6 ms·
I have looked hard at the issues & I'm not on any campaign. Paid prioritization is not a bad thing. Blocking websites is. There is a whole nascent standard c
by droopybuns 12y ago
I have looked hard at the issues & I'm not on any campaign.
Paid prioritization is not a bad thing. Blocking websites is.
There is a whole nascent standard called LTE direct that suddenly becomes pointless if you can't distinquish traffic priorities. There are innovations that are being cut out of the picture for entirely new m2x flows & peer to peer flows because of the desire to remove paid prioritization.
It is not productive to assume that the only reason that someone disagrees with you is because of a coordinated campaign to spread FUD.
- daughart 12y agoAt what point does "prioritization" effectively become blocking? People navigate away from websites that take a long time to load or have advertisements.
- wlesieutre 12y agoNot to mention that companies like Comcast have their foot in the online entertainment content game, and could very easily use paid prioritization to get an unfair advantage over competing services.
- droopybuns 12y agoIf you do a search for the net neutrality position of companies like AT&T, Verizon, Centurylink- you'll find plenty of material on their policy blogs on why this is a bad thing. http://community.centurylink.com/regulatoryblog/ http://community.centurylink.com/regulatoryblog/ http://www.attpublicpolicy.com/broadband-classification/thoughts-on-todays-vote/ http://www.attpublicpolicy.com/broadband-classification/thou... http://publicpolicy.verizon.com/blog/entry/fccs-throwback-thursday-move-imposes-1930s-rules-on-the-internet http://publicpolicy.verizon.com/blog/entry/fccs-throwback-th... It is extremely interesting that Comcast has made no public positions on title II neutrality.
- wlesieutre 12y agoInteresting that Verizon and CenturyLink focused entirely on the fact that Title II regulations are old. I'm curious what the specific issues they have with it are; after all, there are plenty of old laws that still work fine. It's become really hard to find any actual information or discussion though. Seems like every page I find is a bunch of bullet points either saying "Regulation will stop all innovation, drive ISPs out of business, and destroy the internet" or "If we don't regulate, Comcast and TWC will stop all innovation and destroy the internet." Sheesh...
- droopybuns 12y agoIs your use of the Internet limited to websites? If you play video games, use realtime video services like facetime, or use a VPN for mission critical work, do you really want your traffic flows to be identical to something that delivers content to a rendering engine? Different traffic has different requirements & performance characteristics. From a pure engineering perspective, this is a very poor decision and treating all traffic equally provides artificial and often unnecessary constraints. I have never experienced any situation where prioritization has effectively become blocking.
- daughart 12y agoInteresting, but I'm more concerned about prioritization of Skype over FaceTime because Microsoft pays Comcast more than Apple.
- MereInterest 12y agoDifferent types of content have different requirements. Different content within the same type does not. This is not an issue of prioritizing VoIP over a large single file. This is an argument against prioritizing Company A's VoIP over Company B's VoIP.
- tzs 12y ago> Different traffic has different requirements & performance characteristics. From a pure engineering perspective, this is a very poor decision and treating all traffic equally provides artificial and often unnecessary constraints. Banning paid prioritization does not stop the ISP from prioritizing traffic for engineering reasons. What it bans is prioritizing traffic because a content provider paid for their traffic to be prioritized. For example, Amazon would not be able to pay to have Amazon video prioritized over Netflix video, and Netflix would not be able to pay to have Netflix video prioritized over Amazon video.
- novalis78 12y agowhat would be so bad about that? doesn't that mean that we would have to "regulate" every successful business, because they could use their additional profit to gain "unfair market advantages"?
- CamperBob2 12y agoI can't help but suspect that the ultimate effect of paid prioritization would be to stifle future growth. Say you're a mobile carrier in the not-too-distant future. Lots of your customers are addicted to the hottest new social network in the Valley, FaceFlix. Trouble is, you're catching hell from both sides -- your customers are complaining about sluggish performance, and FaceFlix is complaining about inadequate bandwidth. What's your path of least resistance? Do you spend money building out your network, buying expensive chunks of spectrum at auction and erecting new towers all over the place? Or do you choose to make money, charging FaceFlix to drop their competitors' packets into the bit bucket? It will seem like a no-brainer from your point of view, and that's why paid prioritization is a bad idea. It just sounds like a recap of the apocryphal story from the early telephone days, where a local undertaker was said to be bribing the operator to reroute calls away from their competitor. Why is that behavior suddenly OK this time around, in your eyes?
