GPRS basics

I'm looking for some basic info on using cellular modems. I have spent some time googling and I wanted to make sure I have understood what I have found so far.

My understanding is that on 2G GSM networks it is possible to have essentially a point to point data call between two subscribers. This is a CSD call and limited to 9600baud.

Now that we have 3G and 4G, CSD is being dropped by a number of service providers.

On 3G GPRS systems, data comms is basically done over the internet using normal TCP/IP protocols. This means a Client Server approach and makes data calls from subscriber to subscriber (Client to Client) more difficult unless both ends are permanently connected (which may be impossible with service providers preventing it).

The normal method is that subscriber (client) connects to a Server (PC on the net) and data transfer takes place.

In a situation where the remote node is a battery powered device requiring it to be in sleep mode most of the time, it cannot be contacted by the Server or by another cellular modem. The get around for this is to send a text message to it to make it wake and call the server (or a voice call that is dropped and not answered - the ring wakes the device).

So essentially, as 2G becomes obsolete, mobile data comms becomes IP.

Texting remains a simple method to pass small amounts of data about.

Is this more or less correct?

If you have two cellular modems and you dial one with the other, do they connect and provide a transparent data link, like old fashioned PSTN modems did? Or will they only do this if CSD is supported by the carrier?

In my application the remote device with the modem will be asleep most of the time, waking every so often to gather data. At some point the device must be contacted by a central station and it's data uploaded. To do this the central station must ring or text the remote device to wake it and then the remote device establishes an IP connection to the central station to deliver it's data.

Thanks for your help.

Cheers.

richardtheboffin:
My understanding is that on 2G GSM networks it is possible to have essentially a point to point data call between two subscribers. This is a CSD call and limited to 9600baud.

On 3G GPRS systems, data comms is basically done over the internet using normal TCP/IP protocols.

So essentially, as 2G becomes obsolete, mobile data comms becomes IP.

In my application the remote device with the modem will be asleep most of the time, waking every so often to gather data. At some point the device must be contacted by a central station and it’s data uploaded. To do this the central station must ring or text the remote device to wake it and then the remote device establishes an IP connection to the central station to deliver it’s data.

A few corrections and comments:

High-speed Circuit-switched Data (HSCSD) gives 14.4 Kbps per channel and can take up to 4 slots giving a throughput of 57.6 Kbps. However, for operators, it uses a lot of channel resources and so is expensive to provide.

GPRS is not a 3G mode but is implemented by 2G networks. It is commonly called 2.5G. It uses the TDM slots, like HSCSD, but gives up to 21.5 Kbps per slot. In practice, if a complete transceiver is allocated to GPRS, 115.2 Kbps is available to be shared by the GPRS users.

GPRS introduced IP networking to cellular and this has been continued by 3G networks.

For your application I would try to stick with IP comms. Can’t you implement Wake on Lan (WOL) to wake up the remote device?

Thanks for your comments.

I'll look into wake on lan options on my modem.