GSM indicates the network architecture. The’G’s are simply a tool for marketing to help you pretend you are buying something better than the last ‘G’. It used to be governed as a spec from the ITU. Then when 4G service (which was spec’d at 100Mb stationary by the ITU) was being marketed and sold by carriers, and then delivering 11Mb. They didn’t care. So you shouldn’t either. However, on a GSM network, the voice protocol is different than its data counterpart, just as long as it isn’t VOIP. VOIP is an IP based protocol. Voice and Data streams can perform independently of each other.When I liked Microsoft, I went to a school where they taught us about TCP/IP and the OSI model. Anyway, there is a portion of a layer that remembers the order the packets were assembled for transport, but when they go, they get there however they can, show up when they do, and can be reassembled, verified by that same packet header, and used.
Voice packets,on the other hand, need to arrive in the order they were transmitted.Because of this, you can see that voice requires a constant qualitybitrate, but doesn’t need much throughput. They(the voice connections) are easy tomaintain unless a weak signal breaks the stream. That’s when you get a crappy call. Data connections don’t need a constant bit rate. There are different network profiles like aggregation, where if network load is low it will widen its frequency band to allow more throughput. There is burst transmission if it needs to send a lot now, and then slow down and cycle so another connection cantransmit because of high demand. As long as there is relative proof that QoS is intact, the receive buffer will wait for something.So it is possible to have data packets sent and received without the ability to have a voice connection.