x
Video services

SBC, Microsoft Defend Lightspeed

SBC Communications Inc. (NYSE: SBC) is close to announcing delays to the commercial rollout of its fiber-based triple-play offering, stemming from difficulties in trialing its IPTV service, Light Reading has learned from sources close to the situation.

SBC has said it expects to roll out new commercial services by late 2005 or early 2006. The services will be offered over a massive new fiber network slated to reach two thirds of SBC’s customer base, or 18 million households, by the end of 2007 with a capital cost of $4 billion to $6 billion.

Our sources' talk of rollout delays underscores the complexity of SBC’s ambitious plan of rolling out a bundle of new IP-based services via a fiber-fed network in such a short time (see SBC Sheds Light on 'Lightspeed' and Swisscom IPTV Stall Sends Shivers).

“You can expect SBC to announce delays,” a source close to SBC’s procurement process told Light Reading on Wednesday. “They have said they will roll the product out commercially by the end of the year, and they are not even close to being ready for [commercial] trials.”

What's the holdup? Sources say it has to do with how quickly and elegantly the Microsoft Corp. (Nasdaq: MSFT) TV solution will scale.

In trials so far, the source says, a single server has been required to support every 10 users. This is because of the processing power needed to support the Microsoft platform's coveted instantaneous channel-changing capability, one source says.

"Things like scaleability are obviously a big part of the focus of [developing] the platform," says Ed Graczyk, director of marketing for Microsoft TV. "I would argue that one server for every 10 households -- that just doesn’t ring true. You wouldn’t have a solution then." But, he insists, "the scaleability in beta versions is not what it will be in the released version."

Microsoft's IPTV system -- and IPTV systems in general -- deliver video channels differently than traditional cable networks, where the customer's set-top box receives all available channels and channel changes happen right at the box.

IPTV is different. The 200-some available channels are delivered only as far as a network server, and only one channel at a time is sent to the set-top box in the customer's home. When a channel change is made by the consumer, it takes a couple of seconds for the box to ask the server for a new channel and for the server to send it down.

Microsoft's flavor of IPTV, according to industry sources, has a unique way of creating the experience of an instantaneous channel change. The server immediately sends a burst of buffered digital "frames" from the newly selected channel down to the set-top box, then uses buffering again to gradually catch up with the real-time broadcast. There is no latency in the signal, only a brief moment where the user is watching somewhat less-than-live programming.

Graczyk disputes our description of how Microsoft's IPTV channel changing works. He does, however, admit that buffering is involved.

Regardless, it's clear this scheme requires some heavy duty processing power, especially when numerous set-top boxes are making demands on the server at once.

SBC denies any problems serious enough to shift the timeframe of the rollout. “We said that we are going to roll out IPTV commercially by the end of ’05 or the beginning of ’06, and that is the target that we’re aiming for absolutely,” says SBC spokesman Wes Warnock.

But our sources are only piling on what's already been talked about. Jefferies & Co. Inc. equities analyst George Notter has a similar take on how SBC's project may be met with delays. “We understand that SBC is still working through a host of back office issues associated with rolling out video services in commercial volumes,” Notter writes in a brief released Tuesday.

Most people believe that Microsoft (with partner Alcatel) has the intellectual and monetary resources needed to work through any scaleability problems with its IPTV platform (see Alcatel & Microsoft Going Steady). But many believe that the software giant, like its client SBC, has simply over-promised on the time it will take to deliver a commercial-grade product.

SBC has so far completed a 10-user, video-only trial with SBC employees and family, and has moved on to a 250-user customer trial featuring voice, video, and data services, Light Reading sources say.

“We’re told that the carrier is working through issues with Microsoft’s IPTV middleware software in relation to delivery of HDTV services -- the codecs for HDTV set-top boxes aren’t ready,” Notter says in his brief (see IPTV: Microsoft's Window to Carriers).

Notter notes that SBC has already shifted the timeframe for the launch of its triple-play offering. The launch of IPTV services was originally planned for the fourth quarter of 2005, but SBC officials say it may sneak into early 2006.

Notter also believes that the delays result from overaggressive goal-setting by SBC as much as they do from actual technical obstacles. “Like so many other RBOC initiatives we’ve seen over the years, the rollout appears to be slipping from the original schedule,” he writes.

SBC’s Warnock says people within SBC are working furiously to deliver the IPTV product on time, but he is conscious of the skeptical gaze of the outside world. “Sometimes I think these things are played up a little too much,” he complains.

— Mark Sullivan, Reporter, Light Reading

Page 1 / 3   >   >>
cyber_techy 12/5/2012 | 3:12:33 AM
re: SBC, Microsoft Defend Lightspeed
<bold>
<italics>
SBC denies any problems serious enough to shift the timeframe of the rollout. GǣWe said that we are going to roll out IPTV commercially by the end of G05 or the beginning of G06, and that is the target that weGre aiming for absolutely,Gǥ says SBC spokesman Wes Warnock.
</italics>
</bold>


What he meant was "We said that we are going to roll out IPTV commercially by the end of '05 or beginning of '06. We didn't specify if it was 2005 or 3005. We need some leeway as well"
OldPOTS 12/5/2012 | 3:12:33 AM
re: SBC, Microsoft Defend Lightspeed Heck, and I love channel surfing.
In fact I have a rubber controller from Comcast to use when I really want to stay on the same channel.

"Things like scaleability are obviously a big part of the focus of [developing] the platform," "Regardless, it's clear this scheme requires some heavy duty processing power, especially when numerous set-top boxes are making demands on the server at once."

I believe more of these scalability and heavy duty processing power problems are to come, and I have the scars to prove it.

