ALL ISPs oversubscribe the bandwidth that they’ve purchased. In the late 90s, when cable broadband was just taking off, an ISP used statistical multiplexing models to determine the amount of bandwidth they needed to purchase knowing that they were oversubscribing what they had.
The idea behind statistical multiplexing was simple. If you had 25 customers, not all 25 were using the Internet AT THE SAME TIME. For the sake of this discussion, let’s say that 10 of the 25 were actually using the Internet. By the “bursty” nature of web traffic, if those 10 customers were viewing web pages, they would not be pulling in a constant stream of bandwidth – it would come in bursts. And those 10 may not all be viewing web pages. 5 of them could be checking email, which is also “bursty”.
As bandwidth capabilities increased, so did the requirements of the content that was springing up on the Internet; the biggest being streaming video (audio came first, but it really didn’t take that much bandwidth). When YouTube came on the scene in 2005, the load on the Internet started to escalate and the statistical multiplexing models started to fail. When Netflix streaming became popular, the statistical multiplexing model was all but destroyed.
With the model gone, ISPs could no longer oversubscribe like they did before (I’m not saying they still didn’t, but in ratio terms they could no longer do a 10 to 1 bandwidth allocation – more along the lines of a 5 to 1 or 3 to 1). This meant increased costs for the ISP because they had to purchase more bandwidth for the same amount of customers than was previously required. From the perspective of the ISP, their costs had increased without any significant increase in revenue. And the driving force behind this was the increased usage of the customer viewing streaming content, such as YouTube videos and Netflix.
This scenario wasn’t really the driving force behind Net Neutrality – at least not at first. My earliest recollections of hearing about Net Neutrality came from ISPs that were launching their own Voice over Internet Protocol (VoIP) service and how they would throttle another company’s “nomadic” VoIP service (the one case that comes to mind is Comcast versus Vonage) or provide a higher priority for their own. This was deemed anti-competitive and Comcast was ordered to stop.
The Netflix equation came in later and I personally think it shouldn’t even be in the Net Neutrality discussion. Do I believe that [adrotate group=”2″] an ISP should have the right to slow down certain types of traffic because they feel it’s overwhelming their network because their customers are using said service a high percentage of the time? No – just as they shouldn’t do the same thing if they had a competing product to Netflix. In either case it’s anti-competitive.
In the case of Netflix and Comcast, Netflix made a decision to provide a better quality product to a large customer base – namely Comcast subscribers. They did so by peering or essentially connecting straight into Comcast’s network instead of relying on the traffic flowing from Netflix over their backbone provider to Comcast’s backbone provider to Comcast’s customers (which also happen to be Netflix’s customers). On the surface, this seems to fly in the face of Net Neutrality, but on further scrutiny, it’s a good business decision that Netflix made. You could call it the “cost of doing business”.
I don’t believe that we need any type of government regulation or FCC ruling to ensure any type of fairness on the Internet. As real world examples have proven, market forces will make the necessary changes. At some point, when a company is no longer providing the level of service that its customers demand, they will make different choices. And when different choices aren’t readily available, those same customers will make enough noise and provide enough bad press that some changes will eventually be made. Sometimes, competitors will even spring up to provide those customers with better service.
I believe that government regulation has stifled innovation in the telecom industry. Telephone companies have no real incentive to develop new services or take chances on rolling out newer technologies when they are required and bound by regulations on what they can charge. Hell, even if they lose money on providing a service connection for a customer, they have to do it because the regulations say they must.
Some pundits have argued that it’s really not technically possible for ISPs to downgrade the traffic from a company such as a Netflix because the nature of the protocols allow for routing to determine the best paths and process traffic accordingly. This is only partially true. Take Comcast for instance. They probably have multiple entry points into their network that bring in and send out traffic to and from their customers. They CAN downgrade, lower the priority or slow the traffic down at those points for ANY incoming traffic and they can pinpoint the traffic coming from certain companies. If anyone tells you otherwise, then they truly do not understand networking and have probably never worked for an ISP or broadband cable company (I did for 13 years, first as Network Manager and eventually General Manager).
Internet traffic patterns have changed significantly over the last decade. The model of determining bandwidth needs no longer works as traffic has gone from bursty style traffic to almost constant streaming for a large majority of the day. Broadband providers are finding that they have increased bandwidth costs and those costs can be directly associated with a select number of companies that provide services to the broadband providers’ customers. I agree that broadband providers should not tamper with the traffic coming into their network, but we also don’t need government regulation to keep it from happening. At some point, the customer will provide enough pressure on the provider to stop actions such as those. We all vote with our dollars. The main issue is when we don’t have an alternative to give our dollars to. That, to me, is where the real problem lies. We need more competition, not regulation.
Donovan was born and raised in the deep south of South Central Georgia, roughly two hours from the Georgia-Florida line. His father was a guitar player, farmer, and eventually blue color worker for GM. His mother suffered from Scleroderma starting a few years after he was born, so she became a home maker. Growing up as an only child, Donovan’s interest included music (though he really never learned to play anything) and anything dealing with technology, but specifically computers.
He has spent his entire life involved with computer technology either as a hobby or as a career. In his middle to late teens, he ran a BBS (electronic bulletin board system – the precursor to the modern day Internet). He learned about networking computer systems, building computers, and communication technologies as part of his career.
Later in life, he fulfilled his dream of running his own ISP (Internet Service Provider) when he was hired first as the Network Manager and eventually the General Manager of the Telecommunications Department for the City of Tifton, known as CityNet.
Today he runs his own IT business and has been podcasting in some form or fashion since 2011.