UK ISP prioritizing gaming traffic is common sense
A UK ISP prioritizing gaming traffic might sound like common sense to engineers and gamers, but it’s raising the ire of ignorant Net Neutrality proponents who buy into the myth that the “end-to-end” architecture of the Internet requires that all applications are treated equally. Reasonable people are willing to drop the “religious zealotry” on end-to-end when presented with the facts. Even leading Net Neutrality advocate Vint Cerf doesn’t buy into the notion that all applications must be treated equal and that jitter should be an engineering matter.
The UK article which raises the concern over Net Neutrality shows the level of misunderstanding that so many in the media have over Net Neutrality. Too many people wrongly believe that gaming is high bandwidth traffic when in fact it’s very low bandwidth like Voice over IP (VoIP) communications. Gaming and VoIP take less than 0.1 Mbps so if a network carves out 0.2 Mbps for gaming and VoIP when needed, it frees up the remaining 2.8 or 5.8 Mbps of a typical broadband connection for other things without having to worry about the harmful effects of jitter and latency. In fact when network engineers are prohibited from managing the network, users take even more draconian measures like demanding that no one else uses BitTorrent at all even when there is plenty of bandwidth capacity on the network not needed for gaming or VoIP.
Vint Cerf was asked during a debate he had with David Farber at 50:10 about the fact that the Internet is discriminatory by nature since it has to manage jitter. David Farber nodded in agreement to the question, but Cerf initially dodged the question by saying that this was about ISPs blocking access at all. Few would argue that wired ISPs should be allowed to block applications or censor content so it’s not really what the contentious policy battle over Net Neutrality is (wireless ISPs are different). Cerf eventually addressed the question of managing jitter to ensure the proper function of real-time applications and stated:
“With regard to things like jitter and other kinds of technical matters, no, I don’t consider that to be discrimination in the bad sense that I’m worried about. If you have to do certain things to achieve certain service quality, that’s an engineering matter.”
When you manage jitter, that is prioritizing traffic based on content type. Some would argue that ISPs should only prioritize content based on the end-user’s wishes (within reasonable traffic budgets for the priority classes), but the user in this UK example is expressing their wish to have their gaming traffic prioritized when they pay the extra 3 pounds per month for business class service. Furthermore, end-user initiated prioritzation via some sort of application packet tagging isn’t really end-user initiated since that’s set by the application developer.
In the real world, the vast majority of applications don’t tag their traffic with priority bits and it’s up to the network engineers to determine how best to manage the traffic. When I built networks for companies that had to support voice or data, we the network engineers did what was best for the end user since they have no idea what the word “jitter” means or how to manage a network and they just want the darn IP phone to work. We the engineers labeled the packet priorities for the end-user and we sent the traffic to our network providers within the contractual limits we agreed to. It’s simply unrealistic to expect the end user or even most applications to bother with the network management issues.
Outlaw the bad, don’t outlaw the good
Some argue that this could lead to an war of escalation with application developers who will masquerade their traffic or wrongly mark their low priority traffic as high priority. But this has never been an issue because application companies aren’t that blatantly evil. Even BitTorrent which cheats the Internet’s congestion control mechanism and causes massive jitter for other applications won’t go as far as labeling BitTorrent traffic with high priority. Furthermore, any attempt to masquerade high bandwidth bulk transfer traffic as low-bandwidth real-time traffic is blatantly obvious to the network and easily defeated.
Some would argue that we can’t trust the engineers and the ISPs they work for to manage traffic like this, but this too is a nonissue. We the engineers aren’t doing this out of some evil plot and it’s not as if we could get away with it even if we were. If an ISP tried to incorrectly label a real-time low bandwidth voice communication application like Skype as a low-priority class of traffic, it’s blatantly obvious that this is malicious and it would be appropriate for the government to take action. Those who argue that it’s too complicated to know what’s bad management and good management simply haven’t taken the time to understand how the Internet actually work. Good engineering strives to achieve true neutrality in the network where all applications are concurrently made to work as well as possible.
The same goes for the ridiculous argument that ISPs would have the perverse incentive to bump up the latency and jitter on default best-effort traffic because that’s something that could also be easily be monitored and regulated against if necessary. But it would be suicide on both a public relations and political level for any ISP to even attempt this when we saw how badly Comcast got pummeled by the press and the government when their well intentioned but poorly implemented TCP reset management scheme was exposed. The moral of the story is that we should outlaw the bad behaviors without outlawing the good behaviors, but too many regulators and legislators haven’t gotten this message because they’ve been listening to the self purported experts with a poor understanding of the Internet.
Why can’t the website pay on behalf of the end user for priority?
One of the key issue in the Net Neutrality policy debate is who pays. The FCC majority in their proposal for Net Neutrality regulation believes that only the end-user should be legally allowed to pay an ISP for higher prioritization or enhanced access to content or applications. The FCC majority proposed in paragraph 106 of their NPRM that content, application, or service providers (basically all business-to-consumer B2C websites) would be prohibited from paying the ISP on behalf of the end-user which is silly from an economics standpoint.
When the B2C website pays for the enhancement, it’s economically more efficient because it involves a single bulk transaction between site and ISP rather than millions of small 3 pound/month transactions between the end-user and the ISP. The cheaper transactional costs ultimately saves the end-user money and this is precisely how Blizzard operates. Blizzard pays the network operator for an enhanced and prioritized network so that the end user doesn’t have to endure the additional complexity and cost. The minimal costs of the enhancement are embedded into the normal monthly service for World of Warcraft and the end user wins.
