In prepaid and other live billing systems the cost to bill a data session is higher than the cost to bill a phone call. This is due to data sessions having lon

In prepaid and other live billing systems the cost to bill a data session is higher than the cost to bill a phone call. This is due to data sessions having longer hold times (months) compared to phone calls (3-5 minutes). Since a session has a cost to the charging and rating platforms (memory, CPU and therefore software licenses), the more people use data, the more expensive it is for the carrier. As an aside, an SMS has the same cost structure as a phone call. They cost roughly the same on the signalling side, and roughly the same on the charging side. How much data is used also adds cost to the charging system. The charging system will perform periodic writes to persist the money spent so far. While this write will typically be performed once per call, it is performed every X dollars or Y kb or Z min on the data side. If you don't, the carrier loses money when any machine in the network goes down, tearing down the session. With shorter periods between writes, they still "lose" a little money, it's just not several months worth of charges. This is more cost pressure (increasing writes per minute). Historically, data was a minimal portion of a carrier's traffic. That meant that data would fit into the spare capacity of the voice and SMS charging system. Since data is now an appreciable portion of a carrier's traffic, they can't ignore it. Even worse, data usage is increasing faster than Moore's law (doubling every year instead of Moore at 18-24 months), which means that they can't buy newer hardware to solve the problem. They have to buy more hardware. There are solutions. Getting to a cost/subscriber model instead of a cost/transaction one would be a great first start. The hard bit is convincing product managers and salespeople that their beloved architectures and protocols don't fit the new world.