Sending the Right Signals

This holiday season, smartphones will cross the chasm from upmarket to mass market and sales of the new wave of tablet computers, such as the iPad, will likely surge. That is great news for the global mobile industry, but putting such a sophisticated mobile device in the hands of hundreds of millions of people has major implications for the world’s mobile networks.

People love the seemingly always-connected feel of smartphones, be that persistent social networking status updates, the interactivity of turn-by-turn navigation or just the capability to stand in the high street, surf for the cheapest price for a product and then go and buy it without the need to trawl around every shop in town first. Having all of this and more in a device that fits in the palm of your hand can fundamentally change the way people lead their lives and is fuelling an insatiable demand for smartphones.

Global smartphone sales in 2010 will climb to 269 million, up about 55% from 2009, according to Gartner. Moreover, the research firm says worldwide sales of media tablets, with built-in cellular connectivity, will soar from 11 million in 2010 to 166 million 2014.

With the smartphone revolution well under way, the tablet revolution getting started and potentially billions of embedded connections still to come, mobile operators have some interesting times ahead of them. There have already been several high profile reports of networks becoming overloaded thanks to traffic generated by smartphones and USB dongles, sold with keenly-priced mobile broadband tariff packages. Less publicised, but equally significant, is the additional load that smartphones place on the signalling networks of mobile operators.

Many smartphones generate the ‘always-connected’ perception by appearing to have updates and information ‘pushed’ to them by the network, but in reality, smartphone apps tend to use a ‘periodic pull’ model instead – they connect to the network, check for updates, disconnect and then wait for a period of time before reconnecting again. Each connection generates signalling and data traffic, even if no updates are downloaded following the establishment of the connection. The latest tablet computers also tend to run apps designed to regularly poll a remote server via a WiFi or cellular network.

However, the heavy smartphone signalling load on the network could be reduced dramatically. A well-implemented connection and detachment process in a smartphone generates just one quarter of the signalling traffic of a bad process, and the difference in end-user perception being non-existent.

Standards body 3GPP has developed and documented the “Fast Dormancy” feature to optimise the signalling load, and here at the GSMA, we have produced a paper establishing best practice for implementing Fast Dormancy, which explains the nuances of moving between the different connection states a phone can be in – the exotically and impenetrably-named Cell_FACH, Cell_PCH, URA_PCH, Cell_DCH and Idle states.

In addition to implementing Fast Dormancy, there is the potential to better optimise the way smartphone applications use already-established connections. As the developer community is far removed from the nuts and bolts of efficient network operation, it is understandable that most apps aren’t yet designed to make best use of a network’s limited resources. After all, mobile apps are a relatively new phenomenon and many of their developers are relatively new to the mobile industry. But they and other players across the expanding mobile ecosystem need a broader appreciation of the growing symbiotic relationship between the network, the device and the application.

I discuss this topic regularly with handset vendors, but most of them aren’t yet making the changes that are needed to place fewer signalling demands on the network. They argue that their first priority is to optimise devices to meet the needs of applications rather than networks, as if the two are entirely independent. In fact, reducing the signalling load on the network would lead to a better app experience because it would reduce the chances of apps being hampered by network congestion and delays.

In time, I am sure smartphone and tablet manufacturers will become more appreciative of the benefits of Fast Dormancy and developers will pay more attention to how their apps use the network. My deeper concern relates to the whole new groups of connected devices, such as smart meters, in-vehicle telematic systems, environmental sensors and health monitors, which will likely come from a whole new swathe of manufacturers. These embedded mobile, or machine-to-machine connections, could ultimately lead to an order of magnitude increase in total mobile connections worldwide.

If many of these new devices follow the smartphone model of frequently and inefficiently establishing new data connections, then there could be a huge impact on the networks. Even if these new devices don’t connect as regularly or persistently as a smartphone, many of them will have very long life-spans, so it is crucial that they are designed up front to make efficient use of network resources.

The telecoms industry needs to communicate clearly the importance of efficient signalling to hardware manufacturers and software developers across the wide range of sectors that will benefit from tapping embedded mobile connectivity. Essentially, at the very rawest level, the fast-growing mobile ecosystem must treat wireless network capacity as a scarce resource, which should be used in as efficient manner as possible. Otherwise, everyone suffers.