The Artemis pCell: Making Interference Work For You.

It will likely come as a shock to many to find out that Australia leads the world in terms of 4G speeds, edging out many other countries by a very healthy margin. As someone who’s a regular user of 4G for both business and pleasure I can attest to the fact that the speeds are phenomenal with many of the CBD areas around Australia giving me 10~20Mbps on a regular basis. However the speeds have notably degenerated over time as back in the early days it wasn’t unheard of to get double those speeds, even if you were on the fringes of reception. The primary factor in this is an increased user base and thus as the network becomes more loaded the bandwidth available to everyone starts to turn south.

There’s 2 factors at work here, both of which influence the amount of bandwidth that a device will be able to use. The primary one is the size of the backhaul pipe on the tower as that is the hard limit on how much traffic can pass through a particular end point. The second, and arguably just as important, factor is the number of devices vs the number of antennas on the base station as this will determine how much of the backhaul speed can be delivered to a specific device. This is what I believe has been mostly responsible for the reduction in 4G speeds I’ve experienced but according to the engineers at Artemis, a new communications start up founded by Steve Perlman (the guy behind the now defuct OnLive), that might not be the case forever.

Artemis new system hopes to solve the latter part of the equation not by eliminating signal interference, that’s by definition impossible, but instead wants to utilize it in order to create pCells (personal cells) that are unique to each and every device that’s present on their network. According to Perlman this would allow an unlimited number of devices to coexist in the same area and yet still receive the same amount of signal and bandwidth as if they were on it all by themselves. Whilst he hasn’t divulged exactly how this is done yet he has revealed enough for us to get a good idea about how it functions and I have to say it’s quite impressive.

So the base stations you see in the above picture are only a small part of the equation, indeed from what I’ve read they’re not much different to a traditional base station under the hood. The magic comes in the form of the calculations that are done prior to the signal being sent out as instead of blindly broadcasting (like current cell towers do) they instead use your, and everyone else who is connected to the local pCell network, location to determine how the signals be sent out. This then manifests as a signal that’s coherent only at the location of your handset giving you the full amount of signal bandwidth regardless of how many other devices are nearby.

I did enough communications and signal processing at university to know something like this is possible (indeed it’s a similar kind of technology that powers “sound lasers”) and could well work in practice. The challenges facing this technology are many but from a technical standpoint there are 2 major ones I can see. Firstly it doesn’t solve the backhaul bandwidth issue meaning that there’s still an upper limit on how much data can be passed through a tower, regardless of how good the signal is. For a place like Australia this would be easily solved by implementing a full fibre network which, unfortunately, seems to be off the cards currently. The second problem is more nuanced and has to do with the calculations required and the potential impacts that might have on the network.

Creating these kinds of signals, ones that are only coherent at a specific location, requires a fair bit of  back end calculations to occur prior to being able to send the signal out. The more devices you have in any particular area the more challenging this becomes and the longer that this will take to calculate before the signal can be generated. This has the potential to introduce signal lag into the network, something that might be somewhat tolerable from a data perspective but is intolerable when it comes to voice transmission. To their credit Artemis acknowledges this challenge  and has stated that their system can do up to 100 devices currently so it will be very interesting to see if it can scale out like they believe it can.

Of course this all hinges on the incumbent cellular providers getting on board with this technology, something which a few have already said their aware of but haven’t gone much further than that. If it works as advertised then it’s definitely a disruptive technology, one that I believe should be adopted everywhere, but large companies tend to shy away from things like this which could strongly hamper adoption. Still this tech could have wide reaching applications outside the mobile arena as things like municpal wireless could also use it to their advantage. Whether it will see application there, or anywhere for that matter, will be something to watch out for.