Beginning on Wednesday, Comcast, the largest broadband service provider in the U.S., is going to start capping the total amount of data you can transfer using their broadband connection — to 250GB per month. With this move, the cable company will become the symbol of a new Internet era, one that is both monitored and metered. It is an era that threatens to limit innovation and to a large extent, the possibilities for new startups.
I have been very vocal about the short-sightedness of this decision being made by Comcast (and some other carriers), and along with my colleague Stacey Higginbotham, have been covering the story pretty closely. It is a clear and present danger to the way we use the Internet in this country.
In order to give you a better understanding of the issues at hand, I have teamed up with my old friend Muayyad Al-Chalabi, an alumni of Bell Labs and until recently an analyst with The Monitor Group, to release this white paper, “Broadband Usage-Based Pricing and Caps Analysis.”
In this paper, we aim to highlight the possible unintended consequences of such policies, among them the stunting of growth and innovation of web-based applications. And of course, higher costs. Plus:
- The strategy ignores the high degree of dependency “interactions” between power users and the rest of the network. The power users don’t act in isolation and in fact represent the hubs in any scale-free network; sequestering them and overcharging them will result in either low usage or worse, higher costs.
- Given the growth trend due to consumers’ changes in content consumption, today’s power users are tomorrow’s average users. By 2012, the bill for data access is projected to be around $215 per month.
- Strategic pricing involves the recognition that changing prices alone cannot solve the challenges facing carriers. Carriers are taking the easy way out trying to protect the “walled garden” rather than figuring out how to innovate in service delivery and harvesting more value from the overall content and applications opportunity.
If you’re interested in getting a PDF copy of this white paper, please enter your email address below. Otherwise, I have embedded a copy (using Scribd) for you to read it online.
95 thoughts on “GigaOM White Paper: The Facts & Fiction of Bandwidth Caps”
A quick glance at the paper was enough to prompt me to put my email down for a pdf copy so I can read through it at leisure.
Frankly I’ve been sceptical of the view that real innovation is so sensitive to costs. If the only way that some innovation can succeed is if it has zero marginal cost to the user, then it doesn’t seem to me that the ‘innovation’ in question is all that valuable a thing.
I’m going to be interested to see if you make me change my mind at all.
Well, 250g is a huge amount of bandwidth. Who uses that kind of bandwidth? I would bet it is mostly participants in a p2p network. As you know, the vast majority of such people are breaking the law by distributing copyrighted digital works.
I actually think that the failure by our government to police the distribution of such copyrighted works has left the door open to vigilante efforts like this — and the problem is that Comcast does not actually have the end-user’s best interests at heart. What they care about is profitabilty.
We are left with a vicious circle that hurts everybody.
Don’t we have a new name for this kind of management.
Means short sighted greedy without any substance.
For the _nth_ time it’s not about today it’s about the future. See above management description and look at the Investment Banks to what happens when such management practices take place. You bail them out or you pay higher subscription cost, take your pick.
I think it takes either an exaggeration or a rather narrow viewpoint to brand this “a new Internet era” – particularly outside the US, usage based charging is far from new. That said, metering, particularly in the dialup days when every minute spent online represented another penny or more of cost, did have a substantial negative effect on users, both deterring uptake in the first place and deterring users from spending longer than the minimum they could get away with. I doubt Comcast’s cap will have anything like that effect unless either data volumes explode without them adjusting it, or they slash the cap dramatically.
I would prefer a truly unmetered/unlimited service, but there have been far more draconian caps than Comcast’s out there for years without the sky falling on us.
Response to Allen’s comment:
My reference to innovation was not limited to the creation of new service or widget. Innovation can be achieved along multiple axes; Finance (business models), Process, Products and Offers (design, performance, systems, support services) and Delivery mechanisms. Innovation happens when there is a contradiction between two and more elements- in our case end-user demand and carrier cost.
What the carriers need is a mind set built around “creative destruction” (using Schumepeter’s phrase) – a process in which the old ways of doing things are endogenously destroyed and replaced by new ways. Cost-plus only method of pricing will not blend well with the explosive traffic growth and shifts in consumer behaviors. For example, if users had to pay for every web sites visited, then we would not have ad-based business models and users will have stopped using the internet all together.
My general feeling is that Comcast is looking at the future, and seeing their own internal bandwidth needs for content delivery. There is an inherent business conflict in being a content provider and an ISP. They’re most likely to sacrifice the latter to the benefit of the former based on the premium revenue they receive for content.
You couldn’t be more right about describing Comcast (and other BSP) and their failure to address the big question. I think FCC needs to step in and essentially separate services from the pipes – something they should have done in 1996 and instead we are back to the future. I continue to blame the FCC for all our mess.
I think you just made my point… essentially the usage based charging which was common place outside of the US is now going flat rate. And in places where usage based charging was present, the usage was not enough and innovation was focused almost entirely on making use of as little bandwidth as possible – which if you believe broadband is a platform doesn’t make much sense.
I stopped reading on the first page as soon as I saw the line “direct rate increases are hard to implement in a competitive market.” You can’t have it both ways. You can’t argue that the cable companies are being unfair because the market is non-competitive, and at the same time acknowledge that they’re targeting power users because competition prevents them from implementing direct rate increases.
In fact, I’m more sympathetic to the argument that the market is NOT competitive, but the author blew his credibility with that particular line. With the market as it is, I don’t blame the cablecos for trying to be as profitable as they can be. As far as innovation is concerned, usage caps and metered usage should spur entrepreneurs to come up with ways to deliver content with lower bandwidth usage.
As far as the market is concerned, we should focus more on getting regulators to encourage competition via broadband wireless, etc.
Too bad, you stopped reading, because if read further, you might have gotten the points of the paper. It is not just about fairness, but about business decision based on facts on the ground. It is just ironic that it is the same guys who fought À la carte video pricing are installing micro-pricing based on bytes.
Surely, the broadband residential market is not monolithic; some markets are hot and getting hotter (like New York City), but most are NOT competitive at all. It is this mixture of markets, products, national footprints, public scrutiny that prevents carriers from having different policies for different regions at this time. Thus, the tactic of indirect rate increases, initially aimed at soft targets, then affecting every user.
The paper points out the fact that the 5% of users who generate most of the traffic is a phenomena that is not constrained to broadband access, but occurs in natural networks like protein networks. Carriers can chose to fight it, but the results will not be as they expected. Hierarchical structures with different degrees of connectivity
If they are limiting you to 250 gb they should be required to advertise the connection speed as such:
(250 gigabytes) / (30 days) = 101.135802 kBps
What you neglect to mention, in your paper, is that you are one of those who are responsible for the imposition of download caps. By cheering the FCC on as it attacked Comcast’s more sophisticated network management techniques, you were among those directly responsible for the FCC edict which required Comcast to implement the caps.
I live in New Zealand, and we have bandwidth caps. I have the largest plan offered done here, 50GB cap, & pay roughly 150 bucks a month for it.
Typically my non geek friends only go for 5-10gb packages, and only costs them 30-50 bucks a month.
None of them can download like i can though ..
Yet somehow everyone finds it 100% possible to do anything. Although I pay more for internet, and I have had to cut down my usage since moving from Canada. I have only limited the amount of things I do, not stop doing things completely.
Although as a consumer I much prefer no cap + cheaper pricing offered by living in Canada. I honestly don’t know how these telcos ever made money in the first place with flat fees. if the ISPs pay per data, why dont the consumers? There is a direct proportion of the cost to the amount of data. If everyone who used the net paid for how much they used it, my folks back home who use MAYBE 1gig of their net bill wont be paying for the excess usage the neighbor down the street has managed to run up. Because the geek down the street knows he can do this.
Well 250 Gigs sure “sounds like a lot,” until it is taken into context that the average user in 10 years will likely consume that much or more. The advent of DVD for download, legally, along with music, and various high def data streams. I streamed the Olympics live when I could, regularly buy and download DVDs and send out home movies to various friends and family using the bittorrent protocol. I think this is all normal behavior, certainly ethical and legal. Between work and home my average bandwith is over 100 gigs and this is for an extremely small in house tech company. Can you point out where this makes me a pirate? Most of my family is in the same boat, sending out HD copies of home movies and birthday parties, weddings, etc consumes a lot of bandwith in and of itself.
250 Gbytes/month = 250 *10^9 * 8 (bits per byte)/(30 (days)/(24 (hours)/(3600 seconds) ~ 772 Kbps.
It is a nice number to talk about, but does not take into consideration peak traffic during busy periods (which can be 2-4 hours of prime time viewing, or 4 hours for a football game), nor does it deal with multi-user scenario where each person is viewing different or multiple streams of data.
Good point your bring up; looking at international peering data and connectivity from NZ, you re-enforce the point is that the lack of power users locally results in most of the traffic being served outside the carriers domain. Higher international peering with Singapore, Sydney, Tokyo and transit capacity to the US is a byproduct of such policy.
Families with two or three kids who also use computers will eat through 250g simply by the streaming video/music/skype. Then toss on the fact you have Vonage and you work out of a home office.
I wonder if comcast voip bandwidth counts.
I agree with this description. “Means short sighted greedy without any substance.”
I think more importantly this is an issue where regulation has to come into play. the cable and telephone companies are much less regulated than the banks who are playing havoc on the economy. I am sure the BSPs with caps are going to do exactly the same on innovation.
Comcast’s cap has a different objective than Time Warner’s.
Comcast will happily sell you digital HD television, which does *not* count towards their 250GB/mo cap (10x larger than TW’s). So Comcast is not trying to throttle use to protect an under-built system. Comcast is trying to prevent NetFlix and Blockbuster from competing with their cable tv business. And to motivate content sites to enter “collaborations” with them. Raising monthly rates is nice too. They are simply trying to leverage an access monopoly to develop and protect other revenue streams.
Lol, here in NZ we have had data caps for ages. I usually pay for about 30 GB per month. The only people I’ve heard complain about the data caps are those who are pirating movies.
Anon @ 4:16 PM struck exactly the right chord. Metering and caps may be ugly, but funneling capacity to one’s own content delivery business is the subtle ulterior motive.
I don’t think I’d ever even get close to using 250Gigs in a month. Our friends overseas have 1Gig caps. I mean THAT’s a CAP. Either way, I’m happy I’m with Verizon Fios.
Pay per Byte and throttles are just tricks they are using to keep you, the customer, from downloading content in direct competition to them, the cable companies. The scare tactic, the potential for a $200 bill from Comcast, will make many families vow never to download a movie, watch a TV show, or play an online game again for fear of going broke. The Cable Companies are hoping and praying they can kill the fledgling ‘on demand’ business by making it too expensive for the average family.
In the New York Metro market, where there is true competition with VZ in the mix, there are no hints of caps or throttles.
So does this mean that Australia, NZ and a heap of other countries are ahead of the US in terms of this technology?
I think the public at large has little understanding of the actual internal ISP network architecture, as is evident in most of the comments above. Briefly, in any give neighborhood there is one path all the data uses. I would hate to live in the apartment building directly across from the Google campus. The usage there would be crazy. You can’t expect there ISP to put in a 40Gps dedicated line just for them. Verizon is spending over 800$ per customer to install FiOS, and that’s just 30Mps. Every network has a bandwidth weakness. Comcast is spending hundreds of millions right now upgrading to Docsis 3.0 and bigger back bones, as is every ISP. The way I see it the guy downloading a hundred movies a day is no different than the guy driving the Hummer in front of me. One pollutes the environment the other pollutes the internet.
Comcast Business Class internet isn’t capped for little more with faster speeds. Why would a true “Power User” have regular service? You’re downloading hundreds of FREE songs and music; you can’t afford to pay ten bucks more a month to do so?
@everybody from over seas or with caps.
Let’s take a look where we (Internet) comes from. The network of interconnected networks developed because of no caps. I remember uucp storing email and data on my servers to be forwarded at night to somebody else on behalf of somebody else.
Next came the Web, would that have been developed if we would have had caps on our research networks?
Remember all the commercial networks at the time with caps? Or the pain to get anything running on them?
So I feel sorry for you if you have caps, but please don’t put me or my mind in a prison just because you are in one. Btw I don’t even come close to the Comcast cap.
And telling me they will enlarge the cap if more people are reaching it. Is just insult to me, why enlarge the cap when many people are hitting it. That will cost even more, so I must think they think I’m not very smart. Let me say this. I return the favor.
You are going to stop my innovation unless you give me lots and lots of bandwidth and a PONY! Right now!
Here in Canada, Rogers, the cable provider now has a 95GB/month cap on the 10mbit and 18mbit connections.
I don’t watch a lot of regular TV, most of my video is in video podcasts (cranky geeks, half of revision3) and of course audio podcasts. I was living on dial up during a transitional phase, and when I moved in to my new place, I was happy to get the cable ‘net. Until I was 3 weeks in to my first month and had already used 80/95GB. This was late August/early September, so there weren’t even new TV shows to “pirate”, it was all getting the high def/med def versions of the vidcasts, legit streaming video (Daily Show/Colbert and the like) and playing online games(and surfing like a crack addict coming out of rehab of course). When I called them, I was even willing to “upgrade” and pay extra for business class, working on the assumption it was unlimited, but the tier 1 guy didn’t know anything about business service, and the ‘retention officer’ that I had to repeat “if you don’t have high speed, I don’t want your service” over 5 times to wouldn’t even entertain the idea.
Bell is the DSL provider, and they have no limits posted on their site. Checked the info for the various levels, the EULA, and the ToS, and could only find reference to “excessive use” being “investigated”.
Other than a bit of oddness with yahoo messenger, the 5mbit DSL is performing as well and sometimes better than the 10mbit cable service.
The kicker is, 2 weeks after I canceled the account, Rogers called me up to do a customer satisfaction survey. Talk about right hand/left hand mutual ignorance.
It’s not only “pirates” who complain about usage caps. With more and more legal media being available for download, these data-caps are going to prove a problem. Music is already available for download and is proving to be very popular amongst the average user. DVD’s are now available for direct download, which can be 4gb or so. If you’re on a 30gb plan, that’s nearly a sixth of your usage gone. Then you’ve got to take into account games, which can also rack up sizeable chuncks of your cap if you’re downloading them, and then also if you take into account online play. All of these forms are legal ways of how you could use up your download cap.
One method that I think might work for those that are going down the cap route is for ISP’s to also be content providers. My ISP, iinet (iinet.net.au) has been fantastic in developing their Media Lounge, where iinet users get unmetered downloads from iTunes, ABC iView (TV streaming), and also a few select TV channels that they’ve set up – including Barclay’s English Premier League, NASA Space Channel, and WA Symphony Orchestra Concerts. They also have their own gaming servers which also provide unmetered content.
Now, if only Telstra would get around to providing us with internet speeds comparable with the US…
I received the email Sept 12th from Comcast and read it in disgust. I new it was coming, but it has been a huge topic of debate where I work at (a phone company where I do DSL support for the technicians)
Even though this does not ‘directly’ affect me ‘now’ (which Comcast wants me to believe), I understand what it will do for the future and how it IS going to affect me. Personally, I am disgusted – as I have been for quite some time.
Although this per se is not the reason I am leaving Comcast (far many other problems they refuse to fix even though they admit it is their problem when I went on a huge rant on my site), this may be the straw that broke the camels back for me. Enough is enough.
I may not be one of the ones downloading 250GB a month that will get hit ‘Directly’ by Comcast (as they say), but I AM one that a) wants what I pay for (which I am not getting currently) and b) I have interest in the innovation and technology in the future.
If I did not have have interest in that, I would not have made it to where I am now; and, in 10+ years, it says a whole lot! In 10 years we have come such a long way, but because of what this will start, we will not make it that far in the next 10 years or thereafter as we have today.
Comcast needs to open their eyes and see what negative influence this will have on the innocents of this.. instead of just blindsiding them to make them think they are not being affected. It will affect us all. Period. And, it is not going to change the habits of the ones they are trying to punish.. just push them further.
Least, that is my 2 cents.. right or wrong.
If my internet doesn’t let me do what I want to do then I switch providers. The company that decides to not throttle will get all the business once people start using more data.
Not sure what the big deal is all about, other than the fact you might have to call and talk to one of these companies to switch (which is a nightmare I know).
The “Power users improve network efficiency” section has [at least] one critical flaw. It doesn’t demonstrate that power users reduce the amount of data exchanged, it merely shows that a simpler topology can be used to exchange it. But that’s irrelevent to whole bandwidth cap argument, which is “charge-by-the-byte”, not “charge-by-the-connection”.
Assuming each users wants to download/upload 10GB of data, then Figure 1 simply means each connection transports an average of 1GB, while in Figure 2, each connection transports 2.5GB. The total data exchanged is still the same, regardless.
Muayyad Al-Chalab is an alumnus of Bell Labs, not an alumni. “Alumni” is plural.
“Levi Blackman said:
If my internet doesn’t let me do what I want to do then I switch providers.”
Thats the essential problem here in the US. There is very little competition in the broadband market here in the US. As a matter of fact, where there is real competition, like in the Metro NY area, the primary CableCo has yet to mention a ‘cap/throttles’ policy. As a matter of fact, Fios is stressing it’s ‘all you can eat’ for now.
Your argument is true, if all the users are exchanging data at exactly the same time, or network capacity is based on average usage over long time period (bytes per month). However, network capacity is fixed and is designed at peak load, amount of traffic generated during the period I don’t see how the capacity is increased to 2.5 Gbps. In reality, since there is no direct competition on inks, the capacity of each link to the hub cab be less than 1 Gbps. We need to differentiate between peak loads that affect capacity directly, vs. average loads that have no direct effect on capacity or capital needed to build the network.
Perhaps the innovation should be improving the use of available bandwidth. I remember when web design placed an emphasis on loading a web page as quickly as possible, when “lean and mean” was the order of the day. I also remember when an operating system didn’t suck up gigabytes of space on a hard drive, too.
Is it time to rethink what we really need in an OS and the Internet? Does everything have to have a bloated, graphics-intensive interface just to read simple text?
My grandmother raised a family during the Depression. In fact, they owned a small hardware store on the South Side of Chicago before, during and after said economic “event”. Her motto? Do more with less. Maybe it’s time for us to do the same.
At least Comcast is telling you that they plan on doing it, though that doesn’t make it right. When Adelphia was taken over by Time Warner, I swear we got capped even though existing Adelphia users with our plan would not be capped, not good.
One thing not taken into account in these postings:
The true hard-core geek.
Not just for games- what about those of us who download kernel source code, ISO images of distributions, watch (free by license) movies and shows online- not to mention the whole gamut of programs available online?
Back in 1998 when ADSL 1st became available here, there were 3 ”tiers” of speed available- 384/128, 768/256, and 1.544/512 … roughly.
These various speeds ranged in price from $30-$120/mo, no caps.
THOSE prices were just for the connection to the telco circuit.
Then- you had to choose an ISP to connect through, often with similar price scales for the Internet connection.
Nowadays broadband has come ”on the cheap,” with unlimited data usage in the Cincinnati market typically $15-$30/mo. I suppose this is one area where competition has served the market particularly well. It’s an even bigger bargain when you *bundle* services such as TV, landline, cell service, etc … to as low as $10/mo for the broadband- in some cases promotions are offered whereby you get wireless service included in the bundle at no extra charge with typically a 1-year contract agreement.
While it is true that there’s costs to providing bandwidth, I smell a big, stinky sewer rat here. What it comes down to is this:
Always Corporate Greed. Always.
I have been worrying about this lately. Although I use less than a gig per month, I could perceive that 250G is an average for tomorrow users. However, a VERY IMPORTANT CONCEPT that most people failed to see: USAGE DATA ARE GOING TO BE METERED!
Some of you missed the point I think.
TO those who say, “who could use 250GB in a month??” I could, quite easily in fact. I don’t use P2P, I download stuff directly from Google queries. Maybe some of you don’t realize just how much traffic occurs simply by visiting sites – especially media-heavy sites. It’s quite simple to download 10MB just by visiting a site such as…Adobe.com. Or SCEA.com. Or even Nintendo.com. Companies are relying more and more on the “rich content” experience, and the caps are counterproductive to that evolution. Couple that with normal internet usage – downloadable news WMVs, downloadable webinars, etc…and it’s quite easy to reach that amount in a month for someone who essentially lives on the Internet as I do. I do 99% of my business online. If I could gas up my car by going to a website, I’d do it.
All said, caps are definitely not the issue BUT the companies could (and should) really consider something different. The power users are the strength behind the structure, as mentioned. OK, so why not just charge the power users a flat rate up to a certain allotment, then incremental charges thereafter? True power users would be more than willing to pay for what they know they need to use. That’s how the dialup era worked. Instead of blatantly cutting off people who are effectively cash cows that could reveal millions in profits every year.
It’s the fact they want to cut people off that makes capping a bad idea. Most wouldn’t mind paying for excess usage as long as the prices were fair and commensurate on true usage vs. value. For example, if they charged $100 for a 15/2, that won’t fly. However, I’d gladly pay $50 for my 15/2 and if I go over a 250GB cap in a month, $5 for every GB over that. That’s more than reasonable on both sides, and it’s basically free money for the providers.
(CONT. FROM MY LAST COMMENT) These companies are trying to STEER THE ARGUMENT AWAY FROM THE FACT THAT ‘USERS ARE GOING TO LOSE THEIR FREEDOM’ by capped the usage data at 250G, which is enough not to alienate many average users now. When enough of these users do not voice against the practice, the practice generally accepted by the “markets” and these companies are making a “breakthrough.” The boundary is crossed and accepted as a norm. When that happens many “passive” users will suddenly found that all that promise about 250G is an illusion. Companies now have the right to cap the data at any number they like, let say after 10G, and you’ll pay a certain fee per gig after. Aside from the knowledge that in the future thing we depend more on the internet network, which would easily surpassed that 250G, WE FAIL TO SEE THE CONSEQUENCE FAR ENOUGH OF FROM THE CURRENT IMPLICATION. These companies have already seen their profitable future. They are successfully steer us away from the REAL DEAL by creating their own arena with the argument of “250G”. WE’RE FALL FOR IT LIKE LEMMING!! Please, everyone the future is in you hand today!
Although Comcast is probably confident it’s moving in the right direction by capping the total amount of data that can be transfer via their broadband links, on a monthly basis, it sounds like a plan that’s destined to fail… miserably.
Comcast has all the pipeline-power it needs to carry data back and forth from one end of its network to the other.
They’re attempting to cap the amount of data that can be transfered by broadband users in an attempt to further “regulate”, “tax” and “control” what goes on within what they think of as “their realm”, online.
Let’s hope their users have an alternative to escape their arrogant little plan to enslave them within their “limited” version of the world.
If users can’t download as much as they want, they aren’t “losing their freedom.” They never had any right to an infinite amount of bandwidth in the first place! They’re simply acting like spoiled brats.
To Brett Glass, that a pretty brass statement you made. It doesn’t sound. The argument is the data usages the minute you access to internet (www).
Many sites out there are already charging you fees to “download” a file or content. It’s like if you want to read a book you need to buy it, though you have an unlimited access to all the books; with some understandable restriction to another country’s. So that is a no brainer.
Second, your logic sounds like: If people can’t breath as much as they want, they aren’t “losing their freedom.” They never had any right to an infinite amount of air in the first place! They’re simply acting like spoiled brats.
Competition, of course will help fight these caps. Even if the competition is run through local governments.
Who Rules Minnesota Broadband?
I fail to see how someone can write a paper like this with ZERO (0) analysis of the actual costs of running networks. Outside of the operators themselves (and therein it’s usually the CFO’s that trully see both the opex and capex) have a clue.
Everything about running networks costs money. Even our government HELPS TO MAKE NETWORKS MORE EXPENSIVE TO RUN.
If subscribers truly want the best service they can get, they are going to have to learn that bandwidth, like water, is — at any one time and place — a limited resource. I’m not arguing that we cannot add more bandwidth — but clearly that costs money.
Many small towns think water is an unlimited resource — until their wells or reservoirs go dry and they have to dig new wells and pay higher taxes.
I find it disturbing to see policy advocates taking a position without an understand of the facts. There a name for it. It’s called posturing. Very disappointing to see that GigaOM has not done significant diligence to understand network operating costs and broadband usage behavior — and has not accounted for both of these in its reporting.
It’s also quite unreasonable to go after broadband service providers for residential when every other telecom service on the globe is usage based. Whether that’s 800 service, long distance, cellular, SMS, etc. I’m not advocating a pay by the bit model because that’s the other extreme of ineffeciency. But certainly tiered service seems extremely reasonable.
Then again when you are a content creater like GigaOM I’m sure you are a bit biased towards wanting a free distribution network. Too bad we don’t have the newspapers complaining that gasoline should be free because they shouldn’t have to pay costs to distribute their content.
I would love to have a 250gb bandwidth cap! my only option where I live is prolog cable internet, with a cap of 50gb upload and 50gb download per month. go over it and they threaten to terminate the account.
are they charging so much and limiting our usage because they can, and for no other reason? here is something I have trouble understanding.. normally, commercial internet bandwidth costs a ton more, right? I lease a server in a top rated datacenter with high speed fiber optic going to the building. they limit my account at 1200gb a month, and charge me only $80 a month for the use of the bandwidth plus renting the computer, plus electricity, security, maintenance, and tech support (which most people use a servers tech support very much more often they they use an internet provider’s tech support). lets take a quick guess and figure the numbers.. lets say half of what I pay goes into bandwidth. that leaves me at 3 cents per 1gb of bandwidth. prolog is charging me 55 cents per 1gb bandwidth.. more then 18 times more money.
Sticking to public financial data; figuring out the cost of running network(s) can be done by careful analysis of public data available from quarterly financial reports, executive presentations at different forums and specialized journals like IEEE.
Few points I would like to add:
1) Network costs are comprised of two components; fixed and variable. Capital costs and operating costs are fixed, and interconnect charges, buying of content are variable. Incremental internet usage on a monthly basis does not add substantial costs, unless the carrier or provider is buying wholesale transit services at a per byte basis. Several factors (beyond the scope of this comments) enter into capacity assignments – demand during peak periods, modularity of link capacity, costs of link types, configuration (point-to-point, share, or hub/spoke), density, locations, etc.
2) Capacity or bandwidth is added when there is competition for demand at the same time period. In network design terminology, the period is called peak period and usually it is refereed to as the peak hour. In enterprise networks, there are two peak hours; one the AM (around 10 AM) and another in the afternoon (around 2 pm). Residential peak period occurs in the evening. So, if a carrier or provider is serving multiple segments, then the network capacity is NOT the sum of demands (your water analogy), but the maximum of the demand. Bandwidth usage is thus not like water consumption, wrong analogy.
3) No single carrier around the world builds their network based on monthly byte counts, so the premise of increased usage on a monthly basis results in increased costs is false.
4) Traffic trend analysis does not show the imminent shortage of capacity. For that, I would like to refer you to Andrew Odlyzko’s superb analysis of traffic ad capacity.
5) It is not true that only CFO really know what the costs are. Let’s examine the financial reports of the major US based telcos and cable companies and cost components that are impacted by monthly usage:
– EBITDA margins for cable companies and telcos range in the 35-42% – a growth trend due to increased revenue (some segments are doing better than others) and efficiency gains.
– Capex as a percentage of revenue is on a decline and is around 5-15% of revenue- some exceptions are there where capex/revenue is below 5% and very few with around 20%.
– Opex – cost of operating the network is also on a decline and ranges in 25-30% of revenue. Opex here includes interconnect costs.
– SG&A is in the range of 20-25% which has nothing to do with monthly usage.
Bandwidth and Water
Victor Blake said: If subscribers truly want the best service they can get, they are going to have to learn that bandwidth, like water, is — at any one time and place — a limited resource.
Your water analogy is really funny; are you implying that if I consume bandwidth at midnight, then the bandwidth will be gone forever and no one else can use?
Are you saying, that if I use bandwidth at mid-night, the carrier has to restock it by early morning or no one can get to the internet?
victor said: I find it disturbing to see policy advocates taking a position without an understand of the facts. There a name for it. It’s called posturing. Very disappointing to see that GigaOM has not done significant diligence to understand network operating costs and broadband usage behavior — and has not accounted for both of these in its reporting.
Well, I know Messrs Al-Chalabi and Malik very well. I would be careful before accusing them of “has not done significant due diligence”. I bet that they have a combined experience of 50 years in understanding the details of newtorks, costs and business models.
I have serious problems with the concept of capping internet usage at the limits they are implementing. As a very heavy internet user, with a heavy internet use household, I can reasonably acquiesce to paying a bit more to maintain my unlimited plan (currently with Time Warner).
However, with Time Warner testing out a 5gig plan before additional has to be paid, I feel my hackles start to rise. There is simply no way that our heavy use household could financially handle paying per-gig over that ridiculously small amount.
Not only do I have a household of online gamers (WoW), but I run several business websites and most often operate from home. Additionally, as a website developer, I up/download massive amounts of data. Even something as simple as doing a round of WordPress updates on client sites would quickly add up when I’m downloading backups, uploading new versions to each site, etc.. That doesn’t even touch on things as simple as watching YouTube, viewing NetFlix movies online, or anything else that requires a large amount of streaming or one-shot downloaded data. As an estimate, it’s very safe to guess the computers on my home network are in use very close to 24/7.
With the trend toward people using internet browser applications (Gmail, etc…) over installed applications, bandwidth is only going to become more of an issue. I also have concerns for the multitude of very good non-profit / free websites out there offering services / information to the public. Many of those sites are self-run by people using a home server, and I hope that their selfless broadcasting of free content to the world will not be hindered by these changes.
As I mentioned previously, I would not be objectionable to a small price hike to maintain my unlimited bandwidth. I simply see these companies considering much more than a ‘small’ price increase. I do feel that with the trend to doing more and more things online, that they are setting themselves up for additional future riches. In order to maintain our modern internet use, and continue to expand the internet as the main international information gateway – we’ll be forced to pay out the nose. Either they’ll get rich, or the internet will slow down it’s productivity because people simply cannot afford to share their knowledge.
nicole said “Not only do I have a household of online gamers (WoW), but I run several business websites and most often operate from home. ”
If that is the case, then nicole you should think of subscribing to “business class” service for the sake of your customers. You are not a typical residential customer.
[If that is the case, then nicole you should think of subscribing to “business class” service for the sake of your customers.]
If I choose to do updates to sites, send email, handle documents, etc.. from home instead of the office, I don’t see how that makes me so different from the multitude of people out there who telecommute (fully or partly). Perhaps you inferred that I actually host websites out of my home? That I certainly do not do. I simply like to be able to work wherever is most convenient at the time.
When much of the world is continuing their work at home, or telecommuting as much as possible (for a variety of reasons, the least of which is not the price of gas), don’t you think that has created a much larger population of people who fit into the ‘atypical residential customer’ mold of yesterday?
For the naysayers out there – a 250GB cap is NOT a large amount. I filled up my MBP 200GB hard drive in 4-5 months of casual usage – most of that content came from the web, and that doesn’t account for all of the bandwidth that was just temporary files. Remember the time your 3 GB hard drive seemed like a lot? That was just a decade ago. Today’s standard hard drives are over 120GB. That’s a 40X increase (minimum). Back then, a decent internet connection was 56k – today the AVERAGE in America is 2.3 mbps. High end connections are over 10 mbps. It’s not hard to see that within another 10 years we would otherwise see speeds of 100s of mbps if not 1GB per second or more.
@muayyad (re your October 1st, 2008 7:16 AM PT comment):
I wasn’t talking about capacity or peak load, or anything to do with time. I was talking about the metric by which ISPs measure their customers’ bandwidth usage: total bytes up/downloaded. (I said “GB” not “Gbps” in my original comment).
In your example, let’s say a new World of Warcraft patch is released totalling 5GB, and each user finds themselves with a 1GB fragment of that patch. How that happens is immaterial – it’s just the initial state of the system. The problem at hand is, “How do we get each user the remaining patch fragments while minimizing the average data transferred (bandwidth) per user?”
The two diagrams you present are equivalent in this regard. In figure 1, each user downloads 1GB from every other user, for a total of 5 x 4GB = 20GB. In figure 2, the four non-power users first upload to the power user (4 x 1GB = 4GB) and then download the 4 1GB fragments they are missing (4 x 4GB = 16GB), for a total of 20GB also.
ISPs measure bandwidth at the endpoints of the network – the end users – thus the topology of the network between a user and the data they need is completely irrelevent. At the end of the day, if a user needs 4GB of data, they have to download 4GB of data and there’s no way around that.
In short, I think you need a different argument as to why power users are important. Improving bandwidth efficiency isn’t valid reason.
(Tim Lee makes some other interesting points as to why your argument along this line doesn’t seem to be relevant, by the way.)
Nicole, I understand what you mean. I run a fairly large free website, schematicsforfree.com . I don’t run it on a home server. even with the 15mb/s cable I have, it just wouldn’t handle the traffic. I pay a monthly fee to lease a dedicated server just to host that site. the money comes out of my own pocket, and I gladly do so to share the information with the world. I do however, need to do periodic backup downloads. the site is over 15gb compressed, so I would quickly use up the 5gb first tier within the first few hours of my download. while I do put up with the price of gas and pay for my server, having to pay an extraordinary internet bill would be the straw to break the back of my hobbies.
the internet in general has been struggling to become a part of the economy, and be profitable for many businesses. there are other threats to the “free” internet other then capping or tier rates, but any one of them has the potential to severely damage the profit of the internet.
look up network neutrality. it’s another not so smart idea that has the potential to hurt the “free” internet.
Not to be mean, but get Comcast Business. There is no cap, and its a little more.
Also, the caps are impeding what innovation? People say “movies, music, HD, streaming.”
This has nothing to do with anything. Most people waste their time watching all these movies and playing games and none of it correlates to innovation.
Stop whining, and use the internet for work. If you are, get Comcast Business class. If you are downloading HD movies, stop it, because the economy is crashing, and you have no time to be watching Forgetting Sarah Marshall.
I am all about innovation, but innovation is created by smart people and/or people with great ideas. These people are usually in college, a business, or have at least a few bucks more to afford a no-cap service.
If Business Class is capped, then I would totally agree with you. But if you are looking for innovation, go to a school. They all still get their bandwidth for free.
Also, the caps are impeding what innovation? People say “movies, music, HD, streaming.”
downloading movies and music does NOT necessarily mean illegally
“This has nothing to do with anything. Most people waste their time watching all these movies and playing games and none of it correlates to innovation.”
I’m sorry, but not everyone is a workaholic like you are. with the economy the way it is, it gives even more reason to go home at night, sit down, and relax to a good movie. what is wrong with that?? just because someone uses a lot of bandwidth does not mean they need to be running a business, or are wasting time. a lot of people don’t have the extra money to have business class internet access, especially right now.
“I am all about innovation, but innovation is created by smart people and/or people with great ideas. These people are usually in college, a business, or have at least a few bucks more to afford a no-cap service.”
I am a smart guy, with good ideas. I am working on putting my ideas to good use and making a profit from it, but I am not doing it for another business, and I didn’t go to college for years, I am completely self-taught. my hobby uses a lot of bandwidth. I offer free documents to tv repair shops. I already pay for a dedicated server in a datacenter each month, and can barely afford that and residential internet access. I can’t afford an upgrade. not that it’s all about me, I’m using myself as an example.
“If Business Class is capped, then I would totally agree with you. But if you are looking for innovation, go to a school. They all still get their bandwidth for free.”
college IS over-rated. I know a bunch of people who have gone to school for computers for 2 or more years. the whole time they were in school, they still came to me for help, and they still never come close to my skills. I can rip a junk stereo apart, rebuild it from the ground up, and have one better then new. I can wire a house completely from the pole to the last outlet, and pass inspection, and I can change an engine or transmission in a car by myself. my point is, I didn’t have to go to school to learn what I know, I learned it all on my own, and all before I was 20. not some know it all kid either, because I have done it all successfully in the past.
you were pretty rude and ignorant in your post. keep your selfish opinions to yourself. if you have a good point to make, and have done your research, then post.
One of the many groups I know of who are not “pirates” (a favorite accusation tossed at anyone who isn’t interested in corporate tryanny, it seems) are gamers. Bandwidth use can easily exceed Comcast’s cap between the game, the addons, and the voip server players use to coordinate with one another.
Most gamers already pay for fast downstream to support patches and response times when playing. To have to pay for load cap as well would destroy the online gaming market.
Personally? All this paper did was insure I stop using Comcast and tell every gamer I know to do the same. Considering World of Warcraft has damn near ten million (perhaps more by now) subscribers and 80% of them are in the USA, I’m thinking Comcast may be in for a bit of a surprise.
This is not at all a good thing. I have decided to monitor the actual usage of the different aspect that Comcast covered. i.e. video uses more than they say, emails use more than they say. I have now disconnected everything in my house minus my pc until I can find a new provider. I am doing web design and marketing from my cell based carrier. And no, none of this is based on P2P it is extremely easy to cross the 250GB limit in this wonderfully wired and innovative society of free learning.
Just as we are beginning to see the web develop into a place where rich media is not only possible but maturing at a rapid pace, the level of service we are paying for just isn’t meeting the demand. They are actually limiting our growth with in the technmology.
There is no need for the complex analysis offered in this ‘white paper’. As several posters have commented: Comcast has a conflict of interest as both an ISP and a content provider. As online video content delivery technology develops, Comcast has quite correctly seen the advancing threat posed to it’s overpriced, over-commercialised cable delivered video entertainment monopoly and has taken preemptive measures to deny Internet users access to the increasing alternatives to it’s own services offered by the proliferation of it’s online competitors. Today it’s a 250 Meg limit. A year from now 200. Two years from now 150 etc…
For one company to be both ISP and content provider presents a dangerous conflict of interest which will arrest the development of Internet technology and stifle free market competition amongst content providers. There are only two solutions: 1) More competition in the high-speed access market. 2) Tough regulation. Something which seems non-existent today, as the corporate world seems to do pretty much what it wants. Ask yourself: if 99.9% of Comcast’s High-speed customers are not using anywhere near the 250 meg cap, then why a need for a cap in the first place? If the less than 1% using 250 megs per month can download as much as they want up to the 250 limit without affecting the experience of other High-speed subscribers, then why would the use of 260 or 300?
Comcast would have two options if I were running the FCC: 1) Update it’s infrastructure to accommodate the Country’s bandwidth needs. 2) Divest itself of either it’s ISP or cable divisions. That the bandwidth cap was inspired by a desire to ensure a uniform customer experience as claimed in ‘CON’casts terms of service, is, quite frankly, rubbish and a big fat lie. Comcast is protecting it’s own financial interests at the expense of it’s customer base. I will not subscribe to any of Comcast’s services as long as this bandwidth cap is in place as a matter of principle. For your general edification, I am a web designer rather than a p2p junkie. I am also a proactive consumer amongst a Nation of sheep.