Dear community
I am planning a research project that includes monitoring of old buildings over several months or years. I really like the open ecosystem of home assistant compared to our conventional monitoring devices at the university with local storage and cable bound sensors. But before I start getting gear and implementation, I am wondering what experienced users think. The setting I am thinking of is a Home assistant Green with ZWA 2 and Z-Wave sensors to guarantee connectivity. Remote access should be provided by the official HA Cloud. I am wondering if this setup is chosen well and what the best solution would be to store the valuable data securely over longer periods and make it available for convenient analysis. Is there a database solution that I should prefer? I am happy to use a hosted soulution to keep thinks simple.
Thank you for sharing your insights
Jonas
You should really talk to your University IT first (I say this as a person who works in IT at a University here in the US). Depending on the options to connect to the University network, you may find it hard to get HA up and running and connect to HA Cloud. And since the Z-Wave devices may be using spectrum that the University is already doing stuff on, those could cause issues.
When you say âbuildings,â how big and how many. Unless they are very close together, you will likely need a Green in every building, and depending on the number of floors in a building and the placement of Z-Wave devices, you might need a ton of Z-Wave repeaters or multiple Greens in a building.
You also mention secure data storage. HA has been designed as a home automation system. While there is some security, I wouldnât call it enterprise level. That is something, again, you should review with your University IT folks.
Robust (verified) backups, data collection, reliable sensors, and bulletproof connectivity make for reliable data to substantiate your research findings.
Do you want to pin your hopes on having that happen over a period of years based on open source software that changes on a weekly basis, built by volunteers, has help forums full of bugs, and outdated AI slop being offered up as solutions?
What do you do if your data disappears on some glitchy SDcard? Start again? Extrapolate? Make up some data? What if it is stolen, or modified by third parties?
Are you going to spend all your time coddling HomeAssistant to work, or doing the research, and your IT infrastructure is a secondary but reliable incidental?
Have you seen how many reported issues on these forums have the letter âZâ in the title? You would think ithis form of networking comprises a native keeping two tins and a bit of string wet, constantly.
Your peer review will be scathing.
Hard NO!
After you have collected your data, and now have your thesis done, how will it be implemented to benefit humanity? Will the data be archived somewhere safe? Reproducible by other researchers that want to stand on the shoulders of giants?
Sorry to introduce realism into your world.
Donât reinvent the wheel.Talk to your prof before you proceed along this track. There is a reason things are done the way they are. Their role is to gently guide and mentor you to stumble across learning, without making your path extra difficult.
Youâre doing building physics, not CS arenât you? Keep focused on your primary goals. HomeAssistant will drag you down and be a major distraction.
Best of luck!
My entire day job is selling technology in education.(yo Kyle DM me we should talk⌠
)
I GUARANTEE there is a reason those sensors are setup the way they are and whomever said talk to It and your professor /chair/ whatever first is correct.
Now if you go talk and maybe ask hey does anyone know why these are setup this way⌠Youâll probably get reasons or⌠An opening to suggest an alternative but Iâd say if you havenât done that⌠Hard stop about face go talk to someone first.
Because really, education IT is under siege. (this is not an inflated number itâs verified by most major sec vendors and being thereâd I agreeâŚ) 80%+ of ALL verified cyber attacks inthe us are currently being directed at education. Why? Bad guys know theyâre a relatively soft target with a shoestring budget. What that does is makes EDU IT guys VERY anxious. I have been walking halls with one of my clients while the guy is walking around literally ripping âshadow ITâ (like cat 5, routers entire unauthorized servers etc) . Out.
If you show up with unvetted gear wireless gear and no solid security and data retention plan. Youâre done on day one.
And if your university letâs you without asking⌠Then thatâs a completely different problem.
When I put together systems to assist research thereâs long talks about the requirements and what IT is ready/willing to do. If youâre talking about something you say will be instead maybe years⌠You need to have those talksâŚ
Im not saying HA may / may not be your answer but you need to ask questions first before you go in with a new great idea.
Thereâs usually research funding involved. What is your budget? Who pays? Have you done that portion of your thesis submission yet?
Real life example that happened in a previous life in the corporate world:
Hobbyist installs printer buffer for third party unsupported software to make fancy color presentations for branch office boss that take a long time to print. Printer software is not on list of corporate standard IT supported platform. The branch boss hid the expensive color laser printer under a miscellaneous category for corporate purchases at the branch office. Head office deliberately kept out if the picture.
Secretary buys a cheap router from tea money slush fund so she can print from her PC as well as the boss, and of course the hobbyist too. Not too sure if he is a cousin of the boss (they often are). He (wisely) selects a different IP range so not to interfere with the corporate network - you donât want corporate IT worrying about errors and problems - the hobbyist will take care of it. You donât want others in the building using expensive color toner and overhead projector consumables. It is all very territorial. Even the auditors are kept in the dark.
