© ISTOCK BORKA JERMAN-BLAŽIČ Digital Object Identifier 10.1109/MTS.2011.943306 Date of publication: 8 December 2011 IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 1932-4529/11/$26.00©2011IEEE | 39 T he Internet has transformed the lives of billions of people in areas as diverse as democracy, education, healthcare, entertainment, commerce, finance, and civil infrastructure. It has become the 21st century’s fundamental societal infrastructure, comparable to the railways of the 1800s and the roadways of the 1900s. The Internet and its associated services have helped transform the world economy and society, catalyzing new forms of communication, collaboration, creativity, and innovation. The Internet deeply affects human communication, and the way humans deal with information and knowledge. Statistics indicate that the Internet is still growing at exponential rates. According to the last report of the Task Force of the European Commission DG INFSO, Internet connectivity is expanding rapidly in geographical distribution and number of users [1]. Currently there are about 1.6 billion Internet users worldwide (from 360 million in 2000) and 4 billion mobile users (from 2.7 billion in 2006); 570 million Internet-enabled handheld devices are in use. The number of people who use mobile phones for web surfing has doubled since 2006. It is expected that in 2012 mobile and wireless users will outnumber wired ones. In parallel with user growth, stored information is growing as well. In 1998, Google indexed 26 million web-pages; in 2009 it indexed 1 trillion. There are 400 million web pages and 55 trillion links between these web pages. The Web is processing 100 billion clicks per day, and 2 million emails and 1 million instant messages per second. Video traffic over the Internet is growing by 60% every year and will be multiplied by 1000 over the next 5 to 8 years. Web 2.0 and social networks are attracting more than 125 million regular users within just 5 years of existence. The Internet is an indis- 40 | pensable part of most businesses with many business processes having been significantly automated by Internet technologies. The current Internet is the most important infrastructure of the digital society. It is also adapting itself with ad hoc technical solutions that help to meet the demands of users, devices, applications, and services, enabling human activities that were not foreseen in the Internet’s original design. The networking community is aware of the rising number of ad hoc solutions to technical problems, and has come to agree that these problems are of an architectural nature. A general redesign may be needed. The Internet community understands that the design of the “Future Internet” should enable the smooth evolution of the current IP network and should rely on the current practice of patches to overcome existing problems. It is also commonly understood that the structural and architectural problems of the current Internet cannot be solved without understanding how the Internet interacts with the rest of the world, including humans and machines. This article is based on work within the EU FP7 project Evolving Future Internet for European Leadership (EIFFEL) [2]. The project organizes semi-annual think-tank meetings where experts from all parts of the world debate the future of the Internet. Most of the identified agreements and disagreements regarding major problems of the current Internet are provided at the FIPEDIA site [3] maintained by the EIFFEL core team. This article introduces major findings that are presented in detail in the EIFFEL white papers on the future of the Internet. Network Community Debate The Internet network in use today is still based on the best-effort, point-to-point service model, well suited to applications between two endpoints that can tolerate occasional performance degradation. However, many newer applications do not easily tolerate performance degradation, and many applications involve multiple endpoints. This complicates any new design for the Internet. There are several major initiatives considering different approaches to meet these challenges. In the U.S., the National Science Foundation (NSF) NetS research program FIND [4] is the major long-term initiative. FIND encourages “clean slate process” research proposals in the broad area of network architecture, principles, and design of the Future Internet. The philosophy of the program is to enable a network design that is free from the current collective mindset about the constraints of the network. The NSF is considering the Network Science and Engineering Committee (NetSE) report published in mid-2009 [5], which recommends further R&D activities. GENI [6], another U.S. program, focused on a flexible and reconfigurable network of “test-bed” experimental facilities and projects. The EU through the FP7 program funds a wide range of research activities that relate to the Future Internet. A complete, up-to-date snapshot of all related European R&D activities in the area is difficult to provide. The Future Internet Assembly (FIA) was established in March 2008 in Bled, Slovenia. FIA is ensuring appropriate coverage of this large and challenging research domain that includes innovative research in the area of networking, experimental facilities and testing within the FIRE [7] program. Recently the initiative related to the Future Internet enterprise system – the project cluster FInES [8]– was added to the FIA program. In July 2009 the final report of the EU DG INFSO [1] Task Group on Interdisciplinary Research Activities for the Future Internet was