Video services

Telecom Texas Hold 'em

DALLAS -- Light Reading's first ever Optical Expo was a pleasant surprise, with nearly 500 service providers and networking equipment specialists packing the house at the Westin Galleria in the Texas Telecom Corridor.

Judging by the traffic patterns, optical networking seemed to be a lot more happening than whatever was on sale at the Galleria. And unlike the service providers at the local Macy's, most of the attendees of our event actually seemed happy.

There was even strange talk of things such as service-provider profits, end demand, and constraints in network capacity. Optical networks are filling up again, apparently. An AT&T Inc. (NYSE: T) bigwig told us we'll be running out of the stuff again by the end of the decade. So does that mean 2009 will be like 1999? (See AT&T VP: 100-Gig by 2010.)

If you don't believe that times are strange, we even heard service providers lament the lack of optical networking startups. (See Carriers: Show Me the (Optical) Startups.)

There seems to be some truth to this. While, yes, there are plenty of networking startups being tossed out like bagged spinach from California, the fact is that the few survivors, such as Infinera Corp. (Nasdaq: INFN), are drawing lots of service provider interest. Infinera appears to taking orders for long-haul optical DWDM systems as fast as the waitress at Dakota's jotted down our orders for prime dry-aged beef.

The first new optical networking startup in a long time – Matisse Networks – unveiled its strategy with an interesting concept called "EtherBurst." As handily illustrated with fancy animated PowerPoint slides that can make you dizzy, Matisse says it's building a metro optical switch that combines the characteristics of an optical ADM and an Ethernet switch – switching packets by assigning them different colors of light. (See Matisse Primes Metro Ethernet Makeovers.)

The technology looks promising. Now let's see if Matisse can sell some boxes before EtherBurst goes EtherBust.

So what about those service providers?

"Our revenue may be flat, but our profit margins are expanding with next-generation services," said Mike Jones, CTO of Broadwing Corp. (Nasdaq: BWNG).

Whaaa? Service provider? Profit margins?

Yes, he said. New Ethernet and MPLS services are selling like hotcakes, said Jones, a man who generates a fine soundbite. He also set the record straight on the infamous Corvis switches. It goes like this: Yes, they worked. Yes, they are still running. Yes, they are reliable. But Broadwing won't be needing any more of 'em. End of story. (See Infinera Gets Corvis (Sort Of).)

How about another Jones zinger: "The FCC is confused and will stay confused for a while."

By the way, is there anybody out there that doesn't yet have a reference to YouTube Inc. drawn up on your PowerPoint? Don't bother – it's too late. Apparently everybody else has already gotten the memo.

For the first time I heard a reference to a new form of ROADM, introduced by Fujitsu Network Communications Inc. , called a DOADM – a dynamic optical add/drop multiplexer. It was discussed by Rod Naphan, VP for business access and data planning.

DOADM (pronounced doe-dum) sounds like a handy reply for just about anything. What you doing tonight? DOADM. Having fun? DOADM.

Naphan made some interesting points. He says there will be 80 million users by 2010 that will have IPTV. He says GPON's the way to go, because the minimum bandwidth requirement for a "competitive" triple-play service will be 45 Mbit/s. And he points out that the bulk of residences, in North America at least, are beyond the 2,000 feet needed to extract the necessary capacity from DSL.

This, in turn will drive demand for ROADMs in the metro and DOADMs in the core of the network, says Naphan.

Heavy Reading senior analyst Graham Finnie appears to agree that we've got plenty of fiber in our future, and that bandwidth demand will be driven by video.

"HD [high definition TV] has implications for bandwidth, more and more streams, including upstream video. That's putting pressure on the network and leading to a 20-Mbit/s-plus requirement," said Finnie. "That's the main reason why we're seeing this transition to FTTH."

Okay, so 45 Mbit/s? 20+ Mbit/s? Which is it? What exactly do you need for triple play?

"Everybody's talking about 50 Meg for a reasonable triple-play offering," said Jon Baldry, technical marketing manager at Transmode Systems AB .

So, really, the bandwidth required for advanced services via FTTH is starting to sound like a game of Texas Hold 'Em. Anyone care to raise?

The IPTV panelists, including representatives from Transmode, Lucent Technologies Inc. (NYSE: LU), and Cisco Systems Inc. (Nasdaq: CSCO), pointed out that an HDTV and FTTH growth phase only means good things for DWDM – and ROADMs.

"IPTV will be the primary driver of bandwidth growth," said Donald E. Crowe, Director and CTO for multimedia networking solutions at Lucent. "That's driving a lot of this bandwidth. It's amazing the ripple effect it's having on metro networks and transport networks."

Jeff Fishburn, CTO of municipal networking builder DynamicCity Inc. , let this little nugget slip: He said that Seattle and Portland are eyeing municipal FTTH buildouts. And he makes the point that RBOCs likely can't keep up with demand. "We have a business model where if customers want it, we can supply it," he says.

Okay, I'm "all in" on the FTTH stakes. If there's no limit to this game, I call with 100 Mbit/s.

— R. Scott Raynovich, Editor in Chief, Light Reading

Page 1 / 2   >   >>
paolo.franzoi 12/5/2012 | 3:40:06 AM
re: Telecom Texas Hold 'em
A whole lot of this depends upon how much unicast video bandwidth there needs to be.

For an MPEG-2 based system, let's say a Telco provides 1 Gb/s of SDTV and 1 Gb/s of HDTV to the end office. 2 Gb/s is a lot for some places, but does not require ROADMs or fancy dancy things.

1 Gb/s of SDTV @ 4Mb/s = 250 Channels - a fine lineup.
1 Gb/s of HDTv @ 20 Mb/s (massive overkill) = 50 Channels of HD - better than anybody.

With MPEG-4, you might be able to double these numbers.

Once you unicast these streams - in other words make them personal - then the bandwidth shifts from number of channels to number of customers.

So, let's take a small 10,000 home town - well let's call it a medium 50,000 home town which you get 20% take rate. If we make the assumption there is 1 HDTV + 2 SDTV per home and a DVR per TV then you could end up with up to 6 Streams per home and two of them are HD.

This give us (even with MPEG-4) right around 30 Mb/s per home (50 Mb/s per home or so with MPEG-2). Now here is the good part. Unless this is cached or sourced locally - in this end city - then now you are backhauling a significant percentage of 30 Mb/s * 10,000. A number that blows away many 10s of Gigs of traffic.

Anybody think this scenario is likely for a city of 150,000? How much money do you think you will be able to charge consumers for this service? Do you think it is more than they are paying today? Is a combination of download (like a Movie Link) and a live service (like a Cable Network) seem a lot more likely?

yarn 12/5/2012 | 3:40:00 AM
re: Telecom Texas Hold 'em MPEG4 makes things pretty efficient and Telcos can probably ride that wave as they do not have a large installed base of MPEG2.

Your 6 streams per household example is a bit pushing it though I think. Everyone would have to watch a different channel and all DVRs must be recording different programs at the same time, rather than playing. A worse case scenario...

So let's make the pipe half empty: typically a single DVR in a master STB is enough as you can view recordings from other STBs in the house.
The combination of DVR, nPVR and near realtime VoD can be used to timeshift traffic load. If there's a monetary incentive to download movies cheaper during off peak hours for later viewing (e.g., at primetime) or watch your favorite primetime soap at the end of the day with nPVR so you don't have to actually record it the moment programs are broadcasted.

In any case, whether you offer 20M, 30M or more to a home, it always makes economic sense to optimize network utilization. It's not just about the size of the access pipe, but how well you use it...
paolo.franzoi 12/5/2012 | 3:39:59 AM
re: Telecom Texas Hold 'em

Content makers do not yet allow MPEG-4 used as a Codec.

Put up a model for your bandwidth and tell me that it is less than several 10s of Gigs for all personalized content.

Your assumptions mean lots of multicast video and downloaded video. This means there is no need for ROADMs or 10s of Gigs of Bandwidth. That was my point.

frnkblk 12/5/2012 | 3:39:58 AM
re: Telecom Texas Hold 'em If content makers don't allow for MPEG-4, how is Cavalier Telephone doing what they do?

karp 12/5/2012 | 3:39:55 AM
re: Telecom Texas Hold 'em Cavalier Telephone is, most likely, getting MPEG2 content and reencoding it to MPEG4.

Scott Raynovich 12/5/2012 | 3:39:54 AM
re: Telecom Texas Hold 'em I still want 100 Megs. Oh, and here's other things I want:

--A DVR that holds at least 50 hours of HD programming, but doesn't cost a fortune (unlike Tivo's new $800 cadillac of DVRs).

--A way to "plug-and-play" a PC hard drive to the DVR to expand capacity (why isn't it so easy?)

--A way to add/remove a la carte programming from my cable lineup, instantly, via a Web portal

--For James Dolan to confess that he's the worst NBA sports executive of all time (regardless of whether or not he's granting stock options to dead people)
paolo.franzoi 12/5/2012 | 3:39:54 AM
re: Telecom Texas Hold 'em
I don't know about Cavalier Telephone in specific, but IPTV has been around in the IOCs for 5 years - before the use of MPEG-4.

Drew Lanza 12/5/2012 | 3:39:53 AM
re: Telecom Texas Hold 'em Scott:

I didn't realize you were a cricket fan...

Scott Raynovich 12/5/2012 | 3:39:52 AM
re: Telecom Texas Hold 'em http://www.lightreading.com/do...

Ah yes, I remember that!

Maybe it's time to revisit?

I agree with the premise that most of these problems could be solved with hard drives, caching, and software... unfortunately it doesn't appear may people are headed that way!

I have been do a little research, and the RBOCs appear to be totally clueless on the concept of caching content and how it could vastly improve their networks. It's all about the pipe, for them.

I think this actually shows there is promising potential for innovation in the future. Think of where home content networking is now -- how primitive it is. I mean, it takes hours to figure out how to just hook your HDTV up to a computer. Someday I imagine this will all be plug-and-play and standardized. Hopefully.
firstmiler 12/5/2012 | 3:39:43 AM
re: Telecom Texas Hold 'em More things I still want...

--A working wireless DVR/STB base station at sub $300 and slave STBs at sub $50 / unit

--A VOD Head End Delivery system with fully loaded costs working out to allow me to reach cash flow positive on a 1 hr piece of content being rented <=7 times at $.50 per rental (net of content cost, in other words I can load a 1950s public domain documentary on the dangers of marijuana and store it for life with a break even by the 7th rental @ $.50/per). This allows me to maitain a comfortable risk/payback balance in loading the "library of congress", without having to constantly groom the content to optimize rentals per unit (higher opex ruining the model).

--Low cost purpose-built GSM capability in IP telephony CPE to allow for acurate E-911 locator database maintenance, avoiding the clunky manual analog method being imposed on the new digital business model


I still want 100 Megs. Oh, and here's other things I want:

--A DVR that holds at least 50 hours of HD programming, but doesn't cost a fortune (unlike Tivo's new $800 cadillac of DVRs).

--A way to "plug-and-play" a PC hard drive to the DVR to expand capacity (why isn't it so easy?)

--A way to add/remove a la carte programming from my cable lineup, instantly, via a Web portal

--For James Dolan to confess that he's the worst NBA sports executive of all time (regardless of whether or not he's granting stock options to dead people)
Page 1 / 2   >   >>
Sign In