Tuesday, September 22, 2015

Networks of trains and other things...

Networks of trains and other things...  This article is talking about (and bemoaning) the chaos on the Internet of Things (IoT) right now.  It's certainly true that at the moment there is a mish-mash of incompatible devices, APIs, and data formats.  I'm not worried about that, though – from that chaos, I'm confident that useful standards will emerge.  The crowd of early adopters will make that happen as they figure out what things really work, which don't, and which are most useful (or make the most money).

This has happened before, and even with networking.  Back in the '70s and '80s, when the Internet as we know it was just emerging, there was chaos in the world of networks.  Conflicting network protocols, data formats, and even hardware abounded.  The advent of Ethernet 10baseT and the IP protocol changed all of that.  Virtually everything else is gone from local area networks (LANs) today, and that pair of standards rules the world of LANs. 

Something similar happened in the 1800s with trains.  When trains were first developed, every manufacturer and even buyer would make up their own gauge (the distance between the rails), weight standards, etc.  This created all sorts of chaos, and spawned new businesses that transferred cargo from train cars of one gauge onto train cars of another.  A lot of ingenuity and capital were invested in these systems, until it finally dawned on the railroads that things would be much better for all concerned if they standardized on a single gauge so that railroad cars could go anywhere in the U.S.  Once accomplished, this completely eliminated that crazy set of businesses that were moving cargoes between incompatible railroad cars.  Today we'd say that was obvious, but it wasn't so obvious at the time – plus it looked prohibitively expensive and time-consuming to accomplish. 

A little earlier there was yet another example, this time with nuts and bolts.  Every manufacturer used to have their own set of diameters and thread pitches for nuts and bolts.  If parts from two manufacturers could work together, that was a miraculous accident.  The manufacturers had all sorts of fanciful justifications for why their diameters or thread pitches were better, but that was mostly nonsense.  In this case, it was the users of nuts and bolts who drove the adoption of a standard.  The users worked together to define a standard, then told the manufacturers that they'd only buy nuts and bolts that met the new standard.  Almost overnight, the world of nuts and bolts converged on that new standard.

The same sort of thing is bound to happen with the IoT.  Well, there's one other possibility: we'll collectively decide that the IoT is a useless waste of time.  I don't think that's very likely, though...

No comments:

Post a Comment