Given that malicious hackers could use IoT networking to access potentially dangerous devices -- perhaps setting them on fire remotely after disabling safety functions -- maybe "killer app" is not the best terminology.
How about an alarm system, home thermostat, or a sprinkler system? These are the kinds of things that would be useful to control and program with a PC. Right now many of them have tiny cryptic control panels in inconvenient locations.
As mentioned, security needs to be solid.
The potential for security threats, unauthorized access of data, sabotage and tampering, etc., is indeed a serious issue that needs to be addressed whether it is a home area / enterprise networks. To that end, the business model for HAN's is evolving and it explains partly why many consumers are pushing back on adopting M2M nodes for lighting, energy monitoring, security, etc. The payback is simply not there for the investment needed so M2M connected homes are often viewed as luxuries.
Good point @docdivakar...most people are perfectly happy with adjusting their thermostat when they get home not 30 mins before using their iPhone as M2M would enable...in fact there is some human in coming to a cold cottage and having to let the fire manually set...I can't imagine that that most my life activities like this would be per-programmed by me even if the technology was safe and reliable...Kris
The problem with many apps, "killer" or otherwise, is that a business needs customers now, not in the distant future, to justify investment. Many smart-building related ideas will no doubt be wonderful to have, but builders don't see these as making them more money. So those things have to be a consumer appeal, and be priced accordingly. And they have to work "out of the box", with no more programming than an iTunes playlist.
Security is also critical, of course, but hardware can provide that (currently at too high a cost...).