracted the majority of focus due to perceivedimmediate commercial benefit, they are only theforefront of a vast variety of Internet-of-Thingselements to join the net.

1.2 Internet-of-ServicesWhile the Internet has penetrated enterprises for

years, not much of our economy has migrated itscore of business onto the net. Web Services areoften a mere window onto the offerings of anenterprise that remain a hidden vast web of digitaland real assets. The Internet-of-Services goes asignificant step further towards providing a fully-fledged digital equivalent to offerings (i.e., servicesin the economic sense) of the enterprise. This meansthat human- and software-based entities alike canprocess these economic-services: negotiation,bidding and contracting, and global interweavingtowards composite services can now happen in theInternet, it becomes the full-fledged market place of the changing economy.

1.3 Internet-of-HumansThis one is probably the most forgotten variant

of the Future Internet. However, in order for theInternet-of-Things and Internet-of-Services tohappen on a global scale, Humans must beempowered with a service-independent ubiquitous


2/7

Figure 1: Research Challenges to Shape the Future Internet

access to the Future Internet. This access and useconcept has to hold and enforce the users wishesand needs in a fully trusted way independent of anyprovider but compatible with all the services andsmart things encountered on the globe.

1.4 Internet-of-Crowds

As digital devices and services interweavearound the globe, two global network trends comeinto focus: one of them is the combination withsocial networks. Both users and providers of services(and smart things) can draw considerable benefitfrom the inclusion of social networks of all kinds,and social networks, at a closer look, suggest awealth of connection points to Internet-of-Thingsand Internet-of-Service offerings.

1.5 Internet-of-CloudsThe second logical consequence of a globally

interwoven Internet-of-Services/Internet-of-Things

infrastructure is the use of Cloud Computing as anobvious platform of choice. We will briefly sketchthe reasons for which a highly dynamic infrastructure- and hence, the cloud - is indispensible as afoundation for both the Internet-of-Services andInternet-of-Things.

All five variants mentioned above are particularand self-contained enough to require specificconcepts and support (and thereby, dedicatedresearch). Therefore, we will dedicate a separatesection to each of them. On the other hand, theyoverlap considerably, e.g., regarding their need forprivacy, security, and trust concepts. This aspect will

be furthered where appropriate in the paper.

2 INTERNET OF THINGS (IOT)

The Internet-of-Things operates on a widespectrum of computing platforms, vastly different interms of size, mobility and usability. The lower endof this spectrum is marked by computers embeddedinto everyday objects and small sensor nodes with

limited processing capabilities. Most importantly, thereal power of the Internet-of-Things emerges fromthe cooperation of many devices. Hence,communication and federated computing arefundamental requirements.

Because the communication requirements of devices are as diverse, several different wirelessnetworking technologies are in place. To address thehigh degree of heterogeneity of platforms andnetworking technologies, we have created thecommunication middleware MundoCore [1]. It isbased on a microkernel design, supports dynamicreconfiguration, and provides a common set of APIs

for different programming languages on a wide rangeof different devices. The architectural modeladdresses the need for proper language bindings,different communication abstractions, peer-to-peeroverlays, different transport protocols, differentinvocation protocols, and automatic peer discovery.

We also observe today that Smart Items areevolving into Smart Products. While Smart Itemsshow a rather passive behavior, i.e., giving goodsidentity, monitoring, or logging, Smart Products havemore self-awareness and can actively interact withtheir surroundings. They are real-world objects,devices or software services bundled with knowledge

about themselves, others, and their embedding [2].They are motivated by the increased complexity of


3/7


4/7

how to best use the limited available input and outputcapabilities, e.g., by leveraging interaction devices intheir environment.

