Main Site Documentation

Need some suggestion for sending high sampling data to the webserver


#1

Hi,
I wan to send 3 axis acceleration data to webserver.
The sampling rate at least 100Hz for continuously sensing for 3 min.
In first condition, this device will be used outside(such as on bridge), so no internet infrastructure,
Does GPRS fast enough for this or Do I need a 3G module?
In second condition, this device will be used in building.
If I use Ethernet module, what approach is suitable for sending data? (Webclient Post?, use socket? , or FTP client?)
Any suggestions I will appreciate

Thank you.


#2

Does it need to be real time or can you store the 3mins and then send to the server?

Is radio an option for both bridge and indoors?


#3

I think store before sending is better.
Dear Justin, what radio do you mean(Xbee)?, the radio may ok indoors
But in bridge, I need send the data back to server maybe 40 km or more.


#4

Yes Xbee or a similar alternative.
Can i safely assume the bridge is not sitting on a mountain top will nice line of sight to your server? 40km is easily achievable LOS with decent antenna…


#5

@ Justin
Most of bridges are not nice line of sight.
Also, long range 868 or 900 is not allowed in my country.


#6

@ Tzu Hsuan - thats a shame, 2.4 with high gain antenna can go aloooooooong way but needs to be line of sight.


#7

sustained data rates of 100Hz samples is almost impossible without store-and-forward, and for much lower data quantities than you can accumulate in that rapid sense period.

Even with a GPRS/3G connection you’re not likely to send a huge amount of data quickly, so you can’t assume you could push minutes worth of data back home in seconds. I don’t think that there’s any possible way to guarantee how fast that data could be sent because of the dependence of external network services that you can’t control, but you could always test this.

Is your situation where you want to store data in a circular buffer and when a critical event occurs, say the accelerometer threshold is exceeded for a time, and you then need to store and forward it as soon as possible? How close to the time the data is recorded do you need to have it back at home? And while the data transfer is occurring, do you still need to be recording data?


#8

Thank you Brett.
I know it is hard to do that in near real time.
Like you said, I need a store-and-forward approach.
I give up real time, I need re-think my approach.