Alleged Netflix “deliberate throttling” conspiracy is normal congestion
The folks at breakitdownblog.com have accused Netflix of deliberately throttling their users to 480 Kbps (60 KB/sec). Their proof? Downloading a video stream using 10 TCP flows is 10 times faster than downloading videos using the normal single flow. The blogger at breakitdownblog.com concludes that this must mean Netflix is intentionally throttling their users because they assume that there must be extra available bandwidth that Netflix is keeping from their customers. Unfortunately, this is often what passes as credible news these days and the blog has been slashdotted so they’re probably getting on the order of 40,000 hits or so.
Well I hate to break the news to breakitdownblog.com, but this is normal congestion behavior. In fact, they’ve accidentally discovered the multi-flow cheat that works around Jacobson’s TCP congestion control algorithm which rations bandwidth on a per-flow basis and not per-user basis. Even if the server’s capacity is completely filled, asking for 10 times more TCP connections will allow a client to pull nearly 10 times more bandwidth at the expense of other normal clients who are only asking for one TCP connection. Peer-to-peer (P2P) applications employ the same technique to accelerate its own performance at the expense of other users.
So what this sounds like is that the particular Netflix server they’re connecting to is running out of streaming capacity and it can’t handle this many users. This may still be bad on Netflix’s part if the performance problems are consistent, but let’s not attribute occasional inadequacy to malice. I can’t imagine this being frequent since Netflix’s support lines would be ringing off the hook. I also just ran a quick test with my Netflix account and verified that everything is working fine. I’m not suggesting that the alleged problems don’t exist, but we need to have a rational discussion about them.
George,
I may not have made it clear enough in the post (I’ll take a look in a bit) but the indicator to us of the throttling behavior was the very different experience we were seeing during streaming on the Xbox 360 (started in under a minute) and the PC video player (~2hrs of buffering and a capped speed at 50 kb/sec — not enough to stream anything).
These are two Netflix-enabled devices on the same network, providing *much* different behaviors. Please note that the entire 2nd half of the article clarifies that Netflix is providing a much-improved QoS to the Xbox 360 Netflix integration at the cost of PC streaming by throttling those connections — *that* is the claim, now that in general Netflix throttles everything, all the time.
I’d also point out that the section of the article that opens up 10 connections and gets exactly (Cap x 10) bandwidth would indicate there is no congestion or overloading on the server’s side as you mention in your last paragraph. I would have expected the connection speed to scale non-linearly if that were the case, maybe capping out at 200 KB/sec… but to increase the connections to 10, and get exactly 10x the performance indicates there is very intentional throttling occurring here.
Hope that clarifies our position — more comments and thoughts are absolutely welcome. Thanks!
Well, of course they’re doing “evil” throttling. After all, bandwidth is free, right? And all ISPs are evil…. Not.
"but to increase the connections to 10, and get exactly 10x the performance indicates there is very intentional throttling occurring here."
This is NATURAL throttling here. This is Jacobson’s algorithm in effect. I would suggest you read this article in case this blog wasn’t clear enough for you. http://blogs.zdnet.com/Ou/?p=1078.
And does Netflix’ contract give you the right to stream 10 movies at once? I doubt it. Bandwidth is expensive, Netflix isn’t a charity. It has a right to set limits.
IMHO, the problem is much more complex than anyone realizes. Is Netflix throttling?………that is not the real question. Are they employing a QoS to provide a high-demand, shared-resource to many people while trying to give each user their fair share? I’d say, absolutely yes. And so are the level one providers, the CDN’s, and your ISP.
Consumers seem to think they are entitled to unlimited everything for almost nothing. As a small Wireless ISP, I can tell you that the Internet capacity is nowhere near ready for unlimited streaming content. Netflix has a business model that can’t be supported with the present Internet infrastructure, period. We are only starting to see the problem.
The Internet is a shared resource. At the present time it is estimated that Netflix is 20% of the total traffic. I suspect it will be half the traffic in a few years. Think about that……this means over a few years it has doubled the traffic. This means one content provider will have as much traffic as all the rest of the millions of content providers. Will all the level one providers, ISPs and content providers sit still for this while Netflix makes millions for an under-priced content that breaks the Internet for the rest of us and pays nothing for the transportation of their product, I doubt it.
Here is an article I wrote about the specific of a wireless network:
For years contention based ISP business models have made Internet service affordable to consumers. This business model is in grave danger due to the increasing amount of streaming content becoming available. The consumer thinks they can acquire a dedicated service for a contention-based price. We must work to educate them.
I believe the average consumer does not understand the cost of Internet products very well, if at all. The cost from least to most is content, bandwidth and transportation. In my area 60 – 70% of bandwidth cost is transportation. Then it must be transported from the providers PoP to my network and across a couple more links to the AP. In my opinion all of this is still not the real problem.
The one hurdle that is the hardest to overcome is last-mile network capacity. This can only be overcome with technology. Technology like most things in life is a trade-off. We can have somewhat better technology for much more money and much less range. What this means is we can have the capacity to stream video to many but it will raise the price of the Internet service for all on the network. It would easily double the cost of the network in last-mile hardware costs alone. Now factor in the bandwidth. If one-third of the customers on the network stream video, the demand for bandwidth will increase to 360% of normal. This does not account for the fact that with 33% of us watching streaming TV on a 1.5 Mbps connection during prime time we will exceed the capacity of the Access point by 3.6 times. Another drawback to technology that supports more bandwidth is that range will be greatly decreased and coverage to as much as half the customers in our area will most likely be lost. I made these calculations based on an access point serving 100 users. But just 10 customers streaming TV on a 1.5 Mbps connection during prime time can use all the capacity of that access point. This means that the ISP could need as much as ten times the equipment to keep up with demand. This will exceed the load capacity of the tower, not to mention the lack of spectrum to accommodate that many APs. So now I hope you are starting to see the value of your present TV service. It is transportation not content that makes your cable or satellite TV cost so much. I can’t see a way to design a wireless network that will support 30% – 50% of it’s customers watching streaming HDTV, even one channel per household during three hours of prime time viewing, and be affordable enough to remain in business. I know in our household, various family members watch two or three different channels at once.
In addition please check the following links to start to understand the magnitude of the streaming content dilemma:
http://www.businessweek.com/magazine/content/10 50/b4207043617708.htm
http://gigaom.com/2010/12/01/fcc-opens-the-door-for-metered-web-access/
http://www.dslreports.com/shownews/111586
This is a real problem that few understand. I know many will think this is a lie from an ISP to control traffic. Seriously, I started a WISP to provide service to rural Americans with few if any other options. In my service area about 35% of us do not have access to anything but dial-up, satellite and if we are lucky wireless. Now I have to watch the Digital Divide grow because I can’t acquire affordable technology to provide streaming content to my customers.
As more consumers use a streaming content, the quality of everybody’s Internet experience will decline at an amazing rate. Fiber is the only answer, but I honestly do not think we can expect it in many rural areas for at least fifty years. I will continue to pay $80.00 per month for my satellite TV. As the owner of my network I cannot transport the content for what the satellite provider can. In addition, it would be selfish of me to monopolize the shared resources of the network.
Hopefully when you see the contrast of what a large percent of Americans have for Internet service you will appreciate your situation more. And yes, the rural/wireless situation is unique. However all that uniqueness does is make the real problem easier to understand. The Internet is not ready for full-blown streaming content. Another competitor will make things even worse and that is bound to happen. Wait and see.