4.3 Voice-based InterfacesVoice-based Interfaces are an important concept

towards the invisible computer. Users may have theirhands busy while performing everyday tasks, thusseverely limiting their ability to interact withtraditional hands&eyes devices, i.e., mouse andkeyboard. Voice UIs (VUIs) are difficult to build dueto their transient and invisible nature. Unlike visualinterfaces, once the commands and actions have beencommunicated to the user, they are gone. Besideobjective limitations of a voice channel, humanfactors play a decisive role: auditive and orientationcapabilities, attention, clarity, diction, speed, andambient noise. These aspects can be grouped intotechnical challenges and audio inherent challenges

[10].In mobile environments, aspects such as

computational power and memory pose stronglimitations that cannot be handled by conventionalimplementations. To solve this, most of the researchutilizes a speech service running in the network, andemploying traditional client/server architectures [11].The fact that there may be multiple audio input andoutput devices, and multiple text-to-speech enginesand speech recognizers with different capabilities ina given environment remains untouched in most of the current techniques and is the focus of ourresearch towards a ubiquitous computing speech API.

4.4 Pen and Paper InteractionIn the domain of ubiquitous computing,

augmenting everyday objects by electronicfunctionality is a viable approach. Paper, aubiquitous medium, which is used for thousands of years is more powerful than computers in manyrespects [12]. To cite only some examples, paper ischeap, mobile and can be easily written and sketchedon.

Previously researched issues like, e.g., supportfor taking handwritten notes, annotating paperdocuments or mapping paper contents to digital

contents is rather well understood. But there stillexist fundamental challenges of bridging paper withthe Internet. This comprises first how people cancollaborate over the distance using real paper andsecond how digital feedback can be provided on thepaper medium.

Previous and ongoing work of our groupaddresses the first issue. We have developed novelinteraction techniques and implemented a systemprototype which allows users to collaborativelyannotate, link and tag documents, both in printed anddigital form [13]. Users can work locally with paperand share information over the Internet. Remoteinformation is displayed in several novelvisualizations for user-generated content, which is

created on paper. In order to integrate paper-basedinput and display-based output channels, wecontribute pen-based interaction techniques, whichclosely integrate paper with displays. This brings theintegration of paper with the Net a significant stepfurther.

The second issue, how to provide digitalfeedback, has initially been addressed using voice-based feedback or separate displays. Recentapproaches use mobile projectors to overlay visualcontents directly on paper. In our group, we arecurrently developing a software framework whichsupports application developers in flexibly providingdifferent types of feedback, using audio, mobilephones, nearby displays, and mobile projectors.

4.5 Secure Communication & Anchor of trustIn the Internet of Humans, an important aspect is

also how people can securely communicate with

smart environments. As part of the Mundo project,we have created the notion of the Minimal Entity(ME) and as prototypical implementation the TalkingAssistant device. The ME acts as the usersrepresentative in the digital world. It carries theusers digital identity and is able to performoperations such as authenticating the user to otherparties. The ME is designed as a secure terminal andthus allows to securely perform transactions -possibly with legal impact - in an Ubicomp world.We believe that interaction with the ubiquitousinfrastructure will mainly happen through a personaldevice, owned by each user, that may serve as an

anchor of trust. This is because it will represent theuser in the digital world and carry out transactions,sometimes with only implicit consent of the user.Hence, it is vital that a user is able to trust the devicecarrying out these actions. The easiest way to instillthis trust in users is to have them actually own thedevice instead of relying on a networked service,which may remain an abstract concept for manyusers.

5 INTERNET OF CROWDS (IOC)

The social networks of the communicating users

and information on their characteristics can be apowerful aid for any network operation. The socialservices have to be provided in a decentralizedfashion at the same time, since a reliable end-to-endconnectivity cannot be expected. Centralizedservices additionally represent a single point of failure and a possible threat to the privacy of users[14].

The initial challenge in the field is to implementa decentralized online social network providing bothavailability and a minimum level of privacy.Keeping all needed services and applications that areimplemented inside the social networking servicesavailable is a difficult challenge. Especiallyconsidering the fact that parts of the network may


5/7

temporarily be disconnected demands solutions forthe migration and load balancing of services andapplications in an unmanaged environment, usingconcepts of self-organization and distributed control.

Safebook [15] currently tackles the challenge of implementing online social networks without

centralized control, among only four predominantprojects [15-18]. However, none of the approachesproposed so far provides for the distributed executionof applications and added services. The question of how such services would have to be distributed,migrated, kept alive for each subnetwork with atemporary lack of connectivity to the rest of thesystem, and load-balancing between the participatingnodes in a disconnected partition additionallyremains completely unreflected upon.