Ultimately it doesn’t matter who pays because the end user ultimately bears the cost, but the market should determine or allow both economic models to thrive and the FCC or the government shouldn’t interfere and outlaw the more efficient business models that saves the consumer money and time.
Determining Who Pays?
With Services such as World of Warcraft, the customer does pay for a portion of the cost. I don’t see a problem with this model for services. ESPN 360 or which ever content provider should pay and allow the customer access for a price.
I don’t have a problem with this model at all. I can see where this can go awry, but that is where regulation can come into play, not before hand.
@Michael
The Net Neutrality proposals in the FCC and in Congress wants to outlaw payment from the application/content provider. In this case, they would prohibit Blizzard from paying for better network connectivity to the user but not prohibit the user from paying directly. But having Blizzard pay is cheaper because of the lower transactional and administrative costs and those savings are passed to the end-user.
The argument that I keep hearing over and over again which makes no sense at all is that using packet priority pricing for services such as steaming videos. But I don’t think many people understand the costs behind setting up packet prioritization or who would actually be sinister enough to use that method to “bump” themselves in the queue to the point that all other websites become slow.
The Fox News verses DailyKos streaming video Argument had to be the absolute most idiotic thing I ever heard.
Granted, I would be frustrated to find my internet experience being so polluted with prioritized traffic to the point that god forbid Craigslist.org takes more than a couple minutes to load. But we are talking about a significant amount of congestion. I do find my experience frustrating enough when I can’t view content because the ads loaded before the content that I want to view. But I also understand the difference between where the ads are hosted and where the content is hosted. I don’t see http://www.someobscurewebsite.com becoming so slow that no one goes to it.
Now, granted, I will say that messing with site design, I would likely go to a website who has a more stable design and less lag, but sometimes no matter how bad a steam is, I usually go back for content if the content is worth while. Thedailyshow.com is a good example. I have had numerous times where a stream would die at a commercial break, or the show would start three sessions of the sound feed while showing a single session of the show creating an echo that was very frustrating to the enduser who just wants to watch the show. But I went to the site for the content, and didn’t go elsewhere because I wanted to listen to Jon Stuart go on about how Fox News likes to stick its own food in its mouth from time to time.
However, I will say this that is true about Net Neutrality as well as almost every other political hot topic in since the dawn of man. People tend to misunderstand, distort, or oversimplify the matter to the point that best suits them. Usually some political loud mouth will represent either side and manage to look like an idiot while trying to present the argument. Usually the loudest one is the one that has the least amount of clue as to what’s going on, yet also the same person that everyone tries to listen to.
Video streaming isn’t jitter sensitive but it is sensitive to bandwidth starvation for any period greater than the buffer size which is typically in the 30 second range. What videos streams need is a priority on the subscriber’s connection that is higher than all other traffic except for the low bandwidth jitter-sensitive VoIP and gaming traffic. The only thing needing higher priority than that is the network management communications which have to go first or the network doesn’t work at all.
Remember that this prior occurs occurs within a subscriber’s connection and it does not affect other people’s wired broadband connections. Now if enough people in the neighborhood jump on the Internet at the same time, everyone’s connection rate will slow but that’s something independent of the inter-subscriber prioritization.
CDNs already effectively offer a type of router prioritization because the lower latency allows for a faster additive increase in TCP flow rates. That’s not functionally much different than a higher DiffServ priority class.
If Netflix streams were prioritized above everything except real-time gaming or VoIP traffic, it will harm non-Netflix traffic. But what if the broadband subscriber WANTS this higher priority over the connection they paid for? What if they tell the ISP that they want this? I see no problem with this kind of discrimination that the end user explicitly asks for.
The problem is that Net Neutrality proponents just don’t want ISPs to be able to have this kind of business even if it’s completely legitimate because they harbor a deep hatred for ISPs. Yet the same people also want ISPs to invest more and offer better services. There is no rhyme or reason other than some strange personal and political agenda.
@George OU
[Remember that this prior occurs occurs within a subscriber’s connection and it does not affect other people’s wired broadband connections. Now if enough people in the neighborhood jump on the Internet at the same time, everyone’s connection rate will slow but that’s something independent of the inter-subscriber prioritization.]
Funny you should say that, college starts next week and my broadband connection last night managed to get .14 mbps on speedtest.net. While the ISP isn’t completely at fault, 7000 additional users on a network that normally handles 12000 users is subject to a bit of strain.
Just noticed you used a SCII screen shot. I highly appreciate that.
Leave your response!
Twitter Feed
About Us
Digital Society is a digital think tank that believes culture and commerce are inseparable, that the digital economy flourishes when people are free and rights are secure, and that free markets free people.
Digital Society is an independent 501(c)3 non-profit organization, funded by donations from Jon Henke and from Arts+Labs. We advocate for a pro-culture, pro-commerce digital society through research, analysis and debate on emerging technology issues.
Reply Comments
Transparency and interactivity are trademarks of the Internet era, and we aim to foster them here at Digital Society. It is inevitable that some people will disagree with the technology policy positions we take. We want to have that constructive debate.
The Reply Comments feature gives our critics a chance to respond to our viewpoints and the Digital Society audience convenient access to competing arguments. Any time we directly challenge the views of an individual or a group on this site, the party in question may substantively respond in a guest post.
Please contact executive director Jon Henke by e-mail.
Subscribe
Daily Digest Email
Recent Posts
Recent Posts
Most Commented
Most Viewed