Netflix Ranks ISP Speeds

Carnth writes "Netflix will start releasing monthly ISP speed reports for the U.S. Google Fiber ranks at the top. They say, 'Broadly, cable shows better than DSL. AT&T U-verse, which is a hybrid fiber-DSL service, shows quite poorly compared to Verizon Fios, which is pure fiber. Charter moved down two positions since October. Verizon mobile has 40% higher performance than AT&T mobile.' Hopefully this will give consumers a better overall picture on how their ISP performs compared to others."

Comments

Only ranks major ISPs

[MrEricSir]

There's plenty of smaller ISPs that get better speeds than many of these providers. Would have been nice to see them on the list along with the heavyweights.

Re:Only ranks major ISPs

[crow]

I'm surprised that Google Fiber is large enough to get ranked. I would have guessed that there were other regional ISPs with more customers that weren't listed. Perhaps they're listed simply to encourage the others below them to pick up their speed.

Re:Only ranks major ISPs

[Rockoon]

They probably only wanted to show ISP's that the majority of Americans have access to.

That is surely why Google Fiber and FiOS are in there.

Re:Only ranks major ISPs

[Anonymous Coward]

For that matter, I bet that all those "students" who who watch Netflix in the university library could show a 5000% better connection than anything Verizon or AT&T has to offer.

Re:Only ranks major ISPs

[Arker]

If they provided a way to narrow it down and see only the ones in your area, it would be very useful. Big ISPs are killing the internet and any sort of consumer guide that presents their information without that of smaller competitors is ultimately a disservice. That said, this information is very useful and interesting, and I would encourage them to continue posting it - just please make it more inclusive. My provider is a small customer-owned co-op and the service is extremely competitive - it would be helpful for that information to be available alongside ratings for the industry giants.

Re:Only ranks major ISPs

[korgitser]

My guess is that they just want to ack some pressure on the big ISPs who all want Netflix to cough up for outbound traffic.

Re:Only ranks major ISPs

[Black LED]

You might be interested in Net Index. It's run by the guys who run Speedtest.net. You can look at various ISP rankings by regions.

Re:Only ranks major ISPs

[ILongForDarkness]

Small regional players probably don't have all the media hype that Google Fiber has gotten. That said 2.55MBPS how cute. Admittedly I can't benchmark against Netflix since it sucks ass in my country but I get 18MBPS (not Mbps MBps) pretty easily from torrents and large sites like MS, youtube etc. It does get pretty annoying actually to have that much speed at times sometimes I open a streaming video somewhere and download the whole thing before I realize it isn't the video I was looking for where as with a slower connection maybe only 1/10th of the video would be loaded before I can click the next/back button. I'm sure there is some limit to how much Netflix will push to you regardless of your bandwidth (you only need to stream so fast to keep a decent buffer on your video).

Re:Only ranks major ISPs

[icebike]

Given that Google Fiber is not huge leaps and bounds above the top 4 contenders, I suspect in-home infrastructure is the limiting factor here.

Comcast and Fios are close contenders, although we don't know where those were measured. Comcast can be very spotty in some locations and
just great in others.

Re:Only ranks major ISPs

[Anonymous Coward]

the problem with comcasts service is that the netops side is not one unified force. the company is basically a big conglomeration of local markets all marketed under the same brand. how things are done in one market can be radically different than the way things are done in another market. the backbone and the connections to it are wonderfully run, but the closer you get to the edge of the network, the levels of quality start to vary based on how the local markets operate. they have a great deal of autonomy and as long as they make their numbers, they don't get bothered.

(posting AC as im currently a comcast netops monkey, and the internet never forgets)

Re:Only ranks major ISPs

[Anonymous Coward]

It does get pretty annoying actually to have that much speed at times sometimes I open a streaming video somewhere and download the whole thing before I realize it isn't the video I was looking for where as with a slower connection maybe only 1/10th of the video would be loaded before I can click the next/back button.

