Slashdot Mirror
Google Fiber: No Charge For Peering, No Fast Lanes
An anonymous reader writes "Addressing the recent controversy over Netflix paying ISPs directly for better data transfer speeds, Google's Director of Network Engineering explains how their Fiber server handles peering. He says, 'Bringing fiber all the way to your home is only one piece of the puzzle. We also partner with content providers (like YouTube, Netflix, and Akamai) to make the rest of your video's journey shorter and faster. (This doesn't involve any deals to prioritize their video 'packets' over others or otherwise discriminate among Internet traffic — we don't do that.) Like other Internet providers, Google Fiber provides the 'last-mile' Internet connection to your home. ... So that your video doesn't get caught up in this possible congestion, we invite content providers to hook up their networks directly to ours. This is called 'peering,' and it gives you a more direct connection to the content that you want. ... We don't make money from peering or colocation; since people usually only stream one video at a time, video traffic doesn't bog down or change the way we manage our network in any meaningful way — so why not help enable it?'"
Fiber server? huh?
I realize that it's all marketing hooey, but I wish that the director of network engineering for google wouldn't mish mash terminology like that. Keep that for the marketing droids.
Why is there not a slashvertisment tag on this story?
So what do you make money from if I become a Google Fiber customer? That's what I'm concerned about. If it's just the fair-market cost of the service I'm paying for, then that's fine. If your noble stance hides the fact that you attach yourself to the fiber like a tick to suck value by monitoring my use of the service and selling that information to the highest bidder, then we have a problem.
Uuuh huh...
“He’s not deformed, he’s just drunk!”
If your noble stance hides the fact that you attach yourself to the fiber like a tick to suck value by monitoring my use of the service and selling that information to the highest bidder, then we have a problem.
Why do "we" have a problem?
There are plenty of people (including myself) that would happily trade the devil we know (Comcast/Quest/etc) for the unknown of reasonably priced much faster connection speed, which just happens to also give Google some aggregate data.
I'm not really a fan of Google collections - I use their services sparingly for just that reason. But I think the value tradeoff in that case is pretty decent and only Google really has the power to break through local connection monopolies.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
as Google continues its transformation to an ISP. It is just still in the honeymoon phase.
Google wants access to all bits[it is their mission statement]. You should fully expect it act to keep others away from them as soon as it has monopoly over those bits....just like the ISPs it is advocating against right now.
This is just another instance of the enemy of my enemy is my friend.
This is Google's hedge against increasingly higher costs for peering and neutrality breaking ISP's, so why would they then turn around and be hypocrites by ruining the very reason they're moving intro infrastucture to begin with?
That said, an affirmation that they're peering neutral just seems like a puff piece for what anyone should already assume.
Does anyone have thoughts on Google spinning this out as a not for profit and make public backbones that are truly ubiquitous and marginalized?
Bye!
As in traditional Slashdot Fashion, I didn't dig through The Fine Article directly, but the summary makes it sounds like Google is claiming to be doing something "extra" by peering, instead of noting this is more-or-less the way it had always been until Comcast and the like decided to start abusing their status to squeeze other companies.
Something along the lines of "Google fiber will continue to observe true Net Neutrality by refusing to discriminate the various forms of Internet Traffic, and will continue to share access to our network without preference or special prioritization; this reflects our belief that recent changes implemented by other Internet Service Providers to ignore these long-standard practices results in a weaker Internet, potentially denying individuals and small businesses the freedom and potential equal representation which has been the Internet's greatest strength."
Except, you know, decently explained. They have the opportunity to not just emphasise that (for now) they're still doing the Right Thing (Or at least close to it), but that Comcast and their ilk are lying like dogs about their intentions and desired outcomes.
Hats off to google - but not really. This is what they do. This is what I have come to expect from every other ISP until this nonsense with Netflix paying that ridiculous ransom. These big old companies have been laughing all the way to the bank having sold all these speed tiers and now that we are more fully utilizing the service - and at the expense of their other business divisions - its become a tool of extortion. Instead of optimizing their traffic flows (for the benefit of all) they are holding this simple engineering work as a hostage. Regional interconnects used to just do this with a service request.
will work for dragon quest localization
We don't make money from peering or colocation; since people usually only stream one video at a time, video traffic doesn't bog down or change the way we manage our network in any meaningful way
"One person" may only stream one video at a time, but "people" as a whole may stream thousands or tens of thousands of videos all at the same time, and that's what creates the bottleneck in the peering connection. These same "people" are the "people" who currently stream videos over Comcast et.al. and create the peering bottleneck between Comcast and Level 3.
