Is fiber to the home (FTTH) the network of the future, or are there competing technologies?

by on August 6, 2013 · 14 comments

This is the second of a series of three blog posts about broadband in America in response to Susan Crawford’s book Captive Audience and her recent blog post responding to positive assessments of America’s broadband marketplace in the New York Times. Read the first post here. This post addresses Crawford’s claim that every American needs fiber, regardless of the cost and that government should manage the rollout.

It is important to point out that fiber is extant in almost all broadband technologies and has been for years.  Not only are backbones built with fiber, but there is fiber to the mobile base station and fiber in cable and DSL networks.  In fact American carriers are already some of world’s biggest buyers of fiber.  They made the largest heretofore purchase in 2011, some 18 million miles of fiber optic cable.  In the last few years American firms bought more fiber optic cable than all of Europe combined.[1]

The debate is about a broadband technology called fiber to the home (FTTH).  The question is whether and how to pay for fiber from the existing infrastructure—from  the curb into the house itself as it were.  Typically the it’s the last part of the journey that can be expensive given the need to secure rights of way, eminent domain, labor cost, trenching, indoor wiring and repair costs.  Subscribers should have a say in whether the cost and disruption are warranted by the price and performance.  There is also a question of whether the technology is so essential and proven that the government should pay for it outright, or mandate that carriers provide it.

Fiber in the corporate setting is a different discussion. Many companies use private, fiber networks.  The fact of that a company or large office building offers a concentration of many subscribers paying higher fees has helped fiber grow in as the enterprise broadband choice for many companies.  Households don’t have the same economics.

There is no doubt that FTTH is a cool technology, but the love of a particular technology should not blind one to look at the economics.  After some brief background, this blog post will investigate fiber from three perspectives (1) the bandwidth requirements of web applications (2) cost of deployment and (3) substitutes and alternatives. Finally it discusses the notion of fiber as future proof.

Broadband Subscriptions in the OCED

By way of background, the OECD Broadband Portal[2] report from December 2012 notes that the US has 90 million fixed  (wired) connections, more than a quarter of the total (327 million) for 34 nations in the study.  On the mobile side, Americans have three times as many mobile broadband subscriptions as fixed.  The 280 million mobile broadband subscriptions held by Americans account for 35% of the total 780 million mobile subscriptions in the OECD. These are smartphones and devices which Americans use to the connect to the internet.

Ordinary feature phones are additional and not include in this number. This report notes that FTTH accounts for 7.36% of America’s total fixed broadband subscriptions, about 6.6 million subscriptions.  The US falls in the middle of the distribution of fiber penetration in the OCED.  The average penetration is 14.88%, but when one removes Japan and South Korea which have over 60% fiber, the average falls to 8.63%. Germany, an advanced industrial nation, has less than 1% fiber penetration.  Israel has zero.

It is also important to note that the Netherlands and Belgium have less fiber penetration than the US.  These two nations are considered #1 and #2 by the OECD for intermodal competition, as more than 90% of homes have a high speed DSL and cable connection. The US measures #3 because of the diversity of network types: cable, DSL, mobile, fiber, and satellite.

Advocates of any particular broadband technology often like to make arguments that broadband will increase economic growth and that nations can compete on broadband alone.  The reality is more complex, and broadband is only a single input to a complex economy, like the level of literacy. Each country has a particular set of industries and policies, and their effectiveness in applying broadband can vary for many reasons.  Therefore the OECD reports only .64 correlation between broadband growth and GDP per capita, a mild correlation.

For this reason we should pause before investing more in FTTH, the most expensive broadband technology. See the following OECD chart. Only Switzerland, Norway and Luxembourg have higher GDP per capita than the US, and of these countries, only Norway has a higher fiber penetration than the US. More telling is that Japan and South Korea with their high fiber penetrations have a GDP that is a third less than the US. See the graph from the OECD report 1k. Broadband penetration and GDP (Dec. 2012).