published. This report identified the design, IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 implementation, testing and validation platforms as major research challenges for the EU. Cross-disciplinary research activities are an essential part of these platforms. Japan, Korea (KOREN) and India have set up similar initiatives; China has its own research initiative on the Future Internet: AsiaFI. Cooperation between this initiative and the EU FP7 projects recently have been set up. Future Internet discussions about Internet governance and business models are ongoing in other communities: international governmental and non-governmental organizations such as OECD [9], ITU [10], and UN-IGF [11]. In addition, the Internet Society and the Internet Corporation for Assigned Names and Numbers (ICANN) are developing position papers and projects on issues such as the Internet economy, Internet governance, and network neutrality. The recently expired contract between ICANN and the U.S. government is one step forward toward building up real internationally governed cooperation, and inclusion of civil society as its constitutional part. The architecture for the new Internet must be designed in a way that avoids predetermining the outcome of particular conflicts in the future marketplace. These conflicts should be allowed to play out inside the architecture after it is deployed. Articulating the grand challenges of the Future Internet and working towards solutions needs a wider debate as well as concrete work among a growing community of interdisciplinary researchers and major stakeholders. Different views exist with respect to what may be missing from the current architecture or why such concepts are missing. Some of the agreements achieved during the thinktank meetings are presented here. A full report is available in the EIFFEL white paper and on the FIPEDIA portal [12]. Structural and architectural problems of the current Internet cannot be solved without understanding how the system interacts with the rest of the world, including with humans and machines. Evolutionary Mechanisms According to EIFFEL, the Internet needs to be carefully observed before starting the new design. The evolution of the current Internet was compromised [13] because its architecture does not allow legitimate concerns to be expressed after its original design. As a result, users, providers, and business customers solve their problems in ad hoc ways, adding carbuncles that violate the original architecture. Then subsequent requirements are even more difficult to satisfy, because of all the feature interactions that are exceptions to the original architecture. The root of this problem lies deep in the processes used to design architectures and solutions. Currently, much emphasis is placed on the design phase of the architecture, with requirements phases and use case definitions, accompanied by processes of standardization. This inevitably leads to an emphasis on concerns that are important to the players who are deeply involved in this phase, while it neglects concerns of the actors entering the scene after the solution has been fixed. This Newtonian-Descartian concept of system design, relying on requirements and user case definition phases, assumes the ability to capture all relevant concerns and therefore resolve the most probable run-time problems at design time. The widening scope of the Internet beyond mere technology, and the increase in ad hoc solutions after the design of the original architecture bring this design process into question. Some IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 authors propose [14] a shift from a reductionist Newtonian-Descartian approach to Darwinian approaches [15], where the evolutionary kernel is a component that has been proven successful for multiple uses, so it may act as a platform for evolution [16]. Using this approach, design becomes the design process itself, i.e., a process in which concerns of actors are incorporated into the system at runtime, recognizing the inability to cater to all possible requirements during design time. However, this requires understanding what was good in the old design and what should be preserved in the new design. These considerations during the EIFFEL discussions of Internet evolution can be summarized as follows: ■ ■ The Internet, like any largescale system, needs to evolve. This evolution is particularly important considering the evolution of society due to the Internet. We need to understand the dynamics in play and devise an architecture that is suited for these dynamics to commence during runtime. The scope of the dynamics affecting change of the Internet is widening. The Internet has become more than a technical artifact – it has transformed from a network for geeks to a crucial infrastructure for society and business. The virtual and real worlds abide by similar rules, including respect for human rights. Hence, the question of | 41 The Internet, like any large-scale system, needs to evolve. ■ ■ the evolution of the Internet is no longer merely a technical question. The Internet’s evolutionary speed is increasing with advances of technology. For instance, memory is becoming so cheap, in particular when compared to the formative years of the Internet, that solutions for caching vast amounts of content locally is likely to transform the way users and customers deal with content. In that context, another problem needs immediate attention: consumption of energy related to increased memory use and processing power. The Internet has become another area for energy savings and