vol1_gaps

Upload: vdboomen

Post on 09-Apr-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/8/2019 vol1_gaps

    1/37

    Journal of Network Theory Spring 2006 1

    JJoouurrnnaallof

    NNeettwwoorrkk TThheeoorryy

    Spring 2006

    IINNSSTTIITTUUTTEE OOFFMMEEDDIIAA&&RREE//PPRREESSEENNTTAATTIIOONNFACULTY OF ARTS OF THE UNIVERSITY OF UTRECHT

    UTRECHT,NETHERLANDS

  • 8/8/2019 vol1_gaps

    2/37

    Journal of Network Theory Spring 2006 2

    The Journal of Network Theory is apublication of the Institute of Mediaand Re/presentation of the Faculty ofArts of the University of Utrecht.

    Copyright 2006 Journal of NetworkTheory

    All rights reserved. Printed in theNetherlands on acid-free paper.

    ISSN: 2109-1109

    www.networkjournal.nl

    EDITORIAL BOARD

    Thomas Nachreiner Monica NegriF040923 F050157

    Shane Pulmano Loes Vollenbroek0313912 0112828

    The Journal of Network Theorywishes to thank ProfessorMarianne van den Boomen forher contribution.

  • 8/8/2019 vol1_gaps

    3/37

    Journal of Network Theory Spring 2006 3

    GAPS, DIVISIONS AND BOUNDARIES

    Table of Contents

    1. Introduction: Gaps, Divisions and Boundaries .......................................................4Shifting Referentiality in the Use of Spatial Metaphors.................................................4

    Introduction...............................................................................................................4Why gaps, divisions and boundaries?.......................................................................5Invisibility ..................................................................................................................6Oscillation.................................................................................................................8Simultaneously metaphoric and material...................................................................9Concluding remarks................................................................................................10

    2. Finding or Choosing Boundaries?.........................................................................12Exploring the position of the researcher in ANT and SNA ..........................................12

    3. Continuity on the Web ............................................................................................13Beyond Cinematographic Boundaries........................................................................13

    From the Matrix to the Network...............................................................................14The Continuity.........................................................................................................14The Boundary.........................................................................................................16The Gap..................................................................................................................17The Network requires the Matrix .............................................................................19

    4. Domain Name System.............................................................................................21Protocol, Governance and Divisions ..........................................................................21

    Introduction.............................................................................................................21Protocol, Governance and Divisions .......................................................................22Protocological Reality on the Net ............................................................................23Internet Governance: Political Conflicts and Technological Means .........................27Conclusion..............................................................................................................30Bibliography............................................................................................................31

    5. The Last Day of Mankind ........................................................................................35Or an evening with the network society......................................................................35

  • 8/8/2019 vol1_gaps

    4/37

    Journal of Network Theory Spring 2006 4

    1. Introduction: Gaps, Divisions and Boundaries

    Shifting Referentiality in the Use of Spatial Metaphors

    By Shane Pulmano

    Abstract:In this article functions as an introduction to this issues theme and the subsequentarticles. The importance of making place for the discussion of the spatial constructsgaps, divis33ns and boundaries is argued. The article portrays how these concepts aremarginal in the discussion of networks. It is argued that these concepts are either explicitor implicit in network themed articles, but that their function remains invisible and un-debated. Also, their referentiality poses a particular problem, namely that writers areoften unconcise whether the concepts are utilized as metaphors or that they refer tosome substantive material reality. In some cases, a certain oscillation of the reference

    takes place; the reference shifts from a metaphoric to a substantive function and viceversa. In other cases, the reference is both a metaphoric and material.

    Introduction

    In 1995 sociologist Ilana Silbercommented that within the last decade,specifically, sociological theory hasbeen marked by the increasing currencyof spatial (quasi-geographical) imagesand metaphors such as, most

    commonly, field(s), space, action-space, boundaries (1995, 1). Thisdevelopment, she states, has remaineda rather implicit and underconceptualized feature of contemporarysociological thinking (Silber,1995, 1).More than a decade later, as sociologyhas adopted a new metaphor foranalysing society, the omnipotentnetwork, this marginalization of spatialmetaphors is still an issue. Sociology,however, is not the only approach that

    has opened its arms to the newmetaphor. Contemporary media studies,feminist theory and information studiesare also among those applying this newterm. In general, writers from theseapproaches advocate the use of thenotion network to explain certainprocesses or phenomena the sum ofwhich from hence will be referred to as

    network theory - look at the networkscore and its expansions, rather thanthe peripheral areas. This journal issuewill explore the conceptual outskirts,focusing mainly on the concepts of gap,divisions and boundaries which areargued marginal concepts withinnetwork theory. This article introduces

    this issues theme and argues theimportance of making place for thediscussion of the spatial constructsgaps, divisionsand boundaries. It will bemade clear that these concepts areeither explicit or implicit in networkthemed articles, but that their functionremains invisible and un-debated. Also,their referentiality poses a particularproblem, namely that writers are oftenunconcise whether the concepts areutilized as metaphors or that they refer

    to some substantive material reality. Insome cases, a certain oscillation of thereference takes place; the referenceshifts from a metaphoric to a substantivefunction and vice versa. In other cases,the reference is both a metaphoric andmaterial.

    In the following paragraphs, themarginalization of and the function of a

  • 8/8/2019 vol1_gaps

    5/37

    Journal of Network Theory Spring 2006 5

    discussion of these concepts within theacademic approach will be presented.Next, the invisibility of these conceptswill be discussed followed by adescription of the ossilation ofreferentiality. To make these statements

    concrete, some paragraphs will bedevoted to analysis of four differentapproaches to network analysis.Physicist Albert Lazslo Barabasis scalefree network and his notion of continentsand islands represent the computationalapproach towards analyzing networksand will a function as an example of theinvisibility of these concepts. LiliaEfimova and Stephanie Hendricks ontheir observations on weblog boundariesemploy social network analysis act as

    an example of how the concepts areexplicit but whose reference seem toslip from one meaning to another.Alexander Galloways approach ofanalyzing networks via codes andtopology will also be an example of thisoscillation, while the work of sociologistManuel Castells and Dutchcommunication researcher Jan Van Dijkwill be examined for the posibilty ofsimultaneous function in their use ofthese concepts. These examples will be

    followed by an introduction of thearticles in this issue. The function andthe contribution of each article will bebriefly examined. But first, anexplanation of the spatial concepts isneeded.

    Why gaps, divisions and boundaries?

    This relevance of discussing theseconcepts lies in the fact that manyauthors advocating the use of concept

    network in their works indicate orloosely refer to them but hardly reallydiscuss them. If a network is a conceptof systems within systems (Van Dijk,2005, 2), then there must be somedivisions made by researcherssomewhere in order to observe thatthere are different systems. Two thingsare happening to these metaphors in

    use. First they are rendered invisibleand in most cases taken for granted.This is the prime reason why are theynot discussed or reflected upon.Second, their reference is of unclearnature. Are these concepts used as

    conceptual metaphors or do they referto more substantial objects orsituations?

    In the first case, invisibility meansunnoticed penetration, which in turngives the concept a pervasive character.Invisibility allows these concepts to seepinto everything and shape currentthought without being questioned. Silberexplains this as follows:

    on the one hand, they tend to

    be much more abstract than suchmaster metaphors as organism,text, theater or even variouseconomic metaphors such asmarket or capital; on the otherhand, they are deeply intwinedwith and thus harder to distinguishfrom their equivalents in ordinarylanguage (for example, onesfields of interest, giving spaceto someone, respecting onesterritory, transgressing

    boundaries, and so forth) and fromthe many spatial and orientationalcategories (for example, high/low;horizontal/vertical; open/closed;internal/external; and so on) thatinform our taken-for-granted, dailylexicon. Spatial metaphors, as aresult, are less visible and catchybut also potentially more pervasivethan most metaphors hithertoallowed to shape sociologicalwriting

    Silber, 1995, 2

    These concepts shape the path ofacademic discussion without beingacknowledged. For the case of a goodtextual account in the Latourian sense,all relations it must be taken inconsideration. For, to truly understandwhat we are claiming, we must be able

  • 8/8/2019 vol1_gaps

    6/37

  • 8/8/2019 vol1_gaps

    7/37

    Journal of Network Theory Spring 2006 7

    The spatial concepts of gaps,divisions and boundaries are implicit inhis work. In his reference to cores (orhubs) alludes to the presence of aperiphery and outskirts. For somethingto be seen as a core, it must be different

    from to its surroundings, i.e. it has aconcentration of nodes and connectionsthat are far higher. This immediatelyputs the surrounding nodes in peripheralor border position. This notion, that thefurther away from the core, the less thenodes and the less the connection,alludes to the idea that at some point orplace the connections end and nodesare no more. This same idea is echoedin his notion of continents and islands.Real life continents and islands have

    boundaries, i.e. the point where landstops and water starts. Also, continentsand islands are separated from eachother by a body of water. Without these

    boundaries and divisions, there wouldbe no difference in whether a piece ofland is a continent or an island. Therewould merely be one vast land. Thenotion of continent and island would nothave even existed if there was nothing

    to separate these two pieces of land.The same is the case with a core. Acentral core is not a core if there isnothing surrounding it. But vice versa, aperiphery only emerges with a center inposition. This particular issue is whatmakes reflection on the functions ofthese spatial constructs necessary.Academic discussion focusing mainly onconcepts such as core or continentwithout giving proper attention to theoutskirts is like viewing holiday pictures

    with an apparatus that only allows a onedimensional view. Reflection on borderconcepts gives extra dimension to thediscussion.

