Preserving a Universal Internet: The Costs of Fragmentation

3 September 2014 - A Workshop on Other in Istanbul, Turkey

Brief substantive summary of the workshop and presentation of the main issues that were raised during the discussions

The panel on the costs of fragmentation detailed the positive and negative effects that the breakup or splintering of the Internet will have on economics and politics. Currently, there is a single universal Internet at the technical layer. This technical layer, based upon the TCP/IP protocol, allow for the potentially universal communication between all connected devices. At the content level, however, the Internet is actually quite fragmented. This fragmentation is derived from a number of sources: national sovereignty concerns, differing national legal jurisdiction, concerns and actions of the private sector, and even individual users all fragment the Internet in various ways. The precise form of fragmentation varies as well. Data localization, traffic management, censorship and even technical protocols that lack universal interoperability are all forms of fragmentation. Importantly, fragmentation is not all bad and the costs are not evenly distributed. Some forms of fragmentation, such as data localization, can actually improve the privacy of citizens in some ways, provide a clearer tax base for developing nations, and result in preferential market access, even if it will also restrict the content that people can have access to and potentially lead over the longer term to breakup of the Internet at the technical layer. Costs and benefits accrue to every form of fragmentation, what matters is both whether the costs are greater than the benefits and how the costs and benefits are distributed.

Conclusions drawn from the workshop and possible follow up actions

There are several takeaways from the session. The first takeaway is that the sources of fragmentation are multiple, so a single cause cannot be identified. Each type of fragmentation has multiple contributory causes. Secondly, both costs and benefits accrue during the processes of fragmentation, leading to concerns of net benefits and distributional outcomes. Simply saying that fragmentation is costly is overly simplistic. Finally, efforts at either maintaining a more universal Internet or at implanting various forms of fragmentation often have iatrogenic effects when those measures are scaled upwards. What is individually rational at one level might be collectively disastrous. Moving forward, more research on the conditions under which fragmentation is beneficial or costly would be valuable. A careful mapping of the ways in which the costs and benefits of different forms of fragmentation diffuse would also provide a more nuanced understanding of the issue.

Estimation of the overall number of participants present at the workshop


Estimation of the overall number of women present at the workshop

about half of the participants were women

Extent to that the workshop discuss gender equality and/or women’s empowerment

it was not seen as related to the workshop’s theme and was not raised

A brief summary of the discussions in case that the workshop addressed issues related to gender equality and/or women’s empowerment

Note: The choices in in the pull down menu are very limited. In response to "Estimate the overall number of women present at the workshop," we have to choose between (1) all of the participants were women, (2) the majority of the participants were women, (3) about half of the participants were women, and (4) none of the participants were women. In this case, about 25% of the participants were women.

Reported by

Eric Jardine, Caroline Baylon, Samantha Bradshaw