Sarah Bailey: Are current telecoms networks fit for purpose?
Miguel Ponce de Leon: No they are not. Current telecom networks lack an adaptive network architecture that is capable of supporting a wide range of device capabilities, multiple connectivity models and heterogeneous application areas with their specific requirements.
The current architecture is a design from the 1960s, which has been supporting a variety of applications till now. But emerging applications demand better quality, programmability, resilience and protection. Any alterations to the architecture now have become restricted to simple incremental updates and plug-ins instead of radical changes by introducing new solutions.
Fixed access networks are hitting 2Gbps – is the increase in speed delivering more value? What are the downsides (beyond cost) of offering ever more access speed?
The increase in speed is heightening the customer’s expectation that all their devices in the home will be served equally, from those video on demand flows, to the game console flows, to the live communication flows. However, the increasing number of devices in the home is causing major congestion on the fixed line access and thus not ultimately delivering any more value to the end customer.
If not speed, then what should we invest in?
Time and resources should be given to the re-architecture of the network, fundamentally looking at the Internet Protocol (and TCP) to see its flaws for what they are and to build into the software defined networking culture a thought process of rebuilding and reshaping from the ground up.
We need to be giving fully customized network capabilities to the different devices and/or the application areas using the network (such as e-health, self-driving cars, consumer electronics, home automation, content distribution and ad-hoc mobile networks).
Are the Internet of Things and virtualization glossing over underlying network problems?
It would not say glossing over. I would more say that they are exposing network problems. The flat nature of IP (v4 and v6) addressing means in an IoT world routing tables have to grow to accommodate accessible devices. We have already seen one major fault with 512K router table sizes: the network is catching up with Moore’s Law, and there may be nowhere to turn to.
As for virtualization this will also expose flaws in the way virtual network functions in a service chain can be connected. It will be clear that a new approach to facilitating a mapping between virtual network functions and distributed service functions is needed, one that can support a richer and more flexible means to model NFV-aware infrastructures will be needed.
What will be the key new technology or trend that will shift operator thinking towards investing in Network QoE?
Web Real Time communications (WebRTC) and contextual communications for all users. We are rapidly moving away from the world of a simple person-to-person phone call. It’s now an immersive online experience for all consumers where the context of a conversation is shared through audio, video and live media within an application. This world will bring a ton of pain to network providers if the current status quo on network deployment stays in place.
Miguel Ponce de Leon will be at participating in a panel discussion, ‘Evolving QoE Based Networks to Enable the Future Internet,’ on Day One (Tuesday 20th October) of the Broadband World Forum.