    Figure 1. Barabasis continents and islands. Source: Barabasi, 2002, 2

    An example of an explicit mentioning ofspatial concepts is the work of LiliaEfimova and Stephanie Hendrick. Thepurpose of that particular paper is togain insight into methods of finding lifebetween buildings: virtual settlementswhere weblog communities may resideand to address research challenges to

    locating a community when there is nosingle shared space to start an analysisfrom (Efimova & Hendrick, 2005, 3).They argue that while some bloggercommunal spaces may have visibleboundaries more often indicators of acommunity are more difficult todistinguish (2005, 3). They note that

  • 8/8/2019 vol1_gaps

    8/37

    Journal of Network Theory Spring 2006 8

    community participation in weblognetworks is not located in one place, butdistributed both on individual weblogs,as well as in the space between thepersonal weblogs. (Efimova & Hendrick,2005, 3).

    Efimova and Hendrick speak of

    communal practices they refer toreprocipal linking from one blog toanother. From this reprocipal linkingdoes the space between the weblogsemerge which they see like the thespace between buildings full of actionbecause blogging is similar to lifebetween buildings in a real city, i.e.passing vehicles manned by individuals,pedestrians walking, shopping etc(Efimova & Hendrick, 2005, 7).

    The space between weblogs(the

    divisions: that what separates but alsoconnects weblogs) seems to refer to asubstantive development that isoccuring on the Internet. With it thewriters wanted to indicate thatcommunal life on the Net emerges fromreprocipal linking. However, unlikeactivity happening in the space betweenreal buildings, where people walk, talkpublic transport, etc, which are visiblebetween buildings, the interactionsbetween weblogs are not. All results of

    the reprocipal linking is found on theweblogs and not in between. For, unlikelife between real buildings, actionsoutside the weblogs is made invisible tothe weblog user. One does not see theactions by the computer nor thesoftware. Since, it is the nature of themedium to hide the processes thatoccur when links between websites aremade, life signs then, are only visibleON the websites. The list of artifactsmade by the researchers on page 8 and

    9, about the meme paths, readingpatterns, linking patterns, weblogconversations, tribe marks etc, aregenerated by those elements visible onthe weblog. The road or space inbetween is not visible. This means thatresults of these life signs refer to aconclusion taking place outside the Net.The spatial construction here refers tosomething mainly metaphorical.

    The journey itself, although short, fromone weblog to another is that whatdivides or separates one weblog toanother. For although many things arehappening when the weblog moves fromon weblog to another via links, these are

    hidden to the user and actually saynothing about communal practices.Because nothing is visible to theblogger, it may even be argued that thisshort time space between weblogs areexperiental gaps because nothingoccurs for the blogger (aside from theblinking loading balk at the bottom of thescreen). This use of spatial constructsas argued above is symbolic. It is notlike the weblogs are actually sitting orstanding beside each other in virtual

    community. The link, that is purelytechnical not only acts as the connectingentity, but also refers to the divisionbetween the weblogs. The technicalaspect is what strange enough, makesthis use of spatial concepts substantive.The materiality in this case consists ofprotocols and codes that active before,during and after the journey from oneweblog to another. These codes andprotocol that jump into place once theweblog user clicks on a link. The

    reference of the metaphor spatialdivisions in this case is a conceptualimage, the relations that the researchersare refering to are observed viasubstantial means, i.e. technical linksmade in codes and protocol. This is astrange development, because withoutreflected upon, the observations haveshifted from substantial materialobservation to metaphorical use.

    OscillationThis slipping is also visible in Barabasiswork is starting point of material linksmade by technical hubs and nodes,terms which he uses as metaphors todescribe the processes occuring. Theuse of continents and islands aremetaphorical also and thus so is theimplicit presence of boundaries and

  • 8/8/2019 vol1_gaps

    9/37

    Journal of Network Theory Spring 2006 9

    divisions. This metaphoric use becomesproblematic when asked to specify whatdoes the metaporic gap or void thatseparates these continents from islands(see figure 1). Although Barabasihimself does not make this clear, onecan conclude that this gap or division

    between continent and islands is theabsence of electronic connectionsbetween nodes. The hubs and nodes ofone center do not or are limitedlyconnected substantively to the other.What results is a seeming slipping fromone frame of reference to another.

    Alexander Galloways work ontopology and protocol also displays thisoscillation. For Galloway, the core ofnetworked computing is the concept ofprotocol. A computer protocol is a set of

    recommendations and rules that outlinespecific technical standards (Galloway,2004, 6). Shared protocols are whatdefines the landscape of the network.Without a shared protocol, there is notnetwork (Galloway, 2004, 12).Galloways account studies the materialsubstrate of media (Galloway, 2004,16). His focus on substantive elementsof the medium the Internet leads theobserver to conclude that in relation tothe concepts of gaps, divisions and

    boundaries, he sees them as very muchmaterial. For Galloway, the Internet isclearly not a free and uncontrollableenvironment. He argues that once rootservers are taken out, that wholeportions of the Internet immediatelycease to exist (Galloway, 2004, 10).Since the root servers are at the top,they have ultimate control over theexistence (but not necessarily thecontent) of each lesser branch. Withoutthe foundational support of the root

    servers, all lesser branches of the DNSnetwork become unusable. Such areality should shatter our image of theInternet as a vast, uncontrollablemeshwork (Galloway, 2004, 10).Obviously, some critical notes may bemade about this hierarchical thinking.First of all, the idea that the Internet ishierarchically manned seeminglyundermines the very notion of

    distributed network that he proposes toadapt in the first place. Unlike in acentralized network, in a distributednetwork even if one portion isdislocated, other parts may continue tothrive because the power or energy isdistributed among the nodes. Belief in a

    hierarchical structure does not supportthis thought, because arguing that oneroot server has the power to sustain allelse, is turning the root server into acentral hub, thus a centralized networkis the result. Secondly, a critical reactionmay be made concerning the all-important role that Galloway owns toroot servers. From the local PC to theroot servers, a great many local servershelp carry the load. Although theseservers do not have all codal records of

    websites and are dependent on the rootservers for new code content, they arecapable of copying frequently requestedwebsites and offering them even if theroot server would break down.

    The interesting thing aboutGalloway is that although his main tenetis that the Internet relies on materialcodes and protocols in order to functionproperly, he implements metaphor useto explain certain processes, such aswhen he states that to help understand

    the concept of computer protocols,consider the analogy of the highwaysystem (Galloway, 2004, 6-7). Thus,himself explicitly makes clear that hedeparts from a metaphoric stance, buthis own work only points to the materialsubstantive reference.

    Simultaneously metaphoric andmaterial

    An example of a simultaneous

    metaphoric and material reference ofspatial concept use, is the work of JanVan Dijk. His starting point is ametaphoric use of the terms, systemsand networks functioning like neuralnetworks. He defines a network as arelatively open system linking at leastthree relatively closed systems and asystem is a set of interdependent units

  • 8/8/2019 vol1_gaps

    10/37

    Journal of Network Theory Spring 2006 10

    working together to adapt to a changingenvironment (Van Dijk, 2005, 3). But henotes that while, networks carry theimage of being open and free, . theyposses these characteristics only in aparticular way and are unable to becompletely open and free anyway (Van

    Dijk, 2005, 11). For Van Dijk, theInternet have very substantive borders,networks of ICT in particular in requireall kinds of programming, codes andaccess barriers (2005, 11).Thesetechnical regulation is what isresponsible for making the network, inthe last ten years is has been closed,barred and regulated in a rapid way(2005, 11). That boundaries anddivisions for Van Dijk are trulysubstantive, is visible from the following

    statement: We all know the phenomenaof core members, nodes, bridges andisolates, intimates and outcasts, spidersand flies in networks. In all social andmedia networks there are more peopleoutside than inside (Van Dijk, 2005,12). And yet he employs bridges andisolates, intimates and outcasts, spidersand flies as metaphors for those shutout of the system.

    Like Van Dijk, Manuel Castellsalso utilizes the metaphoric and

    substantive reference at the same time.While, his space of flows is theconceptual metaphoric reference,spaces of places refers to substantialreal life situation (Castells, 1996, 423).While he argues that transactions in aglobal level take place in a moreconceptual manner, at a deeper level,the material foundations of society,space and time are being transformed,organized around the space of flowsand timeless time. Beyond the

    metaphorical value of theseexpressions, supported by a number ofanalyze and illustrations in precedingchapters, a major hypothesis is putforward: dominant functions areorganized in networks pertaining to aspace of flows that links them up around

    the world, while fragmenting subordinatefunctions, and people, in the multiplespace of places, made of localesincreasingly segregated anddisconnected from each other. (Castells,1996, 476).

    Concluding remarks

    In this introduction, it was argued thatthe concepts of gaps, divisions andboundaries are generally marginalconcepts in network theory. As arguedabove, these concepts are loosely orshortly referred to in different works. Theconcepts are invisibly present, on theone hand because they are not implicitlymentioned but are logically present andon the other they are made explicit but it

    is not made clear what reference theymake. While the conceptual use seemsto slip from a metaphoric to asubstantive and vice versa, there arealso instances in which simultaneousreference is possible.

