What could you do with fat fiber?

Two years ago, Bob Frankston wrote Why Settle for Just 1%? while in the midst of his ramp-up as a Verizon FiOS customer. The question is still on the table. I'd like us to help answer it by re-phrasing the question: What could we, as Linux developers and users, do with fiber to our homes and businesses?

The answer should look like good business for the Verizons, RCNs, Comcasts and other fiber-deployers to be in. Let's help them imagine benefits to carrier incumbency other than doing the same old thing, only faster.

Here's the gist of what Bob wrote, two years ago already:

The fiber they are installing for FIOS is really a cable TV plant disguised as a network. It is a Passive Optical Network (PON) designed as a distribution system from a head end to the terminals at each home though it does have capacity to send data back. A single fiber has the capacity for gigabits of traffic.

There's so much capacity that they can simply allocate a portion of the capacity to emulating traditional Cable TV. The 15mbps they reserve for their Internet service is less than 1% of that capacity! The big lesson of the Internet and personal computer is that it makes more sense to just deploy simple IP connectivity and then use standard digital technology to convert the IP video streams to analog video when necessary. While I might forgive the Telcos for neglecting the old-line telephony business, it's harder to understand why they are deploying technologies that are obsolete before they are deployed.

My own new apartment, just one town over from Bob, has real competition in both fiber and copper to homes. There's fiber from RCN and Verizon, and copper co-ax from Comcast. (And possibly fiber as well, though I'm not clear on that.) After spending way too much time trying to plumb each company's BS for some indication of what they actually offer to customers who want real Internet service, it's clear that there's room for real differentiation here. I would think there's a huge opportunity for one of them to break loose and offer exactly what Bob, a helluva good customer (he co-invented the spreadsheet and helped home networking happen), is asking for. They could offer simple IP connectivity. I'd understand something less than all-you-can-eat (there can still be tragedies of the commons, even with fiber); but hell, let us eat something more than telephone, television and third-banana Internet service crippled to make room for the other two.

I have no beef with all of them offering television. That's low-hanging fruit, and bread and butter for as long as customers remain addicted to tubes (and now, flat panels with digital pictures compressed to the point where every solid color looks plaid). But why exclude better Internet service? That's not just the freaking future, guys. It's the freaking present, and has so much possibility it's hard to imagine it all. Why not partner with Google or Amazon for offsite backup services? Why not offer business services in the form of data warehousing, server farming and mirroring, and who knows what else?

Look at what Digisense is doing, just to route around the dumb and unhelpful asymmetric service that ISPs have been delivering for the duration. (Disclosure: I've consulted Digisense in the past.) Companies like these can be excellent business partners for ISPs and carriers — including municipalities and regional groups that are offering fiber to premises.

I'll bet the Linux community has some of the best answers to the question of what to do with the rest of that fiber capacity. I wish I could be at LinuxWorld next week to talk with people about it, because I think it's a huge and wide-open frontier for both business and technology. A lot of ya'll will be, though. Post some of your answers here.


Doc Searls is the Editor in Chief of Linux Journal


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I want to upgrade to this

charlie 's picture

I want to upgrade to this pretty bad, but of course there's no fiber optic coverage in my area. It's ridiculous how these big telecommunications companies won't go into more rural areas because they won't make enough money off of it. America's brand of capitalism has reached the point of ridiculousness - I live in Evergreen, CO, and the "high speed internet providers in my area" (look at the webpage and you'll see what I mean) is like the pirates of the carribean ride at disneyland. You wait for forever and then once you're finally on you get kicked off after about 5 minutes.

AT&T U-verse is IP-only

Brian Stretch's picture

AT&T's new U-verse service is IP-only. Their residential gateways plug into their network via Cat5e. The RG has an Ethernet switch, WiFi, POTS ports for a future VoIP service, etc. The Motorola settop boxes plug into the RG via Ethernet. In theory AT&T is on the right track.

Unfortunately they cheaped out and went with a FTTN network, so all you get is a 27Mbps DSL feed from the node to the home RG over copper. 6M/1M is the best Internet service you can get, the bulk of the remainder is devoted to the most hideously overcompressed TV service I've ever seen. The sad thing is that their Internet service is actually competitive with Comcast so I'll probably keep that and drop the TV portion when my two free months are up. If AT&T would only have spent the money for FTTH which they're going to have to do eventually anyhow they could have a fantastic service despite the flakiness of the WinCE-based settop boxes.

