And, in the U.S., Verizon is the only major carrier betting consumers will want 100 Mbit/s to their homes in the near future.
"Just a short time ago we didn't think we'd be selling 50 Mbit/s but we're selling it today," says Verizon spokesman Mark Marchand. "The market is clearly not demanding 100 Mbit/s right now, but do we think it will evolve there? We do."
Verizon may already be preparing its network for one day delivering that 100-Mbit/s bandwidth to each home. For example, Russ Sharer, VP of marketing for Occam Networks Inc. (OTC: OCNW), says he has heard that this involves designing PON splits of only 24 homes instead of the standard 32.
On a 2.4-Gbit/s GPON network, a 24 home split would deliver exactly 100 Mbit/s into each home.
Verizon confirms that there are situations where it does splits of only 24, but it has not necessarily become its standard practice: "We've always said that the network splits are up to 32 homes," says Marchand. "There are many cases where we don't do the split up to 32."
Marchand notes that Verizon could re-engineer the PON at any time to meet changing demands and that they believe this demand will eventually evolve into 100 Mbit/s. So while Verizon is not necessarily installing GPONs with 24 home splits everywhere, it has the flexibility to easily adapt to it should bandwidth demand surge that high.
Verizon currently offers 50 Mbit/s in select markets in Connecticut, Florida, Massachusetts, New Jersey, New York, and Rhode Island. It is currently testing 100-Mbit/s service in employees' homes.
FiOS currently runs mostly on BPON which delivers 622 Mbit/s. On a 32 home split, that comes out to a hair under 20 Mbit/s to each home, which is the standard offering in many markets where FiOS is available.
But while Verizon will likely one day increase this to 100 Mbit/s, AT&T doesn't appear to have moved much from its position over a year ago when it said 25 Mbit/s would be more than adequate for the foreseeable future. (See AT&T on Bandwidth.)
"We have a path to deliver more bandwidth with compression and copper pair bonding," says AT&T spokesman Wes Warnock. "But also the key benefit of an IP multicasting model is that you're delivering video on an on-demand basis."
While the on-demand nature of IPTV does make it possible to deliver more content with less bandwidth, Warnock would not go so far to say that AT&T doesn't have 100 Mbit/s on its radar. But recent comments by an AT&T executive at the TelcoTV conference suggest that the company still has no plans to move beyond the 25 Mbit/s it currently delivers to each home. (See AT&T Shows Off IPTV Tricks.)
At the show, AT&T Labs Executive Peter Hill reiterated the company's stance that advancements in compression and video encoding technology were ahead of schedule and would continue to allow the company to satisfy the bandwidth crunch.
Meanwhile, Qwest Communications International Inc. (NYSE: Q) is taking the most conservative approach of all. It is also rolling out an FTTN network capable of 20 Mbit/s speeds although it said this week that it may be doubled through the use of pair bonding. (See Qwest to Spend up to $300M on FTTN.) As of now though, Qwest has no plans to deliver video over those FTTN lines.
Unless [structural separation] is enforced, the generation of our children is doomed to use a substandard telecom system fraught with fraud and misrepresentations and censorship by corporations owning this last mile.
Some people fear the government will censor more than corporations. They'd prefer market based discrimination over government based censorship. Regulation is perceived as state controls.
You can bypass the access networks. Pay for a server on the real internet and only use the bogus "access networks" to display things to a local screen. If T and VZ roll up all what's left of long haul and colos then agreed things will get really bad.
note that ISPs buy subscriber management platforms and other packet processing platforms when using wholesale facilities, so some regulation should be principle centric as opposed to operator centric.
also just to be clear i was making a technical statement not a political statement. there are many things that could be done better to live with peak rate marketing, and the alternative to peak rate marketing is not without its own set of challenges.
This goes to show once again that the access business (last mile provider) should be kept separate from the switch and ISP provider business. Access is a natural monopoly. We need a Sherman Act of some sort to break up this anti-competitive vertical integration. It was wrong 100 years back, it is wrong now. The 1996 Telecom act did not do it, FCC screwed it up.
Note that this is still a free market --- multiple access providers may provide the last mile connection with the clear mandate to avoid conflict of interests. Now everyone will be able to compare apples to apples their service.
Unless this is enforced, the generation of our children is doomed to use a substandard telecom system fraught with fraud and misrepresentations and censorship by corporations owning this last mile.
The system now is reminiscent of Standard Oil and the robber barons of railways --- history keeps repeating itself.
data communications has a long history of being defined at the service level by concepts such as peak, sustained/average, and guaranteed. these are useful ways of thinking about data networks. these are also the kinds of concepts sometimes incorporated into edge/aggregation equipment sitting behind access networks (even within some types of access technologies).
while an access loop is not the only factor that influences the end to end throughput, the service can never exceed the throughput of the access loop (putting aside clever compression/FEC techniques etc.). therefore it is tempting to view service marketing as being based on access loop speed.
however, the most generalized way of thinking about it, is as peak rate marketing, imo. most Internet services are marketed according to peak rate - regardless of what determines that peak rate. sometimes that peak rate will be determined by the access loop, and sometimes the peak rate will be determined by aggregation ratios and/or explicit packet shaping.
in the case of a PON service, if the marketing was based on the access loop speed why would it not be marketed as 655 Mbps or ~2.5 Gbps - are these not the speeds at which all bits are actually received by the user?
in saying that a 100 Mbps service is being offered, is Verizon not implicitly implying that the peak rate will be determined not by the access loop speed but by a statisical packet shaping function somewhere in the network? does Verizon not have sophisticated subscriber management/BRAS/aggregation/policy enforcement point/whatever systems in their architecture for a reason?. to me it is just another example of peak rate marketing.
potentially they could sell average rate and guaranteed rate services as well. doing so creates implications at least for the capacity of the entire Verizon network. there are also implications for coordination across multiple operators in any given Internet path between a source and destination. for any significant HD VOD concurrency assumption (or similar traffic demand) they will of course need to deal with the issue of their own network capacity. coordination among operators is of course another kettle of fish which is why peak rate marketing for any service that goes through the Internet is the most likely approach in the short term.
One more data point. On a Gig link using a $50 gigswitch and two computers about 2.5 years old I get TCP performance at 455Mbs. So the internet is easily 10x too slow and needs to be upgraded just to catch up with 2 year old, commodity, LAN technology.
Servers aren't the problem. It's monopolists wedded to voice revenues and regulators that will make it such that next generations are screwed unless we start doing something about the problem.
The server(s) at the other end of a connection are more likely to be a limiting factor.
Hey Fred, I've got a few servers on the net in different parts of the country. For grins, I took some iperf measurements between Dallas and the Bay Area. It looks like I'm getting 40mbs full duplex (tcp). This suggests the internet (or at least the internet I can afford and get access to) is still the bottleneck as a single server can easily do more than that.
I really don't think either CPU (or storage) will be a problem for a long while.
Of course, from my home machine it's a mere .5Mbs up and 2.5Mbs down. Been that way for about 7 years and now. No upgrade in sight as T decided u-verse is the thing for access in our part of the hood.
I think people may be missing the issue by accepting the claims that video is driving the bw explosion without question. Video may be but what kind of bw are we talking about? There is a huge difference between broadcast video and unicast video. I suspect it's unicast bandwdith that is experiencing that massive growth in demand. If so, VZ screwed up in their technology selection w/FiOS and now VZ investors will have to eat the stranded costs.
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
