Cloud Mini-Series Part 3: The Costs of Forced Localisation
This post is part of a series on the opportunities and challenges of cloud computing in the developing world. Previous posts covered the following topics:
ICTs are generally among the most restricted of all industries when it comes to forced localisation. They suffer from virtually all types of localisation barriers, from local production requirements to local content requirements. With regards to the cloud, the two most relevant localisation barriers are IT infrastructure requirements and data storage requirements (e.g. forced localisation). The use of those two barriers increased significantly in the past few years.
Local IT infrastructure requirements are designed to force Internet companies to establish data centres and other infrastructure within the country as a condition to be granted access to the local market. A number of countries have adopted such laws or are considering them, including Brazil, China, Indonesia, Nigeria, Norway, Malaysia and Vietnam.
Norway and Denmark, for example, introduced laws in 2010 that stop municipalities from using cloud services unless the servers are located in-country. Meanwhile in Brazil, the original draft of the ‘Marco Civil da Internet’ included a provision to force Internet companies to “install or use structures for storage, management and dissemination of data in the country”. That’s similar to a new law introduced by the Vietnamese Government this summer, which will require Internet companies to keep at least one server in Vietnam to service the local market. Indonesia, on the other hand, intends to go one step further than that: If adopted, a proposed new law will mandate all data carriers including foreign banks operating in Indonesia to establish local data centres.
Local data storage or local data residency laws, on the other hand, compel companies to keep the data they collected within the territory of that country. This type of regulation has been introduced or is actively considered in a number of places, including Argentina, Brazil, Brunei, Greece, India, Malaysia, South Korea and Turkey, among others. In many cases, countries link those laws with local data centre requirements. For example, the Indian government proposed measures that would force companies to locate part of their ICT infrastructure within the country (to provide police services with quick access to encrypted data on their servers) and would also require that data of Indian Internet users and businesses hosted on these servers can not be moved out of the country. While some local data storage laws are tied to one industry (e.g. health care or banking) others apply to any type of data.
The cost of forced localisation
It is no surprise that forcing Internet services to store and process data based on territorial instead of market considerations comes at a cost: It hurts consumers, businesses, entrepreneurs and innovation in general.
In an increasingly ‘territorial’ Internet, consumers have fewer choices since Internet companies, particularly startups and those servicing niche markets, shift their focus to key territories. The end result may look similar to other compartmentalized markets such as the copyright industries, where the high transaction costs associated with licensing significantly limit the choice and availability of content in smaller markets. Even where Internet services are available, the quality would likely be compromised as they could no longer use the most efficient methods to store, process and route their data globally. In addition to that, as services could no longer rely on a globally distributed infrastructure, they would become more vulnerable to power outages and extreme weather events.
Of course, forced localisation also hurts Internet entrepreneurs. In a balkanised market environment, fewer cloud services will be profitable and fewer innovations will be marketable. Large markets, such as those enabled by the global Internet, serve an important function for innovation. The connection between market size and innovation may be best illustrated by the famous quote by Thomas Watson, the former Chairman of IBM, in 1943: “I think there is a world market for maybe five computers”. The smaller the market, the harder to find money, the more difficult to make inventions viable. In other words, Hollywood blockbusters would have fewer special effects if they could not be marketed outside the United States and BMWs would have fewer gadgets if their sale was restricted to Germany. The need for large markets is particularly relevant in the online environment, which typically depends on scalability.
Another aspect to consider is that a large part of the Internet ecosystem depends on the sharing of personal data. Many services use personal data to make their products more useful, or to make the ad models that power the services, more efficient. In addition to that, if services could no longer share that data across territories based upon user consent, they would likely become less compatible with each other and create fewer possibilities for innovation. Imagine you are a Brazilian and you wanted an online subscription of the New York Times for your smartphone. Under the original Marco Civil da Internet, the Times could have been compelled to set up local servers to keep your private data, e.g. your credit card information, in Brazil. Even if the Times did that, the actual content would still be produced and stored in the US. Would that affect the way you browse the New York Times app on your smartphone? Yes, because in order to suggest more relevant articles to you, the app is designed to access your location and other private information such as your interests and reading history, a process which may no longer be legal under that regime.
More importantly, if the trend towards forced localisation continued, it would undermine the potential of the Internet as a general purpose technology that improves efficiency levels across industries. For example, the Internet’s ability to connect anyone, anywhere, has resulted in a truly global marketplace, which, in turn, created unprecedented opportunities for small businesses to specialise in niche products. In other words, larger markets make the long tail longer. To illustrate that, imagine you are a motorcycle aficionado from the island of Malta. The Internet has made it a lot easier for you to find second hand replacement parts for your 72 Harley Davidson. Localisation requirements would make it harder for Internet services to serve a global user base. As a result, the ability to establish global markets would be compromised.
Last but not least, forced localisation may actually make personal data less instead of more secure. Without the economies of scale of large data centres, Internet services may not be able to provide similar security standards in its smaller data centres.
To resume, forced localisation balkanises markets which, in turn, leaves countries with fewer and less performing cloud services. Ultimately, the development of a two-tier Internet would widen the digital divide and slow down innovation for the entire online sector. Instead of trying to lock cloud providers in through legislation, it would be better to let market factors guide the location of data. It would be better for users, better for service providers and better for innovation.
The next post in this series will take a look at the current communication strategy of the tech sector and argue that while the message is good, they may not have the right messengers.
This was the Third in a series of posts on the opportunities and challenges of cloud computing. The final post will examine how industry is communicating to policymakers in the developing world on the benefits and challenges of cloud computing.
Matthias Langenegger is Deputy Geneva Representative in the Geneva office of the Computer & Communications Industry Association.