Most likely, I like a lot of the principle of Net Neutrality, I'm concerned about the impact on QoS and the ability of regulations to keep pace with innovation.
The Net Neutrality regulations as they exist today allow load balancing and what you describe. Nobody is arguing against that. The issue isn't that Mary's VoIP call gets prioritized over Bob's image download, the issue is when Verizon's VoIP service get's prioritized over a WhatsApp voice call. Both are VoIP, but Verizon as the ISP can play favorites with their own service to the detriment of their competitor. This is what people are mostly concerned with.
It allows for load balancing but not for QoS prioritization under the Title II approach. Peter Rysavy wrote about that a while back and what it would mean for network slicing (basically preventing it) on 5G networks. To me this is an important issue and needs to be both allowed and easily amended over time. We're entering the age of autonomous vehicles and although they'll have failsafes to operate when disconnected they will no doubt perform better when they can communicate with other networked vehicles and devices real time with low latency. I want to make sure that as new protocols are defined for those behaviors we're not mired in an 18 month comment period with the FCC where every special interest comes out of the woodwork claim some edge case why these devices don't deserve high QoS. The same could be said for streaming AR (motion sickness demands low latency so higher QoS needs than standard streaming video) or new technologies we haven't even dreamed up yet. The problem is how do you allow for quick adoption of new QoS requirements without just opening the floodgates to other sorts of packet discrimination which may not be so beneficial. I think that could be addressed by leaving QoS decisions to industry but making them apply QoS decisions universally by protocol and not allowing differences between providers of the same service category. That certainly doesn't adhere to what net neutrality purists want nor does it provide the total laissez faire approach that the antigovernment crowd wants but in my opinion something along those lines would be best for the health of the internet. There would be winners and losers per se in that approach (Skype would like that a lot more than Dropbox for example) but they would be reasonable considering the objectives of the various services.
T-Mobile is certainly testing the limits of what is acceptable. The problem is they aren't transparent about how to get approved. There isn't a public API or anything like that, which if you stream video using this API then it will be zero rated. They say they welcome any service, but that isn't really how it happens in practice.
I guess I don't know the specifics of how they onboard partners to that service but in principle I don't have a problem with packages like this.
I am certainly for anything that incentivizes efficiency, but there is a way to do it without letting ISPs essentially pick favorites and punish users of their competitors. You hit the nail on the head with comparing it to electricity - make internet a pay per unit service. Pay for every MB you download and upload, and people and apps will be very incentivized to become efficient.
I'm with you on the first half but we diverge on the second. I do NOT like the idea of an ISP throttling a competitor. Comcast limiting Netflix bandwidth when Comcast has an obvious vested interest in preventing disruption from streaming video is the sort of situation I seem as highly plausible and what I would want a net neutrality program to discourage (I don't think it can be fully prevented in a reasonable sense but certainly made difficult). Now there were subtleties there too (not that I want to hijack the thread) as that was a peering arrangement and the interconnect had limited throughput but I digress. By and large I think that outcome would be the most likely negative outcome of no regulation and that's what I think we need to protect against. The a la carte plans of pick your specific web sites are FUD.
As to your second point, I disagree. We had that early on in the internet (although I was very young at that point and don't really remember this first hand) with services like AOL and Prodigy that charged by the minute to access the internet. From what I've read people were careful to log on and do things very quickly then log off because there were horror stories of AOL bills in the hundreds of dollars (or more) due to high usage. Had independent ISPs not come along and started offering unlimited plans I don't think we'd have seen the massive growth of the internet that we ultimately did in the late 90s. I believe that usage of the internet should be encouraged. I can understand approaches to reduce impact on infrastructure while still providing (nearly) equal effect (e.g. opt in reducing video resolution on mobile devices where highest resolution often isn't the primary concern or transcoding to more efficient video codes or use of CDN and peering to get the data/application closer to the user and reduce backbone congestion etc...). Otherwise I think the access to information provided by the internet has too much societal benefit to make people nervous about their consumption in the name of conservation. I'm willing to talk about renewable energy for data centers and such but just the act of consuming data isn't a conservation concern in the sense of water usage or trash (and yes I understand it consumes electricity but that's why I'm open to talking about renewable energy to power those needs).
Giving one service a zero-rating for streaming in 480p while counting another newcomer service against usage caps is not the best way to get efficiency.
But it is a step. It's not the start up video and streaming services that are taxing mobile networks, it's the big players. Cutting bandwidth by 75% by dropping the Netflix/Youtube/Twitch/Hulu streams from 1080 to 480 is a good way to reduce congestion on cell towers. NotMyName Streaming Inc not participating in that plan but only accounting for 0.001% of traffic isn't a big deal. As my service grows I'd look into participating as well.
As I said above, the current Net Neutrality regulations allow QoS prioritization based on loads and real-world needs. They prohibit paid or anti-competetive prioritization. For very bad reasons, Trump's FCC want's to get rid of that.
It doesn't address the QoS problem though. If we can address that - and critically, in a way that doesn't throw up roadblocks to changes in those QoS categories - I'm fully on board.
And not that this thread needs more variables to consider but I'll also mention this. The Internet doesn't work like the picture I believe most people have in their head. I think most people have an outdated model that says if I connect to Google I'm transferring data from Mountain View, CA to my home in Manhattan and it rides along the "information superhighway" between those two points. That hasn't been the case for big players for some time now (and even modest players more recently). Big internet companies (and remember there are about 25-30 companies that are responsible for 50% of internet traffic) have peering and CDN arrangements. These companies host servers in the ISPs' data centers and connect their own routers right into their networks. When you request a search from Google you are not making that request across the backbone of the internet, you've hit a local data center and received your results from there. This has already then provided the "fast lane" people refer to if not in the specific way they've envisioned certainly in effect and that's completely legit under net neutrality regulations. So, even if we state a packet is a packet is a packet there's still the competitive advantage that larger players can get by geo-diversifying and getting their data closer to their customers than you can. I still think that's a good thing for the health of the internet though as it reduces backbone load and increases robustness. As I alluded to earlier this is also becoming feasible for lower end players as well, cloud computing (AWS, Azure) have geo-diverse options and CDN packages (Cloudflare etc...) are becoming more accessible. Still maybe not the right fit for a small startup trying to bootstrap but at the point they start scaling they can access these tools too. Sorry for the tangent but as others have mentioned (without really explaining) the Internet isn't a "fair" place even with net neutrality, there's more at play here than how you treat packets.
