Reading the latest edition of the IEEE Communications Magazine, my eyes fell on an interesting article entitled “Five Disruptive Technology Directions for 5G” where the authors (Boccardi, Heath, Lozano, Marzetta and Popovski) probe five “hot” technologies that they consider have the potential of giving rise to disruptive changes in the way we build systems and design components (e.g. user equipment). Below are a few my own comments on their potential impact:
1) Device centric architecture (read: less base-station – terminal design, more D2D)
Been there, done that … This is a trend that has been discussed every 10 years or so since the 1980:s, but it simply doesn’t seem to fly on a larger scale. I have been in the MANET business in several of its cycles, and the experience is that it is not just a matter of more, and capable processing – it is a matter of complexity that we have difficulties in handling. Distributed networking has fallen short every time because of its lacking reliability and performance. We are simply much better at scaling up simple systems by adding more brute-force computational power, massive memory and wasting bandwidth (like in the current end-to-end, hierarchical, IP-based systems), than doing smart things involving complex logic, parallel processing with lots of interdependencies, interactions between many devices. This is why we have an explosive growth in applications that are build on cloud services – not because they are more efficient, but they are transparent and simple to manage – and work on any device. The place where it is happening is simple 1-to-1 application level connections in you home – connecting your table with your TV etc.
2) Millimeter wave
Spectrum scarcity – “Spectrum eldorado” … not again! Spectrum is scarce and expensive in wide-area macrocellular systems – since the alternative is to build more, very expensive macro-base stations. mm-waves are not suitable for that application. Indoor, short-range we do not see a spectrum shortage in the same way since we can densify the infrastructure at low cost. Hype – no doubt. Disruptive technology – not likely.
3) Massive MIMO
Promising (at least theoretical) for outdoor wide area systems when there is plenty of multipath to operate on. Even larger antennas are prohibitive, though. But how much can be gained in Ultra Dense Deployment in the indoor “capacity domain”, with several line-of-sight base stations in the same room ? At mm-waves the potential is of course greater due the small antenna dimensions. Hype – yes (=plenty of research papers!), but are there sufficient gains motivating the expense… ??
4) Smarter Devices
No hype, but reality! Moore’s law will keep us going. However, as already mentioned this not likely affecting the lower layers (e.g. D2D – see comment above). Local caching of content and other techniques to enhance the user experience – Yes, by all means! However, where the disruptions based on this will occur, has been notoriously difficult to predict. Where are the “Black Swans” ?
5) Native Support for M2M Communication
Definitely! The requirements for M2M and personal communication services are so different, that it is hard to imagine a “one-size-fits-all” 5G solution. Months/Years of battery time, extreme reliability, low data rates – but short delay seem to be difficult to reconcile with best-effort, high bandwidth systems. The question is how large the M2M market has to be to make a native support architecture viable, business-wise. There is a potential for many, many low-cost, high-volume but low-profit M2M-applications. Can they afford their own infrastructure – or will this in the wide-area domain be limited to applications that can afford LTE and a SIM-card? In personal communications we are now far beyond the need for a “killer app” (it has the IP-based smartphone), but is this true for the still immature M2M market as well ?