Online social network technologies can alsoprovide numerous benefits for Teams Under Pressure.However, this application field poses additional

challenges. First Responders, especially in the caseof large-scale disasters tend to avoid both, novelcommunication technology and software systems andstill prefer their traditional workflow including theuse of pen and paper. We identified three mainchallenges, which need to be addressed in order toimprove the quality and speed of the decision-making process in large-scale catastrophes andimprove the acceptance of new technologies in thisdomain:

(i) Novel and non-disruptive interactionstrategies : Although there is a variety of softwarespecialized on nearly all important tasks of disaster

relief, pen and paper is not yet replaced and isunlikely to disappear in the next couple of years. Thechallenge is to support teams under pressure withnovel IT interaction mechanisms and aggregatedsensor information without forcing them to abandontheir traditional workflows.

(ii) Integration of sensors, communicationdevices and decision support systems : Commerciallyavailable closed solutions do not support incidents of organizationwide or nationwide scale. This requires aclean and smart interface between differentcategories of data sources and sinks, each one fromdifferent vendors and in different levels of

sophistication. In particular, sources and sinks mustbe able to interact and dynamically adapt in terms of prioritization, scaling and QoS. Due to fast changesin information needs and in network quality,adaptation must be possible at much higher ratesthan with common approaches.

(iii) Break the glass security : Imagine a largescale chemical disaster where perilous gas is about toenter the environment. It is good practice to give thefirst responders an emergency authorization, whichallows them, e.g., to enter every office/laboratory,destroy valuable assets, or flood a floor if necessary.Hence, under certain circumstances, it is possible tooverride all security restrictions [19]. Since more andmore systems are based on IT, the demand for an

electronic equivalent of an emergency hammer risesfrom day to day.

6 INTERNET OF CLOUDS

Two driving forces suggest Internet-based Cloud

Computing as a foundation for the Future Internet: i)resource-poor IoT devices must be surrounded bydata and processing capabilities of vastly varyingamount and origin, often owned by other partiesthan those (currently) hosting an IoT entity and ii)global and highly unpredictable demand for IoSservices cannot be reliably and cost-effectively metby an IoS provider, particularly by a new economyprovider.

While the Internet of Clouds will leverage off itspredecessors Concurrent and DistributedComputing and Grid Computing for much of itsstructural components, standards, and engineering

concepts, we strongly believe that it will inherit fromPeer-to-Peer computing with respect to highlyscalable basic mechanisms and from UbiquitousComputing with respect to Privacy, Security, andTrust (PST).

7 PRIVACY, SECURITY AND TRUST

PST can be seen as a key challenge of the FutureInternet, orthogonal to all the ingredients discussedup to now. Thereby, the Internet-of-Clouds is to beconsidered the final deathblow for long-establishedfoundations of IT security: enterprise firewalls

become questionable in view of interleavingenterprises, and classical security means (PKIs,certificates, End-to-End security concepts etc.) are of limited use in view of other Ends in End-to-Endcommunication is a priori not trustworthy. Thesetrends arose with the first four Future Internetcomponents discussed, but now with the Internet-of-Clouds even the entire computing andcommunication foundation becomes unreliable froma PST point of view. In the remainder, we will selecta few of the related issues for more detaileddiscussion, starting with security due to its longstanding history in classical IT (compared to trust

and privacy).

7.1 SecuritySecurity in the Future Internet is clearly beyond

confidentiality, integrity, and authentication, but(e.g.) also on the reliable enforcement of the interestof users and service providers. This requiresextended research on security policies and remotepolicy enforcement - especially consideringdistributed environments. Moreover, multilateralsecurity [20], i.e. providing mechanisms that allowfor balancing and negotiating between conflicting(security-relevant) interests are an important buildingblock for the Future Internet. Through multilateralsecurity approaches competing legal, organizational


6/7

and individual points of views can be taken intoaccount to design solutions that fit the needs of allinvolved parties [21]. Finally, the usability of security mechanisms, i.e., how appropriate are thosemechanisms for practical usage, becomes more andmore important, as IT gets more and more

interwoven with our daily life.

