This is purely speculation on your part though. Also, isn't the entire point of peer to peer that very little data is in fact reported back to the servers? I know people used to be able to hack themselves indestructible ships and whatnot purely because there was no server authentication of the values their client was reporting (though I don't know how much of that is still possible nowadays). This doesn't explain why you brought up playable area either.
It's worth pointing out that the issues with the "netcode" go beyond "I had lag and/or couldn't connect". There was an issue with wings for a while where certain events could cause the game to hardlock for everyone but the wing leader. It's still possible to get that old bug where having two wing members at a station causes one of them to see the low detail version of the station interior and prevents rotational correction from working. I've had a bug occur a few times in multicrew where the helm sees the NPCs in a signal source disappear in the middle of combat (accompanied by the high wake sound effect) but the gunner and fighter can still see them and kill them (though they'll see each other shooting at different locations while striking the same ship). It's not uncommon for multicrew sessions to desynch entirely and have the crew remain in supercruise while the helm drops to normal space or vice-versa but without a disconnection occuring. It's hard to see how any of these issues would be fixed by spending far too much money on a router, to be honest.
Guess that depends on your definition of "too much money" - and I did not say it was the be-all-end-all cure to every problem, only that it can be one of many problems.
And it's not as much speculation as you might think. Run Wireshark some time and read your packet captures. A fair bit of data does go back and forth between us and them. Sure, the individual size of the data itself is fairly small - 10,000 people each carry a penny and throw it into a fountain. There's still $1,000 in the fountain, as compared to 1000 people carrying a dime and throwing it into the same fountain.
Volume of traffic vs size of traffic, it's all traffic generated. The more traffic, the more impact it has. That's the basic concept behind things like ddos attacks.
And yes, I'm aware of all these issues, and have seen more than a few of them. Perhaps my favorite is trying to dock at a station, having the landing pad vanish, the interior turn grey, and being pushed against the left side of the station when around 20m above my pad. Exit the station, try again, it's fine. Do the same thing a million times not in a wing, never have a problem.
But is that a netcode issue? Or an issue with wings? Or a little of both?
Playable Area - 10,000 players, in 10,000 different locations, all in the same mode. Update tick rolls around, all that data has to go back to the server. Let's say they're all parked at different stations. Now they all want to see what missions they have available. 10,000 mission updates go out, each to a different location in the galaxy. Combine that with, and feel free to prove me wrong, updates to the position of planets in 10,000 systems at any given time, as planets do orbit stars, moons orbit planets, and Other Things happen throughout the galaxy. 400 billion systems. Even if each update moves only 1 byte of data, that's in the neighborhood of 400 GB of data, uncompressed.
Working with the Silent Falling Tree Theory, we can assume that all of this data is not updated until someone enters one of these systems. That will reduce the overhead. 20,000 system updating at the same time with 1 byte of data puts us in the neighborhood of 20 kb of overhead. More than reasonable. Of course, there are more than just one server as well - there is an entire pool of servers. So let's assume again.. one server per mode would give us three servers (Open, PG and Solo), we'll call this a cluster. One cluster per geographical region (Euro, US, Asia) that's 3 clusters, 9 servers. These would be the "Matchmaking servers" from which we regularly get disconnected from - not the database servers, which will send and receive data to and from these clusters. There's likely far more than just 3 servers per cluster - there's likely a load balancer in there as well - which may e server-based or appliance based (still basically a micro-server). And there's likely not one singular database server either - there's likely a cluster of database servers - which have to update each other, as well as all the regional clusters.
See how this gets pretty complex, pretty fast? Now we've got all these servers running, updating each other, so the server you connect to can provide you with game content to interact with - and that connection is dependent on a number of things:
1. The connection between your ISP and this server hive.
2. Your connection between you and your ISP.
3. The middleware and infrastructure of the server hive.
4. The net code of the game itself.
And when it comes to complexity - netcode is usually some of the simplest - you want it that way for the sake of performance. Your middleware, load balancing and let's not forget backups (including backup servers) are vastly more complex in comparison. And I'm trying to keep this all pretty simple.
Some other things that can impact network performance:
1. Your OS/OS Environment - Some are inherently better than others.
2. Drivers
3. Firmware
4. Cabling or gods help us, WiFi
I'll break here a minute. WiFi is terrible for gaming. It's innately slower than wired connections, and subject to all manner of interference from cordless phones to microwave ovens to the type of paint on your walls (High gloss enamel will bounce a wireless signal like mirror bounces light), RF interference from anything from a HAM radio at your neighbor's house to sunspot activity. Yeah, sunspot activity can mess with WiFi.
Most of the time, it's not that big of a deal, but it certainly can be.
And wired connections are not immune to issues as well - cross-talk from other cables in bundles, damaged shielding inside a shielded TP cable, the conduit around it (running CAT-5 though your duct work isn't that great an idea), EM interference, corrosion, oxidation - all can impact wired connections.
Ugh...
The bottom line is, since none of us are actually at the data center, nor have access to any of the actual code, hardware configurations, routing tables, and the lot, to simply say "Oh, there's a problem with the netcode" is almost as helpful as saying "the ocean is wet".
Is the netcode perfect? I doubt it.
Is the environment at the data center perfect? Probably not.
Is the internet itself perfect? Definitely not.
Is everybody's home network perfect? Not a chance.
And odds are none of it ever will be either. Things are going to happen. Frontier will do what they can to shore things up on their side. All we can do is our best to shore them up on ours.
The rest.. well, that falls on ISP's, Communications providers and even governments to maintain all the rest of the infrastructure.
And I haven't even really addressed infrastructure issues, as that's a whole other matter, for a whole other thread, in very different forums.