Oh you poor baby! That reminds me. I sure hate when I am standing on the deck of my yacht and I light a Cuban cigar with a $100 bill only to remember that I'd rather have my butler bring me more martinis to drink before smoking. I mean doesn't that just plain suck? Other people think THEY got it rough, well buddy they should try that sometime!

Re:Only ranks major ISPs

[ILongForDarkness]

No the thing is I have a traffic limit after which they charge me. So: wasting traffic still matters to me since it effectively costs me 50c per GB (both for upload and download which essentially makes being a nice guy with torrents cost me double)

Charter plain and simple sucks

[Khyber]

These people lie constantly. When I signed up for Charter, I asked if I could run temporary instances of game servers so I could play my favorite games online. They said yes. That's a big lie, they block pretty much every port. I call to talk about this, I get sent to business class support, which ends up saying "We don't block anything over port 8080, so you should be able to run your games just fine."

Nope. Can't connect or host shit on my PS3, or my computer.

Then to boot, I'm paying for 100 mbit down. I can NEVER get more than 30mbit down.

Charter is a business full of false advertising and sheer incompetence. Avoid these fuckers like the plague if you can. As soon as Verizon FIOS is available here, I'm ditching Charter. Fuck those lying sons of bitches.

Re:Charter plain and simple sucks

[Anonymous Coward]

Charter was great when I signed up. I even got the 100meg service and it was amazing.. For a while.

A few months ago the services started taking a dump. I went from 100mbit all the time, all day, to a laggy 300kbps in the evenings. (300kpbs down and 5 megabits up - How fucked up does your network have to be for that to be true?)

It's not on my end either. I had their techs at my place for eight hours making sure they had they cleanest signal they've ever seen.

I hear it's pretty much system wide. Their whole network is in the shitter and they don't seem to be doing much about it.

Re:Charter plain and simple sucks

[ducomputergeek]

I had the 100MBs charter business line into my house up until a year ago for work (sold my company so they were no longer paying for it) and went to their standard 30mbs connection. I was never hosting any servers, but was involved on a large software project that transferred several gigs of data each day doing repo pulls and pushes, etc.. What I found wasn't that I was having connection problems on my end, but it was the servers that I was connecting to which seemed to be the bottle neck. I tested this from the main office which had a 100Mbs fiber line and found much the same that the most the remote servers we were using would allow us to pull was about 5MB/s sustained. I used to stream movies/tv from hulu on my iPad while waiting for code to download/upload and sometimes while playing my XBox all at the same time. Bandwidth never seemed to be a problem.

Even now on the 30Mb/s connection I don't really notice any problems even if other people are over and using their computers/iPads/Phones and whatnot.

I think the problem with Cable in general is a lot depends on how many users are on your line. I know for a fact that I am one of two houses on this line with cable internet. And the other house on the street is currently unoccupied while being renovated. Everyone else switched to Direct TVa couple years ago and are older and don't use the internet.

Re:Charter plain and simple sucks

[mysidia]

and found much the same that the most the remote servers we were using would allow us to pull was about 5MB/s sustained.

Did you rule out the possibility that it was just the end-to-end latency, because the server was far away?

Remember, with the TCP protocol, as end-to-end latency or distance increases: the maximum possible throughput decreases, and the minimum TCP buffer/window size required to achieve the maximum possible speed increases.

E.g. at 100ms round-trip latency, you have to have a TCP buffer size in excess of 256 Kilobytes, to get a throughput of 20 Megabits/Second; which requires special tuning at both ends of the connection.

If your TCP buffers are stuck at 64K; the best possible transfer speed at that latency will be 5 Megabits; even if you have 1 Gigabit of throughput to the server, and the server has 1 Gigabit of throughput to you end-to-end.

Re:Charter plain and simple sucks

[gravis777]

I've had Charter for years. Never had issues.