    The articles in this issue pose tomake visible the functions of the spatialconcepts. By discussion their positionthe writers contribute to thedemarginalisation of these spatialconstructs. Loes Vollenbroeks work

    compares Social Network Analysis andActor Network Theory by focusing onthe position of the researcher. Sheargues that the very act of doingresearch creates boundaries. MonicaNegri will explore the notion of gaps andboundaries in relation to continuity onthe Web. Thomas Nachreiner willdiscuss the position of divisions for theDNS protocol system. Nachreineremploys the concept in a substantivetechnical manner, while Vollenbroek and

    Negri both explore the oscillation of theconcepts. This issue ends with a shortand sweet narrative by Nachreiner thatreflects on the role of the writer inrelation to his or her textual creation.

  • 8/8/2019 vol1_gaps

    11/37

    Journal of Network Theory Spring 2006 11

    Bibliography

    Barabsi, A. & Albert, R. Emergence of scaling in random networks. Science 286 15October 1999, pp.509-512

    Barabsi, A. & Bonabeau, E. Scale-Free Networks. Scientific American 288. 2003.

    Barabsi, A. The new science of networks. Cambridge: Perseus Publishing 2002.

    Dijk, J. van. Outline of a multilevel theory. Paper presented at the Annual InternationalCommunications Association Conference, New York, 2005

    Efimova L. & Hendrick S., In search for a virtual settlement: An exploration of weblogcommunity boundaries. Paper presented at 2nd International Conference onCommunities and Technologies, Milano, Italy, June 2005https://doc.telin.nl/dscgi/ds.py/Get/File-46041/weblog_community_boundaries.pdf

    Galloway, A. R. Protocol: How control exists after decentralization. Cambridge:MIT Press. 2004. 2-78

    Latour, B. Reassembling the social: An introduction to actor-network-theory. Oxford:Oxford University Press 2005 (pp.121-140)

  • 8/8/2019 vol1_gaps

    12/37

    Journal of Network Theory Spring 2006 12

    2. Finding or Choosing Boundaries?

    Exploring the position of the researcher in ANT and SNA

    By Loes Vollenbroek

    Abstract:While in traditional sciences the researcher was positioned as an external objectiveobserver revealing facts, epistemological reflections have led to a revision of the positionof the researcher. In the analyses of networks, following one of its leading figures BrunoLatour, Actor Network Theory (ANT), calls for reflexivity and acknowledgement and carein the way the process of writing in itself is a method. ANT calls for no a priori notionsand an open minded researcher with no pre-set framework, this paper will explore themethodology and researcher position in Actor Network Theory in comparison with SocialNetwork Analysis. First addressing the similarities and differences between the twoapproaches in method and approach of networks. Following, while looking at examplesof studies where the two approaches have been used, an exploration of the position ofthe researcher in theory and in practice. Concluding that in both approaches researchersare in need of boundaries, because of the extensive amount of data that will otherwisebe generated. In ANT these boundaries are metaphorically called black boxes, while inSNA boundaries are generally placed upon network. In both cases, we see that thechoices researchers have to make create boundaries in the analysis. The question is notif researchers influence their research, but if science is better if researchers reflect uponthe process. The difference between science producing results, like SNA, or sciencedescribing processes, like ANT.1

    1Due to personal circumstances the full article will be available in a couple of days and a new version of

    this journal will be send.

  • 8/8/2019 vol1_gaps

    13/37

    Journal of Network Theory Spring 2006 13

    3. Continuity on the Web

    Beyond Cinematographic Boundaries

    By Monica Negri

    Neo, The One: I know you're out there. I can feel you now. I know that you're afraid.You're afraid of us. You're afraid of change. I don't know the future. I didn't come here totell you how this is going to end. I came here to tell you how it's going to begin. I'm goingto hang up this phone and then I'm going to show these people what you don't wantthem to see. I'm going to show them a world without you, a world without rules andcontrols, without borders or boundaries, a world where anything is possible. Where wego from there is a choice I leave to you.

    From the screenplay of The Matrix - Wachowsky Br.,1999

    Abstract:This article reflects on boundaries as an embedded characteristic of the networkstructure, considering two different levels of analysis: the physical plane where cablesand routers distribute bits along the net, and the content surface which allows thetransition of knowledge among the connected nodes. This distinction between a stagewhere the linkage is established by mechanical support and the layer which allows the

    passage of information directed to the user, implies a consideration of the possibleobstacles to the continuity of the process. Within this analytical perspective, a functionalassociation with the cinematographic concept of continuity will be introduced in order toclarify which is the importance of an uninterrupted connection and what are the potentialboundaries or gaps that can occur. Similarly to Joness theory of virtual settlement(Jones in Efimova & Hendrick, 2005), I will make use of meaningful movie such as TheMatrix (Wachowsky Br.,1999) to provide an analogical metaphor with the networkenvironment, considering the role of boundaries in this cinematographic subject.

  • 8/8/2019 vol1_gaps

    14/37

    Journal of Network Theory Spring 2006 14

    From the Matrix to the Network

    The image heading this article belongsto the last sequence of Matrix, perhapsone of the most well known moviesdealing with new media and informatics.The story plot stresses on the main roleacquired by network connections incontemporary society: it is centred onthe metaphor of a matrix whoseunavoidable linkage kidnap users to aconnected system, where theboundaries coincide with the limits ofsocial life itself.

    With an allusion to AlbertBarabasis (2003) perspective, in theMatrix people are nodes connected toa structure which provides to them

    relations, communication and a role:once that a disconnection occurs, themissed-node gap is replaced by anothersubject whose function is to compensatethe interrupted flow of energy whichmachines derive from the Matrix.Referring to Scale-Free networks,Barabasi suggests how accidentalfailures as the lost of a node connectioncould be easily being compensated bythe system itself, which could send,referring again to our metaphor of the

    Matrix, clones of Agent Smith to solvethe gap (Barabasi, 2003, 52). Instead,coordinated attacks as Neo and theother aware nodes would threaten thestability of the system, undermining thenetwork in its most vulnerable points.

    Many concepts so far have beenintroduced: not only the idea ofcontinuity and its related corollary orboundary, but also the possibleoccurrence of a gap, all notions whichnecessitate a description before our

    discourse develops further.

    The Continuity

    The concept of continuity has gainedimportance following the development ofthe cinematographic medium, withinwhich it covers an fundamental role, due

    to its strict relation with time slide. Theword strict is chosen since the task ofassuring an uninterrupted flow of data isa prerogative to which contemporarycinema can not more renounce to,specially after the passage from silent

    movies to sound era.The fact was that until the

    twenties, the need of dialogue inserts a justified technical manipulation of thesequence connection in order to makethe story understandable, implies thatthe continuity on the plot level wasobtained through a discontinuity on thephysical plane. In fact, the firstdistinction involved when dealing withthe cinematographic continuity relates tothe difference between a temporal and a

    physical level, which both take part inthe final process of the text fruition. Thetemporal flow in a movie consists of thelogical and continuous stream ofinformation which improve the plot: itcan be experienced considering thewhole narrative process and is based onthe causal connection between theevents within the story. The physicalplane relates instead to the material onwhich the data are conveyed, i.e. thevery film and technical effects applied to

    obtain the product in a defined version.Between these two levels we have toconsider the language used tocommunicate by cinema, made of unitsas frames, scenes and sequences,which are structured by the process ofmontage, shot following the perceptionrules and codified on causalconventions.

    Considering the metaphor withwhich I have started this article, I willnow establish a parallel between cinema

    and web, considering the possibleanalogies between their flows of data.As Barabasi (2002) suggests, the unitsof connection inside a network can beidentified with nodes, that pragmaticallycoincide with the cinematographicturning points and catalyst eventscharacterising a story plot: both of them,in fact, represent linking spotscontained

  • 8/8/2019 vol1_gaps

    15/37

  • 8/8/2019 vol1_gaps

    16/37

    Journal of Network Theory Spring 2006 16

    supported by its Latin meaning torestore: each new medium is supposedto justify itself from the statement that itis improving upon a predecessor andsince a new medium appears, a lack ofimmediacy starts being associated to

    the traditional media, that is alsoconsidered inadequate (Bolter & Grusin,1996, 15). In this sense, the process ofcontinuity lays also on the practice ofremediation, that allows an improvementof the network system through referenceto contemporary and previous media:the case of similarity between web andcinema can be considered in reason oftheir reciprocal influence but this is adifferent point of analysis, whichManovich (2001) has exploited in a

    exhaustive book.Finally, dealing with continuity as

    a state of uninterrupted flow ofinformation means referring to Castells(1996) space of flow which, due to thenature of its real-time interactions andthe material infrastructure that makesthese interactions possible, has bothmaterial and immaterial aspects, astemporal and physical level are impliedin the network stability. From physics welearn that for every force exerted, there

    is an equal but oppositely directed one(Newton's third law). Therefore, if wehave a space we should also have alimit (at least beyond which we cannotsee) and if there is continuity, probablysomewhere, well find a discontinuity.

    The Boundary

    Introducing the parallel with cinema, Ireferred earlier to the early silent movieswhere the visual continuity was

    interrupted by inserts with dialogueswhich, if functional to the process ofunderstanding, were at the same timeembedded in an interruption of the routeof flowing. The initial distinction betweentemporal and physical continuity needsto be exploited also in the case ofdiscontinuity events, which can becategorized in Planned Occurrences

    and Causal Blunders. To be moreaccurate, the second circumstanceproperly refers to a taxonomy of errors,which can of course be planned, but notin order to maintain the continuity, whichis instead the focus of our analysis.

    Following it necessitates anindependent dealing which willconstitute the last section of my article.