All working well. Branch boss impresses fellow MBAs with colorful handouts, and overhead slides. Hobbyist gets to print his Pamela Anderson and Lamborghini posters in glorious color for his man cave. Secretary prints posters for kids canteen.
Boss wants to do some last minute updates at a remote conference hotel room. Secretary buys a dialup modem so he can dial in from his home laptop. Hobbyist connects it to spare phone extension that supports direct dial in.
Boss dials in from his hotel room, the night before and on the morning, makes a few changes, and secretary gets the revised items to him just in time for his great presentation. Phew!
Military police contact head of corporate legal in head office. People in uniform are politely ushered into the conference room on the top floor to meet the board. Somebody from our branch office is attempting to hack a naval research establishment. They have the logs to prove it. Both during work hours and, more alarmingly, out of hours. There is hushed talk of hacking, terrorism, and national military espionage. Head office, alarmed, calls head of IT to investigate, PRONTO!!!
We jump in police escorted vehicle and turn up at branch office, unannounced.
Yes, there it is, in the corner of the bossâs office, not even wired neatly, with wires stuck down with duct tape across the carpet connecting everything. Foolishly, the hobbyist, noticing a growing gawking crowd, steps forward and asks what is going on with âhisâ equipment.
Seems like the hobbyist unknowingly selected the same IP address range as the military use and when dialup, local network, and third party router started chatting across the internetz, alerts lit up all over the place.
Heads rolled. Lots of beer and it was brushed over. Hobbyist âofferedâ a job in IT department, with lots of courses on networking, however starting out the back in the warehouse. Branch boss and secretary offered âre-trainingâ.
Boss of corporate IT somehow ended up with all the confiscated equipment, including the souped-up PC with added memory and super fast graphics card. Not sure where the cheap router and dialup ended up, but I made sure they were reset to factory settings before they left the building. Somehow the secretary sent all the consumables they had stockpiled for the color printer and now WE had all the colorful handouts to present at conference time.
Corporate routers, firewalls, and network promptly upgraded so internal routing is locked up TIGHT and no defaults left open. Board approved cost and security upgrade plan we had been pushing for a long time in minutes. Backup and data disastet recovery plan to be externally audited.
Corporate memo about third party computing addons issued to all staff. HR advised to add instant dismissal to corporate policy for breaches.
Yes, this was a sin of ignorance. We shuddered to think if it was a sin of wilful attack.
What worries me is not the ones that get caught outright, but the ones THAT DONâT, and the corporate or research data that is modified, for the benefit of a small number of interested third parties, without anybody knowing, over a lengthy period of time. Purchase orders, student results, budget papers, research findings, payroll data, corporate goals, patents, intellectual knowledge, preferred suppliers, etc, all untrusted. Did somebody slyly mention that catchword BLOCKCHAIN?
You do NOT want that happening at your corporate, research, military, or educational establishment!
[All for the sake of expediency, and a few dollars saved, up front, but not in the long term]
Play by the rules. Document everything. Get it vetted by wiser heads and approved. Stick with your plan and check often if still appropriate.
Thanks a lot for sharing the insights, I had probably done half of that without feeling like doing anything wrong. However, I would like to explain my situation a little better, maybe you can help me to find a better solution. The project is about three old factory building where we want to know how they could be used for other purposes. We want to learn, how to build a low price, low environmental impact and flexible house in house solution that can maintain a certain levels of thermal comfort. I am only handling the building physics part, which covers monitoring temperature, moisture and maybe some radial temperatures etc. and developing guidelines for thermal comport management. I am there tempted of having a home automation system in place for monitoring, that could later be used to implement smart heating or similar.
The devices that we usually work with have a local storage, which means that I would need to visit the sites every now and then (they are quite far from each other) and hope there was no interrupt in the mean time. Online access would be a huge plus, this is what I meant by securely storing. I want to be able to remotely check if the recording still works. May this be over an existing network (I think WLAN is available) or mobile data. I would also be ok with LoraWAN or even old fashion SMS. I consider the data not to be sensitive, we would other wise have the logger somewhere without extra security where everyone could steel or manipulate it. We trust in human kindness 
Cable bound sensor also constraints for monitoring larger object like thees are. Plus we see the problem, that the supplier of our system just has a limited portfolio of sensors, which lets us with a âletâs work with what we haveâ situation. In a similar project I did before, I wanted to use a presence sensor. No chance, with the existing devices, I had to use some raspberry Pi solution next to the other data logger. And there goes my works-out-of-the-box setting.
I am aware that HA is not built for monitoring and especially not for any high fidelity application, but I would love to have a flexible solution that can be integrated with whatever other systems and sponsors we want to use. You are right IOT7712, I am not a computer scientist, just a user.
I am grateful for every hint or input! Thank you again for the feed back so far, I really appreciate it!
No worries. May your IOT device batteries not go flat the week before your annual data collection is my only closing comment.