What keeps the same thing from happening to your gateways? And what keeps the price for your service from going up as you have to add more bandwidth to your peering arrangement to deal with ever-increased levels of streaming? Or will you try charging the data source for the extra bandwidth so you don't have to charge your customers directly?
You say you don't want to make money from the peering, but you also don't want to lose money. The costs have to go somewhere, and the customer is the most likely recipient.
As I understand it, peering is the problem that Netflix is having with the ISPs. The problem is that the ISPs refuse to upgrade their equipment to allow the peering to run at fast enough speeds - basically the peering connection is too slow at the ISP end, and that's where the bottleneck is, not the "last mile". Simply saying that Google allows peering doesn't address the issue at all.
So.....it has come to this....
Is it sad that we've come so far as to have a company make a press release assuring customers and peering partners, that they will continue to abide by industry practices that have existed for decades?
...even if some party has to pay for it. Google is an ISP so their peering traffic is not equal. It is good for them and their customers to peer with as many popular content providers as possible. Connect eyeballs to content. I keep pointing out that Yahoo! did this years ago with huge success. It was reported that Yahoo! only payed for half of their total bandwidth requirements. That is, only half of their total bandwidth requirements were going over transit. This was years ago. "Fast lanes" are not new.
The difference with Netflix is that they had to pay the ISP for their peering. This is new. Even so, it still may work out for them. The the peering costs may still be cheaper than their transit or using a third party CDN. Like Google Fiber pointed out, peering does not prioritize traffic, it just makes links to networks. If peering is an unfair fastlane, then the Internet has always been "unfair" since peering is an integral part of the Internet.
So why does Netflix have to pay? It is called supply and demand. The market pressures are such that Netflix *wants* to pay to get their data delivered directly. I suppose they could have backed off and stopped using any sort of CDN with peering to ISPs. But then their transit costs would have gone up. I suppose Netflix could have done this and really slammed the ISP's transit connections until *every* customer was complaining about terrible performance. Netflix decided it was less expensive and better for their customers to pay ISP's for peering. Is this fair? As the saying goes, "Life is not fair." Deal with it.
The best way to deal with the situation is for cities to encourage new ISP's to build out last mile connections. Make it easy without a lot of red tape. Phone companies and cable companies will yell and scream, but there is nothing they can do legally. It is up to the city to manage right-of-way so that things don't get messy. So instead of complaining to the FCC, go to your city council and see what can be done to encourage Google Fiber to come to your city.
> since people usually only stream one video at a time, video traffic doesn't bog down or change the way we manage our network in any meaningful way
Ahem -- we had 7-8 people cramming into a 3 bedroom house during college time. Yes, we would have streamed only one video at a time per person.
So do they still ban residential users from running servers?
Nope. They explicitly permit non-commercial servers. From the Fiber use policy: https://support.google.com/fiber/answer/2659981?hl=en&topic=2440874&ctx=topic:
"However, personal, non-commercial use of servers that complies with this AUP is acceptable, including using virtual private networks (VPN) to access services in your home and using hardware or applications that include server capabilities for uses like multi-player gaming, video-conferencing, and home security."
Municipal fiber is the way to go. It would change the world and give the US economy a badly needed shot in the arm.
ISP costs have risen four times faster than inflation. We're on the road to having just two national providers. When that happens, costs will go up even faster.
1) Designate ISPs as common carriers.
2) Break up any ISP that provides content.
3) Take a bow for having brought about the digital revolution part 2.
Unfortunately, our elected jackoffs are too beholden to corporate money to do anything like this. Obama, who was supposed to be the first president who "got" the Internet, turned out to be the worst of the bunch, appointing telecom lobbyist Tom Wheeler has head of the FCC, and they're not poised to put the last nail in the Net Neutrality coffin. Obama is a failed president on that count alone.
You are welcome on my lawn.
It's worth remembering that Google also used to have a "Don't be evil" tag phrase or policy, but that changed rather quickly when they discovered that being evil was very profitable.
It's a safe bet that "We don't do that" will meet the same fate, and I doubt that we'll have long to wait.
"So why does Netflix have to pay?"
Because Netflix competes with Comcast/TWC/AT&T's ka-ching buckets-of-money-spinning video distribution platforms. If Netflix gets popular enough, Comcast is reduced to a dumb internet pipe for $50 a month (profit of $5), not a primarily a video provider ($100+ bills, profits of $20+).
Which is the problem. If Comcast *were* an internet-tube provider (only), they'd generally be pro-peering. They might try to charge Netflix some (they like money), if the market would bear it, but mostly it's to their advantage to peer. However, most of the ISPs in the US are not pure-internet providers, so if Comcast video can use Comcast internet to hamstring Netflix, that's a natural reaction.