    In order to explicate the abovelabelled as discontinuity plannedoccurrences (DPO), it is first necessaryto make a further distinction amongwhat I am going to label Proceduresandthose I will refer to as Regulations. It isrequired to delineate that, if the state ofcontinuity adopts a linear behaviour andleaves a certain independency among

    the different levels, instead a conditionof discontinuity leads to an instabilitywith distributes itself on more than thoseplanes we have analysed before(Temporal & Physical, Intertext,Paratext, Intratext).

    In fact, if we consider our moviereference, until Keanu Reeves is MisterAnderson, i.e. a passive programmerwith a common life and no rebelattitudes, the Matrix has a continuitywhich is maintained even if some

    normal nodes die or any Oracle ishyperlinked. Instead, when Neo wakesup from his sleep, a central node causesinstability in the system and thishappens on more than one level. FirstMister Andersons time of life recalls thefield of Temporality, since he virtuallydies while Neo is born as new consciousindividual. Secondly, after hisdetachment from the Matrix cable,Mister Andersons body doesnt give anymore his energy to the machines and

    this can be situated on the definedPhysical level. Then, since Neo canpass from the Matrix text to Zion land,this allows him to change hisembodiment, metaphorically configuringa space of Intertextuality. Also, whenNeo fights Agent Smith, i.e. a cluster ofnodes arranged in defence of themainframe, he is involved in the process

  • 8/8/2019 vol1_gaps

    17/37

    Journal of Network Theory Spring 2006 17

    of Intratextuality. Finally, Reeves is atthe same time Mister Anderson, asAgent Smith calls him, and Neo, forTrinity and Morpheus and thiscontextualization is involved in the so-called Paratextuality.

    As we referred before and usingBarabasis words: just a few key hubs[] splintered the system into tinygroups of hopelessly isolated routersand this is exactly what happens withNeos gained awareness, but also whena boundary raises within the web. Thecategorization of DPO can literary followour cinematographic metaphor and, forexample, those twenties writtendialogues could belong to the Physicallevel, act versus the Temporal continuity

    and being included within theIntratextual ground.

    Following, it becomes clear howboth procedures and regulations canoriginally being located on a single level,but spread their effect on more then onestage. For cinematographic textsprocedures coincides with techniques ofmontage (Physicality) that, creating adiscontinuity of reality, insert boundariesof fades, cuts and effects, building onthe continuity of the movie. Regulations

    are instead part of the paratext, sincethey influences the text without beingembedded in it, with an effectcomparable to the Cascade StyleSheets employed in web programming.

    It is now effortless the passageto the network analysis, where we canretrieve a similar taxonomy for webboundaries. First we considerSubscription & Passwords, that can beincluded in the procedures group: theyare concretized in code sequences

    (Physical), but necessitate a temporaleffort to be exploited, which interruptsthe continuity in the process (ex. ofbuying) pursued by an user. Firewalls,Spy-ware, Antivirusstrictly belong to thecategory of software and, subsequentlythey are based on the code: their task isassociated with the concept of boundaryfor antonomasia. Copyright Laws, which

    belong to the paratext, control the use ofcode and content, both on the form of atemporal and physical level: accordingto the Italian Law, the right of exclusiveright can expires with time, while theproperty can change owner following to

    detainment reasons.In the end, we consider

    Machineson which the code runs, beingvisualized as content or acting asregulator: the physical support is alwaysimplied as source, even in the case ofVirtual Reality (which if doesnt need acable to pass, at least both the senderand the receiver have to be providedwith magnetic or electric interceptors. Itis important to underline howboundaries, due to their feature of being

    planned, have the pursue of ensuringthe temporal continuity which above Idefined of 2nd level, i.e. the stability ofthe system in a future perspective, evenif this means that some nodes arelocally slowed down or turned off (exvirus nodes).

    Concluding, continuity isindependent form discontinuity since aboundary can be a sufficient but notnecessary condition to create a flow ofinformation; instead continuity is a

    passage both necessary and sufficientto which DPO can follow.

    The Gap

    When dealing with the interruption of thedata flow we regarded to the differencebetween DPO and Errors, basing on theintentionality or causality of theirhappening. A Gap can be defined as alack in the sequence of data, which canbe perceived directly, through its visual

    manifestation, or indirectly by its effectswhich differently relate to the space.Also, if the concept of blunder isdenoted with an unplanned feature, thismeans that it can occur on each level atany time and, above all, that thetemporal continuity is, on principle,excluded from each level by itsattendance.

  • 8/8/2019 vol1_gaps

    18/37

  • 8/8/2019 vol1_gaps

    19/37

    Journal of Network Theory Spring 2006 19

    a chronological basis. When a linkagetakes longer, it can be due to themechanical features not enoughpowerful for sustaining the connection,but it is also possible that the samenetwork is too busy to allow any further

    procedure.It has become clear how code

    errors cover an important role indestabilizing the continuity in a relatedsystem, while the levels implied have tobe considered strictly related by theeffects of this gaps, since without aworking code, the communication withina network cannot take place. The mostperceivable consequence of gapsoccurrences leads to knowledge that thenetwork has borders, which means also

    that is physical limited. Temporally,instead, it is continuous due to itsembedded characteristic of being scale-free (Barabasi, 2003): the most powerfulnode needs to be targeted as in ourmovie reference, until the main clusterof Agent Smith is not defeated, Neocannot complete his task and the Matrixwill continueits existence.

    The Network requires the Matrix

    If so far the metaphor of the Matrix hasbeen useful to explain the abstractconnection within the network system,looking forward to correlate theoreticalconcept with visual references, it is timeto reveal the differences between thecontinuity of this cinematographicsubject an the network environment.Foremost The Matrix movie is basedon the necessity of overcoming theboundaries: the real world is conceivedas something beyond them which both

    Neo and Mister Smith are determined to

    reach, the first in order to defend, thelatter resolute for its destruction.Following, Neo is himself a gap: whenMister Anderson node is awakened byTrinity and Morpheus, he leaves hisfunction in the system, whose error is

    immediately perceived on the controllevel and the Agents are sent to repair it.

    The purpose of a networkcoincide actually with Mister Smithhimself, that chases a temporalcontinuity in the system permanence: ifa gap occurs a resolutor is sent to solvethe error. It is important to underline howthe set of solution that Mister Smithuses in order to re-establish the order inthe Matrix primarily consist ofboundaries, which are useful to stop the

    discontinuity created by the rebelsactions.

    In the same way, the collectionof Procedures and Regulations createfor the web system a set of referencesthat can provide stability, at least inorder to prevent a gap event: it seemsthat the network, being built onconnection and nodes, cannot reach astate of continuity without its matrix.Concluding, we can reflect on the afamous sentence by Jean Cocteau:

    Cinema is Deaths working laboratoryas the cinema is told to work on time topreserve it from the past, assuring acontinuity to those data it represents. Atthe same way, a network bases itstemporal continuity on physicrequirements, coded language and nodecommunication, in order to assure anuninterrupted flows of informations: inthis sense, we could metaphoricallyaffirm that beyond the boundaries hidesthe same goal of immortality affecting

    the humankind.

  • 8/8/2019 vol1_gaps

    20/37

    Journal of Network Theory Spring 2006 20

    Bibliography

    Barabsi, A.L. & Bonabeau E., Scale-Free Networks. Scientific American 288, 2003,pp.60-69

    Barabsi A.L. & Albert R., Emergence of scaling in random networks. Science 286, 15October 1999, pp.509-512

    Barabsi A.L., Linked: the new science of networks, Cambridge, Perseus Publishing2002, pp. 143-197

    Bolter G. & Grusin R., Remediation, in Configuration 4.3, 1996, p 311-358.

    Efimova L. & Hendrick S., In search for a virtual settlement: An exploration of weblogcommunity boundaries. Paper presented at 2nd International Conference onCommunities and Technologies, Milano, Italy, June 2005https://doc.telin.nl/dscgi/ds.py/Get/File-46041/weblog_community_boundaries.pdf

    Galloway A. R., Protocol: How control exists after decentralization. Cambridge: MITPress 2004, pp.2-78

    Genette G., 1997Paratexts thresholds of interpretation, tr. J. E. Lewin, New York:Cambridge University Press, org. Seuils, 1987

    Latour B., A prologue in the form of a dialog between a Student and his(somewhat) Socratic Professor peer reviewed journal about digital culture, BrunoLatour's Web Site May 2004 http://www.ensmp.fr/~latour/articles/article/090.html

    Manovich L., The Language of New Media, MIT Press, 2001

    Online resourcesWachowsky Brothers, The Matrix screenplay, 1999http://www.theforbiddenknowledge.com/the_matrix_script/index01.htm

  • 8/8/2019 vol1_gaps

    21/37

    Journal of Network Theory Spring 2006 21

    4. Domain Name System

    Protocol, Governance and Divisions

    By Thomas Nachreiner

    Abstract:In the administration of the Internet technical notions merge with the political ones. Thisis especially true for the implementation of the Domain Name System, which is regardedas one of the central protocols structuring the Internet. This paper portrays theprotocological features of DNS in its technical and political dimensions. Starting from atheoretical perspective the daily practice on the protocological layer of the Internet issketched out with the aim to sharpen the contrast between the notions of protocol and itsapplication in reality. Through this approach two issues are addressed: the assumeduniversality of the DNS on the technical layer and the discussion about it on the politicallayer. Showing that the first is no inherent property of the system, the second isinterpreted as a self-regulating process which is based on a wrong assumption. Thisleads to the conclusion that divisions within DNS are contradicting the assumption ofuniversality, but that exactly these divisions enforce and reproduce this universality.

