And, as someone who, until recently, traveled the country extensively as well as someone with some technical knowledge, I can tell you that from both a theoretical and a practical standpoint, overall neither is better. But, in certain local areas, obviously one will often be better, and that's the one you should use.
I'll agree in real-world practice, but according to base-station designers, there's significant differences.
Granted, they're biased since they don't build AT&T's.
I can tell ya this... AT&T's back-end is at least a decade behind VZ's, having worked directly with both. VZ had an all-VoIP long-haul a long time before AT&T even considered it.
I couldn't discuss any of this in my previous job. But I got direct access to all of 'em. RSA keys and logins to all the large carrier CO control networks and got to see how they all did business. AT&T was awful. Stuck in the 80s. VZ of the traditional telcos was the most advanced. Global Crossing and Level 3 were a long way ahead of AT&T too.
I chased an ISDN problem at AT&T for over six years. I waited while they tried to find someone who could read the traces. Tickets opened, tickets closed, no one cared, no matter how much their own management wanted it fixed, the Union / Bell System attitude was pervasive once you hit the core switching group.
Specialists were great but had no authority nor motivation to push on anyone for anything. Projects halted and never moved again after layoffs and they never re-hired. Many folks would tell ya flat-out they were just waiting for retirement.
Around here a number of areas were grafted on when AT&T switches ran out of capacity in the late 90s, and outgoing calls require a double SS7 dip. 30 seconds for an outbound call is rediculous. They haven't made any effort to fix it.
At some point, they started putting a slight artificial hiss into some cell networks because with the advent of digital, everything was so perfectly silent that people would start saying "Are you there?" or hanging up because they thought the call had already dropped.
Well aware of comfort noise. It made problems for me in teleconferencing for the last decade, since Cisco loves to put it out the trunk side of their devices, which is utterly and definitively wrong, per ITU spec.
This is not comfort noise. If it were comfort noise it wouldn't mute. It's subtle and only on ring indicator.
Comfort noise in a 200+ line conference call from 10 people hanging off a misconfigured Cisco gateway is a significant problem when the CEO is trying to be heard during his/her earnings release conference call.
VZ only has one total gaff in their back-end that I'm aware of. They chose UDP for their SS7 signalling.
That cost 'em a lot of money and time to design a separate QoS'd network nationwide just for SS7. But it's been done and working for almost ten years. Fun to do a trace and show the request never showed up, though.
But their back-end blows away AT&T's...
Our AT&T interface was still a DSU into what was essentially X.25 carried over a point-to-point data network which required specialized hardware for redundancy (V.35 physical relay switch that monitored a heartbeat created by a Solaris kernel module) and was god-awful to configure.
The new job both got me away from ever having to go to Central Offices first built in the 50s for AT&T and better, freedom from their techs who couldn't/wouldn't even configure a PC to do a 9600/8-N-1 connection to fix an outage.
I can't tell ya how many times I explained how to do that to multiple AT&T CO techs, who work daily in a building full of devices that used serial ports as backup.
You'd be shocked at how many times they couldn't figure it out, too.
Ask a VZ tech if they can console into something on a conference call, they'll say, "DTE or DCE pinout?" and then get their laptop connected.
Ask AT&T techs, they'll say "I don't have any way to do that."
Of course if AT&T engineering would have fixed their terminal servers that were racked, powered, and connected initially and then abandoned when it was found that certain Cisco 2600s had a hardware bug, instead of fixing them, we'd have fixed our stuff through those. Seven years. Never fixed it.
AT&T still used telnet everywhere too. VZ would block anything less than SSHv2 and high-encryption.
They're laughably different companies.
At the end of the day it was a motivation difference. VZ was trying a lot harder than AT&T internally.
By far the weirdest was Bell Canada. Super nice folks but culturally, nothing was ever an emergency. I had to explain to a NOC tech once that the outage we'd found (they had little monitoring) was costing them about $500K an hour.
After I asked that he page out to out VP-level contact he said, "Oh, you suppose he'd want to know about this, eh?"
Yeah buddy, I've got his cell number and would have called him myself but thought I'd follow procedure since you guys need to get someone rolling across Toronto since I know that's a dark site...!
Loved their techs though, nicest people you could ever talk to.
Worst tech ever? NYNEX. Called the CO with a massive number of T1s down. Guy actually said, "Ahh, that sounds bad. I better take my Union-guaranteed coffee break now. Hang on." He put the phone down for 30 minutes on a table in the CO. On the dot he came back, "You still there?"
Gave him the circuit IDs, he hollered across the room while covering the handset something about, "Hey what circuit did you jack into about 30 minutes ago?" All the T1s went Yellow then Green. "We don't see anything wrong with the circuit. How does your end look now?"
Telco gets old. Much happier back in the Linux server world.
I had never tried it. I'm sure someone has. What were your results? What altitude could you do it? I know people have successfully had a cellphone conversation in flight (I saw a pilot have a full conversation in 2001 while I was sitting right seat in a C-172).
The cell networks work largely the same in this context. You need signal to one tower for it to work well. AT&T's 3G network is actually a CDMA-based technology, so there shouldn't be much of a difference and IME there isn't - Some networks work better in some places, others in other places.
They all suffer from the "multiple-tower syndrome". Right?
