Machine-to-machine (M2M) communications is making life easy. Our coffee machines, washing machines and printers now have the ability to sense when we’re running low on coffee beans, detergent and ink – and re-order what we need automatically. Automotive companies like Mercedes-Benz have already showcased their self-driving cars at this year’s Consumer Electronics Show – complete with self-route capabilities, automatic detection/ prevention of hazardous driving conditions, and voice/video calling technology directly from the vehicle.
From the business side, M2M is transforming billing systems and big data collection capabilities. For example, beverage companies like Coca-Cola have installed “smart vending machines” allowing them to diversify their payment options, automatically conduct purchase analytics in real time and know exactly when and what to restock.
This, however, is only the beginning, and the future will look much different thanks to M2M. Within the next few years, M2M technology is expected to move from smart devices to smart homes, and eventually to smart cities.
Leading operators around the world are adopting Heterogeneous Networks (HetNets) based on a much tighter integration of cellular and Wi-Fi networks with multiple goals being reducing churn, reducing cost and generating new revenue streams. It may be early days to demonstrate widespread proven revenue gains but there is mounting evidence that the HetNet does indeed take a significant cost out of service delivery and provides tangible results in churn reduction by giving subscribers a continuous high quality experience wherever they are.
According to the respected industry analysts Rethink Wireless, mobile operators in developed economies are now convinced that seamless Quality-of-Experience (QoE) across multiple networks can shave as much as 2.2% off annual churn rates. This represents a significant 10% plus reduction, given that churn for a tier 1 operator is typically 17% to 20% per year. The potential saving scales with the size of the operator but if there are say 10 million subscribers paying an average $50 a month that would be as much as $132 million a year. Even a tier 2 or 3 operator would achieve significant savings.
The era of mobile data offload is morphing into that of the HetNet but is leaving one legacy, which is the fact that up to 80% of all mobile data consumed on mobile devices is carried over Wi-Fi. At the same time, there is a growing user demand for access to cellular services, whether to watch TV or just browse the web, while travelling and often, beyond the reach of a Wi-Fi hotspot. Furthermore this is quickly evolving from being a “nice to have” into an expectation, which is why making it work seamlessly combats churn and also attracts new subscribers.
But for the service to be both seamless and high quality, operators need real time information not just from the cellular network, but also from the devices themselves, in order to gain a holistic view of the user experience across all cellular and Wi-Fi networks. This objective, real-time perspective of the user’s QoE allows the operator to take a more immediate and proactive approach to monitoring and resolving the users' QoE issues. For example a user may be getting high retransmissions over the Wi-Fi network while a highly reliable cellular connection is available. In that case, the operator needs to switch transparently to cellular without the user being aware this is happening, resulting in a much better QoE. Ideally the user should not have to care or know whether the device is connected to Wi-Fi or cellular.
Another key issue is that up to 40% of the users QoE metrics can only be retrieved from the end-device and is not available in the operator’s cellular network. This is certainly the case for understanding how a user consumes services over a Wi-Fi network. Visibility into the Wi-Fi network usage, drawn from correlating device data with network data to analyse the overall QoE enables Operators to retrieve this information, and allows them to identify how best to improve it.
This leads to an ongoing debate in the industry about how to use Wi-Fi to drive further costs out of service delivery. If indeed a consistent QoE can be achieved when transitioning a user session between cellular and Wi-Fi, and the transition is transparent and seamless, then why not become much more agnostic in the RAN, using either cellular or, as policy defines. Certainly, policy can be guided by quality, as we’ve addressed. However, why not use cost as a policy metric. In other words, the transparent selection of the RAN may be driven by matching the consumer’s low price tariff to the lowest cost RAN asset at the time of network selection. In contrast, a user’s high price tariff could be guided by high quality metrics, such that that user will always get the highest performing RAN at that moment in time.
