Game developers demand regulations to ban premium peering
A group of game developers have recently gone to the FCC to advocate “Net Neutrality” regulations that ban premium peering services. These efforts follow an overarching theme of Net Neutrality advocates to ban broadband providers from offering services to content and application providers. But these premium services have blossomed over the last few years to enable on demand video streaming services over the Internet and they could eventually facilitate more reliable real-time communications such as video conferencing or online gaming which helps to drive adoption.
The tactics taken by these game developers fall into the typical strategy of scaring the public about ISP blocking or degrading traffic and then arguing for blanket prohibitions against premium ISP services. They don’t ask for rules against blocking or degrading because those rules already exist. What they really want is much more expansive rules that benefit their own businesses so they conflate the mere existence of premium services with blocking or degrading when this argument is completely misleading. Once we examine their statements closely, their real motives stop other gaming companies from differentiating their services. We can see this entire strategy play out in their meeting notes with the FCC.
Raise the ISP blocking bogeyman
“Mr. Scherlis noted that not all companies have the resources to identify blocks or to persuade ISPs to stop blocking. Without rules for transparency, such companies could not anticipate such blocking, let alone argue against it.”
This is ridiculous because there is only one real example of blatant ISP blocking and that was Madison River Communications blocking Vonage VoIP services. The FCC slapped Madison River with a $15K fine in 2005 under the existing regulatory framework and that practice immediately stopped. Comcast was more of a subtle case where the “victim” Vuze wasn’t ever blocked or even arguably degraded, but it was probably the best thing that ever could have happened to Vuze publicity wise. A company that hardly anyone knew about was now getting a free technology showcase before the FCC and the national media. Comcast on the other hand got slapped hard by the FCC and media for trying to ease congestion for the vast majority of their customers but going about it in a suboptimal manner.
Even the mere accusation of blocking such as the case of Cox and Craigslist caused a media and political firestorm. From what we have seen over the last four years, there is virtually no chance that an ISP wants to think about blocking or degrading content and become the next “example”.
Associate premium services with blocking
“They each can promise potential customers a superior-QoS, in reaching their own subscribers, from the servers that they host. This is an easy promise to make. All the ISP need do is ensure that non-ISP-hosted services suffer appreciable degradation as part of interconnection.”
This is a one of the most common tactics to spread misinformation about QoS. The assertion is that ISPs will resort to underhanded tactics to decrease the performance of the network for all non-premium customers to force more customers to buy premium services is ridiculous. One extreme example is Karl Auerbach who argued that ISPs would impose a permanent 100 millisecond delay on all standard class packets even though this would violate the definition of “best effort” standard class traffic. But the reality is that no ISP does this even though most of them are offering premium QoS services today. Furthermore, the ability to inject 100 milliseconds of delays isn’t even a feature of any production switch or router and it would require some WAN simulation equipment not intended for use on production networks. ISPs would have to go to extraordinary lengths to be able to systematically inject latency or jitter into their networks and none of them do it. Even if it was feasible, there is no way they would want to advertise “worst effort” service to their customers and if they covered it up and lied about it, the FTC would be all over them for fraud.
In reality, we can see that latency and jitter aren’t really a problem for non-premium customers today and they haven’t been a problem over the last decade. I’ve conducted some ping tests yesterday afternoon at 4:30 PM pacific time with two gaming servers that I use on the weekends (see figure 1 through 3). The TN game server is on the West Coast which provides very good local sub-20 millisecond response times for me. The No Heroes game server is on the east coast and has a baseline latency of 89 milliseconds from my home. That isn’t optimum for West Coast gamers but its relatively low 120 millisecond latency for Europeans make it a great halfway point if a server is needed between European and US players.
Both servers are small businesses that offer premium reserved slots to gamers and they would be the first to complain to their ISPs if their jitter (abrupt increases in latency) were high. The jitter that I measured was no more than 7 milliseconds above the minimum latency values over the course of 8 minutes during peak traffic hours (8PM US East Coast time), and typical jitter was below 3 milliseconds.
Figure 1 – Latency getting to the first ISP router at SBC
Figure 2 – No Heroes gaming server on Peer1.net
Figure 3 – TN game server on gblx.net (via swbell.net)
If regulators are really concerned with a hypothetical degradation in Internet latency and jitter for standard transit services, they could simply punish ISPs that behave badly and it is obvious if any ISP did behave badly. Large Gaming platforms like Valve’s Steam service could easily collect large scale anonymous latency and jitter data from their millions of users about the performance of various regions and ISPs and Valve would be the first to complain if ISPs were degrading their customer’s traffic below best effort. There is simply no reason to resort to ban premium gaming services because the two models are coexisting in harmony today.
Argue that QoS isn’t needed if more capacity was available
“Mr. Dyl notes that the need for quality of service should primarily be an issue in an overly congested network. If the network is not saturated, all packets should get through in a timely manner. Currently, a consumer will pay for a 10Mbs connection, but doesn’t always get that performance.”
This is completely untrue. It is very possible to have high jitter problems even during times that appear to have low levels of network traffic. Furthermore, a small percentage of BitTorrent traffic can cause a huge increase in latency levels.
Argue against content or application providers paying ISPs
“The real debate is over the question of charging for premium QoS. Particularly, whether application-developers should be charged, as opposed to consumers; entrepreneurs and application developers seem to find little problem in the ISPs’ desire to offer different levels of service to their subscribers.”
“Mr. Radoff stated that software platform developers like Microsoft and Facebook pose less of a threat to innovation than infrastructure owners. He contended that if the Internet were balkanized, and developers had to negotiate separately with each ISP, that would be a substantial drag on innovation because it would divert resources from development.”
Amazon has demonstrated the benefits of content providers paying for Internet access on behalf of their customers with devices like the Kindle e-book reader, yet Net Neutrality proponents insist that ISPs should not be permitted to charge content or application providers. Net Neutrality proponents have even convinced Ed Markey (D-MA) to write bills lik H.R. 3458 which bans ISPs from charging content or application providers and they’re hoping to convince the FCC to ban premium services as well.
The problem is that this “balkanized” Internet that we have today is anything but hostile to content or application providers. It’s really an open bandwidth market that allows content providers to negotiate cheaper and higher quality bandwidth yet Net Neutrality advocates want to eliminate these faster and cheaper options. Giving content and application providers the option of negotiating directly with broadband providers saves them money so I don’t know how anyone would argue that it “diverts resources”.
Argue that no content providers should be permitted to differentiate
“As a result of this, all application developers face this choice: pay for relationships with every major ISP (and for the complexity of running an unnecessarily multi-hosted service), or instead choose to offer an end-user experience that will be reliably (if not fatally) inferior to their competitors’”
Here we see the true motives of these particular game developers. They simply don’t want to compete against other game companies that want to differentiate themselves with superior products. But differentiation and the desire to attract more customers is precisely what spurs innovation and the last thing we need are blanket prohibitions on new and innovative services. Killing these business models would reduce investment on the Internet.