Hi community! I was thinking it might be fun to hack on some IoT related projects using Tardigrade. I’m by no means an expert in this area, but IoT devices/systems seem to generate a lot of data which may make an interesting use case. Any thoughts?
5 Likes
boosting for notice – this is a thing @jennifer and I cooked up on a phone call and the team is really excited about – so please dont be shy, we would love to hear some ideas, and especially love to collaborate with community members 
I spent a lot of 2018 and 2019 staffing the company booth at conferences. and IoT was one of the most common questions people woud ask about, so am pretty sure there has got to be some great and innovative ideas pinging around folks’ brains.Even if its an idea or something small or large, we would love to hear. the more creative the better
2 Likes
As most IoT devices are light on compute, and network, I’m quite unsure how well this would work with all the crypt functions on the IoT device being done in a timely or power inexpensive manner.
That being said-
I could see things like driver logs or accrued sensor data being packaged up and thrown into tardigrade for LTS/log storage.
3 Likes
Can you name some more IoT applications where Tardigrade could be integrated?
Off the top of my head I can only think about stuff that wouldn’t benefit and doesn’t generate much data, much less data that should be stored somewhere else. In my opinion most IoT data should remain in my own network.
It might make sense for something like a driver log, yes and maybe all those IP-Cameras (if you consider this IoT).
1 Like
Not so much of a problem if your IoT devices “phone home” to a hosted gateway but yes, I’m also unconvinced about this particular use case.
Then again, maybe I just lack the imagination 
1 Like
I see the advantage mostly for anything that would create larger files like video or audio recordings. I’m not sure whether Tardigrade would be the best fit for the tiny amounts of sensor data that a lot of other iot devices collect. I have previously suggested Wyze as a possible partner in this as they have recently run into higher cloud costs than expected and actually went so far as to ask their customers to pay for something they had previously promised would always stay part of the free service. (They didn’t break their promise though, the payment was optional)
Though one of the things they do is analyze videos with machine learning to detect if there was a human seen in them. I’m not sure that could be done from Tardigrade storage as Tardigrade doesn’t provide compute. This service is now just for paying customers though. It would make sense to store data for non paying customers on Tardigrade either way and for paying customers they could still keep it on Tardigrade for long term storage.
Furthermore integrations with NVR systems would be very interesting. Especially for online storage of detected events.
Or simply software that can detect and record video events from an RTSP stream and upload it to Tardigrade. You could run something like that on a raspberry pi perhaps.
Of course it’s probably also possible to do all this on the camera itself for a very neatly integrated single package. Wyze cams may again be a nice option for this as they are very cheap and there is open source firmware available that could be modified to have Tardigrade uploads built in natively. And since you could already stream from Tardigrade directly through the link sharing solution, it would be fairly simple to let you view these events from a web interface as well.
So that’s a couple of suggestions around cameras that sound very interesting to me. If only I had the time to actually try to tinker with this myself. But hey, just throwing it out there. Someone else might want to pick some of these ideas up.
4 Likes
Well, I think tardigrade could be used by companies running the core system of a global IoT service storing small bits of data, but for millions of devices.
Like this service for instance:
2 Likes
I also feel like if it was unpacked sensor data or logging, we could be talking about a not trivial amount of objects that could actually make the object fees an issue more so than the actual storage amount.
1 Like
This sounds like something I had suggested before: Has Storj ever looked into the cctv market?
This market is so huge. This could and should actively be worked on.
2 Likes
Hi @jennifer
I can share my experience with IoT infrastructures, The IoT devices are usually small sensors with very low bandwidth but long-distance channels (for example LoraWAN), these sensors providing data for central collection and processing, then data is displaying on dashboards. For this purpose usually using time series databases like influxdb, prometheus, etc. Also, before useful information will be displayed it requires a lot of transformations with data before it will be a nice dashboard.
All these things require fast storage (low latency, high thruput) in the same locations where data will be processing. Unfortunately tardigrade is not the best choice for this type of task, I know only one type of task that tardigrade can perfectly handle in IoT sector: Video processing from CCTV cams, for example, police CCTV cams on the road or security cams on the airports (finding faces, numbers, etc.).
2 Likes
I have suggested this use case ages ago: Has Storj ever looked into the cctv market?
I don’t know why Storj is not acting on this…
1 Like
Who says they’re not. Companies usually don’t do their customer acquisition in public.
3 Likes
Maybe you are knowing more than I do.
They could do a simple thing like liking the topic or put some ‘working on it’ to it. If you check out the thread I have mentioned there is some answer from @super3 that has suggested, they have not really looked into it.
The market is huge. I checked with one single cam, producing about 3 GB of data per day. No take that by millions. There is an amazing number of cctv cams out there from gas stations to private one. This could produce a huge income for Storjlabs. But not even a word, if they are looking into this market.
@jocelyn: Maybe you can tell something.
This sounds exciting. Let’s hope for the best.