Looks interesting, some comments for you: How do you propose to solve https://en.wikipedia.org/wiki/Byzantine_fault_tolerance which you invoke with your "The system and all its components must embrace the possibility of failure and design for partial operational states." combined with your fully distributed design.
How much benefit is there in smart controllers? Can you measure the 'regret' between a simple thermostat and a smart controller? How much is this worth in $? I presume that the main reason for Nest is for data collection and techporn, rather than any real benefit from a smarter controller. Perhaps there is some value in load shedding? But that's going to require real contracts with real utilities, something beyond an open source project. Is a wired protocol the best choice for adoption? have you considered throwie like esp8266 + ds1822? How are you going to keep data persisted if all the nodes are unreliable? Of course wired is generally much more reliable, but most people are not willing to wire stuff up. I built my father a smart home 20 years ago using technology of the time. Something things I learnt: * Automatic light switching is not really very interesting, and hard to do well. * Traditional mechnical thermostats are more reliable than PCs. * When you have house-sitters your smart systems will break (also when you go away for a conference). * Wires + outside = corrosion. Even when you put inside a UL whatever rated enclosure, with silicone grease and silca desicant. * Everything will stop working in a way you didn't predict. * Electromechanical relays are about the only thing safe to use to communicate with other systems. Don't use opticalelectronic (SSR) ones. * Humidity sensors drift a lot over a decade. * Bird shit is corrosive. * PVC is porous to water. * UPSs break more often than the grid. * There are no battery technologies I can afford that last a decade. * The best user interface is no user interface. * PCs of that era were surprisingly robust (the original machine only died last year, when its 400MB drive died). On the other hand, it did cost more than 100W continuous power. It has been replaced with two beaglebone blacks, one of which only survived 3 months. * sending power and signal in the same cable is not really worth it. Transformers are robust and reliable. * use AC wherever possible to reduce corrosion. * irrigation pipe is much cheaper than electrical signalling. And more robust too. * Don't put anything in the walls, even if it saves you lots in cabling. Having said all of that. He still has it, he still likes that the chicken house door reports its state over the web and that he has a graph of indoor and outdoor temps, solar electricity production etc. But he's also a retired EE/Computer Science Professor with skill and time on his hands. My recommendation is to find a specific problem that you can show you can have a direct effect on with automation, and solve that simply and cleanly. Then find another one. And perhaps combine them. I want you to succeed (we all do or we wouldn't be on this list), I share your dream, but this proposal is too big to survive without such motivating examples. Alternatively, implement a light weight distributed consensus protocol in say Rust to run on embedded systems (e.g. esp8266) and become famous for that. I'd buy that. Good luck, njh On Sun, Feb 07, 2016 at 12:39:49PM -0800, Jerry Scharf wrote: > Hi, > > A while ago I had asked if anyone was interested in working on a new > building automation control system. There were some who were willing > to review and comment on it. I am going to avail myself of those > kind people. > > Here is a problem statement that I wrote up about how to move to > what I think of as a distributed building automation system. I would > love comments on this. This should probably happen off the list so > the spam is kept to a minimum. One of the pieces of this would be a > 1-wire gateway to the messaging system. > > thanks, > jerry > > ------------------------------------------------------------------------------ > Site24x7 APM Insight: Get Deep Visibility into Application Performance > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month > Monitor end-to-end web transactions and take corrective actions now > Troubleshoot faster and improve end-user experience. Signup Now! > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 > _______________________________________________ > Owfs-developers mailing list > Owfs-developers@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/owfs-developers ------------------------------------------------------------------------------ Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 _______________________________________________ Owfs-developers mailing list Owfs-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/owfs-developers
