x
Ultra-Broadband

Verizon Leads the Great 100-Mbit/s Bandwidth Race

With Verizon Communications Inc. (NYSE: VZ) announcing a five-fold increase in the amount of HD channels on FiOS TV next year, it again brings to light that consumer bandwidth is a growing concern. (See Verizon to Quintuple HD Channels.)

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. (Nasdaq: 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.

— Raymond McConville, Reporter, Light Reading

Page 1 / 3   >   >>
paolo.franzoi 12/5/2012 | 2:59:40 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race
A really simple way to think about this is:

Access Line Bit Rate/Number Subs per Line = Bandwidth per Subscriber.

This makes several assumptions that are simply not true in the internet world.

1 - The Access Line is actually divided evenly per sub. Not true. It is actually statistically multiplexed. So, it responds to real time situations.

2 - Number of subs per line represents the number provisioned or engineered for. Again, who is online and how many are actually installed is another question.

What I am leading to is what you have defined is the bandwidth you have defined is the dedicated bandwidth per subscriber (in the downstream direction) with no over subscription.

Now, lets hop to the other side of the OLT or DSLAM. Let's say that a box serves about 500 customers (that is a nice round number and is close to what CO DSLAMs did - not far from a BPON OLT shelf does). How much bandwidth is uplinked out of that box? Is it 155? 622? 1Gbps?

So, if you say 20M * 500 Customers, you need...(wheels turning)....yes 10 Gbps to provide that 20M dedicated through the OLT or DSLAM.

People will now want to argue with me that, well there are a variable number of active subscribers and that the statistical multiplexing leads to over subscription gain. To which I answer, I just said that about the PON Interface (DSL is of course, 1 customer per line).

So, if they are going for 100M per customer are they uplinking out of the OLT with 50Gb/s per 500 customers? Let us extend that to an end office with 5000 customers. Are they exiting the office with 500Gb/s?

I think you need to change your math here.

seven
macster 12/5/2012 | 2:59:39 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race Hi Seven,

Can you clarify the following pls:

1. What do you mean by Access Line actually?
2. What do you mean by sub actually?

I really don't get your reasoning at all. Let's take a current DSLAM which can support 960 lines to end homes (64 lines per card and 16 cards per subrack). Let's ignore the AN frequency plan and all that for a minute. Also, all end users are near this CO DSLAM. If each user gets the ADSL max of 8MB, 960 users = 7680 MB. Doing it like this is silly! For starters, the DSLAM backplane (be it cell-switched or packet-switched) cannot handle 7680MB - at least not the DSLAMs I have worked with!

In general, a contention ratio is used at the uplink (a GigE interface for the usual IP DSLAM). Contention ratios of 50:1 for residential homes are common, and 20:1 for business users.

A PON (let's take GPON) will have the 'standard' 32 splits and an uplink of 2488 MB. Why the comparison to a single subrack/shelf DSLAM that can support 960 users?

Do advise.


Mac.

OldPOTS 12/5/2012 | 2:59:39 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race
I believe what was presented as 100Mbps is actually "DEDICATED bandwidth per subscriber (in the downstream direction) with no over subscription". (seven's point)

This is a shared GPON line of 2.4Gbps that all the Subs (24 or 32) on the Access Line could access and contend for.

So the question becomes how much of that 2.4Gbps is each Sub allowed to access (up to the 2.4Gbps), whether limited through BW dedication or over-subscribtion. This is what the user could expect to see during access line busy hour with no other limitations.

But inorder to be fair, if access is limited by over-subscribtion anywhere in the network, the largest (congestive) limiting BW ratio across the entire network should be identified so the expected minimul BW for busy hour is really known.
Or else this is analogous to cisco's math, best case calculations.

OP

'Fairness in reporting' = Minimul BW during busy hour. (Calculation not always straight forward)
American Indian 12/5/2012 | 2:59:38 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race

Let's forget all the tech talk for a moment and focus on reality.

What does 100 Mbps to a home or business mean?

The answer: T1, DSL and copper related applications will be out stripped in performance based upon bandwidth. Translated: Idiot CLECs that have built a business model on ILEC legacy infrastructure are dead companies walking.

It does not matter what the FCC does or does not do. The reality of physics of glass versus copper will in fact remove the weak players who do not have a long term, sustaining business model because they try to live off the ILEC.

Back to your tech debate that does not matter ... fiber is to copper as Einstein is to Forrest Gump.
macster 12/5/2012 | 2:59:38 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race OP,

Thanks for explaining...

This is the bit where I like to put forth my "Business Drivers Meet Service Requirements" slide.