    Introduction

    Quite recently breaking news shockedthe Internet community: in the aftermathof the World Summit on the InformationSociety (WSIS) China seemed to haveintroduced their own root zone fileswithout coordination with theInternational Corporation for AssignedNames and Numbers (ICANN). Thiswas mainly received as anti-americanpolicy, and furthermore as a threat tothe universality of the Domain NameSystem (DNS) and thus of the Internetin general. The triggered discussioninvolved the structure and architectureof the Internet and its protocols, as wellas Internet governance and its players.

    In my argument, I will start withAlexander Galloways notion of protocol:how the Transfer Control Protocol /Internet Protocol (TCP/IP) and the DNSshaped the Internets topology, imposingcontrol through their on the one handdistributed, and on the other handhierarchical features. I will especially

    highlight the sharing of protocols as acrucial feature for communication, andfurthermore the dangers of DNSshierarchy as Galloway describes them.

    This will be contrasted with amore detailed picture of the DNS, how it

    works and its role in the dailyprocedures of the Internet. The issuesadressed here are: the origin andpurpose of the DNS, the role of the rootservers in the system, and howcounterprotocological features areinherent to the system, pointing at theconsequences these might have.

    In a last step I will discuss thepolitical debate around InternetGovernance: The struggle aroundICANN, and especially how

    technological issues enter the sphere ofpolitics. I will especially highlight howcomplex matters are reduced to fit the

    justification of political agendas.The purpose of this paper is

    twofold: First, I want to show that theuniversality suggested by many notionsof DNS is already an illusion, and under

  • 8/8/2019 vol1_gaps

    22/37

    Journal of Network Theory Spring 2006 22

    no respect a natural property of thesystem, though it was its originalintention. Second, I want to show howthe fragile compromise making thesystem work is achieved through thedistributed character of the Internet,

    and that the perception of threats isheavily shaped by political interest andhysteria.

    This might help to answer thequestion to what extent there is adivision of the net on the protocologicallayer of the DNS, and what theconsequences of the recentdevelopments might be.

    Protocol, Governance and Divisions

    Protocols: The Governing Principles ofthe Net

    Prior to the dawn of the digital ageprotocol described any type of correctbehaviour within a specific system ofconventions (Galloway, 2004, 7),required to establish the essentialpoints necessary to enact an agreed-upon standard of action (Ibid.). Whilethe earlier protocols were supposed to

    govern social or political practices,computer protocols now represent thestandards for the implementation ofspecific computer technologies. InGalloways notion protocol is atechnique to achieve voluntaryregulation in a contingent environment(Ibid.), yet, in the context of computertechnology no longer a question ofconsideration and sense (Ibid.) butrather one of logic and physics 3 (Ibid.).

    Operating on the level of coding,

    protocols coordinate communication

    3For the sake of his argument Galloway is

    confusing regulation with administration. See

    also Becker et al.

    (2002, 40), who see protocol rather as part of

    the architecture and design of the technological

    infrastructure

    than as a tool of regulation.

    among different devices while remainingindifferent towards the content ofcommunication. The Internet, oftenperceived as a chaotic mass of datalacking central control, is organized byprotocols that form a distributed

    management system (Galloway, 2004,8). In this sense protocol is howtechnological control exists afterdecentralisation4 (Ibid.).

    The idea that the Internet is notcontrollable is according to Galloway based on the properties of the TCP/IPprotocol: its structure resembling ameshwork, the distributed technology ofTCP/IP facilitates peer-to-peercommunication (Galloway, 2004, 46). Itallows Internetworking if the connected

    hosts share it as a universal languagewhile it is open to a theoreticallyunlimited number of hosts in differentlocations. As the result of the action ofautonomous agents (computers) (Ibid.),the robust and flexible TCP/IP forms asystem of distributed control (Ibid.).

    The second protocol, which, inGalloways terms, constitutes theInternets mode of organization, is theDNS: responsible for the process calledresolution it translates Internet

    addresses from names to numbers. Itsstructure is that of an inverted tree(Galloway, 2004, 49), and on its top theso-called root servers are located. Onthese in turn, the information is locatedwhere the name serversfor the differenttop-level domains (TLDs) are situated.These again carry the information whatexact name server on the branch belowcontains the information for a certaindomain name. This server then canresolve the asked domain name into the

    numerical address of the computer

    4Note that after in this context is used

    chronologically: when the decentralised diagram

    was replaced

    by the distributed diagram as the mode of

    organization.

  • 8/8/2019 vol1_gaps

    23/37

    Journal of Network Theory Spring 2006 23

    asking for a certain domain.5 Thismeans that a hosts query for a domainname has to be answered step by step,from the top to the bottom of the tree. Ifone server fails to give the authoritativeanswer about where to find the

    responsible server on the branch below,communnication would collapse. Whiledelegated authority in the DNS followsthe chain from top to bottom, eachsingle server only is able to giveauthoritative information about thebranch located immediately below it.That is why the DNS is on the one handhierarchical, and on the otherdecentralized (Galloway, 2004, 49).

    There are two important notionsGalloway connects with the DNS: first,

    the absolute control one branch exertson its sub-branches. Seen from thisperspective, ultimate control over thewhole Internet lies on the root servers:simply by changing the root serversinformation or configuration, whole rootzones like for instance a countrycodetop-level domain (ccTLD), and thus awhole country can be erased from theInternets surface (Galloway, 2004, 10).And second, while it rigidly organizescontrol into a tree-like decentralized

    database (Galloway, 2004, 53), DNS,like TCP/IP, presents a universalizingsystem (Galloway, 2004, 51) being theencyclopaedia of mankind (Ibid.) and asingle, exhaustive index for all things[] DNS is not simply a translationlanguage, it is language (Ibid.). Withoutspeaking the same language, the nodesof a network cannot communicate, andthus without a shared protocol, there isno network (Galloway, 2004, 12). In thissense, incompatible protocols would

    represent the ultimate separation of theInternet.

    5For a more detailed explanation how DNS

    works see for instance Tanenbaum 2003, 579-

    588.

    Protocological Reality on the Net

    However, as plausible as Gallowaysframework of a universal DNS and theinherent control of the hierarchicalstructure might seem, it still raises thequestion of how accurate it is incovering the reality of the DNS on theInternet. This shall be examined in thefollowing chapter, especiallyhighlighting the origin of DNS, itsunderlying infrastructure and the recenttrend towards counterprotocols, whichseem to contest the universality of DNS.

    Origin and Purpose of the DNS

    In the early days of the Internet the DNSwas only a single file named hosts.txt,

    updated by computer scientist JonPostel. Published via a FTP-server runby the non-profit organization StanfordResearch Institute International (SRI), itcontained the information of all hostsconnected to the ARPAnet and later theInternet (Becker et al., 2002, 33). Thissystem, though working quite efficiently,reached its administrative and technicallimits after several years due to theincreasing size of the Net. Additionally,the system changed fundamentally

    when Paul Mockapetris developed anew DNS architecture in 1984 (Ibid.).Now all crucial information was spreadover the whole Internet, instead of beingplaced on one central computer. Everydomains owner provided informationabout it on his own host, and a centralcorporation only maintained recordsabout where this information was to befound (Ibid.). More than twenty yearsafter its introduction this is basically stillthe same system in use. What had

    changed were only the organizationsresponsible for the administration of theDNS and the distribution of the domainnames, but this development will bediscussed in detail in Chapter three.

    In the course of time, DNS hasbecome the primary governor of trafficflows on the Internet. When the DNSstops working, so do all applications: no

  • 8/8/2019 vol1_gaps

    24/37

    Journal of Network Theory Spring 2006 24

    email, no web browsing, no instantmessaging, no FTP, no e-commerce(Vixie, [3]). The reasons for thetremendous success of the system areprobably based on its ambitions whichare mostly human goals, because

    naming systems are really for use byhumans [sic.], who say them, hearthem, read them and type them(Templeton, [2]). The DNS makes thenavigation on the net much easier forthe user, because in the commonhuman perception remembering namesis considered to be easier thanremembering numeric IP-addresses.Presenting a registry of URLs (and theirnumeric translation), the DNS can beunderstood as the address book of the

    Net.Besides its usability the most

    important factor regarding the DNS is itsreliability: due to the frequent use of thesystem billions of times a day byalmost as many people the user wantsit to work quick, combined with thecertainty to get reliable results (Ibid.). Inother words: to reach the same domainsfrom any host on the Net. Thisuniversality is considered to be crucialproperty of the DNS, because it makes

    the Internet not only accessible for avast majority of people, but alsofacilitates the trustful usage bybusinesses and political institutions(Dyson).

    Related to the DNS there areseveral other issues at stake, fromdiscussions over the ownership of thedomains to the symbolic dimensions ofthe names (see for instance Watson),from the actual practice of domaindistribution to phenomena like

    cybersquatting (Becker et al., 2002, 34):the struggle over them triggers heavydisputes thematizing not only theadministration of DNS, but also thepossibility of its replacement6 - because

    6See for instance the proposals ofCornucopia

    (Watson) orDotDNS (Frankston). Lawrence

    Lessig also stresses

    many aspects of TLD naming andmanagement are traditional rather thanessential (Watson).

    However, the system still exists inits original design, holding up its claimand status of being the universal core of

    the Net. The following chapter shallpresent an attempt to get an idea of theDNSs myth of universality in examiningits underlying infrastructure and the dailyfunctions it fulfills.

