Chapter 4. Designing a governance network


Who rules the Net? You and I and 600 million others, in some measure.

  Vinton Cerf
Table of Contents
4.1. Anarchistic
4.1.1. Anarchy and the Internet
4.1.2. Anarchistic Internet governance
4.1.3. Criticisms
4.2. Hierarchical
4.2.1. Bureaucracy
4.2.2. Oligarchy
4.2.3. Meritocracy
4.2.4. Hybrid models
4.2.5. Anarchistic–hierarchical Internet governance
4.2.6. Criticisms
4.3. Democratic
4.3.1. Representation
4.3.2. Consent
4.3.3. Transparency and accountability
4.3.4. Inclusion
4.3.5. Criticisms
4.4. Consensual
4.4.1. Consensus between stakeholder groups
4.4.2. Deliberative consensus
4.4.3. Consensus in Internet governance
4.4.4. Criticisms
4.5. Multi-stakeholder public policy development

To resolve that public policy governance of the Internet should be the province of a network of governments, private sector and civil society organisations does not presuppose that its form should be that which has in fact taken shape in the IGF. The structure and processes of the IGF, which are to be examined in Chapter 5, were not organised spontaneously, nor inevitably. A network can in fact take any of a number of different forms. The main purpose of this chapter is to examine and compare four such forms that might be used to structure a governance network such as the IGF.

As background to this exercise, an instructive analogy is found in the topology of computer networks such as the Internet. During the NSFNET period, the Internet was arranged in an hierarchical (or tree) structure whereby networks connecting to the Internet were required to establish direct links to the NSFNET backbone network (or if they were too small to justify a direct link, to link to larger networks that were in turn connected to the NSFNET backbone). So for example, in order for one university network to reach another, rather than sending its data across a link that directly connected the two universities, the data would be sent by the first university to the NSFNET backbone which would route it through to the second university by reference to an authoritative table of routing information that was maintained by the NSF.[1] This is a “top–down” structure.

The topology of the modern day Internet on the other hand is a distributed mesh network, in which the routing function is decentralised. Any node on the network can communicate with any other node across a multiplicity of possible paths, none of which need include any given central point. The redundancy of any given link in such a network makes the network as a whole more resilient against failure. Although it contrasts starkly with the top–down model, this is not so much a bottom–up structure—since that still implies the existence of a hierarchy, though inverted—as middle–out; or in computer network terminology, peer-to-peer.

Top–down and peer-to-peer computer networks have their equivalents in what we will respectively term hierarchical and consensual organisations. Whilst on the Internet the consensual paradigm is dominant, in organisations, the hierarchical paradigm prevails. This is in part the legacy of Weber, who theorised that a bureaucracy structured along hierarchical lines, with a strict division of labour and standardised procedures, was the most efficient (or rational) form an organisation could take.[2]

However examples of post-bureaucratic organisations, in which decisions are made by dialogue and consensus between peers,[3] have also more recently begun to emerge, and in fact are exemplified by a number of the institutions of Internet governance.[4] For such organisations, electronic communications are an enabling force,[5] not merely making old structures more efficient, but offering new ways of organising of which Weber could not have conceived, characterised by the fluidity of authority and the use of soft power.

These organisational forms, hierarchical and consensual, will be discussed below as two of the possible structures for a governance network such as the IGF. However, two other structures for which no ready analogue is found in computer networking are also possible: the bottom–up form, which will be described as democratic, and the absence of ordering altogether which will be described as anarchistic.[6]

This chapter describes each of these four forms of organisational structure in turn, with consideration of their inherent merits, as well as where relevant their compatibility with the international system described in Chapter 3, and with the culture of the Internet as described in Chapter 2. The aim is that by the end of the chapter we will have settled (if at a fairly abstract level) upon an appropriate organisational structure for international public policy governance of the Internet, which satisfies the criteria developed in previous chapters.



Ford, P S, Rekhter, Y, & Braun, H W, Improving the Routing and Addressing of IP (1993)


Weber, Max, The Theory of Social and Economic Organization (1964)


Heckscher, C, Defining the Post-Bureaucratic Type (1994)


See Section 4.4.3.


Fulk, Janet & DeSanctis, Gerardine, Electronic Communication and Changing Organizational Forms (1995)


It might seem a retrograde step to be considering top–down and bottom–up structures for a governance network, when governance by network has already been distinguished from governance by rules in its lack of hierarchy. However this can be explained by distinguishing between governance by network, which is a mechanism of governance through which public policy issues are addressed in concert by a coalition of affected stakeholders, and governance of the network, which is the coordination of the application of that mechanism through whatever internal structures and processes the stakeholders may devise. There is no reason why governance of the network might not involve a hierarchy of organisational roles.

To put this more concretely, an organisation consisting of governmental, civil society, private sector and intergovernmental stakeholders could equally conduct its affairs through the leadership of an elite subcommittee, or by democratic vote, or through consensus, or without relying on any predefined mechanism at all.