That the internet works for so many people and across so many different technologies can seem kind of miraculous, but the internet’s founders had nothing like this in mind. When you consider that the early-days internet, circa the mid-1970s, was a solution to the problem of non-mobile, centralized supercomputing resources, its effectiveness for a world in which every yahoo lords over a half-dozen IP addresses like they’re pet goldfish is nigh unbelievable. And yet, here we are. Somehow.
While it works admirably, the internet also works suboptimally—this isn’t what it was designed for. So, we’re left with a good bit of room for imagining alternative internet technologies that might be closer to optimal and might come better equipped for addressing some of the emergent concerns of internet super-connectivity, such as privacy and the preservation of free speech. Enter named data networking (NDN).
Simply, NDN replaces IP addresses (locations) with named data (things), wherein a unit of data might be referred to in a way similar to the directory-based naming schemes we’re used to as PC users (as in, /Users/someuser/my_dir/file.txt). The whole internet would be structured like a big filesystem—a hierarchy of namespaces—where the most specific directory (from our perspective) would be our own local computer, while the most general directory (the root directory) would be the entire internet. As we traverse from our local machines outward, we access higher and higher directories as larger and larger subnetworks of the entire internet.
As such, the internet we experience is only as big as we need it to be depending up on the data that we’re after. If that data is on our own machine, we stop there; if not, we check our local network, and then we start checking router caches and CDN stores, etc. This would seem to make a lot more sense for an internet that’s based on providing information rather than one that’s based on enabling communication between network endpoints. It would also make a lot more sense for a network that’s designed to preserve free speech, according to a National Science Foundation-funded research study published in the current Communications of the ACM.
The paper, authored by a team of computer scientists based at UCLA, provides a concrete example of NDN in action via the Internet of Things, the looming internet eruption in which every toaster, car, and coffeemaker comes equipped with a network connection and IP address. We might reasonably question how much internet a coffeemaker actually needs to fully leverage toaster connectivity.
“For example, a manufacturer-assigned name, such as /local/appliance/kitchen/toaster/Black&Decker/serial_number, might be used to address a kitchen appliance from another device in the same smart home,” the paper explains. Another device might then connect to the toaster by broadcasting a sort of packet known as an “Interest,” which is basically just a request for data featuring a reference by name (as above).
“In this case, NDN enables applications to use the network layer directly to discover nearby devices in these well-known namespaces (for example, /local/appliance), without needing the devices to be connected to the global Internet,” the UCLA group writes. “At the same time, they share the same network layer protocol as all other NDN Internet applications, providing opportunities for straightforward integration with local or global Web applications, using data signatures and encryption-based access control for security.”
“NDN makes it easier than IP to share data via alternative communications paths and opportunistic connectivity without global infrastructure.”
Maybe it’s already clear how this might enhance free speech (and privacy). NDN facilitates the development of networks that may be connected to the global internet, but, at the same time, may also allow for insulated data transmission across local networks that don’t require the infrastructure of global internet providers. In this scheme, data packets can be stored and republished by anyone using any device.
“NDN makes it easier than IP to share data via alternative communications paths and opportunistic connectivity (toasters and phones as well as laptops and routers), without global infrastructure or complex intermediate services providing indirection or anonymization,” the researchers continue. “Users moving in cars or planes or people with ad hoc wireless on their mobile devices can exchange data via NDN by leveraging storage on their devices and intermittent connectivity to pass content around, without leaving traces of where the data originated.”
Current internet architecture is well-suited to censorship. It’s just a matter of whoever’s in power limiting access to forbidden websites via the aforementioned global internet providers. NDN, however, provides a mechanism from routing around these global providers by bridging local networks. Imagine that Comcast, in response to some completely hypothetical future American despot, is legally forced to block access to the New York Times website. NDN would allow us to chain together local WiFI networks until we eventually reach some router in Canada or New Free Cascadia or wherever that’s connected to an uncompromised ISP.
Obviously, this is just a high-level view, and actually reengineering the entire internet would be a hell of a thing. But we should take heart that privacy and free speech are not necessarily casualties of hyper-networked society. “By diversifying the nodes that can provide data,” the paper concludes. “NDN will likely improve conditions for free and anonymous speech and information seeking for consumers and producers.”