low-energy consumption devices for infrastructure and applications. terests achieve dominance in the Future Internet’s development. The scenarios are presented in Fig. 1, and illustrate possible designs around two axes that point to different outcomes. The vertical axis designates whether the Future Internet will remain true to the old Open Internet Model (generative, rather than reductive). The horizontal axis designates whether it will become distributed and decentralized (rather than under the command and control of regimes). These axes represent two key areas of external world conflicts (social and economic) between Internet stakeholders, impacting the deployed Internet reality. Together they delineate four quadrants, each of which can be described as an illustrative scenario. The quadrants between the two axes reflect the effects of misalignments in the incentives of the Future Internet. The main incentives and what drives the stakeholders’ actions are presented in Table I. The columns catagorize the major Coping with the changes and with the research agenda were issues discussed and worked on within the think-tank meetings. It was obvious that the old models of development that were mostly based on an engineering approach are not very sufficient. The complexity of the system and its interrelation with society require scientific methods based on facts and measurements to understand and react to the global picture and to the expected evolution. Internet-Society Scenarios In 2009, the Internet Society (ISOC) [17] provided EIFFEL with an illustration of possible forms of evolution based on an “Internet Futures Scenarios” exercise. This exercise produced four visions of the future in which different stakeholders’ in- Internet Futures Scenarios Will the world embrace or resist the open Internet model? What model will be more successful? Command and control? Or, Distributeed and Decentralized? Generative Porous Garden Scenario App Stores and Closed Devices Client Capture Innovation Competition Sticky Services Common Pool Scenario Interoperability Open Access Internet Model Trust Command and Control y av He ive n tio ula g Re rds Ex Content Control Moats and Drawbridges Scenario Re Sta gu nd s clu Censorship ht en Sta Tight Regulation Lig Op a nd lat ion ard Decentralized and Distributed s Security Risk-Mitigation Control-Through-Rules Balkanization No Consensus Multiple Roots Reductive Boutique Networks Scenario Fig. 1. ISOC Future-Internet scenarios. 42 | IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 stakeholders in terms of what they fear and what they are greedy for. Among the four quadrants, the Common Pool quadrant is the most positive with respect to the “generative” and the “distributed and decentralized” properties of the Future Internet. In the Common Pool scenario, all the resources that are part of the Future Internet are made available to the overall community. This scenario can be considered as the ideal for which Internet development and deployment has always striven, though never perfectly achieved. This scenario provides the maximum flexibility, deployment, innovation, and opportunities to all stakeholders. Technologies are planned to be built out “horizontally,” rather than in fullservice verticals. This quadrant is named the Common Pool to suggest a future where information service and application gardens will not be completely walled, but will still be somewhat restricted to particular channels. The Porous Garden scenario is designed around stakeholders’ incentives for increased control over business and revenue. In this quadrant, the application and serviceprovider stakeholders are leading the evolution with architectures that feature increased command and control in vertical services. In this vision of the Future Internet, the networks remain global, but access to content and services is tied to specific networks and associated information appliances. The root of the problem lies deep in the processes used to design architectures and solutions. Financial incentives for content producers and software developers would result in continued innovation within the appliance-based model, but network operators will be constrained to evolve their services to support appliances and not to support general Internet services. Consumers would have to purchase multiple appliances and associated subscriptions to avail themselves of the full range of innovation on the network. This scenario reflects the general misalignment between the incentives of the content producers and those of end users, as well as (ultimately) network operators. The Boutique Networks quadrant represents a scenario where the networks, not the applications, are expected to be the dominant drivers of the future. This quadrant features intense network specialization. This Future Internet is not expected to be a single, general network, but rather is composed of specialized networks that provide “boutique” services. This scenario envisions a future in which political, regional, and large enterprise interests fail to optimize on the social and economic potential of a shared, global set of richly connected networks (today’s Internet). Instead this scenario reflects the outcome of parties optimizing control in small sectors (political and otherwise). While these balkanized networks continue to leverage the benefits of existing Internet standards, they do not collectively provide the basis for generalized application and service development and deployment. In that regard, this quadrant represents the opposite of the Porous Garden quadrant, in that it is network-development interests that are expected to dominate. The Moats and Drawbridge quadrant reflects a future where stakeholders seek tighter command and control and where more reductive, constrained network environments are expected to prevail. The increased (perceived) need to provide security and consistent environments through command and control operations and closed development practices means this scenario is drawing an Internet that is heavily centralized and dominated by a small number of big players who create their own rules in a few “big-boys” clubs. In this scenario, strong regulations can be expected as governments seek to impose some public-interest obligations on the industry, as the users’ interests Table I Internet-Stakeholder Incentives End users Service/Content Providers Network Providers Governments Fear Privacy, Overcharge, Pricing Unfairness Losing their market share because of competition, innovation and regulation Losing their market share because of competition, innovation and regulation Security, Politics IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 Greed Cheap/free services, Cheap/free content, Cheap/free network access Market dominance, Monitor users’ behaviours, Tiered Services (no network neutrality) Market dominance, Monitor users’ behaviours, Tiered Services (no network neutrality) Control of Content, Control of Access to Content, Monitoring (Spying) on Users | 43 The nature and impact of choosing a particular evolutionary path for the Future Internet needs to be made explicit. and incentives will not be natively supported. Control could extend to limiting the equipment that could connect to the network. Content could be proprietary and protected by strong intellectual property rights. This quadrant shows the highest barrier for the entry of new applications, networks, services, and end users. These scenarios illustrate a vision based on discussions at the EIFFEL think-tank meetings. From an economic point of view (the most democratic at least), the perfect scenario would be the one that exists in a perfect market, the one for which the following are present: perfect market information, no participant with sufficient market power to set prices, no barriers to entry or exit, and equal access to production technology. An immediate result of perfect market information should also be perfect pricing mechanisms. Doubtlessly, the Common Pool scenario is the one that mostly resembles the perfect market and is close to the current “ideal.” Is it possible to influence the design of the Future Internet so that it can naturally stabilize itself to this quadrant? The answer is probably “yes,” and the way to achieve this is via a design for change. Special attention has to be paid to the information aspect, which in a networkbased system translates to network measurements and monitoring. Business and Social Demands The particular technical approach to the current Internet created business structures that evolved around it, such as the transit and peering relationships of autonomous systems. Any fundamental change to the current Internet will undoubt44 | edly have an impact on these existing business structures. Too radical a change will cause problems in adopting the change – and so delay the advancement. Hence, technical and economic migration strategies from here to there are crucial when targeting a wide adoption of proposed Internet changes. For this reason, grand challenges in economics must be addressed as well as technical challenges. This needs to start with gathering the right audience for this work and it needs to be driven by a clear emphasis on solving concrete problems. The Internet in its early days was a vehicle for email-based communication – reasonably immediate, but not requiring real-time end-toend connectivity. It was also a vehicle for simply improving the information flow between parties who would have otherwise exchanged the information, but more slowly or in smaller quantities. Along with this beneficial relationship with social structures, the Internet brought opportunities for antisocial misuse. The demands of improving social communication and reflecting social structure are growing, but at the same time issues of privacy and safety in a completely connected world are becoming increasingly problematic. Birth/death records, medical records, banking records and so forth were kept long before there was an Internet, but the Internet made them more accessible to legitimate users, and also made it simpler for malefactors to get at the records. As the information infrastructure became increasingly integrated into, and critical for, our society, attacking the Internet infrastructure also became increasingly worthwhile. The Internet is a reflection of society, but this reflection is always partial. As such, it will evolve to provide increasing aspects of social infrastructure requirements, but it is difficult to accurately predict the next steps. In fact, some of the innovation is likely to come from outside the Internet. Who would have thought that carrying around small wireless cell phones with tiny keyboards would turn into instant messaging and from there make the leap to the Internet and soon into all the different modes of social networking with the user-designed Web 2.0 tools? Innovation will always have an element of surprise for some stakeholders. One of the interesting social challenges the user community is facing today is information overload. There is too much information. There are too many services that want to claim users’ trust. There are too many options and too many individuals who want attention. The challenge will be to evolve approaches that reflect human and social approaches to dealing with information overload. This is already happening in what are probably simple ways in social networking contexts. Users group their friends, create channels for topics, create wikis, and follow their friends via Twitter. The world is being clustered, but this can be understood as the early stage of the social change induced by the Internet. Newspapers were a mechanism for filtering, organizing, and limiting information that would otherwise overwhelm the reading audience. With the demise of newspapers, what elements of the almost infinite flow of bits will bring an order that is reflective of the human mind and human social structure? In the long run, will that also allow each human to retain a somewhat personal view in large social structures? How will individuality and privacy be protected? Another question relates to the impact of governance on the IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011 Internet and vice versa: what is the impact of the Internet on governance? It is clear that the lowcost and pervasive availability of a uniform communications substrate has had a significant impact on the global society that is becoming the digital society. Historically, explorers circled the world and laid claim to other lands, thus beginning political and economic connectedness around the globe. The presence of the Internet has qualitatively changed the nature and degree of that connectedness. In the current economic and political situation, no country can make decisions that will have only a local effect. There is no more pure isolation. Given that, the relationship between the Internet and governance is becoming increasingly important. And perhaps even more importantly, the Internet may change forever the governance of, by, or for the people. Blogging and cell-phone cameras that can transmit photos are having profound effects on the capability of individuals to constrain their governments. This is likely to have an impact, for example, on regulation when considering the growing role of end users in the participation of the Internet. End users may grow into an essential part of the Future Internet, moving away from their current role of mere consumers into a more active role in governance. How this will affect ways to regulate certain parts of the Internet will be important to understand. The current Internet is the most important infrastructure of the digital society. taining levels of performance that render it fit for purpose. Hence, the research emphasis should move from design time to a largely runtime model for resolving potential conflicts. Some of the most important aspects as identified by the EIFFEL Think Tank are: ■ ■ Steps Forward The most important observation of the EIFFEL Think Tank is that future architecture should not be a balance at design-time towards the wanted world. Instead, a minimum substrate should be designed that gives the Internet flexibility to behave in different ways at different times and in different places depending on the outcome of market selection and social regulation mechanisms [18], [19], while reIEEE TECHNOLOGY AND SOCIETY MAGAZINE ■ | Resilience, failure tracking, and management: The Internet’s distributed design is popularly renowned for its robustness to failure. Indeed failures often do heal automatically, but not quickly. The result is an increasingly unreliable service. Also, many failures are not amenable to automatic solution, being due to human errors in configuration and so forth. It is generally believed that today’s Internet does not have effective solutions to these problems. Availability and robustness against attack: The Internet is being used repeatedly by malware to attack both delivered services and the Internet’s infrastructure. Solutions to these problems often block innovative legitimate uses of the Internet as well as illegitimate ones, effectively slowing down the Internet’s evolutionary progress. Proper architectural support to address the root means of these attacks is needed, but there is no consensus about the partial solutions that have been proposed so far. Information security scalability: The state of the art in information-security techniques is sufficiently robust to ensure any form of se- WINTER 2011 ■ ■ curity, except that the techniques do not scale to global proportions in nonhierarchical groups. Another aspect of information security is information accountability. While the Internet can cause information to be shared or not, once it has been shared at all, any control is essentially lost with regard to further sharing. Exposure is dependent on some vague sense of trust in those with whom we have shared. Resource accountability: The current Internet architecture allows everyone to use any resource anywhere on the Internet to the extent that they want. However, at present, network operators are deploying boxes to limit or block communication with certain users or with certain applications. Even if Internet networks were trying to share capacity without making judgements about content, the architecture does not reveal the information they need to make other networks and their users accountable when they overuse stretched resources. The consequent inability to properly limit free riding (or to deliberately allow it) leads to uncertainty over whether capacity investments can be recouped, which in turn negatively affects the whole value chain of the Internet. Network-application coordination: Over the years, the application programming