It's actually fun and trivial to do this yourself with a spider: go to http://www.snapchat.com/username, mark 'best friendships' as followthroughs. A fun thing you notice pretty quickly is that the 'best friend' designation isn't one-to-one, prolific snappers will appear often in peoples' friends lists.
I think a SnapChat API is likely on the way, and once they figure out a way to market this stuff to advertisers (hey, ten seconds of undivided attention towards an ad is probably worth more than thirty seconds of ignoring a commercial) they're going to have a whole new host of problems.
1.) Pay for SnapPro 2.) Sit through advertisements 3.) Let people who want to pay for SnapPro pay, while you don't because you don't want to pay for it. Therefore no ads, and Snapchat is making money.
Option 3 is the best for consumers. Option 3 is possible if the price-point dictates that enough people will pay for it to remain alive and well. Not everyone will pay. But when is that not always the case?
Isn't this the sort of thing that could be implemented as a peer-to-peer networking service?
You can only talk about the "all but 0" marginal cost when you can produce a piece of software that really does have an all-but-0 marginal cost. And, therefore, doesn't require the computer running it to be always on, or run on a dedicated VM, etc. Sure, such a thing may be possible, but you have to produce it first.
Unfortunately.
There are plenty of distributed designs which could be secure and "read once within a time period".
(btw, a distributed network topology can look isomorphic to a centralized topology at the application layer. You seem to have a highly simplistic map of the territory.)
Snapchat has always positioned itself as a temporal messaging app, where anything goes - but only for a few seconds. It's always meant to be the foil to our Facebook, Google+, et. al., culture of heavy-hitting, abundant, follow-you-everywhere data trails. Why they would embrace this functionality - betraying their core value-proposition - is puzzling.
Whether this data has been theoretically accessible for some time now is somewhat immaterial - unless we're collectively proving our cleverness by correcting the author. The take-away is that Snapchat has seemingly flipped it's core promise - a non-permanent trail and ostensible privacy - by including this "top friends" feature.
However, the trust part is extremely difficult. Savvy users know that they can't trust 3rd parties with data and usually talk about encryption, self-hosting, P2P etc (that kind of stuff). i.e. not really 'trust' per se but more that privacy is 'built in'. Normal users (the vast majority), don't understand this stuff so the only thing they can do is have some faith in the company/service and carry on blindly (until they get burned).
Trying educate your market is a sure-fire way to go out of business. I don't know how developers can seize this without users already understanding why security/privacy (at a technical level) matters.
Even if Snapchat isn't making a big network diagram it's certainly possible to create one.
Perhaps the fact that there are top-users lists highlights the tension between what Snapchat says that people use it for (sharing brief moments with friends) vs what people actually use it for.
In fact, I actually use Snapchat for sharing photos with friends, and I still would prefer that this network information disappear. The existence of a top list suggests that there is quite a bit of metadata retained long after the photos are (ostensibly) destroyed.
It implies that the company is building a database that could be used to reconstruct an enormous network diagram.