See UverseUsers and U-talk Forums for more on U-verse.

I'll check out communityfiber.org.

For Doc: do we agree?

eric dynamic's picture

The model proposed at communityfiber.org may sound a bit radical, but seems to me to be what people would wish to use if they could get it.
I'm wondering what objections (both philosophical and pratical) you might have to that model - and if you agree with the proposed model, can you advocate it (spread it around)?

Would have sent the question via email but didn't see an address in 0.1 second, sorry.



Michael R. Bernstein's picture

The things that are possible with a fat pipe diverge a lot dependiong on whether the pipe is symmetrical. Which scenario are you more interested in?

Informative comments at LWN.net

Anonymous's picture

This post was mentioned at LWN.net (http://lwn.net/Articles/244211/ ) and some of the comments make interesting points about the cost of routers being the real limit, not the costs of fiber.

BitTorrent--for legal stuff--is one great usage

Terrell Prude' Jr.'s picture

Yes, I know, it gets heavy usage already, and we know at times what for. :-) But I have something else in mind--something legal. :-)

The reason I got broadband with decent upload speed (2Mb/sec at the time of this writing) was actually to be able to mirror some smaller size, but popular, F/OSS projects (OpenOffice.org comes to mind). By "smaller", in this case, I mean something smaller than a CD-ROM ISO image.

But then along came the BitTorrent protocol. Once I learned how to do basic port-forwarding (it's not that hard), I sucked down the latest CentOS 5.0 DVD and OpenOffice.org Extras CD via BitTorrent. And I am continuing to seed (it's been over a couple of months for CentOS). I'm also mirroring some of the smaller projects, too, as I had originally intended, on my Web server via traditional HTTP download (not OO.o, though).

So, thanks to the way that BitTorrent works, I get to share more than I had thought I'd be able to. That matters to me because I've gotten a lot from the FOSS community.

By the way, CentOS 5 rocks. So do Slackware, Ubuntu Dapper/Feisty, and Debian Etch, by the way, all of which I also use.

'local' apps are one solution

James Thompson's picture

OK, so if the ISP(s) gritch that bandwidth is "too expensive", then take the lead on the buildout of local (metro or even neighborhood) area applications.

Local (city block) IP radio stations, a kind of "micro FM" for the masses
Local weblogs about neighborhood happenings and politics
Carpool planing sites
Local IM (Jabber) servers
Local grid computing, let your neighbors use your CPU cycles for FOSS-based image/video processing (and other CPU-heavy apps that folks might use.)
local video sharing (including 'news'), possibly via mythtv
annotated neighborhood area maps
neighborhood area debian/ubuntu repositories, available via multicast IP

I could go on.

The simple fact is, the "ISPs" won't allow this to happen, either. In order for this type of networking to truly bloom, the citizenry has to own the infrastructure. One method is via "fiber down the fenceline", though there are others, such as metro (city) owned infrastructure.

p.s. sorry about failing you on the searls.com wordpress installation.

frat boys & fat transfers

Anonymous's picture

Yup, Bittorrent and the network will be saturated with everyone downloading everything from everyone else until everyone has everything; until they realize they don't want to do it that way, where everyone has to own 20 TiB of portable storage.

That's an abuse of the network, but there ya go, an example of filling a pipe of any size - with trash.

bidi bidi bidi... the return of Buck Rogers

shmget's picture

"20 TiB of portable storage."

Whoa, precision must have been paramount for you to use this stupid IEC unit.
(for those who didn't get the title, this 'thing' is pronounced tebibytes (te-bee-), the lower units are mebibytes and even kibibytes, which remind me of Twiki : http://en.wikipedia.org/wiki/Twiki)

Resolution x Streams

Phil Wolff's picture

Our video resolution in TV is still at a basic level. Capabilities are growing to gigapixels per stream, allowing for wall-sized imagery. As camera costs continue to fall faster than Moore's Law, watch resolution, color depth, and frame rate grow. Homes will generate video streams from many devices, from security systems, nanny cams, in-fridge inventories, etc.; some on demand and others as continuous streams.

