The rise of the internet from back-room DARPA project to its present omnipresence is one of the great success stories of the 20th century. Roughly half of the world population is online today and some estimates put the web’s contribution to global GDP as high as $4.2 trillion in 2016, and in G20 countries alone, the internet generated up to 5.3% of GDP that year.

Yet, the internet calls out for governance, as a phenomenon that crosses borders and flows over regulatory landscapes with remarkable ease. It’s because of the web’s sheer lack of boundaries or center that its governance is so fraught. Any decision-making process needs to consider a plethora of stakeholders, platforms, and infrastructure providers in order to keep it running smoothly. That the internet is so decentralised is both a blessing and a curse for those seeking to govern it.

A debate over internet governance which had been simmering away in the background, out of sight of all but a handful of insiders and experts finally boiled over in 2016, as American legislators scrapped over whether they could, or indeed, should, stop the White House from relinquishing ultimate control over the of the Internet Corporation for Assigned Names and Numbers (ICANN). ICANN, a Los Angeles-based nonprofit organization, is regarded as the “internet’s phonebook.” Much of what the body does is effectively clerical, as it is in charge of the policies and procedures that underlie the Domain Name System (DNS) at the backbone of the web — ensuring that web addresses point to the right places, and ensure that one machine can talk to another.

Although the handover was made final on October 1st 2016, critics of the decision regard the move as ceding control of the internet over to foreign governments. This isn’t quite right: contrary to a popular belief, the U.S never owned the internet in the first place. But the White House’s critics aren’t entirely wrong, either, because whoever controls the address book has no small amount of power. They can censor the internet by deleting domain names, leading to worries that countries with less respect for privacy and human rights than the US would profit from this new set-up by trying to influence decision-making within ICANN.

That the internet is so decentralised is both a blessing and a curse for those seeking to govern it.

However, ICANN does not work that way — it is hard for any one actor to wield any degree of influence over the organization, which is accountable to a plethora of international stakeholders including technical and industrial committees, internet users, and telecommunications experts. And, like any large organization, ICANN has rules and procedures and by-laws of its own. Although governments do have a voice in the organization, this voice is heard in quangos which do not have the power to make decisions by themselves. Essentially, ICANN is not the UN.

The multi-stakeholder model adopted by ICANN, akin to that of their larger cousins, the Internet Governance Forum, confers no small degree of legitimacy and transparency to their decisions, providing a good model for centralized internet governance. However, consensus-based approach makes agreements hard to reach, slowing down the pace of transformation in a way that frustrates participants and puts the brakes on development. Besides, not everyone is truly comfortable with diluting the influence of states on the future of the internet, leading to calls for truly decentralized control.

No one person, one body, no one state has complete control of the internet globally. States still have the power to determine how the flow of data is managed within national borders, leaving space for countries dissatisfied with sluggish pace of global governance to advocate for internet sovereignty.

Russia, China, Turkey and Saudi Arabia, are examples of such states. They have long since lobbied the UN’s International Telecommunication Union to repatriate powers of decision-making to individual states’ governments, a bid that has been roundly rejected by most Western states. In face of such rejection and the rise of online forces that disrupt traditional national power structures, some states are increasingly looking to defect from global dialogue at both a political and a technical level, opting to implement their own mechanism at national level, to erect firewalls or to create their own infrastructure. The result: a balkanization of the internet that poses significant dangers one of its most crucial factors — its interoperability throughout the globe.

The most extreme example of state control of the internet is the Great Firewall of China. With one of the strongest systems of control in the world, the Chinese government currently blocks access services as innocuous to Westerners as Gmail, Facebook and Netflix. Some behind the Firewall had begun to use virtual private networks (VPNs) — especially foreign ones — to circumvent censoring, but the government got wise to those tunneling under the wall, launching a crackdown on VPN users. But the Chinese are not alone. In 2012, Iran established a Supreme Council on Cyberspace in order to disseminate online propaganda and prevent cyber-espionage, including attacks such as 2010’s Stuxnet. Russia, in turn, currently requires that the personal information of Russian citizens is stored within national databases. Even Germany is pushing to keep data within its borders.

States have been given further impetus to bring the internet under their control lately owing to threats such as the use of social networking in the spread of radicalism, and an increase in misinformation campaigns — “fake news.” But as well as damaging freedom of speech, this fragmentation can disrupt normal business activities and cause serious harm to an increasingly important internet economy.

But something far more disruptive looms on the horizon, and its power is only just being recognised. A disproportionate amount of traffic passes through a small handful of privately owned sites and services such as Google and Facebook, companies which have a great degree of soft power over internet users and infrastructure. Fears are only now being raised that vast private companies are becoming the de facto stewards of the internet, wresting control both from states and from ICANN.

Some companies have used their influence over internet service providers (ISPs) to offer what are known as zero-rated services, allowing end-users to access some corners of the internet without using up their data — without paying, in effect. Such policies allow companies to funnel internet consumption into patterns that best suit their own bottom line, structuring users’ online lives in new and potentially worrying ways.

Corporations can also use terms and conditions to impose limits on the behavior of their users. Social media companies have the power to restrict content to suit their own dictates, and can moderate the content shared on their platforms to suit their own biases. The new prominence of big private actors in internet governance is shown by a recent slew of court rulings. The European Court of Justice, for instance, forced Google to implement the controversial right to be forgotten, a process which allows users to request that historic materials about them be suppressed from search results. The ECJ ruling is a very public admission that the responsibility to implement privacy measures lies with companies like Google, like Facebook, and like Twitter.

Question of internet governance will only grow more pressing in the future. As the internet grows vaster and harder to avoid, so too do problems inherent to governing it. But the sheer multiplicity of actors involved in decision-making only makes things harder. Fragmentation ultimately puts into jeopardy the existence of an invisible complex of rules, regulations and safeguards that allow the internet to function, and that we take for granted. Preventing fragmentation is the most vital of tasks for anyone interested in ensuring the internet reaches its full potential.