20 Comments

Summary:

As AT&T preps its fiber to the home launch in austin, we offer a lowdown on the technologies it plans to use to eventually deliver a gigabit.

Fiber optic cable
photo: Jamani Caillet / EPFL

AT&T is set to launch its GigaPower fiber to the home service in Austin, Texas, any day now but other than wondering if and when my neighborhood would get it, I’ve also wanted to understand how AT&T is rolling out the service. I spoke with Eric Small, VP, fiber broadband planning at AT&T on Tuesday to get the scoop.

First, Austin isn’t a pilot market for AT&T’s fiber to the home effort just because Google has picked the area for its own fiber-to-the-home project (although the government concessions for Google should also help Ma Bell). Small said Austin residents consume 15 percent to 20 percent more broadband than the average AT&T U-verse customer making it a good place to test demand for higher speed services. More Austin residents also have subscribed to AT&T’s fiber-to-the-node U-verse service than currently subscribe to the older DSL technology. The city’s residents are a modern, broadband lovin’ people.

Plug and play makes AT&T’s day

Small said that since AT&T’s fiber-to-the-node installation in 2005, in certain parts of the country, in new neighborhoods and in some apartment buildings AT&T has upgraded to fiber to the premise, so the technology involved in the GigaPower service isn’t completely foreign. (Unfortunately for people who live in those neighborhoods, AT&T only offers the standard U-Verse speeds of 24 Mbps or 18 Mbps as opposed to higher speeds associated with fiber to the premise).

att

Now, in Austin, AT&T will upgrade its fiber to the node technology with fiber to the premise and eventually offer the gigabit speeds that fiber can handle.

To do this, AT&T is using technology that could be described at “plug and play” to connect homes back to their neighborhood cabinets and terminals. It’s also manufacturing the cables for each neighborhood offsite as opposed to building each fiber strand in the field. Small says these two tweaks will shave some costs from the overbuild as well as let AT&T install the service when customers call, as opposed to requiring a batch roll-out like Google has done in Kansas City.

When a customer buys a fiber connection, they are getting an optical termination box as their home and a strand of fiber back to a cabinet in the neighborhood. One option for fiber deployments connects a customer at the cabinet by melding the glass fibers together– called fusion splicing. But AT&T is going to use a more modular port-based system where the customer fiber strand terminates in a connector that is then plugged into the cabinet.

This is a fairly common set up across the industry with Verizon, Sonic.Net and others choosing to use connectors. The downside is this can cost more in cabling because the lengths are pre-cut and the connectors cost more, but you don’t have to have a well-trained guy with a $5,000 fusion splicer connecting each home.

Austin

The other deployment strategy AT&T will use is pre-made fiber that’s manufactured to its specifications offsite and then shipped to Austin. So what AT&T does is choose a street and measure the distance between poles or terminals and note the number of houses between those two points. It does this for the length of a street and then ships that information to a fiber manufacturer (Corning has this technology for example). Then a few months later, a reel of fiber arrive bearing the name of the street and the technicians string it or bury it.

Getting from 300 Mbps to a gig

Small notes that the connector technology gives AT&T more flexibility in terms of reusing ports if a customer drops service or moves. As for the Austin deployment, AT&T has selected four neighborhoods that will have a mix of ariel and underground fiber so it can test its tech in a variety of environments. However, it won’t start service at a gigabit. While Small didn’t say that the upgrade at the last mile also required an upgrade on AT&T’s core network equipment, it’s clear that there’s an equipment upgrade that needs to happen.

That’s one reason why the service will launch in December delivering 300 Mbps instead of a gigabit. AT&T plans to upgrade  to a gigabit in the coming year at no cost customers.

The logical point for the equipment upgrade is at the cabinet or at the customer home. AT&T’s current residential gateway providers, Pace and Motorola (bought by Arris), aren’t offering gigabit capable gear just yet, although it’s coming soon. The other option explaining the pause before delivering a full gigabit, is that the gear in the cabinet isn’t upgraded to handle the symmetrical gigabit service that AT&T plans to offer.

When asked, AT&T sent the following statement:

The transition from 300 Mbps to 1 Gbps will be easy for the customer, and at no extra cost. We’re starting with 300 Mbps speeds while we complete an equipment program upgrade to enable 1 Gbps speeds in 2014.

