Minded Systems

Sorry IoT hubs, its not personal

The “Internet of Things” (IoT) market has been around as a bit of a wannabe-mainstream consumer product category for a while now. Sadly it continues to go through the same old cycle; shiney new up-n-comers wow an audience followed quickly by total non-adoption by consumers. Generally speaking, the main show-stoppers have always been (in one form or another) compatibility and security; or lack there of.

Most IoT systems and home automation kits are built around the general idea of a hub-and-spoke architecture. In recent years there has been near fever-pitch debate over the ideal IoT hub. While I respect that this sort of centralized system works and can be made viable, its fundamentally backwards to the nature of IoT systems and how they [will] exist in real life. IoT is a collection of various ‘things’ that are intended to work together. This means that they will inherently be disparate and unpredictable. This means that any number of devices of any given type may or may not be present at any given time. I personally feel it also provides distinct and important influence to the way IoT systems shoud be designed.

Central command-n-control systems for IoT devices, more commonly called ‘hubs’, are often seen as a necessity. While this works it is not a requirement. More importantly, a single device that has full access to all connected devices is huge security hole just waiting to be exploited. To date the alternative has been to have a slew of vendor supplied apps for all the various devices in place. This is the situation IoT finds itself in; stuck between incompatibility or insecurity.  As is the case with most of the new and “magical” features we have, the solution is rooted in old technology.

The assumption that a central hub is required for good and managable IoT needs to be challenged. I would like to offer the evolution of music sharing in the 90’s and early 2000’s as a comparative example. Early internet music sharing made it’s first real big splash in mainstream culture with the coming of Napster. The model implemented by Napster used a central database for controlling all of the file information required for the service to operate. This centralized system was ultimately taken down by a single crushing [legal] blow. In its wake came BitTorrent, the same basic premise is the same but built in a decentralized way. File information and currently available locations (called seeds) are maintained in individual torrent files that are disparate from each other. Sites then house collections of torrent files availabe to them for use and not necessarily the files they point to, often with many sites housing the same torrent files for redundancy. This created a decentralized network of files offered by a decentralized network of sites to facilitate a very resillient service. How can this possibly have ANYTHING to do with IoT?

Lets assume that IoT device access or permissions are the desired music files (or seed locations), for the sake of comparison. The current IoT landscape that assumes a central hub controller would be comparable to the central database for Napster. The alternative to Napster’s central database was disparate torrent files. For the comparison to IoT suppose that the torrent files are the IoT devices themselves. To complete the example, suppose that our personal devices and services are the various sites housing collections of these disparate torrent files. Reconstructing the scenario, IoT hubs are ultimately going to be crushed by a single [likely security] blow. In its wake will come a decentralized IoT where device information and permissions are maintained on the individual IoT devices themselves, disparate from each other. Personal devices (phones, tablets, etc.) and services (cloud automation, virtual assistants, etc.) then house collections of IoT devices available to them for use, often with many devices and services having the same IoT devices available to them (more for ubiquity than redudancy though).

One of the most important benefits of decentralizing IoT systems is the removal of a barrier to entry. Most hub devices are $200 or more making it a potentially cost-prohibitive environment for most consumers especially considering that an IoT hub is essentially useless without any ‘things’ to ‘internet’ with. Furthermore, it’s becoming quite clear that IoT adoption will not happen overnight. It’s more likely that as old home appliances and gadgets need replacing, IoT capable replacements will be widely available. This means that many homes may be an internet of one thing for quite some time making an entire system hub dedicated to controlling a single smart coffee pot a tough sell. Allowing everyday-users to adopt IoT at their own pace without the need for any complicated home infrastructure seems like an easier path regardless of the technical difficulties.

Please provide a valid email address

Trackback URL

http://www.minded.ca/2016-06-15/sorry-iot-hubs-its-not-personal/trackback/