Slashdot Mirror
Why the Coming Data Flood Won't Drown the Internet
High Waters writes "Ars Technica examines predictions of an 'exaflood' of data that some alarmists believe will overwhelm the Internet. A closer look reveals that many of those raising the alarm about an exaflood are generally doing so to make the case against internet neutrality regulation. 'There's a reason that "exaflood" sounds scary. It's supposed to. Though Brett Swanson's Wall Street Journal piece tried to avoid alarmism, it did have an explicitly political point in mind: net neutrality is bad, and it could turn the coming exaflood into a real disaster'."
Simple. It didn't happen before. The Internet has experienced 'exafloods' before. The amount of data and traffic have skyrocketted exponentially every year since this big major growth spurts started in the 1980s and 1990s. How can the Internet do that?
Because it was designed that way, that's why. The Internet is the largest distributed network in the world. TCP/IP was purposefully designed to be scalable on a massively large scale. Sure, we've improved the technology along the way, but the bottom line is that the routers directing all those tubes aren't going to buckle under the pressure anytime soon, and routing technology is just getting better all the time.
My blog
Predict bad things are going to happen unless people do what you say/buy what you sell/give what you want/etc.
Nothing new here.
...are the people who want to control the internet.
Media companies wanting to shut down distribution of content not authorized by them (not just illegally copied content but content created and shared under licenses that specifically ALLOW sharing)
News organizations and governments wanting to continue to maintain control over what news we read, view and listen to so they can make sure that the "sheeple" stay "sheeple" and dont actually try to CHANGE their lot in life
Telecommunications providers (including providers of cellular telecommunications) who want to maintain profits for services THEY control and not allow the growth of alternatives to the telco-provided services
Churches and other groups opposed to pornography, gambling and other "vices" who want to be able to ban such content (or if thats not possible, at least control it to the point where its effectively banned)
Manufacturers, distributors and retailers who want to control your abillity to buy stuff to keep bricks & mortar stores alive or to keep people from buying stuff from a country where its cheaper than their own (for example, here in australia, a number of online stores were selling Panasonic DVD recorders really cheap due to the low overheads of those stores. Bricks & Mortar electrical stores complained since they couldn't sell at the price the online guys were selling at and actually make any money. So Panasonic stopped selling the DVD recorders to the online stores)
Governments and spy agencies who want to control the internet so that its easier to spy on the people and look for people who might "rock the boat" or that want to use internet control as a way to hang on to power (look at what happened recently in Burma for example where the government restricted internet access to try to stop the world from finding out how many innocent civilians were being hurt and killed in the name of keeping the dictatorship in power)
When traffic increases (overall, or peaky) to handle more video (for example), capacity has to be added or it quite simply will not get moved. Squeezing out/delaying other traffic will not go very far. Dark fiber has to be lit. When capacity is added because there is more traffic, there is also more "gaps" to fit in "low priority" traffic.
The fundamental problem is people think of the internet as a water pipe, with very simple capacity constraints. It is not. You don't care about water latency while data packet latency or jitter are extremely important.
It is beyond annoying that certain commercial entities are exploiting this misunderstanding to further their own interests at the expense of their customers. One cannot help but see them as grasping and acting out of malice.
Dire predictions are usually followed by a project/business plan on how to fix things. In other words someone wants money to fix something that won't need fixing.
How many times did people predict that Usenet would collapse due to the massive amount of data being passed around on the old modem network? It never did happen.
I thought this paragraph from TFA was especially interesting:
But the growth in file sizes is made worse by a concurrent increase in the use of P2P as a delivery mechanism. Distribution gets pushed from the center of the network to the edges as users increasingly become both the consumers and providers of content, so the tubes could be clogged in both directions.... The [US Internet Industry Association] describes this transition as a traffic shift "from the Internet backbone to a peered system in which content is streamed directly to consumers," and the group notes that it will require ISPs to upgrade the most expensive part of their networks to keep pace: the last mile.
Wasn't the Internet designed from the ground up to be "peer-to-peer?" Yes, I know we started with client/server technologies and "the Internet backbone," but fundamentally every machine with a public IP address is, and has always been, the peer of all the other millions of machines with public addresses. That's what makes the Internet so profoundly democratic and so profoundly threatening to established interests.
I suppose cable operators weren't used to seeing the world in those terms, but telcos certainly were. Voice/data services were always interactive, not unidirectional broadcasting. Why should anyone be surprised that the Internet is being used for the purposes its designers envisioned?
Oh, and why is a system where "content is streamed directly to consumers" described as "peered?"
I'll take a data slowdown over that any day.
One swallow does not a fellatrix make