3 ms·
To me, this seems totally backwards. The corporate presence on the (web|internet) isn't really taking advantage of IP (or TCP or any other internet transport-i
by PaulDavisThe1st 5y ago
To me, this seems totally backwards.
The corporate presence on the (web|internet) isn't really taking advantage of IP (or TCP or any other internet transport-ish layer protocol). It's taken over the web, loosely defined as a protocol with HTTP at its core.
Switching away from "the internet" buys you only one thing: for now, most corporate online presence will not follow you. But most of the problems that the author and most other people have with this corporate presence relies only on the HTTP+above aspects of the current web. If we switched the web to run on, I dunno, DECnet, and corporations followed, not a single problem (ok, maybe 2 :) would be solved.
All the chit chat in the article about IPv4 vs IPV6, about NAT, about cable ownership: this is just noise. What's wrong with the web, if indeed there is actually wrong with the web at all does not originate in the internet. How many ways does the author think there will ever be of getting a packet across an ocean? The answer is destined to remain in the low single digits (i.e. a very small number of cables, controlled by something, plus perhaps a couple of satellite routings).
The author identifies NAT as the primary cause of some computers not being "first class citizens" online, but a much more fundamental issue is asymmetrical connections. Your computer can pull data at 100MB/sec, but can only push 5MB/sec ... ergo, it can't sensibly act as a server to any widely popular material. You can consume, you cannot publish (widely popular material), without using someone else's network connection. (This asymmetry is not universal, but is still typical).
The "content-centric" alternative proposed by the author is a nice distributed caching scheme. It's fundamentally a technical detail that is of little actual consequence to anyone, other than any impact it has on apparent performance when accessing information. When I want to read an article that a friend just told me about, I do not care where my computer gets it from, as long as it gets it. We already have this in a totally implicit way, via CDNs. The scheme proposed just makes this explicit, and in so doing adds an entirely new layer to content addressing. Anybody familiar with cache architectures is going to see the kinds of problems this can cause, and the potentially insane levels of complexity required to avoid them.
How about we stop trying to "save the web" at all? How about we recognize that "the old web" continues to exist for those who care (and most do not). How about we recognize that the technical aspects of a distributed, hyperlinked content system are probably not likely to be the best substrate for other kinds of networked communication, and that therefore it doesn't matter if the web cannot "do it all".