The 100mbit is really stupid for most end users to have. The issue isn't with charter, but with who you are connecting to. You are not going to get 100mbit downloads from anywhere. The 100mbit line is really for people who have like 5 kids, and everyone is trying to stream HD Vudu or download torrents or play games at the same time.

I had 30, and it was overkill for me. I finally reduced it down to 15mbit, and I was more than happy - I was still able to view Netflix or Vudu HDX while I was downloading on two PCs simultaniously without a hit in speed.

People just don't really seem to understand - faster connections on your end doesn't mean you have that connection to every single server in the world.

Also, Charter doesn't block any ports that I know of. Open up a DMZ on your router.

Now, I just moved, and went from Charter to Time Warner in the new area. Now, you might think "OMG, Time Warner is number 7, they must be horrible" but notice that if you actually look at the speeds on the chart in the article, Fios, which is number 2, is 2.19Mbps, Time Warner is 2.12. That really is not that big of a difference. Even Uverse, at 1.94, that is only a 200k difference, which is roughly a 10% difference.

Now, you start getting below UVerse, and the speeds get horrible. However, I think CentruyLink and Windstream tend to be providers for apartment complexes and such, so you are probably on a shared connection. With the DSL speeds, that is probably average, and I am willing to bet that many of those people have the $15-$20 a month 768kbps package.

This chart is pretty much useless. It doesn't show the potential that an end user can get, it just shows what the average speed of a Netflix stream on that connection is. I am willing to bet that many people have the slowest broadband speeds their ISP offers. 6Mbps DSL is plenty fast enough for a Netflix HD stream. Before I got Charter, I streamed over 6Mbps DSL. Worked perfectly.

Averages with how much deviation?

[redelm]

One of my pet peeves as a numerate person not impaired by statisto-phobia is the [ab]use of averages. Sure, the mean contains some information. But the standard deviation contains just as much, if not more! Very seldom do I see anything from which sigma could be inferred, yet whenever you collect data for averages, you can easily calc sigma.

In this case, network averages are useful only for advertising and not much use at all for consumers, with the possible exception of some large corporations who might reasonably suppose they have enough users spread evenly so they _on_average_ will see the average.

For individuals, what matters is the service you will see. And that depends with any carrier more on the neighborhood loading and upstream provisioning on that node.

The only real info you might guess from averages, provided you can make some reasonable assumptions about wirespeed, is what percent of a providers customers are under-provisioned. If cable is commonly 6 Mbps and DSL is 3 and they both net 2, cable is horribly cramped in spite of higher bandwidth.

Re:Averages with how much deviation?

[ArsonSmith]

The reason this is all released is a way for Netflix to fight the chance that their service gets throttled. It's a free market solution to anti-netneutrality legislation. I like it.

Re:Averages with how much deviation?

[girlintraining]

Sure, the mean contains some information. But the standard deviation contains just as much, if not more! Very seldom do I see anything from which sigma could be inferred, yet whenever you collect data for averages, you can easily calc sigma.

I can understand and appreciate your frustration; I share it. But let's be honest: The average person doesn't understand sigma, standard deviation, margin of error, or any of those other statistical concepts. They do like "Top 10" lists though and rankings. And for these things, averages are usually the best metric, even if they don't tell the whole story, or even a particularly accurate one.

The other thing is, most of the ISPs on that list are using some variety of traffic shaping. Internet users don't care whether their download takes 5 minutes or 5 minutes and 30 seconds... but they're going to throw a hissy fit if their video starts in 10 seconds instead of 4, even if the remaining duration plays without a problem. Needless to say, ISPs aren't blind to this -- they prioritize traffic to sites like Netflix. Or, in the case of mobile providers... they throw it under the bus. But network neutrality doesn't exist in the United States or the UK, which is where Netflix operates... so even a detailed statistical analysis wouldn't be terribly useful.

We can't expect the average person, with the attention spa--oh look a kitty! ... to be bothered to see the deeper truth that a full statistical exploration would reveal. Bluntly, they're too stupid to either know, or care.

