ADSL, VDSL, Plus, 2: What do consumers really need?

The VDSL2 standards process is making progress - as of late May, the standard was "consented," meaning that there is a technical freeze and six months to cross all the 't's and dot all the 'i's. The number of carriers committed to deploying VDSL2 systems is growing. And there's a fair amount of worldwide momentum behind VDSL in general - the folks at Ikanos tell us deployments have grown from 2 million to 6 million ports in just a bit more than a year.
 
So VDSL is gaining and VDSL2 seems to be about ready. At the same time, ADSL2+ has arrived – at least on the chip level and in hardware, if not in terms of higher speed ADSL services. Any price difference between ADSL2+ and plain old DMT ADSL chips is vanishing, and (much like the move from 802.11b to 802.11g) the foundries are gearing up to crank out only chips supporting both standards.
 
There's been much discussion in the press, among analysts, and within the industry in general about the suitability of both ADSL2+ and VDSL2 for the kinds of networks telcos are building alongside (or in lieu of) their fiber-to-the-home deployments.
 
Everyone takes it as gospel, for example, that VDSL or VDSL2 is best for fiber-to-the-neighborhood or basement node deployments, where you can really crank up the bandwidth on super-short runs. Alternately, it’s widely accepted that ADSL2+ is the only choice if you've got mainly 8,000+ foot "ast mile" runs to the customer.
 
Vendors and technology proponents for either side will argue to these "truths." For example, VDSL2 proponents can show charts illustrating that their systems support higher bandwidths than ADSL variants out to a large percentage of customers, and not just to the lucky few within 2,000or 3,000feet. ADSL2+ proponents have similar charts, of course, showing something completely different.
 
What's missing from many of these discussions, however, is enough emphasis on the upstream side of things. We've been talking about the application demand side of the broadband equation for a long time now – after all, supporting customers' applications is what broadband service is supposed to be all about.
 
And the upstream side of application demand is what's really outstripping bandwidth supply these days. Work at home, home entertainment, and "prosumer" applications make short order of the upstream of today's existing and short-term roadmap broadband offers.
 
After all, it takes only one session with iMovie or Windows Movie Maker and a DV camcorder to realize that sharing large image-sized video with family members around the country is still not an application that's truly supported by any domestic broadband services.  Uploading such a file is bad enough.  Don't even try to stream video from your own PC in real time with today's networks  it's just not possible. Whether 256K bit/sec, 384K bit/sec, even 512K bit/sec or a 1M bit/sec upstream – none of these will really cut it for many bandwidth-needy users these days.
 
And home videos are only one of dozens of apps driving the upstream side of things. Digital photographs have gone from  2 megapixels to 8  in just a couple of years, and show no sign of slowing down. Combine these file sizes with the profusion of online photo storage/archive/printing services, and you've got a recipe for upstream bandwidth demand. Try uploading  100 3 megabyte pictures from your baby's first birthday party, and you'll quickly realize that with today's networks this is a "walk away from the computer" background task; not something you can do "real time".
 
Beyond video and photos, you'll find applications like email,where attachments are exploding in size. Or online games, which are growing more popular and more complex – and bandwidth is essential to a good experience there. Heck, even low bandwidth apps, like the upstream half of VoIP and conferencing services, are becoming heavier and adding to the load.
 
Which brings us back to the carrier's technology decisions.  Not all solutions under consideration support the kind of upstream bandwidth that the applications are demanding.
 
Some do however. For example, the VDSL2 folks have a pretty compelling story to tell about the upstream side of the equation. Their distance-bandwidth curve shows upstream speeds in the tens of megabits per second out to several thousand feet – while many competing systems claim upstream speeds maxing out at a few megabits per second at best.
 
That's a huge difference for anyone working with video, audio, digital pictures, or simply working with large files. Imagine sending that 10 Megabyte PowerPoint 45 seconds before the start of the conference call and having it get to all recipients in time! It’s quickly becoming an upstream if not symmetrical world.
 
Now we want to emphasize that we've not yet had the opportunity to set up some line simulators and get some production chipsets to verify these claims ourselves yet. But neither have we heard any strong rebuttals.
 
We think that any telco making decisions about future network technologies needs to do exactly that testing, and to design networks that can support what customers really need in terms of the upstream side of things. Cable vendors, CableLabs and the MSOs certainly are thinking along these lines. Shouldn't telcos?