Broadband and GDP

Bandwidth Requirements of Web Applications

My institute the Center for Communication Media and Information Studies at Aalborg University in Copenhagen made a report about broadband needs in 2020.  It includes some scenarios about a family of four, providing extreme and a normal usage. In the extreme example, each family member is  in the midst of a bandwidth-heavy activity. Mom is a on a video conference, daughter is watching HDTV, and son is playing a video game.  The bandwidth needs for this scenario are 40-130 Mbps download and 10 Mbps upload. For the “normal” scenario the recommendation is 30-70 Mbps download and 10 Mbps upload.

While these scenarios are interesting, they fall well under the 1000 Mbps (1 gigabit) threshold that FTTH offers.  They require that the family upgrade to some serious hardware including devices that can properly render HDTV and 3DTV. Moore’s Law has helped the price of hardware decline tremendously but such a television costs a few hundred, if not a few thousand, dollars.  Furthermore the scenarios are less applicable to the fastest growing household segment in the US, the single person living alone, to whom a Wi-Fi network in a public place may be an additional appreciated location for broadband activities, rather than only at home.

The single largest source of traffic on American wire line networks today is Netflix[3]. The company has some 29 million subscribers[4] in the US and appears in roughly every third American home. Crawford’s book provides an example of the performance reports that Netflix publishes of how well its service runs on different networks, noting that 2.5 Mbps is sufficient for a high quality experience. Additionally Netflix is constantly making its service more efficient, and it has developed its own content delivery network to cache and speed content to its users. As for the leading websites, Google, Facebook, YouTube and Amazon; they want to have as wide exposure as possible, so they are not necessarily trying to make their applications more bandwidth intensive. Even YouTube, which takes up a disproportionate share of network traffic, continues to make its platform leaner.

Bandwidth needs for education

Crawford asserts that without FTTH we will not be able to take advantage of important applications in education and health. Let us review some of the leading modalities for online education and their bandwidth requirements.   The most bandwidth intensive modality is massive open online courses (MOOCs). These have been available on existing networks for years from many of America’s leading universities as well as some startup ventures.  Many enjoy MOOCs for its ability learn on a wide range of subjects.  Some education experts, however, find MOOCs less than ideal. They see MOOCs as an extrapolation of a large classroom without individualized attention and note that it works well for some kinds of learners not others.

The adaptive electronic textbook may be a format better suited to student’s needs. It is an ebook with interactive features as well as content that adjusts based upon the student’s level.  As textbooks can be downloaded or offered in chapters, they need not be high-bandwidth applications.  As for other modalities such as games, online social learning, tablets and independent certification, there is nothing inherent that requires they have FTTH. It depends on design parameters, and all of these modalities are alive on today’s networks.

The extent to which students use video and in what framework is an important question.  The flipped classroom model is one in which students watch lectures on their own (MOOC) and do homework during class. The student and teacher may meet in a video conference, but they may opt for mobile or VoIP as well.  Skype suggests 1.5 Mbps down/up for high definition video one to tone calls and 4/512 Mbps down/up for high definition video calls for 5 people.  Again, this requirement is well within the capacity of today’s networks.

The promise about online education is about more than a pipe. The point is not just to send canned high definition videos across the wires, but rather to provide intelligent customization to each student. The greater part of the value and engineering need is  upstream in the algorithms, less in the network delivery itself.  There is nothing inherent in online education that requires FTTH.  Indeed if the job is to educate millions, having light, low-bandwidth applications improves the efficacy of the business model.

Bandwidth needs for health

The Norwegian Centre for Integrated Care and Telemedicine, the world’s oldest and leading institute for telemedicine, notes that most applications run fine on average broadband levels (for example, video consultation), and even the most advanced app would require no more than 10 Mbps[5]. Indeed the limiting factor for telemedicine is not broadband deployment but rather health care providers who are resistant to change.  The other requirements for telemedicine are mobile networks and devices, so investing exclusively in wire line networks is not necessarily an enabler for telemedicine.

Bandwidth needs for entertainment

While education or telemedicine may not require large amounts of bandwidth, ever increasing high definition entertainment could consume much bandwidth.  Games and movies on HDTV and 3DTV are the killer apps for FTTH.  Consider that 60% of traffic on American networks is entertainment.[6]  To be sure FTTH can facilitate rich entertainment experiences.  However I can’t find good arguments for why taxes should subsidize FTTH if the key use is entertainment.  Furthermore it is not clear how to avoid the unintended consequence of subsidizing piracy by subsidizing FTTH.  While online video platforms such as Netflix have a powerful effect to lessen piracy—people don’t trouble to pirate movies if they get get them at a good price—for the most hard core pirates, bandwidth is a boon to their activity.

Innovation and Mobile Broadband

In spite of the assertions that FTTH is essential for future applications of education and telemedicine, the greater part of experimentation and implementation is on mobile networks.  This is not just in the USA, but around the world.  For example, I study with a dozen PhD students from Ghana.  They are engaged in knowledge transfer from Denmark to Ghana in some of the most exciting applications of mobile technologies from intelligent transportation systems, education, social networking security, banking and so on.  Additionally some Indian colleagues are working on low-bandwidth video conferencing .

The world is being remade for mobile faster than we can adjust to it. All of the major websites and applications we use today have mobile versions, and those continue to improve with better usability and more modest bandwidth requirements.  That process, along with declining prices for mobile devices, is narrowing the digital divide. Even internet companies such as Yahoo! are remaking themselves to be mobile first.  Application developers have mobile on his mind when designing for the web. We can see that Google excelled on advertising for using its search engine with a pc. They reformulated that model to mobile.  

Cost of Deployment

Crawford notes that America doesn’t have a plan for fiber and that European and Asian nations are marching ahead.  The fact of the matter is that the EU government does not have a plan for fiber either.  The sources  that Crawford provides are from Europe’s Fiber to the Home Council, a trade association that lobbies the EU government subsidies for fiber.  The EU government has the wisdom to have a technology-neutral policy about broadband. Thankfully this is also the case for the US.

Crawford attempts to shame the US by mentioning the fiber build out in Bulgaria, Moldova, and the Baltics. It is important to understand the history from these former Eastern Bloc countries.  When communism fell, they were two generations behind in telecommunications. Carriers invested heavily in both fiber and mobile networks to help these countries leapfrog to modern era.  The leading broadband based company of this region is Skype in Estonia. While this notable, there is still a brain drain from this region to other parts of Europe and the world where there are better education and job opportunities.  I visited this region in 2012, and it is clear to me that it will take more than FTTH to lift these countries out of the past.

In Denmark in 2005, 14 local utility cooperatives attempted to create their own fiber networks, arguing that there is little difference between bringing fiber or electricity to homes. Their business case never worked because the price of broadband on other networks plummeted. Today, fewer than 240,000 Danish homes subscribe to these fiber networks, a number that’s small even for Denmark.  This case demonstrates the danger of considering broadband as a utility akin to electric service when broadband services – and needs – are so diverse.  Norway has a similar story.

Naturally I am keen to see how things fare next door in Sweden where the government has made huge investments in FTTH.  A series of reports from Acrea, a Swedish government owned consulting firm conclude ”It is difficult to estimate the value of FTTH for end users in dollars and some of the effects may show up later”.  They note positive but weak outcomes. However, those results may be even less strong when adjusted for the government’s devaluation of the Swedish currency. As such, in Danes are lucky that no new taxes were levied to pay for broadband, nor were the citizens made to bear the brunt of private investments that didn’t work out.  Nearly 100% of broadband investment is private in Denmark.  Carriers, not the citizens, bear the risks.

The OECD reports that more than 60% of Japan and South Korea’s broadband subscriptions are fiber.  What many overlook about the countries however are the important political, cultural and historical factors that allow them to deploy fiber.  Compared to the USA, these countries have more collective societies and cozy relationships between business and government.  While the zaibatsu and keiretsu systems no longer exist in name, both of these governments want to ensure that their incumbent telecom companies survive, and business plans of 200 years or more are not uncommon.  Thus any national fiber plan is certainly good for the incumbents.

I suspect that most Americans would not be keen about a national FTTH plan that expressly rewarded AT&T, Verizon, Comcast or TimeWarner.  Indeed Americans value the more decentralized nature of their government where communities have more flexibility to determine their broadband needs.

The greater metropolitan areas of Seoul and New York City have roughly the same population, but Seoul is eight times as dense as New York.  This is an important fact, whether the government or a private company is bearing the risk for investing in FTTH. The Japanese improved their case for fiber by using wires above the ground, similar to telephone phone lines of old.  This certainly helped to lower the deployment cost, as well as the fact that most people live in apartments.  Carriers were responsible for the cost of fiber to the building; the landlords are required by law to take it over from there.  Interestingly many Japanese youth are quitting fiber for LTE only broadband plans.[7]

We need not to go to abroad however to evaluate the business case of fiber.  There are important examples in the US. Chattanooga, TN has a municipal fiber project with some controversy. There may be different interpretations on how successful this project is, but the limiting factor in is that not every municipality can get a $100 million grant from the Department of Energy.

Plenty has been written about Google Fiber and the various concessions made by the Kansas City government to win the project. Recall as well that the $300 subscriber sign-up fee had to be nixed in order for the project to get off the ground, showing that consumers balked even for a small fee relative to the life of the subscription.  In the case of Provo, UT, for $1 Google took over a municipal fiber network, once $39 million had been sunk in the project.  The network was financed by a $5.35 monthly fee levied on all the households in the town whether they subscribed or not.  Now that Google takes over the network, only subscribers will pay, but if it doesn’t work out for Google, they can sell the network back to Provo for $1. It is interesting to note that Mountain View, CA, where Google is based, declined to make concessions for the company to build a fiber network.

Business model for broadband networks of the future

There is no doubt that FTTH can enable rich video and entertainment experiences, but for the needs of education and telemedicine, these applications don’t require the gigabit speeds that fiber provide.  Even with our knowledge of future scenarios, there are still have many important and unanswered ethnographic questions about how people will use networks.  Future proofing may make sense theoretically, but there is no reliable empirical or mathematical model for it. Many of the companies and governments that invest in FTTH as a future proof strategy found that their models didn’t work out.

The idea to throw the baby out with the bathwater—get rid of all of America’s networks and start over with FTTH is overkill.  Certainly there are ways we can make broadband network deployment more economical such as improving the process with local government.  A case in point is New York City. It can be difficult to get permissions to dig up the streets, and people working in buildings don’t enjoy the walls being ripped out.  To be sure, these disruptions can be streamlined.  Verizon would like to add more fiber, but the conduits are already full of copper wire, and by law Verizon is required to maintain this infrastructure.

Google Fiber in Kansas City proved that lifting restrictions can translate into more investment.  Standardizing the rules for infrastructure rollout so that carriers don’t have to negotiate with each and every town and landlord would go a long way, so would improving the regime for cable franchising.  Another area for reform is spectrum.  And there is no doubt that companies will continue to innovate, whether it is DSL companies transitioning to IP switches, cable companies upgrading to DOCSIS 3.0, or R&D in mobile.  Ericsson, NSN, Alcatel / Lucent, and Qualcomm are just a few of the companies working on 5G standards for mobile, technologies that can download an entire movie in a minute.

The fact of the matter is that many technologies are competing to be the network of the future.  We should encourage this competition. Consumers only benefit from this dynamic interplay. As for the US having a low-average penetration of FTTH, if the argument is nationwide FTTH rollout for economic development, it seems to me to be prudent that the US has not invested more in FTTH, given that the global data does not show that countries necessarily improve their GDP by investing in FTTH.   Maybe that will change in the future, but that is what the data shows today.

People can fall in love with a technology and become blind to its shortcomings.  Thus we need to be careful about these silver bullet solutions, such as FTTH for everyone. There are many things to consider: the speeds of applications, the needs of users, the costs of deployment, and the price of substitutes.  Broadband at any cost is not a worthwhile investment.  If Americans can get access to the bandwidth they want at a fair price, they will care very little what kind of network it is.   The next blog post investigates whether there is a cable/mobile duopoly in broadband.

[1] CRU International Ltd, CRU Monitor: Optical Fibre and Fibre Optic Cable (London, September 2012),
[2][2] Please see the reports titled 1c. Total fixed and wireless broadband subscriptions by country (Dec. 2012) and 1l. Percentage of fibre connections in total broadband (Dec. 2012)
[5] Interview with Sture Pettersen, Department Leader for Innovation and Implementation, Norwegian Center for Telemedicine.  February 21, 2013.

  • Michael Elling

    We have to anticipate a future of multiple full-duplex 4k HD sessions to a given location, in addition to significantly larger file sizes. (For instance when I am mobile I may want to check on a couple of HD security and maintenance webcams). Web 1.0 was narrowband (<1 meg speeds and file sizes). Web 2.0 was broadband (<10meg to 50meg). Most email systems only recently began to allow for higher file size attachments. Web 3.0 (mobile) with the advent of the smartphone small screen and applications ecosystems momentarily “plateaued” the bandwidth race for the majority of people and applications. We’re constrained by 4G performance, availability and congestion and wifi offload. So I believe the article’s perspective of future demand is as flawed as much as people’s expectation for data in the 1980s and wireless in 1996, before the advent of competitive digital business models in both cases.

    As with the Wintel model’s push/pull dynamic between faster/cheaper processors and greater software functionality, a similar dynamic will occur in networks between upper and lower layers. The only problem is that there are no settlement systems in the middle layers. Bill and keep actually leads to monopoly, stifles new service creation and prevents centralized procurement/subsidization. If you want to solve a lot of the “internet’s” ills today, then implement a balanced settlement system in the middle layers whose transaction fees and pricing reflect competitive marginal cost. That’s the only way to broadly clear rapidly obsoleting supply across quickly growing and changing (nearly infinite) demand.

  • Bengie25

    Typo? “limiting factor in is that not every municipality cannot get a $100 million grant from the Department of Energy”

    FTTH on average costs about 20% less to operate than FTTN, which would literally save companies like Comcast 2-3 billion per year. It would pay itself off in short time.

    FTTH will also scale into the future, like terabit and petabit speeds. Copper has an inherent practical limit of around 10Gb/s, which is just around the corner for fiber. Once the next upgrade comes in about 8 years, copper will be unable to keep up because of physical limitations.

    Studies have shown a 0.7% increase in GDP for each doubling of Internet speeds. Nation wide, this means a minimum increase of $100b/year for the entire USA, and up-to $500b/year. The cost of rolling out the fiber? Around $140b.

    It could pay itself back in less than a year, theoretically.

    And why do people keep saying stuff like “Most people don’t use make use of their current speeds”. May that’s because there are too many limits and many copper networks are borderline broken.

    I had cable before and it was a constant disappointment when I would randomly lose my Internet because of issues with signal strength. Calling tech support did nothing expect put you in a 30min queue just to have them tell you to see if it goes away on its own. Fiber does not do that crap unless the ISP really tries to be bad.

    How many people don’t use their full Internet speeds because they don’t want to waste their time waiting for buffering from bad signals cause by copper? Kind of like and ISP saying people wouldn’t make use of an unlimited minute cell phone plan, but then fail to mention they constantly drop calls. Of course people aren’t going to be wasting their time calling if they just get dropped.

  • Robert M. Enger

    Limited as it may be in terms of absolute fidelity, Bluray may be used as a yardstick of the bandwidth required to deliver good fidelity conventional HD. Bluray uses relatively efficient modern codecs (H.264/AVC and VC1). It is a variable bit-rate system, with average rates of 20Mbps and peak rates of 40Mbps or so. For the consumer to get his/her full money’s worth from their investment in large-screen display equipment, they should have access to content delivered with Bluray level of image fidelity. Even incumbent cable-TV operators are loathe to provide that, because their existing 860Mhz HFC networks have limited capacity. Satellite providers are similarly restricted in their capacity as are LTE (even LTE-advanced). Vector-DSL won’t help much, because the distances are too limited. So far, only FTTH seems to be qualified to deliver even this basic video service with good fidelity.

    Suggestions that next generation codecs (eg H.265) will enhance efficiency are offset by observations that next-generation delivery formats will demand commensurate amounts of additional bandwidth. 4k and 8k systems provide enhanced spatial definition; high frame rate (HFR) provides enhanced temporal definition, deep-color provides enhanced luminance/chrominance resolution: all require more bandwidth – easily consuming any efficiency improvements of next-generation codecs.

    It is worth noting that no compression system enhances the viewing experience for the viewer. They are all lossy compression techniques – sacrificing some component of the image (spatial resolution of gray-scale, spatial resolution of chroma, temporal resolution, quantization/decimation of signal values) to reduce bandwidth consumption. This only benefits the network provider, allowing them to squeeze more revenue from the network without making investment needed to deliver more bandwidth.

    As we move from an era of linear programmed network viewing into an era of VoD unicast viewing, the requirements for bandwidth will be greatly increased. Each active viewing session will consume a dedicated chunk of bandwidth from the program’s storage/playback location all the way to the display system. Wireless systems cannot achieve sufficient spatial re-use to deliver such massive bandwidth. Their only option is to dumb-down the pictures using transcoding and deliver crappy-looking youtube video. The situation is even worse for satellite providers. Probably their only hope is to pre-push high bandwidth content to local storage systems adjacent to each display system (and this requires a considerable leap-of-faith from content providers fearful of piracy). Unless you are willing to buy a tent and live on the street within a few hundred meters of the central office, you’re not going to get any high bandwidth from any form of DSL. Their only hope is Fiber-to-the-Curb. This forgoes the cost of bringing fiber to the side of the house. FTTC provides no benefit to the consumer; it just saves the incumbent money (which history shows they will spend on empire building, not on improving service to the public). The only present-day technology that can deliver very high levels of spatial reuse and wide dedicated bandwidth is fiber optic transmission.
    Even FTTH is not a monolithic technology, Passive-split (PON) architectures compete with home-run fiber and other variants. There is room for competition even between FTTH providers, if we ever evolve to the utopia where a market has multiple FTTH systems! Another notable differentiation will be the quality of the wide-area connectivity (backbone and peering/private-interconnect/transit connections).

    Without fiber infrastructure, consumers will be locked-out of bandwidth-intensive technological advances. Assertions that reduced-bandwith work-arounds can be developed invariably result in undesirable compromise of the end-product supplied to the consumer (and reduction in cost for the entrenched incumbent, which is always the unstated goal).

    There are competitive middle-ground positions that do not substantively favor incumbent or newcomer. Notably, government can install duct bank and/or home-run last-mile fiber. Either of these can be leveraged by a spectrum of businesses (incumbents and competitors). It lowers the cost barrier for the entry of new comers and it helps all parties (including incumbents) to afford deployment of modern outside plant. LTE, and especially LTE-advanced, rely on terrestrial fiber optic transmission to each tower location. So, even LTE operators benefit when fiber is plentiful. Abundant dark-fiber and high-data-rate WDM will allow LTE operators to provide 100% backhaul capability to each tower site, allowing them to maximize the capability of the site.

    None of the carriers want to think about this future. But for most of us, we just want the Internet to be big giant dumb pipes. Like the water system (but with more redundancy and resilience to failures). TV networks and cable-tv companies will erode, like brick-and-mortar retail. We’ll buy our content from the studios or sports-leagues for a while. Eventually, the studios may fade away too, and we’ll buy directly from directors/producers (they will hire a CDN or other infrastructure to source the content; and the big dumb pipes will carry it to us). No carrier wants to live in that world. To forestall their demise, they will attempt to strangle the distribution infrastructure. And they will attempt to manipulate political leaders and public opinion. The telecom industry has resorted to astroturfing in the past. Today they seem to rely on social-media sock puppets and hired-gun think-tank analysts (PhD doctors of spin).

  • Roslyn Layton

    Here is the comment from my family friends in Sweden regarding fiber: A fiber network is the best option, but as the alternatives have developed, it is becoming more doubtful whether fiber is economically justifiable. In Johannesdal in Stockholm, where we now live, we have a mature network of copper wire. Our villa compound has discussed a switch to fiber optics with Com Hem, but it has proved so expensive that most of the residents are not prepared to make that investment. I would think that they think it is just as easy to use the current solution or switch to Satellite or Boxer.

  • Pingback: dortgurtgiretl()

  • Pingback: paleo beef recipes()

  • Pingback: cumbria plumbing and heating()

  • Pingback: furniture for you()

  • Pingback: New York City Tax Fraud Defense()

  • Pingback: moving companies Houston()

  • Pingback: speak up opinie()

  • Pingback: inatantly quote()

  • Pingback: CD duplication Scotland()

  • Abdul gh

    Interesting analysis and arguments.

Previous post:

Next post: