Improbable Icon

Interested in using SpatialOS to process and store Iot/AR data

general

#1

We are interested to use SpatialIO to store and process IoT and AR data, for example point cloud building data, depth capture data, sensor data and to use SpatialIO for distributed computing on those data. Basically using this infrastructure modules for storage and distributed computing, and to develop workers to visualize the data in AR/VR.

Here is an example that use AR to show IoT data from a machine, the same could be done for a building for example

We specialize in AR and have tons of tech demoes and ability to quickly prototype (www.arcortex.com) and can handle IoT offerings as well, but we have very limited experience and time to absorb all of the concepts of SpatialOS to test such a concept quickly, so we are looking for guidance to take the optimal path or some partner interested in this concept that might be well versed in SpatialOS but not in AR

Ideally we would be looking for some Unity templates for the worker and the games modules where the game module would store and serve data and the worker would be the client to source and request those data. We also could use some expertise in SpatialOS for finding out if there are major issue / deadlock to achieve this

thanks for chiming any advice you might have and sorry for the newbie questions

Yohan


#3

I know that we’ve already spoken briefly about this but thought I’d post my answer here as well so people can see my initial thoughts!

IoT data isn’t an issue. Essentially, anything you were using a single server for before we can absolutely supplement and increase by an order of magnitude. In terms of AR data, we can’t do any GPU rendering if that’s necessary so you’ll still be limited by the client and wouldn’t necessarily want to take info from a server in that case regardless since it introduces mild lag which is very bad for an AR/VR environment. However, anything that isn’t directly related to the image would be fine.

Essentially, think of SpatialOS as a much, much bigger server than what you could traditionally use and extrapolate performance improvements from a boost there.

Still, I’m not an engineer so can’t go too in-depth on the above but I have a feeling one of our team might get back to you with more techy talk :wink:


#4

Hey Yohan thanks for your post. I’m a developer in our solutions division, so we specialise in applications of SpatialOS outside of gaming.

First of all this sounds like a very good use case for SpatialOS, particularly with regards to visualizing all this data.

Would this data be streamed in live? Or are you thinking more of processing and visualizing existing data, eg. that is currently stored in a database?

There are a couple of ways you could do either. Basically all you would need is some adapter that can read your data. This could be done offline, to create a “snapshot” of the world that you can spawn a simulation from. Or you could create a worker that runs at runtime in the simulation, which can dynamically read your data (either from a live source or a database), then create entities in the simulation world based on this data.

Then all you would have to do is develop a Unity Client that can visualize the world spawned from your data. Or if you wanted to do some processing on this data, you could create a worker to do this as well.

We have worker SDKs in Java, C++ and C#, so you could write a worker to get your data into the simulation in any of these languages.

We have done similar work to process open source map and mobile cell data and spawn this into a simulation.


#5

Thanks for the post! Could the simulation be containing the data without having a database first? For example some worker could create some data from devices and sensor and send them to the simulation, the simulation would keep those data while running and serve them to client worker visualizing those data in AR. Could those data be point clouds?

Would you be available for a call to have a quick back and forth about this?