[An4ximandros]

Screw net neutrality.

[Nuke]

i hope you all brought your ak-47s

[Klaustrophobia]

this is the mother of all tl;dr

[Torchwood]

To sum it up, the obligation of several major US internet providers to treat all traffic passing through them equally is cancelled (!!!). This would enable them to give preferential treatment to some (i. e. pay more for more speed) or ban certain traffic entirely (suspected filesharers). Don't let the length of the document obfuscate how disturbing it is.

[Ghostavo]

I often wondered how can you have QoS with Net Neutrality.

[jr2]

QoS is prioritizing based on packet type (some packets are time-sensitive more so than others, e.g.:

1 Gaming
2 Video Conferencing
3 VoIP
4 HTTP / browsing
5 FTP
6 Peer-to-Peer
7 Updates

Versus, net neutrality is more saying you can't prioritize your own VOIP service (say, with Time Warner Cable) and trash Vonage's so that their service barely works (or doesn't at all).

[Ghostavo]

QoS is prioritizing based on packet type (some packets are time-sensitive more so than others, e.g.:

1 Gaming
2 Video Conferencing
3 VoIP
4 HTTP / browsing
5 FTP
6 Peer-to-Peer
7 Updates

Versus, net neutrality is more saying you can't prioritize your own VOIP service (say, with Time Warner Cable) and trash Vonage's so that their service barely works (or doesn't at all).

While that is a major concern of net neutrality, by distinguishing packet types when dealing with traffic, you are in fact ignoring net neutrality. Which is why this issue is less black and white than it seems.

[Aardwolf]

Apologies in advance for being too lazy to look it up myself, but...

Does QoS actually categorize and leave prioritization up to the tubes, or is it a priority value set by the application/machine sending the packets? Because if it's the latter, there's no conflict here.

[Ghostavo]

Apologies in advance for being too lazy to look it up myself, but...

Does QoS actually categorize and leave prioritization up to the tubes, or is it a priority value set by the application/machine sending the packets? Because if it's the latter, there's no conflict here.

Both.

[Aardwolf]

So I got off my butt and looked up QoS on Wikipedia, and apparently there isn't just one accepted scheme for guaranteeing QoS by the name of "QoS" (as I had previously surmised upon seeing jr2's post).

Thus the answer "both" makes sense. Ok.

Well then, IMO the way to go is to let the application or machine set some sort of prioritization metadata for the packets, and discourage or forbid ISPs etc. from using "deep packet inspection" or alternate ways of prioritizing (unless the metadata is not present?). I add the bit about "discourage or forbid" because IMO packets should be afforded the same legal protections as printed mail (never mind whether those protections exist in practice IRL).

[Ghostavo]

Well, while in an utopian world that could be an interesting way of dealing with it, in the real world, it just encourages everyone to say "Hey, see all these packets I'm sending? They are really important, honest!".

In a nutshell, you are putting a level of trust on the application/machine that may be unreasonable for the network to cope with.

It can be useful with other QoS measures, but solely as a way to prioritize that same user's packets, but not against other users.

[AdmiralRalwood]

Well, while in an utopian world that could be an interesting way of dealing with it, in the real world, it just encourages everyone to say "Hey, see all these packets I'm sending? They are really important, honest!".

In a nutshell, you are putting a level of trust on the application/machine that may be unreasonable for the network to cope with.

It can be useful with other QoS measures, but solely as a way to prioritize that same user's packets, but not against other users.

Yeah, I say it should default to highest priority, and applications that know that they don't need really low latency can optionally let the network know that (any other scheme sounds like it's just asking for trouble).

[jr2]

For what it's worth, I've used cFosSpeed software quite successfully in the past (when I had good Internets!) to prioritize my own traffic - you would be quite surprised at the performance gains.  For example, I could use BitTorrent to download something at my maximum download speed, and yet still successfully be able to browse the Interwebs with little if any noticeable impact on browsing performance.

That's QoS on the end-user side of things, however, not QoS on the ISP side.  Everything of course would still go to pot on lunch break and when all the kiddies got out of school at 3:30.

[Akalabeth Angel]

This video sheds some light on the problem:


https://www.youtube.com/watch?v=fpbOEoRrHyU

Definitely worth checking out (&sharing) for the laughs alone.

Also all americans who care about a neutral internet should direct their anger and comments here:
http://www.fcc.gov/comments