Another data point on this. Today at FL100, between Valencia and Zaragoza, temporary 3G connectivity was possible for long enough to get a full set of METAR updates in Garmin Pilot. Direct from iPad Mini with its own SIM – no need to fire up a wifi hotspot on my phone.
A pretty sparsely populated area so didn’t expect it to work, but there is terrain to 6,000ft so only 4,000agl.
What is really needed here is an iOS/Android app that does all its own smart monitoring of the connectivity state and fetches METARs when it gets the chance, even if that chance is only a few seconds of connectivity. And then doesn’t bother again until they have expired 30 minutes later. And powers off the radio (flight mode) in the meantime to avoid interference.
Garmin Pilot does an OK job of fetching METARS when it sees connectivity but it could be more optimised.
Android (unlike iOS) also provides a low level SMS API, so it would be possible to fall back to that transparently if no GPRS/3G/4G was available. Using a nice compact binary format which should get you several METARS in one 140 byte SMS.
A pretty sparsely populated area so didn’t expect it to work
Nope. That’s exactly the type of area where it will work. Sparsely populated = few towers = no ‘lock-on’ problems.
In the past, most phones would not handle the situation of getting hundreds of responses to their cell tower broadcasts. That alone was enough to prevent them from functioning in the air. With today’s phones it’s much better. Last week, I had connectivity at FL200 above Munich — not really a sparsely populated area. It very much depends on the phone and when there is LTE, you’ve basically won.
Probably should have said “unpopulated”. No roads no towns nothing, but obviously a tower somewhere.
But anyway, what is really needed is an app that identifies these short opportunities and exploits them automatically. Which is back to the WhatsApp point. There was another 100nm that followed that was equally populated/unpopulated and for whatever reason gave no service (at least not when I was paying attention). And I don’t want to be continuously looking at my phone/tablet ready to pounce when I do get service. I’m sure the phone hardware is better than it used to be but that’s only part of the equation.
OK… I do know the one not very well known fact which is that the GPRS/3G/4G networks don’t get involved in TCP/IP (or even UDP) traffic. They act as dumb packet carriers, more or less.
So if you are on Vodafone UK and you are in Mongolia and you hit say euroga.org, all the server here sees is somebody in London, Vodafone. The client IP was issued (via DHCP) by Vodafone UK. Between Mongolia and Vodafone’s servers, the data is carried over the cellular network and over various fibre or whatever links.
So the whole thing should maintain the allocated IP fairly well, through breaks in the connection.
It has to work that way for various reasons, one of which is secure comms (HTTPS) which would break all the time otherwise.
But… the reality on doing this during flight is different, most of the time. There seem to be some tower handover conditions which do comprehensively break the IP protocol.
I think there is considerable room to exploit the situation, but most phone apps don’t do it.
Sorry about the thread drift since this is about 3G rather than 4G, and in the US. I’m not at all sure about the protocol issues described here, but it has been my experience that the availability of 3G at VFR levels in this US has steadily improved in the last two or three years. This May, for instance, 3G was available for a large part of a Coast to Coast routing, allowing email and critically, Foreflight weather. Here’s an example captured in flight:
I should say that this functionality is not 100% reliable throughout every trip, but on legs of 3-4 hrs was available more than 50% of the time seamlessly updating the weather picture and possible diversion plans. The weather picture displayed in Foreflight invariably matched what is seen out of the window and described by ATC and flight service. Metars and Tafs also updated seamlessly.
All this on iPad2 with internal 3G using a UK 3 PAYG Sim, (£15 from Toddington services on the way to LHR after the pre ordered sim failed to arrive), using the AT&T network with no extra charges for roaming.
How is the SMS delivery funded? They can’t do it for free – not in Europe for sure.
You can choose to make Hangouts the SMS app on your device. It substitutes itself for the vanilla SMS app.
I looked into Hangouts buit it doesn’t have retries so if you press Send when there is no signal, the message is never re-attempted.
My Q as to how free SMS can ever work is a different one. It used to be possible in the USA, IIRC, years ago. Never in Europe. There used to be advert-funded gateways; I recall one run by Lycos.
Quote
OK… I do know the one not very well known fact which is that the GPRS/3G/4G networks don’t get involved in TCP/IP (or even UDP) traffic. They act as dumb packet carriers, more or less.
Yes that’s correct, GPRS/3G/4G is layer 1+2 in terms of the OSI model, IP and TCP/UDP is on top of (tunnelled inside) those protocols but could be something else that is non-IP based.
Quote
Nope. That’s exactly the type of area where it will work. Sparsely populated = few towers = no ‘lock-on’ problems.
I think there is more to it that this. Cells are grouped into clusters – and a frequency is never reused in a cluster. So you could be high enough that you get a signal from two cells in two different clusters (same frequency and interfering with each other), or you could be high enough that you can’t a signal from anything. 200ft AGL you would presumably get great coverage of cells in one cluster and it would work just like on the ground. 200,000ft AGL you would presumably get zero signal. Somewhere in between there is a point where you are high enough that you have “too much signal” across clusters, then eventually as you climb you get to zero signal. Cluster size varies with population density (capacity requirements).
A sparsely populated area could mean no cluster frequency interference and therefore a good signal, or it could just mean zero signal because there are no cells in range at all. Certainly I have been in sparely populated areas with no service at all, then gained service as I passed a populated area.
“Too much signal” across clusters could presumably be solved with a directional (straight down only?) antenna. But “not enough signal” either requires a bigger antenna, or a lower altitude, or just waiting until you move to a different area.
I think that “No Service” on your phone means exactly that (no signal received at all) whereas “five bars” but zero data transfer probably indicates interfering cells. Not sure how to verify that though.