Re:Averages with how much deviation?

[redelm]

Averages of averages certainly are subject to the Central Limit Theorum and have diminished deviations.

But a simple global sigma would give the informed a probability of being better than someone else. Rather than assume a better average always applies.

Re:Averages with how much deviation?

[symbolset]

They have to have 5Gbps of Netflix traffic. Based on the figures in TFA, maybe a couple thousand Netflix users. They're load balancing, so target is 5Gbps per box - the boxes can do a peak of 8Gbps. Netflix makes them available for free to reduce the cost of networking and improve the customer experience. Network operators take them for the same reason. There is more here, including an install guide and BOM.

As of consumers can do anything

[guspasho]

Each of us only has one, two, or maybe three (if we're lucky) options to choose from, does it really matter if some ISP that doesn't serve my area is faster than the ones available to me?

Re:As of consumers can do anything

[alen]

either way the differences in speed between the top ISP's aren't anything to get excited about

blu ray quality is around 30mbps. a difference of 2.2 to 2.55 won't be noticed

Vroom Vroom!

[rueger]

OK, I've owned enough fast cars to understand that faster is always better, but in practical terms just how much download speed does anyone really need?

We're using Shaw Cable in Canada, the budget plan, and thus far it does everything we want, including downloading distros and (surely paid for, not pirated) movies in a reasonable amount of time, and streaming video via our Sony BluRay player.

Maybe I'm just an old fart that remembers 300 baud, and the amazing jump to 56k, but really folks, what in God's name are you doing that requires more than cable Internet speed?

(awful rich for Netflix to pretend to be looking out for consumers when their own service rips off Canada customers by offering 1/4 the choices at the same price)

Re:Vroom Vroom!

[Midnight Thunder]

I have a 6 MBit connection, which is good enough for HD video from Netflix.

The only reason I keep it is because, beyond the odd throttling, I don't have cap. If I have to go capped with a higher speed connection, then I need to have 300GB minimum. Its not so much I download that much, but I don't want to be stressing about going over the limit and be gouged.

As for the selection with Netflix in Canada, is you aren't really paying for a US or Canadian account. You are simply paying for Netflix and then getting the selection for the geographic region in which you are currently located. This means if you go on holiday in the USA, then you get the USA content selection and if someone from the USA comes to Canada, then they get the Canadian selection. This aside, I don't know whether the poor selection in Canada is down to Netflix not making the same effort to negotiate the rights or whether it is down to the content companies in Canada just being more difficult.

Re:lol @ your shitty speeds in the US.

[hawguy]

If we lived in a tiny country like Sweden or Japan, it would be easy to have infrastructure in place for good internet speeds for all. "Unfortunately" for us, we just have too much room here in the USA.

You can't blame it all on geography, I live in a small, densely populated city (with density exceeding many Japanese urban areas) located very close to Silicon Valley and my only options are Comcast Cable internet or "up to" 3 mbit DSL.

Remote speeds the same as local speeds

[Sycraft-fu]

I want data to stream just as fat from a remote site as it does from my local drives. That way, where something is stored isn't relevant, it is all the same speed.

That would take in the realm of 10 gigabit.

Or maybe fully uncompressed video, that could be nice, particularly for games but in general for having a more simplified receiver. Well that's over a gigbit for 1080p 24fps, 8-bit. Going to 1080p 120fps, 8-bit is near 6 gigabits per second. Gets even worse if you want to go 10/12 bit and/or 4k resolution.

Something less ambitious? Ok how about just better HD streaming. Blu-rays are generally in the realm of 25mbps for video, often another 10+mbps for audio. I'd like to stream stuff in that quality, it looks noticeably better than the Netflix HD streams.

Speaking of video streaming I'm hoping to see some better content some day, that'll require more. I'd like a 4k 60p stream. Going to need a lot more bandwidth for that.