We're still learning what people will do with the ability to produce video, especially at work and school. As we climb through a generation of experience, we'll come up with many useful forms that blend real life video, our onlives, and our data auras for a variety of activities and interests. A current example: one of the most popular Skype plug-ins is for screen sharing during a call.

We haven't even talked about p2p voice and video calls, file sharing, television/movies, immersive worlds or, that ultimate in bandwidth consumption: smell-o-vision.

All that and more

Doc Searls's picture

We need to have that "generation of experience" with more than just passive reception of hundreds of "live" TV channels (most of which are not) crammed into a pipe. We're producers as well as consumers now. I think lots of us would be willing to pay a bit more for service supporting that.

Doc Searls is the Editor in Chief of Linux Journal

No waiting for the Telcos

Eric Dynamic's picture

It's time to wire ourselves up and forget the telcos and cablecos. We do NOT WANT any further commercial involvement with our networks. WE will build the network we want to use and operate it as a utility, and put an end to: commercial fiddling with rate tiers; any remaining threat to network neutrality; and the telcos and cablecos themselves.

As an ISP watching the big players successively expropriate the market from us, the answer is obvious: the transport must be publicly owned and operated. It will be easier to simply bypass the telcos by building our own network in parallel to theirs, rather than try to force them to share, or try to nationalize their infrastructure.

DO NOT consider involving a for-profit corporation in the community's fiber network: that is EXACTLY what corrupts serving the community's needs.

See http://communityfiber.org for a description of a preferred model for a community fiber network. Remember, either you build and control what you want to use, or you'll be just as screwed in the future as you are now. We do not NEED corporate involvement in OUR network and we must AVOID any such entanglements.

Can't route around it

Wes Felter's picture

I don't see how Digisense is routing around slow ISPs; it still sends data over the Internet.

nothing is free, who is footing the bill?

philanthropist's picture

private contributers, tax payers, etc

Waiting on ourselves

Doc Searls's picture

The political and economic reality is simple: the transport we have today will not be publicly owned and operated, much less controlled.

That's not to say that another network, not owned by the telcos and cablecos, can't be built. I'm all for it.

But I would like to see something useful in my lifetime. I turned 60 last Sunday.

Getting "the public" interested in a Net-native, Net-only infrastructure is not easy. Getting government interested is even harder, especially in the absence of voter concern. To most of us, the Net is gravy on top of our phone and cable services — and bills.

For "the market" to work, we need a real marketplace, and not a regulatory protectorate for telcos and cablecos. Good luck finding that.

Meanwhile, I'm trying, at least for now, to help the carriers improve their offerings, while also working to do what they won't or can't.

In Santa Barbara, we have a situation where the carriers have no interest in providing fiber-grade service. So there we're working to find ways of providing what they don't. And we're open to many possibilities.

In the Boston area, we have a situation where the carriers have already provided two, and in some cases three, fiber connections to homes. That's where the antitrust "rule of threes" (which says you need at least three competitors to have real competitive market effects) at least suggests some differentiation. One possibility is working with ISPs (like yours, presumably) and other third parties to provide Net services that the telcos and cablecos normally wouldn't on their own, but which can bring them some additional income.

Doc Searls is the Editor in Chief of Linux Journal

Correcting the incorrigible while waiting for Godot

eric dynamic's picture

The notion of "assisting" the telcos to improve service makes sense to some extent. While they're around and if they can get "there" faster, then we can use them - if they cooperate.

My hardened opinion is that the large telecom companies cannot be trusted, period. They show too much easy interest in distorting the market to make more money from it - they didn't think twice before deciding they could expropriate the public-commons aspect of the Internet from us in order to ARTIFICALLY create tiers of service (and to be in a position to strangle any target they desire.)

What you do with greedy boys is take their toys away from them. Yes, in no short term will they reform or cooperate. So it will be a separate network - we will build AROUND them.

