Google backs SLAs for residential ISPs
Thrax
🐌Austin, TX Icrontian
Grandfather of the internet-gone-Googler Vint Cerf came out swinging yesterday at the obtuse approach ISPs are taking with bandwidth management. While most are looking to throttle and meter customers in any way they can, Cerf suggests that a bandwidth guarantee would mean everyone wins.
Cerf suggests that users should be guaranteed a minimum bandwidth without traffic limitations. I would like you to raise your hand if you believe that's what you're already paying for. If you raised your hand, it is unfortunate that this is probably not the case. Today's ISPs do not guarantee a maximum or minimum bandwidth for your line, and instead choose to qualify bandwidth statements with "Up to." This charming phrase guarantees that your line will be able to access content on the internet, and implies nothing further. You are not entitled to, nor may you ever come close to, what was advertised on the box.
Cerf's musings seem absurd on the surface, but a closer look at the state of bandwidth in the U.S. reveals a bleak picture. Provider nodes are typically over-saturated; they service entire neighborhoods with an insufficient capacity to deliver as advertised. This irritating little problem leads to the rapid jump in speed that many users experience between the hours of five and seven in the morning when most sane users are rather asleep.
Vint Cerf smartly points out that we have a chance to fix the real problem, rather than masking it with caps and penalties. Guaranteeing each user a set amount of bandwidth means the ISP knows precisely what equipment and provisioning is required for each user and neighborhood. One hundred users paying for 8Mbps/2Mbps service would require a node capable of handling 800Mbps/200Mbps. Miraculously, the congestion issues would disappear as no user is battling another for the slice of the pie.
Ars rightly points out that both DSL and cable are capped by the DSLAM or the cable modem. This leads to well-known bandwidth requirements, but ISPs have oversold the promise of speed, under-delivered on the back-end, and rode that overburdened hog to the bank.
This tectonic paradigm shift in broadband provisioning in the United States would require significant investment in the improvement of existing infrastructure. Unsurprisingly, telcos and cablecos are not in any hurry to fix the root of the problem. It's easier to blame your customers for your inadequacy. There's profit in that.
Cerf suggests that users should be guaranteed a minimum bandwidth without traffic limitations. I would like you to raise your hand if you believe that's what you're already paying for. If you raised your hand, it is unfortunate that this is probably not the case. Today's ISPs do not guarantee a maximum or minimum bandwidth for your line, and instead choose to qualify bandwidth statements with "Up to." This charming phrase guarantees that your line will be able to access content on the internet, and implies nothing further. You are not entitled to, nor may you ever come close to, what was advertised on the box.
Cerf's musings seem absurd on the surface, but a closer look at the state of bandwidth in the U.S. reveals a bleak picture. Provider nodes are typically over-saturated; they service entire neighborhoods with an insufficient capacity to deliver as advertised. This irritating little problem leads to the rapid jump in speed that many users experience between the hours of five and seven in the morning when most sane users are rather asleep.
Vint Cerf smartly points out that we have a chance to fix the real problem, rather than masking it with caps and penalties. Guaranteeing each user a set amount of bandwidth means the ISP knows precisely what equipment and provisioning is required for each user and neighborhood. One hundred users paying for 8Mbps/2Mbps service would require a node capable of handling 800Mbps/200Mbps. Miraculously, the congestion issues would disappear as no user is battling another for the slice of the pie.
Ars rightly points out that both DSL and cable are capped by the DSLAM or the cable modem. This leads to well-known bandwidth requirements, but ISPs have oversold the promise of speed, under-delivered on the back-end, and rode that overburdened hog to the bank.
This tectonic paradigm shift in broadband provisioning in the United States would require significant investment in the improvement of existing infrastructure. Unsurprisingly, telcos and cablecos are not in any hurry to fix the root of the problem. It's easier to blame your customers for your inadequacy. There's profit in that.
0
Comments
That's not entirely true - the telco's lose. They have to pay more to deliver a higher level of service without actually being able to charge more to do so. Further to the point as the technology progresses and competition increases consumers will expect the costs to go down.
While I would love to see this level of service agreement put in place I don't think we'll see anything even coming close to that within the next 5 years. There has to be a much more significant paradigm shift before that will take place.