10-20mbps Internet works fine these days for most things, but that doesn't mean I can't come up with a lot of uses for better Internet speeds. Until it matches local speeds (which it isn't ever likely too) there is room for more speed.

Don't Know What We "Need" Until We Have It

[cmholm]

I was one of the first DSL customers in Hawaii. At the time, my non-technical circle didn't see the point. "Always on? 10x faster? 2x as expensive? Whatever for?" Indeed. Based on your (I suspect) tougue-in-cheek comment, I'd note that neither distro d/ls or streaming video would be possible without it.... but we didn't know until we could *could* do it.

In Australia, they're busy debating whether the proposed National Broadband Network of fibre optics links is "worth it". What would you run over it that we can't run now?

It hasn't been invented yet.

more importantly

[Anonymous Coward]

how many HD movies can be watched each month on each provider without horrific overage charges?

Re:Google Fiber?

[Bengie]

Flush a fiber end down your toilet
http://www.google.com/tisp/install.html

This is has always been a lie

[symbolset]

According to the US Census bureau, 4 out of 5 Americans live in an urban area. Yes, we have some wide open spaces. But that's what they are: open spaces with no people in them. The vast majority of humans live in clusters that would bring the cost of broadband down.

Re:This is has always been a lie

[thetoadwarrior]

All of Europe isn't a city and we manage to get broadband to rural people. I suspect NJ is more than comparable.

Re:This is has always been a lie

[symbolset]

Look, I'm going to give you an example of what you would consider "not urban": Grant County Washington, US, served with Internet by Grant County PUD. 91,000 citizens. 2,679 square miles. 35 citizens and 10 homes per square mile. 64 acres per home. It is almost the least populated county in the state per square mile. Seriously cow country out here. It turns out the homes are still clustered in nexuses, and the cows roam in the vast areas between. This is the kind of place where your neighbors don't bother you about your personal firing range.

They've had gigabit-capable fiber broadband in Grant County to every home for over a decade, and turned an embarrassing profit at it as they're a nonprofit PUD. And they got into it accidentally, with technology that was then as dear as unicorn blood and has since become as cheap as rice. It was actually originally a project to save money on power meter reading labor using SCADA power meters that didn't work out because the vendor folded/deprecated the device.

Tell me again how population density is an issue. If Grant County WA US PUD can wire their 35 people and 10 homes/square mile folk with gigabit broadband fiber 10 years ago - accidently, surely there's money in giving that to people who don't live in a vast desert wasteland now, given advances in technology that have improved network performance over fiber over 10,000x in the meantime.

For comparison, the population density of Los Angeles County California is now 7,544 people per square mile, not 35. It is over 200 times as dense - and this is now when the tech is cheap, not before when the tech was expensive. How could you NOT make money at that?

Re:Google Fiber 2.55 MB/s?

[edjs]

Netflix says for the best quality setting to expect about 600 kB/s of traffic for HD programming (or 5 Mbps), so only households with multiple streams going will exceed that.

If Google Fiber is a "Major ISP", what about epbFI

[JohnA]

Google Fiber has nothing on Chattanooga's epbFI footprint and speeds... this is the message I sent Ken:

Hi Ken,

I saw your latest blog post, and noticed that you included Google Fiber, but not EPBfi. Chattanooga's fiber network is much larger than Google's pilot, and as a customer, I know that I have never received anything less than your highest XL level stream.

Please extract statistics for epbfi.com and epbfi.net to show our statistics.

Thanks,

John

Re:lol @ your shitty speeds in the US.

[thetoadwarrior]

Bullshit. The states with higher population densities than European countries still have worse broadband. Americans just accept crap for a high price.

Re:Define "rural"

[BorgDrone]

Yes, my parents live in a thinly populated area of the Netherlands and they have fiber, also all the houses and farms outside the villages are getting fibered up, some several kilometers from any community (the whole area is mostly farmland).