As it turns out, charging for faster content (the opposite of net neutrality) results in slower content:

The model revealed the benefits of bandwidth expansion with respect to speed are less than they might at first seem because larger bandwidth attracts more traffic. In essence, attempts to speed delivery of select content to paying-customers only leads to additional congestion. Consequently, if ISPs try to speed time-sensitive content to consumers, the research shows consumers will purchase more of the quick delivery content, and as a result, re-congest the information superhighway’s lanes.

In other words, it has the same dubious effect on “congestion” as caps. The problem has always been peak capacity, and that’s what AT&T and other providers have to understand.

Netflix recently added the ability to reduce video quality for their Canadian users. You may or may not know that Canadian ISPs have much lower caps than AT&T’s, so that very little Netflix usage would result in reaching their caps. They added the following three options:

• “Good” – The default setting with good picture quality and lowest data use per hour (about 0.3 GBytes/hour)
• “Better” – Better picture quality and medium data use per hour (about 0.7 GBytes/hour)
• “Best” – Best picture quality and highest date use per hour (generally about 1.0 GBytes/hour – or up to 2.3 GBytes/hour when streaming HD content)

I wish they make this available to everyone, especially since US users will sadly probably need to use this soon. Does this mean Netflix is conceding the fight against Internet caps?

I was at a meetup catering to people in the startup community and metered Internet actually came up. I got a few things out of the conversation:

• Yes, AT&T’s metering is an issue.
• Yes, AT&T is probably doing this to make up for cord cutters.
• A sense that the only people who’ll be affected are people who use BitTorrent (in other words, the “2%” AT&T referred to).

The one thing I didn’t get out of it was a sense of urgency, a sense that this was actually going to affect the Internet in a negative way. It almost felt like when you know something’s going to happen no matter what you do, so there’s not much point.

History is filled with people who, against all odds, fight for what’s right. How do you rally people when the arguments resonate, but the emotion is falling flat? Granted, it was probably not the best type of meetup to discuss this, but if I’m having problems convincing technologically proficient users, how do I convince everyone else? The answer is there somewhere–I just need to find it.

Mistakes happen with any corporation, but AT&T having network packets bound for Facebook routed through China?

Typically AT&T customers’ data would have routed over the AT&T network directly to Facebook’s network provider but due to a routing mistake their private data went first to Chinanet then via Chinanet to SK Broadband in South Korea, then to Facebook. This means that anything you looked at via Facebook without encryption was exposed to anyone operating Chinanet, which has a very suspect Modus operandi.

Not good. We don’t know for sure if it was an AT&T screwup, or if it was something worse, so I can’t speculate. However, if AT&T continually made mistakes like this, it would be nice to be able to switch to another provider with competent network admins.

(BTW, this is why I love the Internet’s inherent transparency and its ability to readily display information like this. And you should probably be using Facebook in SSL mode anyway, but that’s a different subject for another time.)

Do you know how much bandwidth you’re using? To be honest, I didn’t really know for certain, so I decided to get a better idea. Since I live alone, I only needed to measure the bandwidth my MacBook Pro’s using, vs. what my router thinks it’s using. Luckily, there’s an easy to use Mac OS X application called SurplusMeter that keeps track of your bandwidth usage as you’re using it, so I installed that on my laptop.

At 11:40am, I took the following screenshot several minutes after rebooting to clear the meter. All that’s open at the moment is Adium (IM client), Safari with 12 tabs, Terminal and iTunes. I disabled Time Machine for this test because Time Machine traffic gets counted by this application:

At 12:50pm, after one TV episode in standard definition on Netflix (~45 minutes) and a bit of Web browsing:

I took this screenshot 5:10pm, after 6 Netflix episodes of ~45 minutes each in standard definition plus some light Web surfing and IRC usage. This works out to 4.5 hours, or a bit more than the average of 28 hours per week that Americans spend watching TV:

At 3.8GB per day and 31 days at most in a month, this works out to ~118GB for the month. On regular AT&T DSL, I would have gotten at least one warning from AT&T during the month, and come close to getting one on U-Verse. If I had housemates, used the Web extensively, or spent that time watching HD content instead of standard def content, I would be subject to AT&T’s overage fees. It sounds even more ridiculous when you take into account the fact that AT&T won’t be counting U-Verse TV as part of the cap. This is why it’s so important to do your part to let AT&T know how much of a bad idea its plan actually is and to make our legislators and regulators ensure fair, equal competition in the Internet space.