Peering/Transit is a complex economic topic but it is never free. You have to pay for the hardware. I can say that everyone is free to peer with me for free. They *only* have to bring a Gigabit Ethernet cable to my basement. By the way, I live on a tiny island away from the civilization. Also, if my 8 ports switch is full, they have to buy me another switch.
> the existing infrastructure (yes, copper) works well for speeds up to what Google Fiber is offering and more (100Mbps - 1Gbps).
Explain how you can do that and we can both become billionaires.
> probably a regulated monopoly if the local laws require.
Which means it would take, on average, nine years to approve a service improvement. That's exactly what Google does not want.
Silly me. I first glanced at the headline and read "No Charge For Peeing" and I thought, "Oh, good. They're doing away with pay toilets.
Not enough coffee. Yeah, that's the ticked.
They can take my LifeAlert pendant when they pry it from my cold dead fingers.
Sigh.. its not even on their "considering" list.
Nor any other city in Michigan. BAH.
Peering is a prime example of an argument that would not even be taking place if there were any real competition among ISPs. I know there are differing opinions about it even among net neutrality advocates; but the way I see it, if an ISP advertises a particular connection speed, it is their responsibility to ensure that users paying for it are getting that speed for any service or website they access; assuming that said service or website is providing sufficient capacity on their end, which Netflix is obviously doing. On top of that, Internet providers have no right to complain about bandwidth usage when content providers are creating a huge demand for the service that they are selling. It would be interesting to poll customers to see how many of them shelled out for faster connections purely because of streaming services such as Netflix.
Hosting servers on your own network instead of peering achieves roughly the same goal; all it changes is which network the server resides on but that server still needs to connect to the network and effectively acts as a substitute for extra direct-peering links to the content provider.
By skipping the common NNI between CDN/peers/content-provider, the on-net content servers/caches effectively bypass congestion at the NNI layer and act as a fast lane. Different spin, same net effect.
Google knows that song fairly well since they provide content cache appliances to ISPs for Youtube.
It's not 1Gbps, but around here I can sign up for 250Mbps over copper. The local telco only goes up to 260Mbps on their fiber offering.
Write the laws such that the municipal utility must meet standards on congestion, speed, packet loss, etc. Set the fee levels high enough to allow for continual infrastructure improvement.
Around here the phone, electricity, gas, sewer, and water are all municipal or provincially-owned utilities. They seem to do just fine.
If it was only about equal traffic in both directions netflix could just have all their clients send random data back to their servers and then just drop all the data. That would increase the overall network load, but it would be "balanced".
Seriously, it makes no sense that increasing the overall network load would reduce the fees being paid. That's ridiculous.
Good to hear this kind of policy but they're still pretty draconian with their home-based server policy
I would support a model that actually reflects the real costs involved...that is a fixed monthly cost for the physical connection, and a variable per-GB charge.
The reason why most people don't like bandwidth metering is that the ISPs charge way too much per GB at the retail level. And if you lump the connection costs in with the bandwidth costs then the high-usage people end up subsidizing the low-usage people. It's much more fair to break out the fees separately (the way my gas/electrical/etc bills do it).
I think if end-users were charged a per-GB rate that was more in line with the wholesale rate plus a reasonable amount of profit then there would be minimal complaints.
Uverse maxes as 45 Mbps and requires a minimum of UTP drop. It does not work over "the existing infrastructure" (untwisted pair) unless that infrastructure has recently been upgraded.
All it takes to restore net neutrality is enough people to boycott those businesses that seek to compromise it.
That was the turning point of my life--I went from negative zero to positive zero.
This has probably been repeated several times, but Google Fiber is good because it's a real threat of competition, which should cause the ISPs to rapidly upgrade their networks.
I should actually have said that slightly differently. It's SEVEN years for the approval process itself. Two years is a reasonable estimate of the additional deployment time to meet regulator demands re deploying to less populated areas, etc.
> why does that necessarily follow?
There are 50 states each regulating several utilities, and thousands of cities and counties doing the same. What DOES happen is that approvals for service upgrades take, on average, seven years. That's just the fact. Why is it necessary? I don't know, maybe because in a republic, government bureaucracy is designed to be fair, not fast? Whatever the reason, it's a SLOW process, that's consistently true.
> Around here the phone, electricity, gas, sewer, and water are all municipal or provincially-owned utilities. They seem to do just fine.