To answer your question below, let's take the example of an ATM DSLAM. Here, we can set parameters such as SCR and PCR to mimic VBR or CBR services, e.g. SCR = PCR for a 1:1 uncontended CBR service.

In other cases and examples, we can make use of traffic shaping and policing functionalities (at the UNI, e.g. admission control) to each subscriber, i.e. service differentiation - pay more and get a bigger slice of the pie. I have a document to explain this. Lemme know if you want it.

Apologies for my bad English, but I did not understand your last paragraph, the "But in order to be fair, if access is limited..." part. Can you explain again in plain English for simpletons like me... thanks.

To conclude, Manchester United is gonna teach Arsenal a thing or two about football (soccer) this weekend.

Have a great weekend guys and gals!
paolo.franzoi 12/5/2012 | 2:59:37 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race
Macster,

You have hit the point. What has happened is carriers have moved to selling Access Line Rate (i.e. the line rate of the line connecting you to the carrier). As a sub (subscriber), the Access Line Rate is one of just many parameters that represents your performance.

Plenty of DSLAMs have had more than 8G backplanes. The problem is nobody was willing to pay the cost of those. On top of that, if you had a 10G backplane and used a 1G uplink then in fact you are actually sharing the 1G at a rate up to your Access Line Rate.

50:1 is extraordinarily low from what I see. I see over subscription ratios at the DSLAM/OLT of 200:1. It gets bigger, of course, as you get deeper into the network.

A PON port will have 32 ports at 2.488G. But it will be in an OLT that will have hundreds of ports and oversubscription calculations just like a DSLAM.

seven
macster 12/5/2012 | 2:59:37 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race AI,

Yes, the bandwidth presented by fibre is limitless compared to copper.

If that's the case, why are players like Hatteras (copper EFM) surviving... and surviving well???

The gist is... the transition will be evolutionary, and mid-band players do have a market.

Dragging fibre as far out as economically feasibly (for now) and playing with the copper frequency range will be around for a while.


Mac.

fgoldstein 12/5/2012 | 2:59:37 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race Indian,

You have a rather simplistic view of the value proposition brought by different carriers. Bandwidth is not everything!

By your model, cars would simply be rated by horsepower. Big V8s would always be better than anything else, save maybe V12s. Sorry... I lived through the '70s and horsepower ain't everything.

CLECs survive because they offer a better value to their subscribers. Maybe they don't have the highest peak bit rate, but they may offer better customer service. They may offer better price plans. They may be less obnoxious, or be less likely to block a customer's application, or delete a customer's email, or be less likely to allow Big Brother to illegally wiretap their phone calls or Internet traffic.

For most end users' applications, high bit rates are rarely needed anyway. The server(s) at the other end of a connection are more likely to be a limiting factor. A decent cable modem or DSL connection is adequate for today's Internet users. HDTV streaming might eat a lot of capacity, but the Internet itself isn't ready for it; that's still done better via video-on-demand services like those of the cable companies.

And recall that when the ILECs feel comfortable about their monopolies, they don't behave very nicely. If you think Comcast's stupid Sandvine was bad, wait until you see the Bells' Bytemobiles.
paolo.franzoi 12/5/2012 | 2:59:36 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race
Except that FiOS uses the RF overlay.....

seven
American Indian 12/5/2012 | 2:59:36 PM
re: Verizon Leads the Great 100-Mbit/s Bandwidth Race

Macster:

Hatteras? Not a household name. If copper was the play, Cisco, Nortel, Alcatel, Ericsson, et all would buy them. Some more reality for you: bonded copper only works if there are copper pairs available - the ILEC will NOT build out more copper pairs .. More reality ... as the ILEC brings fiber into homes and businesses, before they leave they are taking the copper out -- that's right, removing it. It is hard to bond copper that is no longer in place. Decommissioning copper is totally legal
.

Goldstein:

Your car analogy is dead wrong. Bandwidth is like gasoline that fuels any car ... depends on your type of car (application) of how much gas you need for speed, on-time arrival and reliable transport. Bandwidth is the gas that makes things go. On "Clecs survive ..." --- are you aware over 900 CLECs went bankrupt working the ILEC copper renters business model? The few survivors have bough BK assets to lower their costs temporarily. Forbearance will change all that.

Bottom line is bandwidth speeds that out strip copper physics are omnipresent and will only go higher. I don't feel sorry for CLECs who sat on their fingers since the Comm Act of 1996 relying on the ILEC. Do CLECs need another 11 years to figure out they need non-ILEC partners or non-ILEC network.
Page 1 / 3   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE