dslreports logo
spacer

spacer
 
   
spc
story category
Netflix: Broadband Caps Not a 'Successful Business Model'
by Karl Bode 10:30AM Tuesday Oct 08 2013
Back in 2010 Netflix claimed they weren't worried about the impact of bandwidth caps on their business. That changed dramatically a year or two later, especially as the company expanded their streaming offerings into Canada, where low caps and high per byte overages are the norm. Post Canadian launch, and Netflix began talking about caps as an anti-competitive weapon and a form of price gouging completely untied to economic or network realities.

This month, Netflix CFO David Wells argued that while the company allows users to scale the quality of their connections to manage bandwidth caps, they still don't see usage caps as a successful business approach for ISPs:
quote:
"I don't think that through an ISP, usage caps are necessarily a successful business model, outside of particularly low income and very value oriented consumers," he added, while highlighting that charging for premium broadband service could be a more effective strategy for ISPs. "I think the successful business model is much more, 'let me sell you a faster broadband and I'll charge you more for it,'" Wells said.
Granted the entire purpose of bandwidth caps is to protect TV revenues from Netflix streaming while cashing in on an explosion of Netflix streaming at the same time. That's technically a "successful business model," it may just not be a consumer friendly one -- and it's one only made possible by limited competition. Netflix recently stated they should start offering 4K video streams next year, with the lowest-end streams requiring 15 Mbps connections.

view:
topics flat nest 
rradina

join:2000-08-08
Chesterfield, MO

1 recommendation

Charging Based on Amount vs. Speed

I agree with NetFlix -- to a point. I know I'm on dangerous ground with the 640K-RAM-ought-to-be-enough crowd but will speeds eventually plateau? In other words, will there always be demand for faster and faster speeds to justify the more expensive plans?

For instance, in Google Fiber areas, what speed beyond symmetrical 1Gbps is enough to entice a lot of folks to pay more? 1.5Gpbs is 50% more speed. Is it worth 50% more cost or would it have to be 2Gpbs or 5Gbps before a lot of folks get interested. I know I keep saying a lot but if you only entice 100 folks to pay more for 2Gbps, that business model is going to fail.

battleop

join:2005-09-28
00000

1 recommendation

Re: Charging Based on Amount vs. Speed

The plateau is going to be at 1G. It's going to be a very long time and some what expensive to jump from 1G to 10G in the residential area. Hell it's going to take us 10 years to figure out how to max out 1g to the home let alone come up with 2 to 5gb.
--
I do not, have not, and will not work for AT&T/Comcast/Verizon/Charter or similar sized company.

TAZ

@qwest.net

Re: Charging Based on Amount vs. Speed

Agreed. GigE interfaces are a commodity at this point, so for residential users that's going to be the natural "endgame" for now. 10GigE is still too expensive and LAGs aren't going to be a common setup either.

Also, offering those speeds won't work with the shared bandwidth model right now. You offer GigE now because that's the highest a normal residential user can make use of right now; sell 1000 1 Gbps connections and you can safely assume that most of your users won't be using the full capacity of that link. Sell 1000 10 Gbps connections (or, for that matter, anything in between - say 2 GigE links in a LAG) and you know the usage levels will be much, much higher because the user has gone out of their way to obtain hardware capable of utilizing that link.
rradina

join:2000-08-08
Chesterfield, MO

Re: Charging Based on Amount vs. Speed

If the the CPE is capable of 5GigE and provides a switch with with 5 1GigE ports...
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus
"The plateau is going to be at 1G. It's going to be a very long time and some what expensive to jump from 1G to 10G"

I mostly agree. I see 1gb being around for a while, but what is going to happen is 10gb is going to come down in price to the point that when rolling out new services or replacing failed gear, 10gb will be cheaper.

Just look at the desktop. A 16port 1gb switch costs the same as a 16 port 100mb switch. The same thing will happen to 10gb, and that will probably happen in the next 10 years.

This will apply to ISPs, where the choice between plugging in a 10gb fiber ONT or a 1gb ONT will cost the same, and eventually 10gb will phase in and will be a selling point.