    Infrastructure and Procedures

    Having in mind Galloways remark thatthe root servers [] have ultimatecontrol over the existence [] of eachlesser branch [of the DNS] (Galloway,

    2004, 10), it hardly surprises that mostmyths in the political controversy aboutInternet Governance are groupedaround them.

    Basically, the root level of the DNSis structured as follows: there is one rootzone file for each of the approximately260 TLDs7, consisting of a unique orderof numbers. The role of the root serversin the system is to enable thecommunication between the differentTLDs, thus if they do not contain the

    root zone file of a certain TLD or awrong version of it, this TLD cannot bereached from another root zone(Kleinwchter). Avoiding that andgranting the universality of the system,the 13 root servers8 of the authoritativeroot mirror the data on the so-calledhidden serverseveral times day. And sodo more than 60 other root servers, runby different organizations anddistributed all over the world9.

    the possibility of a DNS independent from thehierarchical TLD naming.7

    243 ccTLDs and about 20 gTLDs

    (Kleinwchter).8

    The original set of root servers, 10 of which are

    located in the USA, two in Europe, one in Asia

    (Becker et al.,

    2002, 30)9

    Like for instance Frankfurts mirror of the K-

    Root, run by RIPE NCC (see Rendek) or the

  • 8/8/2019 vol1_gaps

    25/37

    Journal of Network Theory Spring 2006 25

    While providing universality, thesystem of mirroring one single mastercopy offered by one single sourcefosters an often expressed fear: thedeletion of a ccTLD and thus thedisappearance of a whole country from

    the virtual landscape. According toKleinwchter, this is only half of thetruth: by no means would this be acommunication blackout for the affectedcountry. And it would only disappear ifall root servers would update the mastercopy, but as long as one root server,that still contains the countrys root zonefile, exists, its domains would be stillreachable. Furthermore it has to beadded that by far not all connections onthe internet have to be established

    through a query to the root servers: thecomputers along the path (of aconnection) remember the questionsasked recently, thus the service of theroot servers is basically needed toestablish new or rather uncommonconnections (Templeton, [1]). Theimportance of this aspect is illustratedby a study of the CooperativeAssociation for Internet Data Analysis(CAIDA), which suggested that morethan 98 % of the root server traffic are

    unnecessary, due to wrongconfigurations of some ISPs or to usersentering non-existent addresses(Unknown, [5]).

    While, theoretically, such anetwork division would only occur withsome delay, or even not at all, theparticular case of Libya in 2004 presentsthings differently: on April 9th LibyasccTLD .ly stopped working. Whether ornot this incident was the result of amanipulation by ICANN or rather an

    accident, Libya was cut off totally, atleast from the perspective of theWestern users (McCarthy). RegardingKleinwchters notion that a countrycannot disappear as long as one root

    system of root servers run by the European

    Open Root Server Network (see Vixie, [2] and

    Unknown, [1]).

    server (or one of its mirrors) does notupdate the question is raised: how likelyis this total break-off when the wholenetwork is automatically updatedseveral times a day?

    The dependence on one single

    source of input is often perceived as apotential threat, and the possibility ofseveral sources is a steadily recurringtheme in the discussion over DNS. Thenext chapter will explore DNSs actuallevel of universality while pointing at itsinherent characteristics that can contestit.

    Splitting the Root?

    Due to the strength of the system that

    [root] servers are operated by a widelydiverse group of organisations10(Rendek), it is perfectly possible foranyone, who owns a root server, tointroduce his own root zone files. Thustheres no reason why you couldnthave multiple root systems (Lessig).The question rather is: who wants tohave them? On the technical level theInternets ISPs and backbones couldeasily configure their main routers withnew routing tables addressing different

    root servers (see Templeton, [1]), butdue to the benefits of DNSs universality

    above all reliability they usually haveno interest in doing so. Becauseconsistency is important so that theinter-operability that we have grown toexpect from internet technologyremains (Vixie, [1], Comment made byJonathan Frakes), one should be able toexpect internationally operatingorganisations to have no interest insplitting root.

    Seen from the technicalperspective one might also argue that

    10The various root servers are run by various

    kinds of organizations, among them: the US

    military, commercial

    enterprises, non-profit organisations,

    universities, ISPs, and research institutes (Becker

    et al., 2002, 30).

  • 8/8/2019 vol1_gaps

    26/37

    Journal of Network Theory Spring 2006 26

    coherency was designed into thesystem as a basic assumption (Vixie,[1]): rather than coherency which isbasically an assumption of everyprogramme the essential postulation issingularity, in the sense that DNS does

    process addresses, but is not designedto ask in which namespace theseadresses are located. The dangers ofalternative roots being introducedwithout central coordination might turnout as some hell breaks loose andyoull see lots of extra network traffic,lots of false negative responses, andlots of junk in everybodys syslog files(Ibid.).

    Statements like Paul Vixies wouldsuggest that there are no alternative

    roots so far. However, there arehundreds of variations of the root file inuse[] and the[] real root zone is notexactly what the NTIA11 believes it is:there is a TLD pollution introducing newnameservers and even more :-)[sic.](Auerbach, Comment made by JFCMorfin). The question remains of who isinterested in splitting the root anddefining his own namespace? Basicallythere are two groups making policiestowards alternative roots: on the one

    hand governments and privatebusinesses on the other.

    The latest example illustrating thecase of private businesses is theannouncement of Neustar and GSMAssociation to provide private root-services for their members (Unknown,[7]). Introducing the TLD .gprs forroaming operators triggered theconcerns of the Internet EngineeringTask Force(IETF): the spread of .gprs-requests into the official root zone,

    which could severely affect theperformance of the root servers(Unknown, [6]). Such attempts ofbusinesses to provide their customers

    11National Telecommunications and Information

    Administration, a sub-department of the US

    Department of

    Commerce.

    with exclusive content appear to bewidespread and are often criticised forbeing shortsighted (Vixie, [2]). Otherexamples are the ISPs that redirect rootserver traffic to their own servers. Theirpurpose is more likely to deal with the

    misconfigured DNS clients of their usersthan to make adjustments on the rootzone files, but there would be thepossibility to add or delete informationwithout the real domains consent or theusers knowledge (Levine).

    This strategy is, for instance, alsoapplied by Chinese or Arabic groups,with the variation that their motivationdiffers: since their scripts containcharacters outside the traditionalEnglish ASCII character set,

    Internationalized Domain Names (IDN)where introduced. In order tosynchronize them with Unicode, ascheme called Punycode wasintroduced by the IETF: it encodesUnicode characters as ASCII stringsstarting with xn-- (Ibid.). However, theproblem with this easy technical solutionappeared to be that the IDNs madepossible to register different domainnames that look the same. Theintroduction of IDNs and Punycode

    simply splitted the universal relation ofdomain names and Unicode. Theconsequence was the increasingregistration of Arabic or Chinese domainnames which had Unicode equivalents,and thus the question which one of thepossibilities could be accessed becamethe question of which root zone fileswhere used. The ISPs in the countries inquestion realized that and set up theirown modified root zones and introducedtheir own modified root zone files. This

    means that at least in parts ofPunycode countries people can getdifferent content than people in Westerncountries, though they might type in thesame words (Ibid.).

    This practice seemed to be takento its next level on February the 27 th2006 when Peoples Daily mistakenlyreported that China is launching a new

  • 8/8/2019 vol1_gaps

    27/37

    Journal of Network Theory Spring 2006 27

    DNS. The introduction of Chinese-character TLDs for .cn, .com and.net was first assumed to mean thatInternet users don't have to surf theWeb via the servers under themanagement of the Internet Corporation

    for Assigned Names and Numbers(ICANN) of the United States.(Unknown, [3]). First investigations soonrevised this assumption and spread themessage that China simply introducedan alternative root, due to their use ofIDNs (MacKinnon, see especially thecomments on it by Murdoch, Mueller,Morfin and Watson). Eventually onMarch 3rd, the China Internet NetworkInformation Center(CNNIC) clarified thesituation, stating that China did not set

    up new TLDs (except .mil.cn), but onlycontinued a policy that had alreadystarted in 2002 (Unknown, [2]).Furthermore, this policy never intendedto break away from ICANN or the DNSbecause CNNIC had committed itself tomaintain the stable operation of theDNS (Ibid.). This made the DNScommunity wonder again: while theexperience was that two differentsystems were running in China, CNNICcommitted itself to the stable operation

    of the DNS? The solution Murdoch andMorfin finally came up with was that thesituation occurring was no root split, butrather an Internet split (MacKinnon, seeespecially the comments on it byMurdoch and Morfin). China was andstill is running at least two namingsystems: One is for the externet12internal users (resolving the ChineseTLDs) and the other, as youdocumented uses proximity machinesand plug-in to offer a good quality

    service. (Ibid.). Indeed, there is noviolation of the universal DNS root inthe case of China, yet the Net is dividedalong another line.

    Given the described situation thata large variety of different and

    12For a more detailed explanation of an

    externet see Morfin, 2006.

    heterogeneous organisations is as amatter of principle able to construct itsown root zones while a certain degree ofuniversality is maintained bycompromise for the sake of usability, itdeems necessary to investigate how this

    fragile compromise is achieved andwhat role the notion of universality playsin the negotiations over InternetGovernance.

    Internet Governance: PoliticalConflicts and Technological Means

    There is nothing especially remarkableabout them [domain names] from atechnical perspective, but from a social

    and political perspective they are allsorts of fun. We can have argumentsover control of the DNS root, argumentsover whether names are property,arguments over innate rights to specificnames, arguments over a registrarsright (or lack thereof) to exploitunregistered names for private gain, andmany more arguments besides.

    Brett Watson

    What can in ICANN?

    As seen above, the universality of asingle authoritative root is a basicassumption of its design and consideredreasonable by most participants in thediscussion, at least as long as thereplacement of DNS is not set fordiscussion. But in practice, theuniversality becomes more and morecontested, and the consciousness aboutthat fact is steadily increasing. The

    DNS, carrying around Tim Berners-Lees stigma of the Achilles Heel ofthe Internet, is often portrayed as thetool for controlling the Internet, becauseof its image of being the only source ofcentralized power in the virtual realm.Quite consequently ICANN - in its role ofthe administrator of this single source ofcontrol - is probably the internet-related

  • 8/8/2019 vol1_gaps

    28/37

    Journal of Network Theory Spring 2006 28

    organisation which draws mostattention, discussion and criticism onitself. To understand InternetGovernance and the flexiblecompromise it presents we have tounderstand the players involved. In and

    around ICANN is the site where theymeet, and for understanding theconditions of the discussion over ICANNand DNS it first deems necessary totake a brief look at the history of ICANN.

    Since 1992 crucial tasks inadministrating the internet wereperformed by the Internet AssignedNumbers Authority (IANA)13, due to itscontract with the US government: themanagement of IP address spacelocation, protocol parameter

    assignment, DNS management, androot server system management wereexecuted under its authority while theadministration of domain nameregistration was in the hands of theprivate company Network Soltuions Inc.(NSI) (Castells, 2001, 31). This changedin 1998 when the Clinton administrationcalled for the further development of theInternet as a free trade zone by privatecorporations and the NTIA expressedthe need for more international

    competition regarding the administrationof domain names (Becker et al., 2002,35). For achieving this goal arepresentative non-profit organisationwas meant to be created, and after longnegotiations between the differentgroups of interest the ICANN wasfounded (Ibid.).

    Originally proposed by Jon Postelit would embody the spirit of opennessof the Internet community,decentralization, consensus-building,

    and autonomy that characterized the adhoc governance over thirty years, whileadding a global orientation to itsmembership (Castells, 2001, 31): its

    13The organisation itself was under the

    supervision of theInternet Architecture Board

    (IAB) of theInternet

    Society (IS).

    supervision board is composed not onlyby representatives of the companiesrelated to the technical aspects of theNet, but also by so-called at-large-directors, who represent the interests ofthe Internet community on a

    geographical basis (Becker et al., 2002,37). Though ICANN is trying to servetechnical functions in the narrowestpossible way (Lessig), every singledecision finally gains a politicaldimension, simply because it is relatedto an infrastructure of sensitiveimportance.

    While there are indeedconsiderable concerns whether theICANN provides democraticrepresentation to the Internet community

    and it is often denounced as amarionette of the big corporations toshape the Net to serve their interests(Becker et al. 38), the focus of thistreatise shall lie on the struggle overICANN by the political players in thefore- and aftermath of the WSIS held2005 in Tunis. The central point of thecontestation of ICANN is its relationshipto the USA. While ICANN is officiallyindependent, it has agreed to cooperatewith the country. This is neither special

    nor intriguing since most of the worldssovereign countries agreed more or lessto cooperate with ICANN. However, fortwo (fr zwei was? reasons?) the role ofthe US is indeed an extraordinary one:first, the organisation is located in theUSA and most of its members areAmericans. And second, though ICANNis responsible for the administration ofthe DNS, still the NTIA has to authorizethe introduction of any new TLD in theauthoritative root list (Kleinwchter). At

    least, this is what most people think:according to John Crain, ChiefTechnology Officer (CTO) of ICANN,this is only a merely formal procedure,because the government official whohas to approve the new TLD has nocompetency to decide (Ibid.). Instead heonly has to check whether the formalcriteria to introduce the new TLD were

  • 8/8/2019 vol1_gaps

    29/37

    Journal of Network Theory Spring 2006 29

    fulfilled. Thus, he is fulfilling an auditfunction rather than an authorizationfunction (Ibid).

    Still, a broad (and quiteheterogeneous) coalition of countriesfrom China over Europe to Egypt and

    Brasil is uncomfortable with thissituation. As EU commissioner VivianeReding asks whether it is appropriatefor one government alone to supervisesuch an important part of theinfrastructure (Reding) while othercountries can exercise influence onlythrough the advisory committee toICANN, the USA have the sole right todecide when a new Top Level Domaincan be introduced into cyberspace,whether it be a new country-code or a

    new so-called "generic Top LevelDomain (Ibid.). The USA on the otherhand insist to be committed to taking noaction that would have the potential toadversely impact the effective andefficient operation of the DNS and willtherefore maintain its historic role inauthorizing changes or modifications tothe authoritative root zone file(Unknown, [9]). In short, the struggle isbased on a binary opposition: on theone side there are the USA, which

    regard the Internet as their owncreation, and though their actualprivilege in its administration is ratherlimited, they do not want to give it up forstrategical reasons. A traditionallynegative attitude towards trueinternational organisations like theUnited Nations(UN), combined with thefear that international competitors likeChina might instrumentalize it for theirpurposes makes them reluctant totransfer functions to them. On the other

    side there is an assembly of variousstates which feel a general concernabout the dominance of the USA, notonly in the issue of cyberspace. Thoughthe administration of TLDs is bound tonegotiations on several levels involvingmany different actors, the central role ofthe USA represents an affront towardstheir own sovereignity.

    What we see in the case of DNS ishow an originally technical issuetranscends into the political realm,gaining its importance and meaning ona symbolical level. Technically, the DNSis far from being a tool of total central

    control, because it can be applied andshaped by everyone for his ownpurpose. But in the practical situationand in its symbolical political importanceit is defined exactly that way.

    How Protocol is Controlled afterDecentralisation

    This can be understood only against thebackground of the increasingimportance of the Internet in economy,

    society and culture. While the USA hadan online dominance in user numbersand infrastructure throughout the 1990s,this dominance is eroding: there arealready more root server mirrors outsidethan inside the US (Rendek), and thenumber of Non-american users isgrowing steadily. This means that theother countries dependence on theInternet is increasing as well, and thattheir interests in shaping it are anexpression of their assumed role as the

    major actors on the international terrain.Due to this threefold importance

    of the Internet, most actors involved areverbatim spell-bound by the need for itsuniversality. The universality of the Net

    merely an illusion under manyrespects (see Barabsi, 2003) is clearlyidentified with the universality of DNS.The technical possibilities to dissolve itare not only a theoretical aspect, butalready reality on the Internet, yet theyare not used on a broader scale, but

    rather a side aspect in the daily routineof the Internet. The decision to splitaway from the authoritative root is aquestion of benefits: is it better to agreeon an uncomfortable administration ofthe one and only root, or are thereadvantages deriving from controlling anown root?

  • 8/8/2019 vol1_gaps

    30/37

    Journal of Network Theory Spring 2006 30

    The core of the dilemma mightbe described as follows: on the Internet,everything is done by consensus. Itmightbe possible to create an alternateroot server as a competitor but if thereare few domains on that server, then

    people are unlikely to ever access it. Iffew people access it, then companiesare not likely to register domains. Theinternet is full of tipping points where nostandard is legitimate in any form exceptby the sheer weight of de facto numbers(Mueller, 2004, 266 f.). Basically adivision of the DNS does not occur as itcannot be afforded, rather than theassumption that it would not be possibleon the technical level.

    Picking up and obscuring

    Galloways notion of meaning(Galloway, 2004, 52) we might say: theprotocol itself is indeed indifferenttowards the content communicatedwithin and it is a sheer fact that nocommunication is possible within in theprotocol if one does not follow it.However, it is possible to adopt theprotocol, to shape and use it, and thusestablish another way to communicate.In the case of DNS, this gives meaningto the protocol itself: on the one hand it

    is seen as an unavoidable central andcentralized tool of domination, and onthe other hand its daily applicationshows that it can be its exact opposite, away for anyone to establish his ownalternative mode of organisation.

    Human agency adds a horizontaldimension to the vertical organisation ofDNS: reality undresses it from the notionof a universal hierarchy, but at the sametime reality reproduces it again, finallyexpressed in the very human need to

    establish and maintain communication.Divisions are perfectly possible in DNS,it can be interpreted to fit variouspurposes. But as long as it is the majorpurpose to stay connected to themajority, the divisions will not threatenthe general universal strand of the majorprotocols.

    Conclusion

    We can clearly see that protocolsindeed structure every action on the net.Once implemented, they enact totalcontrol until they are replaced byanother one. Regarding the DNS,Galloway is correct with his notion ofhierarchical control insofar as theorganization of the DNS-tree is one oftotality from one branch to the next.

    However, the technical reality ofthe Internet shows that his simplifiyingnotion of the DNS is limited, becausehierarchical control is no universalproperty of the DNS: First, not everyinteraction does need to be submitted tothe total tree structure. Often usedaddressing paths are gaining stength

    and cross the DNS's vertical structurehorizontally. This means that also themost rigid protocol can be, in a sense,interpreted.

    Second, the implementation ofthe DNS root zone lists is mainlycentralized (from the A-root on) but itcould be decentralised as well. Takeninto account the importance of therouting tables and the attached groupsof interest, the picture is becoming morecomplicated: there are very different

    modes of how DNS could beimplemented, as several examplesshow.Thus the DNSs hierarchy cannot onlybe crossed horizontally, but also be splitalong vertical lines.

