The "Balkanisation of the Internet" trope has reached critical mass. The bogey of the "splinternet" has been around the tech scene since year dot, and the Balkanisation metaphor too, but this largely centred on the pressure to replace open technical standards with proprietary ones - i.e. the drive for commercial lock-in. Some in the early 90s even saw the phenomenal growth of the Internet, and the lack of international management of connectivity, as likely to lead to Balkanisation by way of anarchy. In fact, it led to the "dark net" (or dark address space), those Internet resources that were unreachable by conventional means due to design or accident (the unindexed "deep Web" of dynamic or password-protected content, and deliberately hidden sites like Silk Road, are subsets of this). The Internet manages to be simultaneously a supra-national world-state and the Balkans, which is a neat trick.
Over the last 10 years, the Balkanisation metaphor has been used to describe three distinct developments in respect of the Internet: the growth of national regulation and censorship (China being the prime example); the push towards integrated services by the technology companies (e.g. bloatware on PCs that "manages" your Internet access); and the undermining of net neutrality - i.e. the desire of ISPs to charge different rates for different data traffic (e.g. you pay more for reliable streaming), which is a reflection of the cartelised market for network access (the local loop being a natural monopoly). With the NSA/GCHQ revelations this year, the frequency of the term's use has swerved decisively towards the political realm and flowered as a full-blown trope: the nationalised intranet.
The use of "Balkanisation" in a geopolitical context, which implies both the petty and the internecine, originates in the nineteenth century patchwork of states that emerged in South East Europe from the crumbling Ottoman empire. Following its comeback tour during the Balkan Wars of the 1990s, and its association with the ideology of "liberal interventionism", it took on a second life as a key neoliberal term of dismissive censure, the opposite of (positive) globalisation and shorthand for reactionary backwardness.
Many of the "Balkanisation of the Internet" stories can be read as simplistic neoliberal propaganda, both in its corporate and supra-national state forms. Thus US commentators emphasise the threat to the "free and open" Internet, which they equate with "modern economic activity", while Europeans emphasise the need for regulation: "protecting and promoting human rights and the rule of law also on the Internet, alongside its universality, integrity and openness". This "defence of freedom" versus "need for oversight" dichotomy is currently being played out in the Obama-Merkel spat.
One thing the Snowden affair has reminded us of is that the vast majority of espionage is focused on industrial and commercial advantage, not national security. It's about acquiring non-lethal technology and establishing negotiation vulnerabilities, not James Bond and poison-tipped umbrellas, despite the foaming nonsense in the UK about "gifts to terrorists". The fuss over the NSA in Brazil is more about spying on Petrobras than politicians, and the interest in Merkel's "handy" ultimately reflects German economic dominance, not her indiscretion. Of course, the term "national security" is routinely elided with "strategic economic interests", and there are plenty of stooges who will justify espionage as necessary when dealing with corrupt foreigners (conveniently forgetting the role that US economic imperialism has in encouraging that corruption).
The suggestion that the Snowden revelations may lead to the "break-up" of the Internet is absurd. The Internet is not a single public network, but a loose collection of private networks using common protocols (there's a clue in the name: the interconnection of networks). Short of abandoning TCP/IP, it isn't going to happen. Brazil's proposals to "nationalise" the Internet simply mean privileging its own corporations in-country, and securing regional domination across South America, if possible, much as the EU seeks to do in Europe. It does not mean your email server will be administered by a civil servant if you live in Rio, or that Brazilians will become "isolated". Even in China, the network is a joint venture between private companies and the state, and the censorship of content owes more to a social norm of self-repression than to technical impediments (which are not difficult to circumvent).
Predictably, conservatives have started wailing about "Soviet-style" repression and "national firewalls" crushing online freedom. Some neoliberals prefer the management consultant's argument that this will lead to inefficiency, raising the spectre of redundant bandwidth and hardware ("This would be expensive and likely to reduce the rapid rate of innovation that has driven the development of the internet to date"), but this is illogical. If nations decide to build their own server farms and lay additional cables, this will increase network capacity and spur domestic technology firms. Super-fast broadband would reach rural areas in the UK a lot quicker if we actually had a nationalised network.
The Balkanisation of the Internet trope is interesting because it comes at a time when the tech companies are increasingly driving users back towards the "walled garden" model of the 1990s. The rapid adoption of smartphones and tablets means that increasing numbers of users have little control over their Internet access. For most, this is a boon - they don't have to worry about that head-splitting tech stuff and their devices just work - but it means that more and more exist in a bubble: "What people are choosing is less an iPhone 5s over a Moto X than an entire digital ecosystem that surrounds and permeates their life, and which will affect every other piece of technology that they buy". Even ostensibly "open" platforms like Google (i.e not tied to particular devices) seek this control through the "filter bubble" of personalisation.
The major technology companies, from verticals like Apple through pure-plays like Amazon, are engaged in the "capture" of users and the partitioning of the Internet at the application layer, not at the network layer. But their ambitions go far beyond mere brand loyalty to exclusive control of the shop, which is (ironically) nearer to the "Soviet-style" monopoly of old. The EU's antipathy towards this, expressed in concerns over privacy and data protection (i.e. consumer rights), is a straight tussle between competing neoliberal power blocs.
The real threat to the Silicon Valley companies is the decoupling and fragmentation of the cloud, in other words, the repatriation of data assets. National big data is where the money is - or is presumed to be. The parasitism of the security services is incidental. While the companies could resist this pressure, the more likely outcome will be a deal with national governments. The state will regulate access to national data assets, which in practice means dividing the spoils between global providers, privileged national providers, and the security services of the nation state and its allies. This means the Internet will no longer provide such a pronounced "home advantage" to the US - though it will be a long time (if ever) before it becomes a genuinely level playing field - while preserving the relative commercial advantage of US corporations. The Internet will be no more broken tomorrow than it was yesterday.