When you can advertise, "10gb Internet for $150" when everyone is charging $70 for 1gb, and 10gb is sooo hard to use, no one will really use it, then it's pure profit.

elios

join:2005-11-15
Springfield, MO

Re: Charging Based on Amount vs. Speed

as most users change over to smart phones and tablets Wi-Fi speeds is what will hold things back
likely its holding things back now as are the speeds of said tables
over 1Gbps your limited by the speed of your HDD bus
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus

Re: Charging Based on Amount vs. Speed

I find 1gb hard to use in my network. Even large file transfers are fairly fast, but that's the best part.

The end user will be the bottleneck. Their router, or their computer, or their WIFI will be the bottleneck. This means ISPs will be able to sell 10gb and not worry about anyone actually using it.

The benefit I would get as an end user is I would KNOW that the bottleneck is my home, so if I want something faster, I just upgrade at home.

I know quite a few people who will pay $150/m for the 100mb tier of cable when the $100/m 60mb tier would be perfectly fine for their 16KB/s for Black-OPs.

People will pay for "piece of mind" of not having to worry about stuff.

I don't see home routers handling 10gb anytime soon. The top of the line $200 home routers have gone from 500mb/s of WAN-LAN to 1gb WAN-LAN in the past 5 years.

My router has been speed tested for 480mb/s of routing, but if I let BitTorrent seed with too many connections, it takes a crap and starts dropping packets. Even with a lowly 2mb/s throttle on BT, my router will give me packet-loss if I seed too many torrents at once.

Try doing that with a 10gb connection. Short of a custom-built router, you will not be able to work a 10gb connection.

FreeBSD has some code in the works that will drastically help with routing speeds, but it will take a bit for it to get accepted into general appliances.
Skippy25

join:2000-09-13
Hazelwood, MO
You bring up one of the best points which is that the desktop/device becomes the choke point, not the connection.

There are few things that require or would be better with such a connection speed today, but tomorrow is another day. Video would improve, downloads would improve and response times would improve (latency) but music doesnt need much more bandwidth. Webpage graphics dont need much more bandwidth.

Single desktop drives can't sustain a 1GB throughput and would require disk stripping to do so. I personally do not know of a single VPN appliance that would allow a single connection at 1GB, even at the enterprise level. Could that become the norm or will drive manufacturers make that the defacto speed some day? It's coming, but wont be to the masses.

I personally think they should let it all run at line speed. Get the vast majority of people on the network and off the network as fast as possible and address those that have X+ % of utilization individually.

elios

join:2005-11-15
Springfield, MO

Re: Charging Based on Amount vs. Speed

RAID 0 with SSDs using SATA-3 would do the trick right now

bigger issue is Wi-Fi speeds
most 'joe and jane sixpack' types dont use a wired network at home
Skippy25

join:2000-09-13
Hazelwood, MO

Re: Charging Based on Amount vs. Speed

Did you miss the "single desktop drives" and the "won't be to the masses" part?

I know it can be done which is why I made a clear point that it is not common place today and wont be for quite some time.

TAZ

@qwest.net
said by elios:

over 1Gbps your limited by the speed of your HDD bus

That's not true. Most users are going to have at least SATA 3 Gbps, and many have SATA 6 Gbps. Even a Caviar Green will probably be bottlenecked by a GigE interface (and of course those are meant more for bulk storage than general-purpose use, so they're slower than what most people will be using) after taking into account packet overhead, and SSDs will absolutely, no-freaking-doubt, be bottlenecked by it. (Modern SSDs are actually bottlenecked by the SATA 6 Gbps interface, let alone a 1 Gbps interface, so consumer-level PCIe flash storage is finally becoming available - and THAT would be bottlenecked by even a 10 Gbps interface!)

tl;dr: local storage is almost always faster than the network
rradina

join:2000-08-08
Chesterfield, MO
If "no one will really use it", margin doesn't matter if it doesn't move the revenue needle, agreed? Perhaps there's limited merit as pure marketing but enough customers have to use that plan for it to generate substantial revenue. Otherwise it won't even pay for the production costs of promoting it unless there's some halo effect for slower plans.
Pervbear

join:2013-08-20
I really doubt it will take 10 years to find a use for 1gbit, with 4k res coming down in price and more devices becoming internet connected such as TV. 3g will not go away so when 4g 3d stuff becomes streamable and just 4k stuff period. And there are games getting large and soon we won't be even downloading our games for all devices they will be streamed much like on live did just better. It's all just a matter of time. Most kids have TVs and other inter connected devices, then parents room and living room, not just kids being the primary game players these days. I really do not find it hard to believe there will be a struggle for most family households to saturate most of a 1gbit line. Not all the time but most households are maxing current connections either. Each house will have a peak. The only issue I think many will fins is having a single service to max out a connection but that is quit different.