    Yet, in the political discussion wesee that the DNS is indeed interpretedin Galloways terms: mostly as a totaltool of control. In the eyes of the USAthis is the reason for not giving up theirstrong position in the protocols

    administration, and for its opponents thisis the reason why they are contestingthe United States. However, politicaland technological reality shows thatDNS is already splintered. Though thecore of the system is still built aroundthe A-root, many groups or institutionsare providing their modifications of the

  • 8/8/2019 vol1_gaps

    31/37

    Journal of Network Theory Spring 2006 31

    original root zone files. In this sense,divisions on the protocological level arereality, but by being a threat to theuniversality of the Internet they force thegroups of interest to maintain it. Bound

    to universality for several reasons, thepolitical players come to believe in thestrict and rigid hierarchy of DNS, takingit from a technological question into thediscussion about political power.

    Bibliography

    Barabsi, Albert-Lszl. Linked: The new science of networks. PerseusPublishing: Cambridge, 2002.

    Becker, Konrad et al. Die Politik der Infosphre. Schriftenreihe der Bundeszentrale frpolitische Bildung (386). Leske + Budrich: Bonn, 2002.

    Castells, Manuel. The Internet Galaxy. Oxford UP: New York, 2001.

    Galloway, Alexander R. Protocol: How control exists after decentralization. MIT Press:Cambridge, 2004.

    Mueller, Milton. Ruling the Root. Internet Governance and the Taming of Cyberspace.MIT Press: Cambridge, 2004.

    Tanenbaum, Andrew S. Computer Networks. Prentice Hall PTR: New Jersey, 2003.

    Online Sources

    Auerbach, Karl. About those root servers. CircleID. 01-08-2005.http://www.circleid.com/posts/about_those_root_servers/ (Last visited 13-04-2006)

    Dyson, Esther. Online Registries: the DNS and beyond. CircleID. 01-10-2003.http://www.circleid.com/posts/online_registries_the_dns_and_beyond/ (Last visited 13-04-2006)

    Evron, Gadi. Chinese Alternate Root as a New Beginning and Real InternetGovernance. CircleID. 28-02-2006.http://www.circleid.com/posts/chinese_alternate_root_as_a_new_beginning_and_real_internet_governance/(Last visited 13-04-2006)

    Forbes, Dennis. Interesting Facts About Domain Names. Yalfa Consulting andSoftware Development Services. 29-03-2006.http://www.yafla.com/dforbes/2006/03/29.html#a302 (Last visited 13-04-2006)

    Frankston, Bob. Ye olde DNS. CircleID. 26-08-2006.http://www.circleid.com/posts/ye_olde_dns/ (Last visited 13-04-2006)

    Kleinwchter, Wolfgang. Back to the Internet Roots: Macht oder Mythos?Heise. 02-02-2005.http://www.heise.de/tp/r4/artikel/19/19349/1.html (Last visited 13-04-2006)

  • 8/8/2019 vol1_gaps

    32/37

    Journal of Network Theory Spring 2006 32

    Lessig, Lawrence. Seven Questions: Battling for Control of the Internet. Foreign Policy.08-11-2005.http://www.foreignpolicy.com/story/cms.php?story_id=3306 (Last visited 13-04-2006)

    Levine, John. Splitting the Root: Its too late. CircleID. 02-12-2006.

    http://www.circleid.com/posts/splitting_the_root_its_too_late/ (Last visited 13-04-2006)

    Mc Carthy, Kieren. Libya disappears from the Internet. The Register. 13-04-2004.http://www.theregister.co.uk/2004/04/13/libya_falls_off_net/ (Last visited 13-04-2006)

    MacKinnon, Rebecca. Chinas New Domain Names Lost in Translation. CircleID. 28-02-2006. http://www.circleid.com/posts/chinas_new_domain_names_lost_in_translation/ (Last visited 13-04-2006)

    Morfin, Jean-Francois.Notes on the Root Name Principle. Personal Website. 02-03-2006. http://jefsey.com/rootname.pdf (Last visited 13-04-2006)

    Mueller, Milton. Internet Governance. CircleID. 01-03-2006.http://www.circleid.com/backgrounders/view/Internet%20Governance/ (Last visited 13-04-2006)

    Rendek, Paul. New Instance of DNS root server makes internet history. CircleID. 28-01-2004.http://www.circleid.com/posts/new_instance_of_dns_root_server_makes_internet_history/ (Last visited 13-04-2006)

    Reding, Viviane. Opportunities and challenges of the Ubiquitious World and some wordson Internet Governance. European Commission responsible for Information Society andMedia. 17-10-2005.

    http://europa.eu.int/comm/commission_barroso/reding/docs/speeches/ubiquitous_world_20051017.pdf (Last visited 13-04-2006)

    Rtzer, Florian. Wem gehrt das Internet?Heise. 23-10-2005.http://www.heise.de/tp/r4/artikel/21/21192/1.html (Last visited 13-04-2006)

    Seok-hwan, Kim. The Blackout: Preventing Network Domino Effects. CircleID. 20-08-2003http://www.circleid.com/posts/the_blackout_preventing_network_domino_effects/ (Last visited 13-04-2006)

    Templeton, Brad.

    [1] How DNS works. Personal website. Date unknown.http://www.templetons.com/brad/dns/howworks.html (Last visited 13-04-2006)

    [2] Goals for a domain name system. Personalwebsite. Date unknown.http://www.templetons.com/brad/dns/goals.html (Last visited 13-04-2006)

    [3] WSIS and the Splitting of the Root. CircleID. 14-11-2005.http://www.circleid.com/posts/wsis_and_the_splitting_of_the_root/(Last visited 13-04-2006)

  • 8/8/2019 vol1_gaps

    33/37

    Journal of Network Theory Spring 2006 33

    Vixie, Paul.[1] Putting Multiple Root Nameserver Issue to Rest. CircleID. 28-06-2005.http://www.circleid.com/posts/putting_multiple_root_nameserver_issue_to_rest/ (Last visited 13-04-2006)

    [2] Why I am participating in the ORSN project. CircleID. 01-10-2005.http://www.circleid.com/posts/why_i_am_participating_in_the_orsn_project/

    (Last visited 13-04-2006)

    [3] DNS gets a formal coordination system. CircleID. 20-10-2003.http://www.circleid.com/posts/dns_gets_a_formal_coordination_system/(Last visited 13-04-2006)

    Unknown Author.[1] About us. Open Root Server Network. Date Unknown.http://european.nl.orsn.net/about.php (Last visited 13-04-2006)

    [2] Statement on the Inaccurate Report on the Domain Name System in China.China Internet Network Information Center. 03-03-2006.http://www.cnnic.net.cn/html/Dir/2006/03/07/3623.htm (Last visited 13-04-2006)

    [3] China's Ministry of Information Industry revamps Internet domain namessystem. Interfax. 27-02-2006.http://www.interfax.cn/showfeature.asp?aid=10411&slug=INTERNET-POLICY-MII-DOMAIN%20NAME-DNS (Last visited 13-04-2006)

    [4] Answers from Vint Cerf: The Road Ahead for Top-Level Domains. CircleID.13-03-2006.http://www.circleid.com/posts/answers_from_vint_cerf_on_top_level_domains/

    (Last visited 13-04-2006)

    [5] 98% Of Internet's Main Root Server Queries Are Unnecccary: Should You BeConcerned?CircleID. 26-02-2003.

    http://www.circleid.com/posts/98_of_internets_main_root_server_queries_are_unnecccary_should_you_be_conce/(Last visited 13-04-2006)

    [6] Diskussionen um alternative Roots. Heise. 16-10-2006.http://www.heise.de/newsticker/meldung/64654 (Last visited 13-04-2006)

    [7] GSM Association and NeuStar Sign Agreement to Offer Root DNS Servicesto More than 680 Global GSM Mobile Operators. Neustar. 28-09-2005.

    http://www.neustar.biz/pressroom/files/announcements/ns_pr_09282005.pdf (Last visited 13-04-2006)

    [8] China adds top-level domain names. Peoples Daily Online. 28-02-2006.http://english.people.com.cn/200602/28/eng20060228_246712.html

    (Last visited 13-04-2006)

    [9] U.S. Principles on the Internets Domain Name and Addressing System. NTIA.Date Unknown.

  • 8/8/2019 vol1_gaps

    34/37

    Journal of Network Theory Spring 2006 34

    http://www.ntia.doc.gov/ntiahome/domainname/USDNSprinciples_06302005.htm (Last visited 13-04-2006)

    Watson, Brett. Whats in a Name?CircleID. 27-12-2005.http://www.circleid.com/posts/whats_in_a_name/ (Last visited 13-04-2006)

  • 8/8/2019 vol1_gaps

    35/37

    Journal of Network Theory Spring 2006 35

    5. The Last Day of Mankind

    Or an evening with the network society

    By Thomas Nachreiner

    Dusk set in over Utrecht. The KPNtelecom tower remained just a darksilhouette against the sky covered inpurple by the last sunbeams on thiscalm spring evening. What the heck Imdoing here?, Manuel Castells thoughtby himself while he turning away fromthe windows towards the table aroundwhich the crowd was already waiting.

    He knew them, somehow, fromsomewhere there was Barabsi, witha bottle of vodka, as always, and Latour,but since when did he smoke pipe?And the others, he had the feeling heshould know them, but they looked sosimi