- droopybuns 12y agoLTE direct allows peer to peer routing of traffic over high capacity, licensed spectrum. In rural areas where a device has shoddy coverage, you could conceivably route traffic from devices with poor coverage to devices with strong coverage and then out to the Internet. You wouldn't want to do this for streaming traffic, as it could negatively impact the experience of the device with the strongest coverage. But you could do this for traffic that has lower performance requirements. Paid prioritization has existed over the course of the entire Internet. I happens every day- and it is known as peering agreements. Would you characterize the history of the Internet so far as "stifled growth"?
- CamperBob2 12y agoThe thing is, "net neutrality" arguments are not really about peering agreements. They are a consequence of a deeper, unfortunate reality: anyone who uses a packet-switched network to emulate a circuit-switched one is in a state of sin. Streaming applications like to pretend that they have a dedicated circuit connection to their peer, whether that peer is a centralized service like Netflix (or their CDN provider), or another end user's cable box or handset. That's not the case, but we have to maintain the fiction that it is, and treat packets that are part of a Netflix stream just the same as transient ones to or from Twitter. Just how to preserve this unnatural state of affairs is what is being debated. If we don't understand this fact -- if we create laws or technical standards that favor one type of packet over another -- we risk foreclosing all sorts of interesting and valuable applications in the future that no one has even imagined yet. So no, I don't think it's a good idea to promulgate standards that work well with transient messages and datagrams but not streams. If that's true of LTE Direct (and I'm not familiar with the spec, so I don't know if it is), then that's a problem. We've already baked that behavior into the architecture of the Internet itself, and it's only going to become a bigger issue as time goes on.
- ceejayoz 12y ago> Paid prioritization is not a bad thing. Blocking websites is. Paid prioritization can be effectively used to block websites at the will of the providers. "We offered that website our paid tier, but they refused our $1,000,000 demand and are now capped at a bit per second."
- droopybuns 12y agoThe thing I have hated the most about the last 20 years of the internet is how often I have been blocked access to a website due to caps. This has happened exactly 0 times.
- ceejayoz 12y agoWe've already seen companies with Netflix competitors wielding caps against Netflix. It's hardly a crazy concept that Comcast could wield them against speech they don't like too.
- droopybuns 12y agoThe caps were already there. That is what peering is about. If you can offer the service internally on your network, then by all means you should be able to allocate as much bandwidth as can. The peering points are the choke points. This is where things get nutty. Netflix has developed boxes that can be deployed in operator networks to avoid the peering choke points. They didn't want to pay Comcast to operate these boxes (They have to be operated). It's not like netflix didn't have a choice in this situation.
- ceejayoz 12y ago> The peering points are the choke points. http://qz.com/256586/the-inside-story-of-how-netflix-came-to-pay-comcast-for-internet-traffic/ http://qz.com/256586/the-inside-story-of-how-netflix-came-to... > According to Cogent’s CEO, “[f]or most of Cogent’s history with Comcast…[as] Comcast’s subscribers demanded more content from Cogent’s customers, Comcast would add capacity to the interconnection points with Cogent to handle that increased traffic.” After Cogent began carrying Netflix traffic, however, “Comcast refused to continue to augment capacity at our interconnection points as it had done for years prior.” This was a deliberate strategy to bring Netflix to heel, and it meant Comcast screwing over their own customers, who were paying good money for their internet service, to win the battle. > Netflix has developed boxes that can be deployed in operator networks to avoid the peering choke points. They didn't want to pay Comcast to operate these boxes (They have to be operated). Do you have a citation for this? Netflix eventually paid the peering agreements, which are dramatically more expensive than paying Comcast to operate OpenConnect boxes (unless they're taking the same strategy with those and pretending they'd cost millions of dollars to plug a few 4U servers into a rack).