A couple of years ago I played around with some Mac OS X clients and a packet analyser in an attempt to gain a better understanding of what was going on between the clients when using AirDrop. I was intrigued to see if there was any interaction with local wireless infrastructure. We had been receiving assistance requests from School IT Teams regarding wireless performance where they attributed the issue to student use of AirDrop. So I took a peek at packets and frames to see if I could uncover anything that might be useful from a vendor level to better co- or inter-operate with AirDrop.
I was surprised by how efficient the data transfer was. Inter device file sharing seemed so smooth and transfers were fast. Should I have been so surprised? Of course cutting out an AP (or a hop in a single duplex network) will improve performance.
Apple’s laser focus on end user experience bypassed the thought that vendors might struggle with the concept of co-inhabiting an airspace where clients have direct wireless interaction. The implementation of AirDrop was relatively user friendly and has only improved with the integration with iOS devices. It’s this user centric vision that lead me to my next thought...
What if Apple took this inter-client interaction a step further and made the AP redundant? If there was enough bandwidth, how clients interacted or "meshed" together a network would be largely immaterial. One or multiple devices might backhaul to a WAN or local infrastructure link some how but the access layer could largely be made up by clients directly connecting to each other. It just so happens that the future of 802.11 may have that bandwidth already in the pipeline with 802.11ad, ax and ay (and whatever may come beyond those). There are potential lessons which MU-MIMO may bring to the technology and the full capabilities of 802.11ac may never be realised (8 spatial streams, 160MHz channels) between a client and an access point.
Inter-client meshing overcomes some of the issues which are inherent in hyper-dense situations, such as perfect AP placement (physics vs aesthetics) which may counter channel contention and other problems. The limited range of higher frequencies (e.g. 60GHz) would be beneficial in that micro-cells would be a given. Of course it adds new complexity and completely changes the paradigm away from the current centralised, IT managed access layer. You would hope for some smart channel management and shortest-path intelligence but there is merit in the concept when you think of it from a consumer benefit.
MU-MIMO as it stands would be almost impossible to implement in a mobile device to device scenario because of the potential for change. But we know Apple are looking in to their own wireless technology chipsets which gives them $billion-R&D scope for proprietary innovation.
Li-Fi had us excited as a concept for the future of wireless communication. Such a major shift in the way our industry's fundamental technology works would be awesome!
We build and optimise networks. Continuous learning is our secret to being good. Along the learning journey we will share things here...