One thing this video (and every other pro-net neutrality video ignores) is that there are valid technical reasons to have some content faster than others. In an ideal world, we should be able to mark some content as more or less latency tolerant.

Static webpages for instance; if your online banking or facebook or even videosift takes half a second longer to load, no-one really cares. But if your video stops for a second or an online game lags, that can ruin the experience.

Right now, if two servers send two packets, they are treated the same, even if one is time dependant (video/game content) and one isn't (static text).

That is the core argument against net neutrality. The problem is that it should be the content providers who decide what data is time dependant, not ISPs. Ideally, ISPs shouldn't even know what content is going through their pipes.

Just in case it's not clear, I do not in any way support the idea that ISPs should be allowed to slow or speed up content on their networks, any more than a construction company should be allowed tell drivers who drives in what lane.

1000Mbit doesn't necessarily equate to lower latency anyway, and then Up/Download speed isn't even that important for online gaming, unless you want to stream/torrent/uploadselfies while playing. It is however a nice number that's easy to stick on marketing material. I had a huge fight with an ISP a few years ago after they pretty much flat refused to do anything about crippling packet loss of 10ish% at peak times because it had a "negligible effect on bandwidth". Pretty sad that I get much more reliable net over ADSL copper lines at 10ish Mbit than I had on fiberoptic.

Perhaps they should have actually done an ad about lag, instead of one about latency. If an ISP can't get that right, probably best to look for another ISP.

*backup=[...snipped...]

Thanks. I'm from Melbourne, and it does seem to depend on the day whether Vimeo will be awesome or sucky. Even on bad days though, it does have the advantage that vimeo embeds are very rarely region blocked... even if they might take a while to buffer (and then mysteriously fail and go back to the start 20 seconds before the end). Also, there is some really great stuff there.

The reason that youtube and dailymotion seem better in comparison here is that they have caches in our local ISP's, iiNet in my case, and Vimeo doesn't seem to have the same agreements.

why the hell is it a big challenge? I realize that yes, Internet connections can be flaky and you can't always control it,

but to me, this is easily solved by simply WAITING and let the buffer fill up. But in certain cases, the buffer will only fill if the video is playing.

when you hit play, throw up a screen saying please wait while we download your video so that it can play in an uninterrupted fashion.

but nah, they'll never do that because we've allowed attention deficit disorder to flourish throughout the land so we are simply unwilling to tolerate waiting a bit.

I used to do this ALL the time when apple quicktime trailers for movies first became big. Hit play, but then immediately hit pause and wait for that bar to fill all the way to the right. Sure, sometimes I got impatient and hit play sooner when I thought the buffer was going to fill before I finished watching the video, and yes, sometimes I got burned on this, but hey, it was on me, not on my ISP, or the video I was watching.

You can't even do that with youtube videos though, If i hit play, and immediately hit pause to wait for the buffer to fill, it never will, because they've got some stupid rule in there to only fill the buffer if the video is playing. Get rid of that stupid rule and allow the buffer to fill and I guarantee you the buffering symbol will go away for a lot of people.

Very good speech from Hyppönen, once again. It's funny that thru the years, his english pronunciation hasn't improved a lot.. For those that don't know, he's F-Secure spokesperson (and i guess innovator too, he's been there from the start, too bad their products are crap but with out them we would have no security at all..) They made some important inventions in the mid-90s.Some of them CIA fought with tooth and nails like 128bit encryptions claiming it's a security risk if USA can not intercept every signal they get (yes, this problem is OLD...) but F-Secure and other companies, ISPs, everyone were united in this issue and those security measures are now a commonplace...

Heya @eric3579, thanks for getting Lucky's attention for me! Obviously been gone a little while if I've forgotten the basics

Thanks for taking the time to look into it @lucky760, appreciated After reading your post I thought I'd check the site on my phone, also running IE10 - same behaviour occurred, page loaded without comments, progress indicator stopped moving. Hit the refresh button and the page loaded with comments visible. Went to the site on my tablet that's just been updated to Windows 8.1 and has IE11 - on this device the page loaded fine including comments.

All 3 devices are connected through the same ISP, desktop is connected directly to the ethernet switch on the router, the 2 mobile devices are connected via wi-fi. I'm switching to another ISP next week, curious if that'll make any difference.

Thanks again for checking into the problem, that's why I like this place

It would be correct to say that it's more complicated than that...

With the exception of the Tier 1 ISPs, who were able to use their market position to force everybody else to pay to peer with them, ISPs will generally pay for all traffic that leaves their network. There are no data caps for ISPs, they pay for usage and then take the average data costs and use that as part of their pricing determination for their customers.

