With UMTS now potentially available on all the frequency bands traditionally allocated to GSM, why are we still operating GSM there while UMTS offers nothing but improvements over it and all contemporary handsets support it. The question is particularly pressing since data traffic has for quite a while accounted for more than 90% of network usage in volume and grows faster than backhaul can be deployed and cells made smaller while spectral efficiency has become awfully close to theoretical optima. GSM data modes such as GPRS and its incremental improvements have their purpose well, but they are hacks shoehorning data into a TDM voice world – nothing like the native capabilities of UMTS. Of course, modern marketing knows the value of nostalgia as an advertising vector, but I suspect that the market of users who insist on GSM for nostalgia’s sake may not be sufficient to justify its cost.

Some manufacturers nowadays offer unified RAN infrastructure that supports both UMTS and GSM on a single piece of equipment – and many antennas are now multiband, but there is still an awful amount of specific equipment with the associated duplicated costs… And then there is the effort of maintaining the software for two entirely independent systems, each with its own bugs, quirks and yearly upgrades attempting to squeeze more throughput out of a slice of spectrum that is not going to expand – a single large operator typically has dozens of people whose workload could be cut in half overnight. I for one would love to spend more time on GIS software for the fiber optics infrastructure and less dealing with the Jurassic park.

So what are we waiting for ? Don’t we understand that frequencies are too precious to be wasted on obsolete protocols ? Let’s recycle ! Let GSM retire ! Taiwan’s ministry of transportation and communications is already working on it…