battleop

join:2005-09-28
00000

Re: Charging Based on Amount vs. Speed

I'm talking about individual services that REQUIRE 1Gb, not benefit from 1Gb service. All of your examples would certainly benefit from a 1Gb connection but none of them would require it. For any of the mentioned services 50-100-250mb would be very usable.
--
I do not, have not, and will not work for AT&T/Comcast/Verizon/Charter or similar sized company.
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus

Re: Charging Based on Amount vs. Speed

At some point the technology becomes so cheap that the overhead of managing different price tiers becomes too costly and causes confusion for the end user.

I fully agree with your speed assessments. Lets say the most amount of tiers that you want is 3. If you can easily support 1gb, then that is a no-brainer. So what are the other two tiers? 250mb? No, that will compete with your 1gb tier, so maybe a 100mb tier. Now you need a low-end for people who can't afford much at all. Lets say 10mb, just enough to do basic streaming, but will make gaming a pain if someone tries streaming Netflix HD.

So now you have 10mb, 100mb, and 1000mb. $30 for 10mb, $70 for 100, and $150 for 1gb.

They all cost the ISP the same, on average, as the bulk of the cost is the infrastructure, not the bandwidth.

I could see paying $150 for 1gb, and I'm obsessive about Internet speeds, but I'm not willing to pay much more than that because I am also practical and I have limits on luxuries.

BillbyByte

@comcast.net

4 recommendations

Sooner or later, the ISPs will figure out the only way to handle content providers stuffing GB streams down the pipes will be to start charging by the byte and not by speed. As regular broadcast TV channels dry up and point to point IP streams replace them, the best way to recover lost revenue will be to charge by the byte.
rradina

join:2000-08-08
Chesterfield, MO

Re: Charging Based on Amount vs. Speed

What revenue is an ISP losing? I agree that cable ISPs will lose significant revenue if that model dies but Verizon and ATT don't have a lot of video revenue and does the margin in what video they offer really offset significant plant expenses?

buzz_4_20

join:2003-09-20
Limestone, ME

Not in the Long Run

Of course it's not a good long term business model.

Too bad stockholders don't seem to care about the long term anymore.

Grecco

@myvzw.com

2 recommendations

Seems a little self serving

No offense, but monotizing things that have real costs is part of doing business. Netflix expecting free peering AND no usage charges is pretty self serving. Next they will expect to get network titles to serve for free

A business model built entirely on others infrastructure $$ and assets will never fly. This is why they resort to PR sound bites vs business logic. Good luck
music4praise

join:2006-02-22
Big Rapids, MI

Re: Seems a little self serving

The customer is paying for the pipe to use the bandwidth. If the customer uses the pipe for Netflix, it is his/her choice.

Probitas

@teksavvy.com

Re: Seems a little self serving

Cable companies and their partners would prefer you have NO choice at all. Hence their anti-competitive behavior.

Grecco

@myvzw.com
Isn't that an argument for usage since not every customer is the same or is a NF customer. Or should we socialize it?
dublin00

join:2005-12-29
Dublin, CA

Re: Seems a little self serving

Usage based billing will be the death of these businesses. The wholesale cost of bytes is very low and dropping all the time as technology improves. The experiments by TW and the like are based on hugely inflated per-byte charges that sound more like text-messaging revisited. The dumb pipe is a commodity, even Cisco is slowly collapsing because networking gear just can't command a premium in todays world.

WHT

join:2010-03-26
Rosston, TX
kudos:5

Re: Seems a little self serving

said by dublin00:

The wholesale cost of bytes is very low

That's not where the problem is. It's the last mile that costs. If I let everyone try to use Netflix at the same time on my 20 Mbps and 40 Mbps access points, the inter-node backhauls would choke. That means I would have to upgrade each one of my nodes' backhauls with a 24 GHz 1 Gbps radio pair. $4,000 per pair x 27 modes = $108,000. Who will for that?

You need to stop buying into those "cost is going down" sound bites and learn what it really costs to run an ISP.
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus

2 edits

Re: Seems a little self serving

My ISP can handle every user using all of their allotted speed at the same time. Their Internal network is fully non-blocking all the way to their backbone.

The speeds that they offer: $50 for 10mb, $70 for 30mb, $100 for 50mb, $200 for 100mb, $300 for 200mb.

All symmetrical speeds.

So, no, the last mile doesn't have issues unless you did it wrong(edit: Assuming recent technologies in the past 3-5 years).

As you can tell, prices jump up fast after 50mb, but I assume it costs more to have a full-non-blocking network at high speeds.

As for $108,000, that will get you about 2.1tb/s of dedicated bandwidth at an IX. It's really not that expensive.

edit: " That means I would have to upgrade each one of my nodes' backhauls with a 24 GHz 1 Gbps radio pair"

It just dawned on me that you're using wireless? btw, my ISP doesn't use nodes, all direct fiber, so that probably makes a difference.
Expand your moderator at work
Skippy25

join:2000-09-13
Hazelwood, MO
You clearly have no clue how the internet works.
DanteX

join:2010-09-09
kudos:1
Actually Netflix could get titles for free. Think of the vat amount of titles studios are sitting on or have been sitting on for many many years not doing a damn thing with them no profit to lose if you just hand over the rights to Netflix to stream if you have no intention of releasing them yourself anyways. All Netflix would have to do is cover the cost of shipping and producing streaming copies.

tshirt
Premium
join:2004-07-11
Snohomish, WA
kudos:5
Reviews:
·Comcast

Re: Seems a little self serving

said by DanteX:

Think of the vat amount of titles studios are sitting on or have been sitting on for many many years not doing a damn thing with them no profit to lose...

And the incentive to the studios to "give away" their libraries is?
Netflix could have a lot more users if the service was free.
Perhaps they should include Blu-ray disks with free shipping too?
Skippy25

join:2000-09-13
Hazelwood, MO

Re: Seems a little self serving

His point is simple: They are sitting on titles and thus making no money. They could throw them away and still make no money or give them away and allow someone else to.

You and I know a business would rather sell something even if they were going to throw it away. I personally do not have that mindset, but each to own.

tshirt
Premium
join:2004-07-11
Snohomish, WA
kudos:5
Reviews:
·Comcast

1 edit

Re: Seems a little self serving

They aren't making money on them today, They may at some point. And they certainly aren't throwing them out.
but not releasing them can be a strategy too. to "let the rest, before a remake" or to release it later to a generation that never saw it on the first run.
If it was really worthless, Netflix wouldn't want it.
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus
Soon gas-stations will not only charge customers for the fuel, but they'll also charge oil companies access to the customers! Want to sell your fuel? Pay us money! but... but... the customers are already paying.

In the end, Netflix will continue to pay rock-bottom prices for transit and the ISPs will have to also pay for transit. Instead of working together, they'll fight.

Who wins? Transit providers! Wait, this is a good thing. I want a stronger Internet backbone, not some CDN peering crap, with special routes for different services. The best kind of bandwidth is that via brute-force of a strong backbone.

I'm willing to pay $15/m for Netflix if it helps them purchase more bandwidth to throw at the ISP, forcing them to purchase more transit. Screw ISPs and their attempts to double-dip by charging for peering.

pjcamp

@spelman.edu

1 recommendation

successful is relative

The success of caps as a business model scales directly with the amount of competition ISPs face.
Bengie25

join:2010-04-22
Wisconsin Rapids, WI
Reviews:
·Solarus

1 edit

Neutrality

ISPs should be forced to enforce data caps again ALL data going to the customer, including the 4gb/s of TV data that is broadcasted to everyone 24/7.

They should not be allowed to differentiate between TV and Internet. It should all be "data".

edit:
We see that you have gone over your cap. We have a cap of 12,656,500GB. We will now charge your $10 per 50GB over.

This is essentially what is happening with a 250GB cap on your Internet, when the overwhelming bulk of data usage is TV.

•••••••••