While it won't be as cheap or easy as I imagine, it won't be so expensive or difficult it can't be done, and it only has to be done reasonably well in a small locale to demonstrate the potential benefits. There is a serious drawback in that "many people" need to be convinced, and that won't be easy. Excuse me for chiding some posters here, but the first thing to overcome is the notion that their dribble of speed is "good enough for me so why does anyone else need anything better", fueled by an unexamined fear of paying money through the nose (when one primary goal for the fiber project is to SAVE THE USERS MONEY MONTH AFTER MONTH.)

The communityfiber.org model for the network is massive multicast: everyone becomes an accessible channel for as many audience members as wish to attend. ABC News has a channel and you have a channel and you have the potential to attract more viewers than ABC News.

There are two sides to the communityfiber notion. The downside is that the work has to be done before the network exists, and people will argue how much to spend on the network and so forth. However, the upside is to let your imagination run with what the network will do for you when you've built it as described: (1) The fees are to be utility fees. There will be as much ''profit'' as needed to recover maintenance and development costs but otherwise the network is NOT run to produce a PROFIT for anyone. (2) Symmetry is presumed - at least adequate symmetry is presumed. If I can pull in 1gbps, I should be able to send back at a similar rate. The built-in headroom should provide for each end-user to be able to submit a video stream TO the network.

If it's our network, we have no boardroom jerks in small numbers deciding what we can and can't do with our OWN network. Let them dick around with the consumers they still have in their pockets until those consumers look longingly at what WE have and say - ya know - that community fiber looks pretty good to us too.

Megacorporations are a nuisance and are obsolete. The community fiber project is about getting a fiber network but it is also about re-empowering the little people (us) to finally get what we want and need at a reasonable cost in a reasonable time. If we can get past the LOCAL naysayers and produce a large enough prototype network, then the writing will be on the wall for the telcos. At that point either they "get nice" and start competing on similar terms, or we just let them vanish as they lose customers, region by region. I will shed no tears for the demise of Comcast or SBC (don't grace them by calling them AT&T.) So in a malevolent fashion I agree with using corporations along the way: use them as you need to then drop them the moment you don't need them. That's the way money thinks, so they should be expecting it, apart from deserving it.

So, Doc, let's look forward to seeing the model emerge SOMEWHERE as a demo of what is possible EVERYWHERE. It might take a Herculean effort but I am targeting 2012 for the completion of (much of) our network here.

Read Joel Bakan's "The Corporation". Then you'll know why I have tuned out on monopoly capitalism. Capitalism is ONLY any good when the capitalist is not bigger and stronger than his market of clients; and Capitalism intends nothing BUT to become a monopoly, which our broken laws allow and incent. I have every intention of NOT negotiating with large, blind, careless corporations. They do NOT primarily care about the consumer - that is largely propaganda. They do only what they are forced to do. I don't want to spend anyone's time negotiating with a legally-mandated psychopathic organization for any reason. That has to remain my position until the CEOs themselves rush to Congress to get the laws about "profit first or jail" stricken. They won't. They love their money and power too much. So take their toys away.

15Mb/s isn't exactly third banana Internet service

Richard Bennett's picture

I'd be happy with a true 15 Mb/s second Internet service. The much-touted Japanese service that has a raw rate of 100 Mb/s is typically no more than 20 Mb/s in actual file transfer rates, so that would be pretty much as good as it gets.

At the end of the day, the thing that counts once you have FTTH is the infrastructure; supporting an actual constant rate of 100 Mb/s is real expensive, and if everybody isn't going to use it, why spend the money? That will just make all of our bills go up.

I can't think of any compelling consumer need for a greater than 15Mb/s service, but I'd sure like to hear about one.

Let's have a whole 'nuther fruit

Doc Searls's picture

I'm not sure what's "as good as it gets".

I am sure I'm not alone in wanting 5Mbps or better on the upstream side — especially when two or three carriers have fiber connected to my house.

I've never heard a good answer from a carrier about why those upstream speeds can't be supported. And I don't count "because the demand isn't there" as an answer. You can't measure what you don't allow.

if all suppliers waited for demand before doing the work required to create markets, we would have many fewer markets. Especially in technology. Necessity isn't the only mother of invention. Sometimes invention is the mother of necessity. Pretty darn often, in fact.

Many of the people we call "consumers" are also producers now. And there are plenty of producers who could produce a lot more than they do now, if the net supported it on the upstream side.

Why should "triple play" (with the Net as the third banana) provide the only three ways for last-mile connection owners to make money? How about "hundred play" or "thousand play"? Or (any number) play? How about turning that infrastructure into a real market where all kinds of fruit can grow?

There's no need for these carriers to give anything up, or to submit to greater government control. They just need to find benefits to incumbency other than the three obvious bananas. That's how I'm trying to help here.

Doc Searls is the Editor in Chief of Linux Journal

why impose speed caps?

eric dynamic's picture

There's no reason for the service to operate more slowly than the interface. Within a fiber-wired municipal zone, point-to-point transfers should proceed at 1gbps. Data cached at the edge of the zone will also fly into the users at 1gbps. And once fiber zones connect, the 1gbps point-to-point applies to any pair of users in the merged zone.

Things get cheap when we want a lot of them produced. The expense lies primarily in the switches and routers, but it's not data speed that makes an interface expensive - it's total throughput of the device supplying the connections.

I claim that the differential cost of providing full speed versus degraded speed is not worth concern. The cost will be in the hardware either way. Expecting to wire everyone up at 1gbps speed (with a set committed rate) will make the hardware cheaper as the number of units to be consumed grows. That is why people should expect a gigabit connection to begin with: set the bar where it deserves to be. You won't turn down a 1-gbps connection if it works as well as its name implies.

Any concession to speed based on price caters to continuing control over our communication networks by the incumbents. The question, I say, is not whether some number N of megabits, 10 < N < 20, can satisfy a single given user, but what the point is to sweat a speed cap to begin with. 1gbps fiber interfaces are common now. There's no architectural reason to constrain the speed of the connection, whatever flows are guaranteed for the connection. So, install a 1gbps connection now, even if the end-user can't get better than 10mbps, because one day (soon), they will get better speeds.

The savings per end-user using fiber over copper or cable will easily be enough to offset the cost of running the service. Discussions about cost recovery will be open public discussions, and not something left to people in some boardroom in a skyscraper a thousand miles away.

Routers aren't free

Richard Bennett's picture

Right dude, the costs of networking aren't in the cables, they're mainly in the installation of the cables and the installation and maintenance of the infrastructure. Routers aren't free, and they all have finite capacity. You want more than one user to hit his Gigabit connection at the same time, you need routers than can forward packets at Gigabit rates. And it turns out that it ain't as easy as all that. When I worked at Cisco I developed code for their second Gigabit adapter, and we were unable to get it to pump traffic at full line speed. That's just one adapter in one router. When you have thousands of users hitting a common router at Gigabit speeds, stuff is going to fall on the floor.

And it doesn't matter who owns the network or how pure he is, moving packets still costs money, and the more you want to move the more you got to pay.

One of the reasons Verizon wants to subdivide their fiber between TV and Internet is that the TV connections are much cheaper because they aren't switched.

routers are cheaper with no profit motive

eric dynamic's picture

1. Whatever routers will cost, someone has to pay for them.
We can have SBC pay for them and pay through the nose for the privilege, or we can pay for them ourselves. If we do it ourselves we will save money, and that's just a fact.

2. Switches, etc. are available for high-density connections. Whatever the eqp cannot do well easily now, demand will help cure. If we could not distribute between 6 and 100 mbps to end users on a reliable basis, then neither can the telcos. Since they're doing 6-8 mbps now on copper, I can't imagine we can't do at least as well on fiber - and there is no reason to slow down the interface, because it's not the speed of the interface that is the problem.

My bottom line is that whatever the eqp can do, we can let it do its best if WE operate it ourselves. We WILL get gigabit connections at our houses - even if on day one we can't run every user at 100mbps committed, for example.

If you already know what routers can and can't do to date, then tell me what their limits are. I have to bet that if I approach Cisco to say we're thinking of wiring up our town, they will tell us they can fill the bill with equipment on hand (forgetting cost for the moment.)
I ought to prove that for myself, I suppose.

Then Cisco shouldn't sell routers with Gig interfaces!