Even aside from that though, due to contention (see @charliem's comment above) there simply isn't enough kit installed to cope with all users downloading as much as they can all at once. You could look at a single fixed cost gigabit line (lets say it costs $1024 pcm) as actually having a cost of 0.31c/gigabyte transferred, whether or not you actually use it. Extra data transfers over that line don't actually cost any more until you exceed the link capacity, then you have increased costs.

What your ISP is selling you is actually a share of a link. They don't have to charge for usage, but they have to pay for it regardless... all that changes is how they bill their customers. Since 1% of residential customers generate approx 90% of the traffic, it is a bit fairer to charge for usage, otherwise 99% of the customers are subsidising the 1% who generate most of the costs for the ISP.

That is true. But. Peak transfer speeds are not in any way linked to total amount of data downloaded over time. This is what i meant that data caps are artificial, designed for two purposes: to coax users to download less in hopes that it will help peak transfer rates (it will) but the main reason is just more profit.

Correct me if i'm wrong but ISPs do NOT have data caps in their contracts, it's flat rate as no one can predict the total amount of data transferred in n amount of time. Data caps belong to the 90s.when infrastructure couldn't cope with many users.

Data transmission in modern worlds follows rules of contention.
ISP's typically provision their networks with 1:50 or higher contention ratios, because they know that 50 people on any normal day wont use up 1 torrent-user's typical bandwidth consumption under full saturation.

The ratios are calculated from the average access edge speeds....say avg speed is 3mbps, they could provision 50 customers to share 3mbps in the core network before they consider upgrading.

Contention....noone uses the entire networks bandwidth the entire time its available. Otherwise networking equipment and access costs would be 50 times greater!!

Its just not efficient to plan for everyone to saturate everyones links at the same time and not have a bottleneck....because 99% of the rest of the time, noones using shit on average. What a waste of power, space, equipment, service fees, support contracts, engineers to maintain the equipment / network etc..etc...etc...

This is by no means universally true, and I doubt it's even universally true in the U.S.

Even in a world with no peering charges, if everyone starts downloading all the time then additional kit is required to cope with the increased peak bandwidth requirements. No ISP's have enough capacity for everyone to use all their bandwidth at once, the better ones are just closer.

Where did you get the idea that data transmission is free?

The ISP manager (Tom Heatherington) probably has millions of dollars now... the talk show host does not.

Was just looking on "an internet" to find out more about him.

Nice @critical_d. An even more layman explanation:

When your computer tries to load a domain name like "videosift.com" it asks your local Domain Name Servers (usually provided by your ISP) a question like "What is the numeric address of videosift.com?"

The name servers check whatever their records state and that's what they return. In this case, they were returning an old address that is no longer functional, so your computer says, "Okay, I got the address. Let's connect to it," but it fails and your browser goes kaput with "Can't connect!"

When we make a change to our DNS records, it's at a single place: the authority who has control over our domain (e.g., GoDaddy [though, that's intentionally not who we're with]). Once we've made a change, all the DNS servers around the entire planet need to get updated with the new domain-name-to-IP-address translation. Some get updated quickly while others take a very long time (depending on factors we won't cover in this seminar).

For those whose DNS servers take a very long time, there are two loopholes:

1) Before asking a DNS server to translate a domain name, your computer looks at its "hosts" file. If you've specified the address for a domain there, that's the address it will use. (You could really screw with someone by changing the address of google.com, for example, to some other address.)

2) You can change your computer's DNS servers to something like Google's, which are 8.8.8.8 and 8.8.4.4. They tend to get updated pretty quickly in many cases, so when your computer asks for the IP address of videosift.com, it will get the right answer.

Clear as mud?

Errors 502, 503 and 504 are very often from a server OTHER than the one you want to connect to, so you've actually made a reasonable case for @gwiz665's original assertion... but I agree that it would be far better to have a unique code like 451 to indicate that this has happened.

When I said that the proposed 451 counts as a special case of 403, I didn't mean that it should return 403 just that I saw it as a related concept.
>> ^critical_d:

The site is blocked at the ISP level so the request never makes it to the destination, making the 5xx error invalid. Yet the 403 error being server in this instance is also misleading as it is describes te request as valid but the server will not respond to it. To me the "server" means the destination (host server) and not the ISP server.
However unlikely, I do hope that there is a unique identifier created to flag such events.

The site is blocked at the ISP level so the request never makes it to the destination, making the 5xx error invalid. Yet the 403 error being server in this instance is also misleading as it is describes te request as valid but the server will not respond to it. To me the "server" means the destination (host server) and not the ISP server.

However unlikely, I do hope that there is a unique identifier created to flag such events.

>> ^oritteropo:

It's a form of permission denied, not a server error, so it sits more logically with the others in the 400s:
401 Unauthorized
402 Payment Required
403 Forbidden
Really it's just a special case of 403.
>> ^gwiz665:
Since it's a server error, it should be in the 500 region. Nerd fail!