Almost two decades ago, I attended a lecture by an Internet engineering pioneer, Scott Bradner, justifying the Internet’s implementation of IPv6 (an IP address scheme capable of assigning an astronomical number of addresses per every square inch of our planet’s surface) to prepare for a massively connected world. I remember him offering pacemakers and toasters connected to the Internet as examples of what to expect. Two decades ago this talk solidified in my mind the then nascent Internet’s promise, to become the Internet of Things.
That was twenty years ago. And yet, my toaster is still not connected to the Internet and the world has yet to adopt IPv6, limiting us to IPv4’s modest 20 or so addresses per square mile. Many people think the Internet of Things is approaching quickly, but I wonder why it isn’t already here?
Today, the Things market is about $30 billion, so the Internet of Things must have arrived for someone. Until recently, industries have been the main consumers of Things. The Internet of Things world has arrived, just not yet for retail consumers. This, of course, is now changing as more consumer products make their way to the retail market, as standalone products or added features to existing products. Even still, my expectations of the pace of the Internet’s evolution to the Internet of Things over the last 20 years is that it would have been more rapid.
There have been real roadblocks slowing the pace of its evolution:
For one, while the protocols supporting the Internet are based on open standards developed through collaboration among knowledgeable stakeholders, the Internet of Things is not. Already, the industries and organizations making Things have developed over 400 protocols, many of them proprietary and siloed to their specific products. Some Things do leverage the sturdy Internet HTTP/REST/JSON paradigm, but many manufacturers rely on protocols developed specifically for their suite of products. Until more open standards are embraced, the evolution is going to veer more toward an Internet of Silos.
Another roadblock is security. The success of the Internet is in part due to the transparent collaboration to build secure protocols. These protocols, of course, are not perfect and require constant tweaking to keep up with the latest threats. Building secure protocols is not a trivial task and it is easy to inadvertently expose exploitable vectors. However, relying on a large international community to maintain the protocols, and the widely accepted Open-Source APIs behind them, has proven to be a robust and flexible response for the open standard community to uncovered defects.
Organizations and companies have their reasons for developing their own protocols and APIs. SInce the Things may not always have power (batteries rationing, solar powered or RFID tags, for instance), running a REST web server on the Thing may not be practical. The technology they develop may simply have requirements not met by open solutions, for instance more opportunistic Thing-to-Thing communication. But without the benefit of a large collaborative support community, it easy to see how a Thing can have a security defect for a long time before detection.
Security would not matter as much to retail consumers if it were not for privacy. Many would-be Thing consumers are concerned about other people and institutions, either maliciously or seemingly benign, gathering data from and about us. I suspect that many consumers, at least in the U.S., are wary of having a toaster and refrigerator communicating their frozen pizza consumption data back to the appliance’s manufacturers. This is a more of social rather than technological obstacle and is part of a much larger ongoing dialogue between individuals and those collecting data about them.
If the dam built of standardization, security and privacy finally come down, we will have another problem: a flood of data. Our Data is already Big, but releasing the torrent from massively connected devices will make it Enormous. It is likely the Internet of Things will finally deliver the IP-to-square-inch ratio we disserve and require a much wider adoption of IPv6. Even with motivation of being the Internet in the Internet of Things, it will take some time for ISPs to adopt IPv6 and provision the additional bandwidth required to collect Enormous Data.
So, when wilI the Internet of Things arrive in my world? All sorts of retail consumer Things can now be consumed, from automobiles to thermostats. If I wanted, I could now buy an Internet-connected toaster after all. But the problem is I don’t want to. I don’t need a kitchen appliance with a web server that compares notes nightly with my toaster and a data center on the other side of the world. Well, at least I don’t think I need one.