7.2 TrustTrust concepts become even more vital when

systems become more decentralized. With anincreasing number of service providers the customerhas the choice to select her preferred service providerout of many. However, especially a first timecustomer has little information about the quality of service that those providers offer as traditional hints,as they are known from the real world, are missing.Here, the concepts of trust and reputation have beenshown to be excellent concepts in order to support

the customers as well as the service providers. Inprevious and on-going work, we are developinginnovative concepts which are the basis forintegrating reputation and trust information in (semi-)automatic decision making processes. Thus, werecently developed a novel representation forBayesian trust models that can be adjusted to thecharacteristics of different application contexts andthat can easily be integrated in graphical userinterfaces [22]. Furthermore, we provide a newapproach to limit the impact of Sybil attacks [23], i.e.,an attacker tries to increase the influence of herratings being represented by a high number of

seemly independent entities.

7.3 PrivacyInformational Privacy is related to the persons

right to determine when, how and to what extentinformation about him or her is communicated toothers. The upcoming of the future Internet has adeep implication on how personal data is collected,handled and eventually used. In Ubicompenvironments, many seemingly invisible devicescapture information about users to identify theirdemands and to personalize services. Nevertheless,service providers operating within the EU need to

comply with the regulatory obligations of the EC Privacy and Electronic Communications Directive2002/58/EC and the EC Data Protection Directive1995/46/EC if personally identifiable data isprocessed. It is thus fundamental to identify theprivacy threats in the emerging Internet and designsuitable privacy-enhancing technologies [24, 25].Furthermore, it is necessary to analyze if the digitalidentifiers and identity management systems usednowadays are sufficient.

8 CONCLUSION

This article could only open the door to each of the Future Internet variants, to relevant researchtopics, and to interesting relationships between thesevariants.

In summary, the most substantial insight gainedby the authors over recent years is the fact thatdespite the broad spectrum of applications andchallenges that characterize the five Future Internetvariants, three core research areas turn out to be thepersistent backbone of Future Internet research (forwhat was called the camp two viewpoint of theFuture Internet):

1. Distributed Computing Mechanisms,Platforms, and Development Aids

2. Human-Computer-Interaction, and3. PST (Privacy, Security, and Trust).

Notably, these three areas were formerly ratherseparate. While the first and third became relatedsince several decades, it is only recently the HCI -with the Future-Internet relevant topics mentioned -started to team up with the other two fields. Thisobservations leads us towards a strong quest forinterdisciplinary research in the three fieldsmentioned - and (in the light of the Internet-of-Humans and -Crowds) towards a quest forinterdisciplinary research beyond Computer Science,i.e. with both Humanities and Economics.

ACKNOWLEDGEMENTS

This work was partly supported by the Center forAdvanced Security Research Darmstadt (CASED),by the SmartProducts project funded as part of theSeventh Framework Programme of the EU (grantnumber 231204), and by the Theseus Programme,funded by the German Federal Ministry of Economyand Technology under the promotional reference01MQ07012.


7/7

REFERENCES

[1] Aitenbichler, E., Kangasharju, J., Mhlhuser,M.: MundoCore: A Light-weight Infrastructurefor Pervasive Computing. PMC 3(4) (2007)332361

[2] Aitenbichler, E., Lyardet, F., Austaller, G.,Kangasharju, J., Mhlhuser, M.: EngineeringIntuitive and Self-Explanatory Smart Products.In: Proc. of the 22nd ACM Symposium onApplied Computing (SAC), ACM Press (2007)16321637

[3] The SmartProducts Consortium: ProactiveKnowledge for Smart Products.http://www.smartproducts-project.eu/ (2009) lastvisited: 11.11.2009.

[4] Theseus Pressebro: Theseus Programme.http://theseus-programm.de/ (2009) last visited:11.11.2009.

[5] Schmidt, W., Fleischmann, A., Gilbert, O.:Subjektorientiertes Geschftsprozessmanage-ment. HMD - Praxis der Wirtschaftsinformatik (266) (April 2009)

[6] Behring, A., Petter, A., Mhlhuser, M.: Rapidlymodifying multiple user interfaces of oneapplication. In: ICSOFT 2009, INSTICC Press(Jul 2009) 344347

[7] Behring, A., Heinrich, M., Winkler, M., Dargie,W.: Werkzeuguntersttzte Modellierung multi-modaler, adaptiver Benutzerschnittstellen. i-com6(3) (2007) 3136

[8] Behring, A., Petter, A., Mhlhuser, M.: A

domain specific language for multi user interfacedevelopment. In: Modellierung 2010. (2010) toappear

[9] Petter, A., Behring, A., Mhlhuser, M.:Constraint Solving in Model Transformations. InPaige, R.F., ed.: ICMT, Springer (2009)

[10] Schnelle, D.: Context Aware Voice UserInterfaces for Workflow Support. VDM VerlagDr. Mller (2008)

[11] Mhlhuser, M., Gurevych, I.: Handbook of Research on Ubiquitous Computing Technologyfor Real Time Enterprises. IGI InformationScience Reference (2008)

[12] Sellen, A.J., Harper, R.H.: The Myth of thePaperless Office. MIT Press, Cambridge, MA,USA (2003)

[13] Steimle, J., Brdiczka, O., Mhlhuser, M.:CoScribe: Integrating Paper and DigitalDocuments for Collaborative Learning. IEEETransactions on Learning Technologies 2(3)

(2009) 174188[14] Cutillo, L.A., Molva, R., Strufe, T.: Safebook: a

Privacy Preserving Online Social Network Leveraging on Real-Life Trust. IEEE Comm.Magazine (December 2009)

[15] Cutillo, L.A., Molva, R., Strufe, T.: Safebook:

Feasibility of Transitive Cooperation for Privacyon a Decentralized Social Network. In:WoWMoM. (2009)

[16] Baden, R., Bender, A., Starin, D., Spring, N.,Bhattacharjee, B.: Persona: An online socialnetwork with user-defined privacy. In: ACMSIGCOMM. (2009)

[17] Buchegger, S., Schiberg, D., Vu, L.H., Datta,A.: PeerSoN: P2P Social Networking. In: SocialNetwork Systems. (2009)

[18] Graffi, K., Mukherjee, P., Menges, B., Hartung,D., Kovacevic, A., Steinmetz, R.: PracticalSecurity in P2P-based Social Networks. In: LCN.

(2009)[19] Brucker, A.D., Petritsch, H., Weber, S.G.:

Attribute-based encryption with break-glass. In:WISTP. (2010) to appear

[20] Rannenberg, K.: Multilateral security a conceptand examples for balanced security. In:NSPW 00: Proceedings of the 2000 Workshopon New Security Paradigms, ACM (2000) 151162

[21] Weber, S.G.: Harnessing pseudonyms withimplicit attributes for privacy-respecting missionlog analysis. In: Proceedings of InternationalConference on Intelligent Networking and

Collaborative Systems (INCoS 2009), IEEEComputer Society (2009) 119 126

[22] Ries, S.: Extending Bayesian Trust ModelsRegarding Context-Dependence and UserFriendly Representation. In: Proc. of the 24thACM SAC. (2009)

[23] Ries, S., Aitenbichler, E.: Limiting SybilAttacks on Bayesian Trust Models in Open SOAEnvironments. In: Proc. of CPI-09. (2009)

[24] Martucci, L.A., Andersson, C., Fischer-Hbner,S.: Chameleon and the Identity-AnonymityParadox: Anonymity in Mobile Ad HocNetworks. In: Proc. of IWSEC, Information

Processing Society of Japan (IPSJ) (2006) 123134

[25] Martucci, L.A., Kohlweiss, M., Andersson, C.,Panchenko, A.: Self-Certified Sybil-FreePseudonyms. In: Proc. of WiSec08, ACM Press(2008) 154159

UBICC_Aitenbichler_477

Documents

Transcript of UBICC_Aitenbichler_477