OldPOTS
desiEngineer 12/5/2012 | 3:12:32 AM
re: SBC, Microsoft Defend Lightspeed How long are you willing to wait? How long do you wait with digital satellite? If SBC TV takes roughly 1/2 second to switch channels, is that fast enough? I know I had a change in experience when I moved to digital satellite (and cable).

I don't know if these kinds of user interface tests are done rigorously, but we have always adapted to worse behavior for some better behavior.

It's been a long time since my wireline phone service has given me the quality that I have come to expect from my mobile service, but I have stopped complaining about the mobile service because of its other benefits.

I think SBC TV has to give us something to allow us to overlook the deficiencies of the service, if/when they show up.

-desi
BigBrother 12/5/2012 | 3:12:32 AM
re: SBC, Microsoft Defend Lightspeed How much it is going to cost the subscriber or SBC if the subscriber want to have multiple feed into the house, since changing the channel is on the server side? Is it going to be competitive with broadcast?
mclejc 12/5/2012 | 3:12:31 AM
re: SBC, Microsoft Defend Lightspeed DesiEng -
I think the idea is that one doesn't "wait" at all. I've seen this stuff and it's basically instantaneous. Much faster than what I get in my home from the cable guys today. The issue is exactly what the original poster alluded to - there are other "costs" and the network architecture will likely change as the experience factor grows.

M.
desiEngineer 12/5/2012 | 3:12:31 AM
re: SBC, Microsoft Defend Lightspeed cybertechy: What he meant was "We said that we are going to roll out IPTV commercially by the end of '05 or beginning of '06. We didn't specify if it was 2005 or 3005. We need some leeway as well"


Just posting to get my banana :-)

Hey, cut them some slack!

It took the phone company a hundred years - killer app = Mother's Day :-)

It took the internet 50 - killer app = porn :-(

It's going to take IPTV 25 years - killer app = ?tele-Superbowl party :-)

-desi
desiEngineer 12/5/2012 | 3:12:31 AM
re: SBC, Microsoft Defend Lightspeed OldPOTS: "I believe more of these scalability and heavy duty processing power problems are to come, and I have the scars to prove it."

And are you happy you have the scars - i.e., was it worth it?

It's unlikely that SBC's campaign is going to be something like "It's like TV, not quite as good, it doesn't scale well, and the only advantage is that you get a single bill."

Just like AT&T when they came out with phone service way back when.

But if putting the POTS system was worth it, despite all the glitches, the miscalculations, the oversimplifications, ..., then, heck, the scars are worth it.

Likewise, IPTV, if it lives up to its promise. I ain't saying it will, or that it will save the ILECs, or that it is well-designed yet. Just that it seems to be a cool thing that everything will be IP. And if we're not too sanguine about the nature of the problem, we're likely to deliver yet another medium over the IP infrastructure.

Data, mail, phone, and now, video, more complex, more demanding, and just more of it. Sounds like a heck of a problem to sink your teeth into.

-desi
paolo.franzoi 12/5/2012 | 3:12:30 AM
re: SBC, Microsoft Defend Lightspeed
There are 2 factors in determining channel switch time: Channel Availability and Buffer Management.

As long as people don't get nuts and want to oversubscribe video networks a lot, channel switch times using IGMP can be pretty fast. It takes some work to be able to deal with several of the real life issues, but it is not too bad.

Microsoft has wanted to keep channel change times low so they have kept the buffer size inside the set top box pretty small. This is a good news/bad news scenario. Channels will switch pretty quick, but any jitter will basically screw up the picture.

This is one of several hidden scaling issues. I have no idea about what Microsofts scaling problem is, but the biggest one will end up being DSL performance. There will be lots of noise issues that impact individual lines. The worst part will be adding customers and having other customers start to have problems.

seven
rjmcmahon 12/5/2012 | 3:12:28 AM
re: SBC, Microsoft Defend Lightspeed How much it is going to cost the subscriber or SBC if the subscriber want to have multiple feed into the house, since changing the channel is on the server side?

I don't know the details of the SBC/MSFT network design but I can't see the costs being significantly weighted by channel changing implementations. And the consumer price will be driven by things like broadcasting rights, audience sizes, advertising subsidies (or not), quality of production, release window, etc. These things have little to do with technology.

Is it going to be competitive with broadcast?

Probably not.

The history of cable suggests anybody trying to provide a new mechanism for television delivery will have to enable new sources of content. The monopolists (SBC/VZ/MSFT, etc.) don't have that mind set and aren't in the business of building new markets. (Note: MSFT has been trying to get into video for decades and has had no luck.) From a business perspective, IP/TV is going to be harder than VoIP by at least an order of magnitude.

The model that has the *potential* to succeed requires customer investment into the infrastructure in a manner that enables separation of content producers from the bit distributors. Unfortunately, this requires people changing their mind set which is very difficult.

PS. You might find a book on IP multicast interesting. It is interesting technology. "Channels" (for lack of a better word) are replicated by the network devices and not by the server. It scales much better than a server centric model.
desiEngineer 12/5/2012 | 3:12:28 AM
re: SBC, Microsoft Defend Lightspeed mclejc,

"The issue is exactly what the original poster alluded to - there are other "costs" and the network architecture will likely change as the experience factor grows."

Sure, and from my other post - it sometimes takes scars to make it work. That's why it's called bleeding edge. I'll grant that there is difference between bonehead deployment and due diligence in network engineering. But you can't take care of all aspects in your lab simulation, or even in a localized trial. In a sense, this is the part of engineering that distinguishes it from, say, mathematics. I find that fascinating, and I ask OldPOTS and others if they feel the same.

If there is a will and money to back it up, then it's expertise and grunt work, scars, etc. to make it happen. If you've got some frightened executives who don't see IPTV as a vision, and when the going gets tough, they cut your legs out from under you, watch out!

-desi
Page 1 / 3   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE