Dear ITU, please don't bill Internet use like phone calls

http://www.arstechnica.com/business/2012/12/dear-itu-please-dont-bill-internet-use-like-phone-calls/

by mtgx • 14 years ago

80 32 14 years ago

hn-miw-i• 14 years ago

Despite efforts of the EU and the ITU to put the Internet genie back in her bottle, she grows louder and prouder every day. She is still in her formative days, as the old world money tries to tame her radical free spirit. Let's hope decisions made by old men in new Dubai doesn't keep her from constantly changing humanity.

mtgx• 14 years ago

It's not the "EU" as in the EU Parliament. In fact they've passed a resolution to oppose this sort of changes at ITU's WCIT. The article talks about ETNO, an European telecom association. So it's the ISP's and carriers from Europe pushing for these changes. Nothing to do with the EU Parliament or EU governments (well, Russia and China's censorship proposals are entirely different from this). Just corporations being greedy. The EU Parliament has actually legislated their roaming prices twice in the past (downwards), because they thought roaming prices should be much smaller than they were, within EU's borders.

hardik9882• 14 years ago

I'm pretty sure I'm exposing my extreme ignorance on the matter, but how does the ITU have the authority to take decisions like this? Shouldn't W3 have a say? Or somebody else?

eridius3• 14 years ago

The W3 controls the standards used for the web. Web ≠ Internet.

nitrogen4• 14 years ago

For one thing, the ITU is a governmental body (that is, its members are governments; it is an agency of the UN), while the W3C and IETF are open membership organizations that anyone can join.

In other words, the ITU is run by people whose job is to run things. The IETF and W3C are run by people whose job it is to do things. Groups of the former category tend to want to control those of the latter.

Edit to add: the different outcomes of the ITU vs. IETF/W3C can also be attributed to the purpose of existence of their members. People who run things (e.g. governments) want to keep running things. Thus, their work tends to result in new laws and treaties. People who do things (e.g. engineers) want to keep doing things. Thus, their work tends to result in voluntary, consensus-based agreements.

hardik9882• 14 years ago

Thanks. That clarifies it up a bit. So could this summit result in a real threat to net neutrality?

nitrogen4• 14 years ago

With regards to Quality of Service, what's wrong with letting the packet flags decide QoS, with customers paying for a guaranteed percentage of high-QoS bandwidth during peak network load?

In other words, when the network is underutilized, all packets would be transmitted with high priority and low latency, with QoS flags allowing packets to jump ahead in packet queues. When the network is heavily loaded, high-QoS packets beyond the customer's quota would be relegated to best effort.

eridius3• 14 years ago

In this situation, what's stopping all traffic senders from marking their packets with the QoS flag? After all, they don't want their traffic getting stuck behind some other sender's traffic.

nitrogen4• 14 years ago

I had a sentence addressing this issue which I deleted before posting, because it seemed that the situation is the same as if the network is heavily loaded: if the upstream provider's high QoS bandwidth is exhausted, customer traffic beyond their guaranteed QoS allocation reverts to best effort (with the number of packets reverted proportional to the total upstream overage and the individual customer's overage).

belorn2• 14 years ago

Each customer only has a limited amount of QoS high priority traffic. If someone flags every packet with high priority, all the QoS mean is that the first packages get improved priority and everything else get default low priority.

That mean a customers software can balance the use of QoS for time sensitive traffic, and maybe even flag some data as below default to get bonus high priority traffic points. If its the software that runs on the consumers device that decide priority, the software developer can pick the exact needed QoS his program or service need.

The bonus effect here is that if the QoS fail and the program breaks, its the software developer that are liable and not the ISP. Currently, if a ISP QoS software breaks, the ISP is actually quite open for lawsuits from both the user and the service that get effected.

wmf3• 14 years ago

QoS generally gives ISPs incentives to deliberately cause congestion by not upgrading their networks. Also, pretty much nobody is talking about customers paying for their own QoS; the idea is for Netflix to pay for it.

barrkel• 14 years ago

If Netflix pays for it, where does Netflix get that money? Customers again. The customer always pays, whether directly or indirectly.

The only difference is in the mix and distribution of customer money flows. Some customers will be freeloading to a certain extent, others will be paying over the odds.

lsc• 14 years ago

I've worked on a consumer network that did that. It was a company wiring up apartment buildings in a college town. They would drop 2 T1s into an apartment, then plug them into a FreeBSD box, with a proprietary QoS kernel module (this was the early aughts) and then they'd use 802.11 to connect up the people in the apartments. My job was handling the FreeBSD box and writing ipfw rules to identify and prioritize different classes of traffic. (I am blanking on the module name. It was a long time ago.)

Obviously, the kids are going to absolutely trash that 3Mbps.

They were killed by DSL, which gave you way more bandwidth for way less money vs. the T1s. Doing the same thing with DSL at the head doesn't work because QoS works by reordering the packets in the queue. To get packets to queue, you need to know how big your pipe is, and you never know exactly how big your DSL pipe is. (or, at least, that's how it looks to dumb sysadmins like me who only get to play with a server connected to the dsl modem through a ethernet port.)

Anyhow, my belief is that it's turned out to be cheaper to just buy more bandwidth. Bandwidth is incredibly cheap, and it keeps getting cheaper every year.

dgregd• 14 years ago

> customers paying for a guaranteed percentage of high-QoS > bandwidth during peak network load

This and all similar ideas are very dangerous.

Compare pure commercial networks (mobile SMSes, ISDN) with Internet which has academic background.

What is the right direction for Internet? Why we talk about paid QoS packets ideas and not about Google Fiber?

nitrogen4• 14 years ago

I would definitely have more bandwidth rather than traffic shaping, but I also want my telemusic, teleconferencing, and gaming packets to arrive at their destinations ten milliseconds ago, and not get dropped the same way my bulk HTTP transfers should during heavy load.

Even with plenty of bandwidth, there's still the buffer bloat issue. If you decrease the buffer sizes and use random packet dropping, your VoIP and gaming packets could get too many dropped. If you leave the buffers as they are, you could have very high jitter. So, either we need a way of saying we want high-priority packets not to get dropped when the buffers are full, or we need a way of saying we want high-priority packets to bypass bloated buffers.

jrockway• 14 years ago

I can't help but wonder why they didn't try this 20 years ago. As it stands now, it seems hopelessly out of touch with reality. Why would people start paying more for online services when the price of everything tends to decrease over time? And why do governments think that the big internet companies are going to pay for this? They have shareholders too, after all.

derleth• 14 years ago

> I can't help but wonder why they didn't try this 20 years ago.

20 years ago, nobody cared about the few, the unusual, the networked. 15 years ago, people damned well did care, but everyone was in 'run in circles, scream and shout' mode and business plans were more-or-less optional. 10 years ago it was definitively too late. Does that seem like an eye blink to you? It does to a multinational bureaucracy.

My life fades... the vision dims... all that remains are memories. Looking back over the history of the Internet, from about 1991 to 2001, I remember a time of chaos, bizarre dreams, and the growing suspicion that the average people more-or-less snuck online too quickly for anyone who could have been said to have been 'in charge' to notice.

zb• 14 years ago

The whole concept of "sending party network pays" is just an elaborate wordplay. Relay-switched telecommunications networks are inherently symmetric. The concept of a "sending" party is meaningless; the originating party pays. This makes a lot of sense, since the originating party is the one in a position to decide whether to place the call and therefore should bear the cost.

By renaming the originating party to the "sending" party, they can redefine YouTube/NetFlix/whoever as the "sending" party even though the user is the originating party. Thus they attempt to justify denying the user the service which they already have paid for unless they receive the appropriate kickbacks^W "fees" from whoever is serving the data.

mariuolo• 14 years ago

I really don't understand. I thought the planned ITU involvement was supposed to be limited to the ICANN/IANA bailiwick.

KaoruAoiShiho• 14 years ago

A bunch of different corporate entities getting together with governments to jack up prices, how is this any different from a cartel.

smsm42• 14 years ago

I'm not sure I understand this idea at all. To me, it sounds a bit like gasoline companies asking from Walmart or Costco to pay them because a lot of people drive to Walmart and Costco using their gasoline. That may be true, but they pay exactly for that! Moreover, the only reason why people buy gasoline is to drive to places, so where the idea that somebody owes them extra for that comes from?

In the same way, internet access is a commodity and the only reason I pay for it is that I need to get to Google or Netflix or such. Where the idea that Google or Netflix get "free ride" comes from? I pay for my access, Google and Netflix pay for their connectivity. If the connectivity is not enough, customers would complain and depending on where the problem is, end up paying for better local access (if they know it's slow for them but fine for the neighbor next door that uses different provider) or have their provider (i.e. Netflix) upgrade their connectivity by telling them to improve delivery or they stop paying/visiting/clicking on ads. This all should work as it works now, where the idea of a "free ride" and additional payments due comes from?

ef4• 14 years ago

While we're at it, don't bill phone calls like phone calls either.

Dear ITU, please don't bill Internet use like phone calls

Related Stories