It's already hard enough to run a reliable mobile data network built with industrial-grade systems on exclusive spectrum. Unfortunately, moving to a network like you describe, built on commodity gear doesn't translate to quality, redundancy, and scalability like it does with distributed systems.
Major cell companies are already doing what you describe. Example: both Verizon and AT&T have a fleet of mobile cell towers mounted on the back of trucks they park at large events if they think capacity is going to be an issue. If you look around at building tops in cities, you'll see arrays of panel antennas on many of them. These are for cell networks. Every site is intentionally designed and every single antenna is specifically angled and tuned, and their network engineers collect quality statistics from every call at every site, which are fed into massive data warehouses for quality analysis. They literally have fleets of people like "the Verizon guy" continuously drive testing their networks.
If anyone knows how to run mobile data networks, it's them. Seriously, network engineers at companies like Verizon are treated like software engineers at Google and Facebook. They keep the core business running and they are the rockstars. Everyone else, including the jerk in customer service, is basically a nobody.
From what I can tell, there is a sort of strange, pervasive myth on the Internet that the large cellcos have their heads up their asses and that somewhere, somehow, there is a kind of wild band of rebels with hacked 2.4GHz/5GHz mesh network gear and cantennas doing it for dirt cheap.
Please show me even the smallest evidence that anyone, anywhere is running a true mobile data network using a mesh network on commodity gear. This means continuous coverage for a metro area with low power, cheap, antenna-integrated MOBILE consumer gear, seamless site-to-site and channel-to-channel handoff from users moving at 80mph, high quality indoor coverage in large buildings that is basically taken for granted, ramp-ups/downs of 10k+ customers within hours (sports events, etc)... and on and on.
You'll find a few WISPs doing fixed site-to-site links and (poorly) blanketing areas with WiFi coverage, but nothing on the order of what cellcos do. The fact that one can hold a uninterrupted voice conversation using a $5 phone with no visible antenna while driving down the interstate at 75mph for hundreds of miles is a pretty awesome modern marvel.
Yep, I think that the prospect of trying to achieve real cell phone like qualities out of mesh networks is a no hoper. The real question however is what the end consumer will accept. Often times products are over-engineered / overpriced (whether deliberately or accidentally) beyond what consumers really demand and it's not until a low quality competitor shows up that everyone realizes it. Think how the vast majority of consumers turned out to be happy with low bit rate mp3s after recording studios spent years pushing CDs, and how very few people really care about HD vs SD very much (at least, not on smaller screens). So perhaps there's room for a network that just doesn't work while travelling at 75mph (or even 30mph). So what? Maybe consumers just don't care that much if they are getting it for free.
The 75mph isn't really the hardest part of that. Really, it's indoor coverage with dirt cheap consumer gear that's hard, followed closely by call and data connection handoff. Even if it's free, there's a minimum amount of service quality to even be considered useful, and I think handoff and indoor coverage are both essential for mobile wireless to work. In modern networks, handoff is also deeply tied to availability in high density areas.
Major cell companies are already doing what you describe. Example: both Verizon and AT&T have a fleet of mobile cell towers mounted on the back of trucks they park at large events if they think capacity is going to be an issue. If you look around at building tops in cities, you'll see arrays of panel antennas on many of them. These are for cell networks. Every site is intentionally designed and every single antenna is specifically angled and tuned, and their network engineers collect quality statistics from every call at every site, which are fed into massive data warehouses for quality analysis. They literally have fleets of people like "the Verizon guy" continuously drive testing their networks.
If anyone knows how to run mobile data networks, it's them. Seriously, network engineers at companies like Verizon are treated like software engineers at Google and Facebook. They keep the core business running and they are the rockstars. Everyone else, including the jerk in customer service, is basically a nobody.
From what I can tell, there is a sort of strange, pervasive myth on the Internet that the large cellcos have their heads up their asses and that somewhere, somehow, there is a kind of wild band of rebels with hacked 2.4GHz/5GHz mesh network gear and cantennas doing it for dirt cheap.
Please show me even the smallest evidence that anyone, anywhere is running a true mobile data network using a mesh network on commodity gear. This means continuous coverage for a metro area with low power, cheap, antenna-integrated MOBILE consumer gear, seamless site-to-site and channel-to-channel handoff from users moving at 80mph, high quality indoor coverage in large buildings that is basically taken for granted, ramp-ups/downs of 10k+ customers within hours (sports events, etc)... and on and on.
You'll find a few WISPs doing fixed site-to-site links and (poorly) blanketing areas with WiFi coverage, but nothing on the order of what cellcos do. The fact that one can hold a uninterrupted voice conversation using a $5 phone with no visible antenna while driving down the interstate at 75mph for hundreds of miles is a pretty awesome modern marvel.