To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.

The VDSL2 standards process is making progress - as of late May, the standard was "consented," meaning that there is a technical freeze and six months to cross all the 't's and dot all the 'i's. The number of carriers committed to deploying VDSL2 systems is growing. And there's a fair amount of worldwide momentum behind VDSL in general - the folks at Ikanos tell us deployments have grown from 2 million to 6 million ports in just a bit more than a year.
 
So VDSL is gaining and VDSL2 seems to be about ready. At the same time, ADSL2+ has arrived – at least on the chip level and in hardware, if not in terms of higher speed ADSL services. Any price difference between ADSL2+ and plain old DMT ADSL chips is vanishing, and (much like the move from 802.11b to 802.11g) the foundries are gearing up to crank out only chips supporting both standards.
 
There's been much discussion in the press, among analysts, and within the industry in general about the suitability of both ADSL2+ and VDSL2 for the kinds of networks telcos are building alongside (or in lieu of) their fiber-to-the-home deployments.
 
Everyone takes it as gospel, for example, that VDSL or VDSL2 is best for fiber-to-the-neighborhood or basement node deployments, where you can really crank up the bandwidth on super-short runs. Alternately, it’s widely accepted that ADSL2+ is the only choice if you've got mainly 8,000+ foot "ast mile" runs to the customer.
 
Vendors and technology proponents for either side will argue to these "truths." For example, VDSL2 proponents can show charts illustrating that their systems support higher bandwidths than ADSL variants out to a large percentage of customers, and not just to the lucky few within 2,000or 3,000feet. ADSL2+ proponents have similar charts, of course, showing something completely different.
 
What's missing from many of these discussions, however, is enough emphasis on the upstream side of things. We've been talking about the application demand side of the broadband equation for a long time now – after all, supporting customers' applications is what broadband service is supposed to be all about.
 
And the upstream side of application demand is what's really outstripping bandwidth supply these days. Work at home, home entertainment, and "prosumer" applications make short order of the upstream of today's existing and short-term roadmap broadband offers.
 
After all, it takes only one session with iMovie or Windows Movie Maker and a DV camcorder to realize that sharing large image-sized video with family members around the country is still not an application that's truly supported by any domestic broadband services.  Uploading such a file is bad enough.  Don't even try to stream video from your own PC in real time with today's networks  it's just not possible. Whether 256K bit/sec, 384K bit/sec, even 512K bit/sec or a 1M bit/sec upstream – none of these will really cut it for many bandwidth-needy users these days.
 
And home videos are only one of dozens of apps driving the upstream side of things. Digital photographs have gone from  2 megapixels to 8  in just a couple of years, and show no sign of slowing down. Combine these file sizes with the profusion of online photo storage/archive/printing services, and you've got a recipe for upstream bandwidth demand. Try uploading  100 3 megabyte pictures from your baby's first birthday party, and you'll quickly realize that with today's networks this is a "walk away from the computer" background task; not something you can do "real time".
 
Beyond video and photos, you'll find applications like email,where attachments are exploding in size. Or online games, which are growing more popular and more complex – and bandwidth is essential to a good experience there. Heck, even low bandwidth apps, like the upstream half of VoIP and conferencing services, are becoming heavier and adding to the load.
 
Which brings us back to the carrier's technology decisions.  Not all solutions under consideration support the kind of upstream bandwidth that the applications are demanding.
 
Some do however. For example, the VDSL2 folks have a pretty compelling story to tell about the upstream side of the equation. Their distance-bandwidth curve shows upstream speeds in the tens of megabits per second out to several thousand feet – while many competing systems claim upstream speeds maxing out at a few megabits per second at best.
 
That's a huge difference for anyone working with video, audio, digital pictures, or simply working with large files. Imagine sending that 10 Megabyte PowerPoint 45 seconds before the start of the conference call and having it get to all recipients in time! It’s quickly becoming an upstream if not symmetrical world.
 
Now we want to emphasize that we've not yet had the opportunity to set up some line simulators and get some production chipsets to verify these claims ourselves yet. But neither have we heard any strong rebuttals.
 
We think that any telco making decisions about future network technologies needs to do exactly that testing, and to design networks that can support what customers really need in terms of the upstream side of things. Cable vendors, CableLabs and the MSOs certainly are thinking along these lines. Shouldn't telcos?