As for the cost of all this, Small was mum. But it does look like AT&T is really getting fiber-to-the-premise in Austin, something I wasn’t entirely sure it would ever do.

  1. Now, can anybody at the FCC under the new Chairman figure out what these cost savings mean in terms of competitive prices and start coming up with policy that addresses the digital divide intelligently and efficiently?

    Share
    1. Michael, then once they have a progressive policy, maybe they can figure out what AT&T means when they tell Stacey it’s coming “any day now.” Ditto for the Google Austin announcements.

      Seems that there’s lots of PR talk and very little infrastructure installation progress. At this pace, it also seems highly unlikely that the overall impact to the Austin broadband market will be significant in 2014.

      Share
  2. Sounds like ATT currently runs 300Mbps fiber to the uverse boxes, explaining why that is the initial “Up To” speed they’re willing to offer. So whole neighborhoods currently share 300Mbps, HHHHMmmmmm. Apparently though, the fiber is up to the task of much higher speeds, they just chose to throttle it or use inferior optical transceivers.

    A bit of trivia to chew on: the only difference between “classic DSL” and uverse is that ATT chooses not to offer the full range of speeds on DSL. Uverse is just DSL over shortened wires. The shortened wires improve reception over deteriorating copper wire pairs. They don’t run “classic DSL” for miles to the central offices any more. Folks who still have it are wired right into the same nearby curbside boxes as their uverse neighbors. I confirmed this by chatting with the tech during our last ATT DSL repair. There’s no technical reason “classic DSL” can’t provide the highest speeds that uverse does.

    I’d have liked to see some journalistic investigation to verify that ATT’s fiber to the home movement hasn’t indeed just been a response to Google’s fiber in Austin. You have simply written what they told you. Is that truly good enough? So when did ATT start offering ftth in Austin? Where else do they offer it? Not business fiber mind you, just comparable residential fiber. Is it simply so successful that they never needed to advertise it? Unusual for ATT to be so modest.

    I suggest you to ask Google Fiber whether they use those futuristic “Plug ‘N Play TM” connectors as well. Saying that ATT’s difference is that they use Connectors, suggests that Google’s pole-mounted fiber switches look like a box with numerous fiber stubs sticking out, which need to be fused to the strands leading to the customer’s house. I’ve never heard of such a thing, and can’t find an example of it online. Every switch I see, uses plug-in connectors. Point out a counter example if I’m wrong.

    Another critical point is the “upgrade” to convert to symmetrical gigabit service. Traffic coming from the customer is not being beamed back to the central office on, say, a subcarrier of the same fiber that delivers their service, like cablemodem. It goes back on a second fiber. Both fibers are identical, and run at the same speed. Asymmetrical fiber is simply deliberately throttled fiber. This upgrade they speak of, represents nothing but a software configuration change. Fiber to the home itself may be a single strand, but not the backhaul.

    Lastly, ATT assures that the transition from 300Mbps to 1000Mbps will be easy and at no extra cost. I’m very interested in fact-checking whether indeed ATT will be providing gigabit fiber for the same price as the initial 300Mbps service. Or does their interesting choice of words simply mean there will be no service charge for those who request an upgrade to the gigabit tier, with it’s higher monthly fees. No fees to upgrade also suggests that the home optical termination box already supports 1000Mbps both ways, and that again, asymmetry is just a deliberately-imposed configuration that they can simply reset.

    Share
    1. elfonblog- There are more inaccuracies and presumptions in your comment than I’ve read in any AT&T bashing post ever. Ever. You need to do some more research there buddy. Source:I’m a Uverse NOC employee.

      Share
    2. I stopped reading when you said no reason for “classical dsl” to have high speeds like Uverse. Uverse uses VDSL2 and old dsl uses ADSL, here you go and read – http://en.wikipedia.org/wiki/Very-high-bit-rate_digital_subscriber_line_2. You need to learn the DSL standards and with each new standard means new customer equipment that goes with it.

      Share
    3. Holy cow! Do you have any idea what you’re talking about? I suggest you do your research on digital subscriber lines and how that technology has evolved since it’s inception and into the many different flavours available now. You are a very ignorant person.

      Source: I’m an at&t uverse engineer.

      Share
    4. Oh, yes. I DO know what I’m talking about. I worked for 3 “classic DSL” ISPs in the Austin area and have a thorough understanding of how it works, and more importantly, why it often does not work. I was educated by both my employer’s engineers and ATT technicians clearly and consistently. Also, I’m not a corporate sock puppet, so I’m tellin’ it how it is.

      The “classic DSL” provided by ATT in my neighborhood is ADSL2+ (24Mbps). Hardly the crusty old original spec. It is rated for it’s highest speed, over far longer distances than the 1500′ uverse-branded DSL circuit now runs from house to curbside box. Somehow, they feel that 3Mbps is the “classic” limit, and nationwide, fail to exceed 85% of their advertised speed in any tier (2012). In uverse neighborhoods, “classic DSL” should certainly perform as well as uverse-branded DSL, considering how much shorter those wires are now. ATT simply won’t sell you the faster speeds which “classic” is well capable of, because it would weaken their uverse sales. You can call that a “presumption”, but I call it educated reasoning.

      Improved protocols are certainly important, but in the case of uverse, VDSL2 is only required for the top tier (48Mbs). It’s advantages are instead used to extend the number of years that the modem can tolerate increasingly degrading lines. VDSL2 is actually rated for speeds 5x above what ATT offers in uverse tiers. ATT is reserving that headroom to conceal future degradation. This means that if you’ve contracted for 24Mbps, they can “overclock” your modem’s actual speed all the way up to 200Mbps to overcome interference. Also, their “just over 50% is good enough” policy probably remains in effect, so if they can only ram 13Mbps throughput to your home, they’ll consider their job done. Enjoy your long contract.

      If anyone is going to claim inaccuracies, please list them and provide your corrections. Character assassination, including blanket mockery and disrespect, is the tool of the desperate. Difficulty level: knit your retort without reliance on a drawer full of hypothetical and unmentioned caveats. Go shake your fists and practice your cross faces elsewhere. And really, it’s hilarious to suggest that only biased corporate insiders are qualified to comment.

      Stacey: still eagerly awaiting your answers to the original questions. 1) doesn’t Google also use connectors at the optical switches, 2) detail ATT’s secret prexisting FTTH program, 3) what is the monthly cost of ATT 1Gbps residential service over the initial 300Mbps service.

      Share
  3. Too many typos in this article. Seems poorly and hastily put together.

    Share
  4. We have Google Fiber here and it’s free.

    Share
    1. Free for an ADSL speed service that you paid Google $300 in advance to install?

      Share
  5. Will there be a bandwidth cap?

    Share
    1. AT&T confirmed today that yes, there will be a one terabyte data cap with overage fees of $10 for each additional 50GB, up to an additional $30 per month.

      Share
  6. I wish they would concentrate on rolling out gigabit where we have no option from anyone else. Do they seriously think that given the option between AT&T and Google that people will willingly pick AT&T? Meanwhile I have an AT&T fiber to the curb network in my neighborhood that they will only sell me a POTS line from. Brilliant!

    Share
  7. This article mentions that 24M is the highest speed of U-verse available to customers in Austin that are not part of the fiber build. That is inaccurate, the 45M Power tier has been available in Austin since July of this year.

    Share
    1. It really depends where you lived, 18 was the fastest I could get in Austin, I was too far away for 24.

      Share
  8. I’d sign up for the 300M service here if I could get it. But I can only get 18M, if that. I’ve already voted for my area on AT&T’s U-Verse GigaPower page.

    The question is whether AT&T or Google will make it to my area first. If AT&T does, I’ll probably sign the contract and get their service, then sign up for Google as well when it arrives…I don’t mind having a redundant 300+ Mbit connection for awhile since the total cost won’t be much above what I’m paying TWC for 50M right now. Of course, if Google gets to my area first then I’ll skip AT&T altogether.

    For what it’s worth, AT&T backhauls their VRADs with gigabit Ethernet. Or they did awhile back. My guess is that they’ll upgrade the Austin ones to 10G and then start offering 1G FTTH service when that’s done. They shouldn’t have any issues offering 300M to start with though, since that’s about the same oversubscription model as TWC selling 50M service on a four-channel DOCSIS 3 plan, except the speeds are much higher so percent usage on average will be lower.

    Share
  9. “GigaPower” is an interesting name. Why didn’t they go with “OmPower?”

    Share
  10. I’d be interested in knowing where they’re going to use GPON and where they’re going to use Ethernet. Google has a clever way of finessing this in KC, and they’re willing to discuss it with enterprising young bloggers.

    Share

Comments have been disabled for this post