US infrastructure has some real problems, but yeah it's normally okay to use sewer pipes that are 80 years old. The utility has a full crew that goes around patching in new sections when they break. Same for phone - untwisted copper that was installed in 1950 still works just fine. Using 1950s wiring for internet doesn't work, because that would mean no internet. Yes, those old POTS phone lines still work, exactly the same as they worked 20 years ago. I don't want internet service from 1994. If you're old enough, you may remember it took over 25 years from the introduction of touch tone to completely transition from pulse dialing to touch tone. That didn't even begin until touch tone protocol was approved, 20 years after it was defined. So you had the touch tone protocol defined in 1943, approved in the 1960s. and the upgrade finally complete in the 1980s. That's kind of okay for phones - they haven't changed much in the last 100 years. Internet technology is still in the stage of innovation and improvement. Waiting a total of 40 years from the time a new protocol is developed until it's deployed would be a problem.
Sure, right now, Google is just starting and will do anything and everything to make this work because they're not in a position to do otherwise. What happens if Google Fiber gets big, and can exert pressure on peered CDNs the way established ISPs can?
To add one more house, providing the same old 1980s cable TV service, sure. That shouldn't take more than a few weeks to approve. I'm talking about new types of services, such as Google fiber. What happened with cable TV when it was new? On August 1, 1949, FCC secretary T.J. Slowie noticed that a company was developing cable TV and launched an inquiry. In 1965, finally issued it's first approval.
Dear Google,
If you're reading this, (haha, I know you are!), please, come save me from the Comcast-Time Warner monopoly and their slow as molasses high-speed internet.
-- sudon't
Air-ride Equipped
What can you expect from someone whose soul has been killed?
> Single mode fibre is for Internet what copper is for phone lines. It works even if it is 50 years old
Predicting that fiber optic technology won't change in the next fifty years is pretty audacious. Are you the same guy who decided "noone will ever need more than 256k of ram"?
That gives you a maximum theoretical bandwidth of 900 Mbps from your house to the curb, assuming your coax has been upgraded in the last several years (RG6 with compression connectors, not crimp connectors). If they've installed multi-terabit lines to the curb you're all good. That's not exactly "the existing infrastructure (yes, copper)", is it?
To use "the existing infrastructure" means your neighborhood has a coax line capable of no more than 900 Mbps theoretical max. Let's see, divide that by 100 households ...
I'm not a big fan of some of the cable companies, but to say the pre-existing infrastructure from the 1980s is capable of providing gigabit speeds to each house is plain silly. I'm glad I don't have Comcast or Time Warner (Suddenlink has been good to work with) and I'm anxiously awaiting the arrival of Google fiber or something similar. I sure wish they'd all upgrade to fiber everywhere. There's a reason the upgrades are done by installing fiber, though. There's no magic wand they can wave to turn old coax based equipment into multi-terabit equipment needed to provide each customer with gigabit. If there were, nobody would be installing fiber.
That should read:
If they've installed multi-terabit fiber at the curb you're all good. That's not exactly "the existing infrastructure (yes, copper)", is it?
Pushing theoretically almost a gig on coax doesn't get there when there's one coax line line serving a hundred households. New infrastructure is needed.
since people usually only stream one video at a time, video traffic doesn't bog down or change the way we manage our network in any meaningful way...
Either they are delusional about a typical modern household usage patterns (4 people where everyone watching their own content), or the way they manage their network is mostly that are big winners who (like Netflix/Youtube) who have the resources to co-locate, and small-fry losers (folks that rely on generic backbone provisioning because they don't have enough money to pay to co-locate everywhere google has a drop).
Sounds great. Now hurry up and build it out.
referred to 2 parties in an imaginary conversation
Then "we" still do not have a problem, because Google does not have a problem.
YOU have a problem. You can say that because it's clear. Do not fear clarity.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Google can't possibly give us fast peering with no fast (or slow) lanes without Net Neutrality. They're a company, not a government! I can't believe any of you believe this could happen without a large bureaucracy enforcing arcane rules, written and administered by people who've never been network engineers.
It's not about making any connection faster; It's about making all other connections slower.
If I stream a Netflix video over a connection from an ISP who hasn't strong-armed them into a paid peering agreement, and packets are dropped until they pay up, it's not any faster once Netflix agree; It's as fast as it should have been before.
Finally had enough. Come see us over at https://soylentnews.org/
DOCSIS 3.1 from 2008 has 10Gbps/1Gbps links. You don't need to divide that by 100 households, these services are being oversold 1000:1 if not worse already. Also, you only need that to the distribution point, after that it's already mainly fiber. Even fiber services don't necessarily bring FTTH - many are still copper for the short distance to the distribution point.
Custom electronics and digital signage for your business: www.evcircuits.com