interfaces (APIs) at the top of the TCP/IP protocol suite have become ossified and | 45 stale. More importantly, they have become almost impenetrable. In the downward direction, middle boxes (e.g., firewalls and network address translators) only recognize the protocols that existed when they were deployed. So they block out all attempts by applications to use new APIs to new lower-layer protocols and services. In the upward direction, applications cannot find out about the network or their paths through the network in order to create richer services themselves – services that could exploit a more accurate knowledge of network topology, network failures, and traffic characteristics. ■ Scaling for more extreme dynamics: The dynamic range of the current Internet architecture is hitting its limits. Increasingly, the inter-domain routing system cannot converge quickly enough following a change, leaving longer periods of disconnection. More sites are connecting to the Internet through multiple links to improve resilience, but the inter-domain routing system is designed so it then has to treat these sites as distinct networks rather than as stubs of a single-provider network. This makes the routing system appear much larger without the Internet growing at all. Also, the Internet’s congestion-control mechanisms have hit the end of their dynamic range, since higher bit-rates require higher accelerations to reach them. The nature and impact of choosing a particular evolutionary path for the Future Internet needs to 46 | be made explicit, and needs to be understood by all the stakeholders. There should be clear technical, ethical, legal, and business reasons for making any choice. Since such choices are inherently constraining, the establishment of an orthodoxy that results from making a constricting decision must be balanced by inviting challenges and weighing evidence. It is crucial to pay attention both to the current problems and to the evolutionary mechanisms of the Internet. Those mechanisms will determine how evolution progresses, and if it progresses at all. Decisions made at this point must remain relevant and fresh for at least as long as the current Internet has proved valuable, in a world in which Moore’s law continues to apply. The investment of time and effort in widespread changes to the whole system will not occur unless such changes deliver results quickly enough for practical cost recovery. The changes must also continue to give returns over many decades in a constantly evolving technological, economic, and social environment. The EIFFEL think-tank intends to stimulate discussion on the major points of why and how the world will be going about the Future Internet. Author Information The author is with the Jožef Stefan Institute and Faculty of Economics, University of Ljubljana, Slovenia, and with the Computer and Systems Science Department, University of Stockholm, Sweden. Email: borka@e5.ijs.si. Acknowledgment This article is a result of the work of the think-tank of the EU funded FP7 project EIFFEL. The contributions of all the members and caretakers of the project are appreciated. References [1] Draft Report of the Task Force Interdisciplinary Research Activities Applicable to the Future Internet, version 4.1; http://forum.future. Internet.eu, accessed July 13, 2009. [2] EU FP7 Project EIFFEL; www.fp7-eiffel. eu, accessed 6.1.2011. [3] FIPEDIA; www.fipedia.org, accessed Dec. 12, 2010. [4] FIND (Future Internet Design). www.netsfind.net, accessed Sept. 8, 2009. [5] Network Science and Engineering Council, “Network Science and Engineering (NetSE) Research Agenda;” http://www.cra.org/ccc/ docs/NetSE-Research-Agenda.pdf, accessed Sept. 8, 2009. [6] GENI; www.geni.net, accessed Dec. 10, 2010. [7] FIRE Future Internet; www.future-Internet.eu, accessed Jan. 10, 2009. [8] Future Internet Enterprise System (FInES) Cluster, Pos. pap. vers. 1; accessed May 15, 2009, www.cordis.eu. [9] OECD, “The Future of the Internet Economy OECD Ministerial Meeting Report;” www.oeced.org/futureInternet.pdf. [10] ITU Technology Watch, Rep. 10. [11] Forum (IGF) - The First Two Years’ Report; www.intgovforum.org/, Accessed Dec. 5, 2009. [12] D. Trossen, Ed, “Starting the discussion,” white pap., EIFFEL think tank, July 2009; http://www.eiffel-thinktank.eu, accessed Dec, 12, 2009. [13] B. Jerman-Blažič “The future of the Internet: Tussles and challenges in the evolution pat as identified by EIFFEL think tank V: BERNTZEN, Lasse (ur.),” presented at the Fourth Int. Conf. Digital Society, Feb. 10–16, 2010 (St. Maarten, Netherlands Antilles). In Proc. ICDS 2010, pp. 25–30. [14] R. Hollingsworth and K. Müller, “Transforming socio-economics with a new epistemology,” Socio-Economic Rev., vol. 6, pp. 395–426, 2008. [15] C. Dovrolis, “What would Darwin think about clean-slate architectures?” ACM SIGCOMM Computer Commun. Rev., vol. 38, no. 1, pp. 29–34, 2008. [16] D. Trossen, Ed., “The core-edge story,” white pap., Core-Edge Dynamics Work. Group, Communications Futures Prog., M.I.T., 2005. [17] Internet Society; www.isoc.org, accessed Jan. 6, 2011. www.itu.ch, www.itu.int/dms_ pub/itu/oth/...T230100000A0001PDFE.pdf., accessed Sept. 5, 2009 and Apr. 5, 2010. [18] D. Clark, K. Sollins, J. Wroclawsk., and B. Braden, “Tussle in cyberspace: defining tomorrow’s Internet,” IEEE/ACM Trans. Networking, vol. 13, no. 3, pp. 462–475, June, 2005. [19] D. Clark and M. Blumenthal, “The end-toend argument and application design: The role trust,” presented at Conf. on Communication, Information, and Internet Policy (TPRC), 2007. IEEE TECHNOLOGY AND SOCIETY MAGAZINE | WINTER 2011