03-14-2020, 03:47 PM
Officially, we support one connection. This is mostly because we have implemented the remote access such that the server (phyphox) does not report back state changes (start/stop of the experiment) back to the client (web browser) if these were requested by the client. This means, that if there are two users connected, one of them will not necessarily notice if the other one pauses the experiment (although usually this just feels a bit laggy).
The real practical problem, however, is that the webserver is not multi-threaded. This means that phyphox cannot handle a second (or third) request by other clients while already answering the first one. On a fast device and network, this usually works for multiple users as each web browser patiently waits until phyphox eventually takes its request and if everything is fast enough, this works fine. But still, the number of failed request will quickly increase and at some point it will not qork properly anymore.
In practice, I would say that up to five connections work quite well. This of course depends on the experiment, the amount of data, the performance of the device running phyphox and the network. For this reason, we are also not considering a control panel at the moment.
The real practical problem, however, is that the webserver is not multi-threaded. This means that phyphox cannot handle a second (or third) request by other clients while already answering the first one. On a fast device and network, this usually works for multiple users as each web browser patiently waits until phyphox eventually takes its request and if everything is fast enough, this works fine. But still, the number of failed request will quickly increase and at some point it will not qork properly anymore.
In practice, I would say that up to five connections work quite well. This of course depends on the experiment, the amount of data, the performance of the device running phyphox and the network. For this reason, we are also not considering a control panel at the moment.