Anonymous's picture

If Cisco's selling routers with Gig interfaces but can't (or won't) actually do Gigabit across said interface, then no wonder the telcos have Juniper all over their plants! Geez! Cisco needs to get off its butt and actually produce what they're advertising.

Respecting real costs, and real opportunities

Doc Searls's picture

The carriers need to operate their businesses, and using one business (TV) to subsidize another (e.g. Net) is legitimate and sensible. I'm glad you're pointing that out.

What I'm looking for here is openness by Verizon and the others — especially in markets where there is real competition, and not the customary duopoly — to supporting more uses of the Net than mostly-passive consumption of "content". There is a chance to build real marketplaces here.

Think that can be done?

Doc Searls is the Editor in Chief of Linux Journal

death to corporations

eric dynamic's picture

death to corporations: that's my answer, sorry for such a terse answer.

Verizon, AT&T, Comcast ... these entities ARE NOT OUR FRIENDS.
THEY are looking forward to a balkanized (non-net-neutral) internet, and their sole interest in existing is to make as much money as they can get away with making. And they will screw the public eight ways from Sunday if that's what it takes to make the money: witness EVEN HAVING TO DISCUSS network neutrality to BEGIN WITH. Consider that the best the dems could get in the FCC was a two-year hands-off agreement from AT&T.

My urgent pitch to all readers is: QUIT THINKING YOU'LL GET WHAT YOU NEED FROM FOR-PROFIT TELECOM COMPANIES. The time is NOW to start wiring OURSELVES up so we can get RID of: promos, profit grabbing, artificial tiers, artificial speed caps, and ANY FUTURE THREAT to network neutrality.

See http://communityfiber.org and see if you would NOT want what is described there. If anyone thinks there's a better alternative, then they can explain how we (the people) will be able to control whatever alternate solution is proposed.

Follow the money

Richard Bennett's picture

I figure Verizon will do what they judge will make them the most money over the long term. They've shown a greater willingness to invest in infrastructure than other telcos.

To get back to your question, about the only thing I could do with 100Mb/s that I can't do now is access a file server in NFS or Samba mode. That would be useful, but there probably aren't enough people wanting to do that to make it economical to upgrade the plant.

If we build it, they will come

eric dynamic's picture

If we build it, they will come. That was in fact the subtitle of an article by Robert X. Cringely on the PBS site, where he said, "time to build our own fiber last mile."

I am NOT worried about costs for a superspeed network that we own, and I do not want to bother arguing whether "we could use" 100mbps speeds or not. It can EASILY be taken for granted that the capacities of the network will be exercised when people get on board.

* File transfers will just zip by at full speed. Savings? Time.
* The municipality can run a VOIP gateway and defray telco costs to users.
* Build for speed and capacity and the speed and capacity are there when you need it.

There is no risk to build a community superspeed network, apart from doing it poorly. The network will save its users up to $1000/year per household on charges from telcos/cablecos, and we can ask for some of that back.

The ARGUMENTS against the project are these vague "it'll cost too much" arguments - I argue the up-front cost is worth it and others say it isn't. One thing's for sure: the naysayers will work against our ever seeing the promised benefit of a community fiber network, if they insist that "everything's fine as it is." It is not fine. I personally might like to get access to cable TV, but I REFUSE to deal with Comcast and by golly, I want an alternative. As an ISP I have watched Verizon and SBC gut our business via their friends at the FCC - and since the incumbents won't play fair, then I have decided they should no longer exist. I firmly believe we can do as well without them, and that we should wish to prove so to ourselves. I'd love to pay my city for my connection and know the billing office doesn't have a vested interest in giving me a phony bill I have to verify thoroughly each month before I pay it.

If you want to continue preaching the excessive-cost and nobody-wants-it arguments, we'll have to get into real numbers, otherwise we're both just waving hands. I want a gigabit connection; all things being equal, why not? And I claim things are largely "equal", in this context. Ultimately I intend to prove my claims or see them proved. Will there be hassles? Oh sure. Always. And we get past them. That's the life of any ISP.


who is the "us" and "we"?

Anonymous's picture

I would gladly contribute to help this but I refuse to do so at gunpoint with taxes nor am I arrogant enough to force someone else.