Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
Pulpit Comments
December 21, 2007 -- The Once and Future King
Status: [CLOSED]

I ought to be outraged, why am I bored? ISPs want to break the internet, somehow I am not surprised.

Dave | Dec 21, 2007 | 3:25PM

If there's such a bandwidth shortage, maybe cable providers should start doing what everybody wants them to do, namely unbundle cable services. I don't want 70+ channels. I only want ten or so. So let me pick the ones I want and save the extra bandwidth for other uses.

Rich | Dec 21, 2007 | 3:37PM


Multicast is definitely used at large financial companies for financial apps and for Tibco. Also, I have personally seen it deployed for video in a number of enterprise customers. Do not forget all of the hotels who use it for video to guest rooms as well.

It has certainly been in use for a number of years. It is only natural that service providers use it for video on their IP networks. From the pixilations I used to see when I was a Comcast customer, I was under the assumption they already were streaming shows via multicast.


Andy | Dec 21, 2007 | 3:58PM

You don't understand the problem and reason multicasting hasn't been deployed on open networks.

The reason is that if you deploy multicasting on your network but the network above you has not you (ISP/NetOP) are left paying for unicast-non multicast upstream bandwidth while you are effectively forced to give away the multicast bandwidth you provide.

This is obviously not economically viable so, hence, no multicasting. In essence when you have an open and randomly interconnected set of networks either they are all in re multicasting or none are in.

Rob | Dec 21, 2007 | 4:17PM

"Multicast solves this problem because it allocates no bandwidth to channels that aren't being watched."

Cable companies are already doing this through switched digital broadcast. If they can do it through SDB, then why would they use multicast?

Josh Ferguson | Dec 21, 2007 | 4:34PM


I pity you having Comcast as your ISP. I have them as well. But you might be even better off than I am because in this tiny little corner of the Comcast world (88,000 subscribers in Tucson, AZ), we get everything LAST. Our On Demand came only recently, and with 1/5th the number of free movies back East subscribers. We pay as much as you do, for far less service. How about the long promised Tivo upgrade for the wretched Motorola 6412 III DVR's we have to suffer through (no Scientific Atlanta boxes here to be found)? Don't even ask! If it arrives here a year from now, that will be a surprise. And of course, Comcast is in the forefront of blocking entire ISP's to cut down on spam e-mails, and doing everything else they can (including lying about it) to prevent you from actually using the bandwidth you feel you purchased. With no fiber to the premises competition in sight, our bills are going to remain as high as Comcast can manage into the foreseeable future. As for multicast, I'm sure we'll get it last, provided we get it at all. So why didn't you just title this piece, COMCAST SUX (AND LIES)!

As for multicast, our local cable loop has over 200 Internet accounts, and certainly more TV only subscribers. Since they've got to send the entire program to the whole loop even if only one person is watching it, we could have 200 or more different programs being multicast at the same time. At what point does the local loop saturate?

Btw, Comcast here is trying to retire the entire analog tier and replace it with normal quality (looks worse) digital tier within the next year. I wasn't planning on dying that soon, although it will kill my mother's new Windows XP MCE machine with analog TV card.

David B | Dec 21, 2007 | 4:34PM

I don't think the description of IP multicast is entirely correct, at least as it pertains to video distribution. Multicast data is basically just broadcast data that A) can cross subnet boundaries and B) gets directed only to the subnets that have at least one subscriber for the stream. But it is generally just that--a stream. IP multicast doesn't provide a "local cache" of packets, waiting for a subscriber to fetch them. There's generally no higher-layer ACKs or retries. There's just a bunch of packets whizzing by, hoping that someone on the subnet cares to hear them at that moment.

So while I agree that multicast could help reduce the number of streams flowing to a given subnet, thus reducing the required intranet bandwidth for distributing N channels of live video (like ESPN), it's not going to work so well for the on-demand archived content (like Junior Mint) that unicast P2P provides. If you made your subnets small enough to get benefits for live streams (say, because nobody on my block is watching "C-SPAN3" right now), you just reduced the chance that any two people on the subnet would be clamoring for the Junior Mint archive at the same time. In which case, you might as well just unicast the archive, since unicast is more reliable.

Though, if you really want to replace the "service" currently provided by unicast P2P, I can envision ways to use multicast as part of a solution. I just don't think its quite as simple as described...

Casey Barker | Dec 21, 2007 | 4:48PM

I work for Internet2, which runs a research & education network backbone in the US. We run multicast widely on our network.

My opinions here are my own, however.

the reason network operators hate multicast is not because it's a resource hog OF BANDWIDTH, as others have said.

It is generally NOT a horrible resource hog in the way he's suggesting; the whole point of it is that only one copy of a given stream is only sent ever.

The reason network operators hate multicast is twofold.

1) the "local cache" bob refers to is sort of true - it's a local cache not of data but of what sources are active sending to what groups. Traditional multicast has been multi-sender on a given group number, and the network needed to keep track of what senders where there. THe protocol used for this is called MSDP and it is in fact a resource hog - not of bandwidth but of router memory. It's also fragile and by definition not scalable.

2) current best practice does away with MSDP and goes to a model of SSM - source-specific multicast (rather than ASM, any-source multicast, what i described above). So that solves the MSDP problem. The rest of the protocols are, in truth, quite reasonable and not horribly resource-intensive.

We still hate multicast, however, because the vendors do a poor job of making it work. Both cisco and juniper routers have on and off problems making multicast work, and many vendors' ethernet switches just don't handle multicast at all. Many others offer limited support or don't work right.

So whenever it comes down to actually deploying the stuff, ESPECIALLY in a mixed-vendor environment, you run into lots of problems and it is hard to keep it running well.

dan pritts | Dec 21, 2007 | 4:58PM

How can multicast handle "zapping"?
Or we have to change our way of "knowing that there is nothing worthwhile to see in the cable"!

Luis Alejandro Masanti | Dec 21, 2007 | 5:12PM

I honestly think the future of video is downloaded video, not streamed. The main reason is streaming has no ability to resend dropped packets.

Even at current compression ratios a few of the right dropped packets can ruin several seconds of video/audio, making a show unwatachable very quickly (I find I tolerate a poor analog signal far more than a poor digital signal).

Increase the compression with MPEG4 and every packet becomes that much more important. Drop one and you ruin even more of the signal.

downloads though can resend dropped packets. My TiVo has already trained me to wait before watching a show (so I can jump commericials!) so waiting for a show to download, or partially download, before watching isn't a big deal.

kevinv | Dec 21, 2007 | 6:16PM

How much of Comcast's bandwidth problem is due to the unexpected success of their VoIP offerings?

Bob | Dec 21, 2007 | 6:17PM

As was mentioned earlier, multicast IPTV is alive and well on Internet2 ( ). If you happen to be at a college or University that is connected to internet2 and your network is multicast enabled, there are plenty of IPTV channels you can tune into from all over the world.

To see an example of some of the stuff that is available, take a look at or Both of these sites generate lists of IPTV streams that are currently being broadcasted.

While multicast is not perfect, it is definitely working well enough on internet2 to be useful.

Also may colleges and Universities use multicasted IPTV to transmit all sorts of channels over their local network. Oh, and don't forget that programs like Limewire and some bittorrent clients also use multicast to some extent.

Dan | Dec 21, 2007 | 6:37PM

A paragraph then you want us to click MORE? What for, a web page can hold more than a paragraph.?.

I guess some nerds are so weak they can't carry an ounce let alone a pound of sense, ya suppose that leaves nary a lick of sense amongst thee ears?


Mr. T | Dec 21, 2007 | 7:17PM

So... Comcast is going to save their 420MHz of bandwidth by multicasting, yet Nellie still wants to plug the coax into her TV? Is Comcast going to put a whole-house IP-Multicast to Analog box in the basement/garage ala FIOS then, or on the pole? Sorry, usually you only make me thing about ramifications, not mechanics. :)

Bill McGonigle | Dec 21, 2007 | 7:38PM

Maybe the reason IPTV failed was not for the reasons you describe but because the product was garbage. I worked with it at the time, a time when plenty of computers were, thank you, capable of playback back standard MPEG-1 streams.

A particular problem with streaming media is how to handle timing. The time stamps that are associated with the audio and video streams and at the system level all have to be synchronized, and the decoder has to slave its timing to the timestamps in the system stream. IPTV got none of this right, and didn't even understand the conceptual issue, let alone the specific details of what they were doing wrong. The net result was audio that drifted from video, with both stuttering every so often as the decoder simply gave up on trying to make sense of meaningless timestamps and resynched.

It's not always the case that a product fails because it sucks, but it is sometimes the case; and this was one of those cases.

Maynard Handley | Dec 21, 2007 | 9:25PM

Cringely, what do you possibly mean when you say that "the Congressional law for all TV's to support digital by Feb 09 doesnt mean much"? (paraphrase).

Of course it affects cable companies, and of course it means a LOT. It signals the dying out of analog and will begin to let Cable companies taper off supporting that. In fact part of that bill/law stipulates that vouchers can be given to customers to purchase set top converters that can take digital signals and make them analog.

These facts seem to lessen the grimness of the picture you painted about all ISPs having to spend a majority of their bandwidth transmitting analog signals. While they may not be relieved over night, there is a pretty wide path for the conversion to happen in a reasonably short period after the law goes into effect.

Am I missing something?

Tim | Dec 21, 2007 | 9:36PM

I'm surprised that you seem to have overlooked AT&T's IPTV video rollout that they're calling U-verse. Sure, it's only available in limited areas at the moment, but it's expanding like crazy. It's multicast IPTV.

Wes Larson | Dec 21, 2007 | 11:42PM


The February 2009 deadline doesn't mean much to cable companies because it applies only to over-the-air broadcasts. In fact, most cable companies are required to provide a certain level of analog service (either through state laws or through the franchise contracts that they sign). And apart from outright requirements, cable companies have to deal with multi-TV households, where they might be willing to pay for digital cable for the main family TV but do not want to have a set-top box for every TV (and definitely do not want to pay for having a set-top box for every TV). Further, unlike over-the-air situation, where the feds have a financial interest in cutting off analog broadcasts and selling the bandwidth for other uses, the feds have no such financial interest in how the cable companies transmit their signals. If anything, the FCC would rather keep things the way they are.

BTW, the set-top converter program that you mention is for ATSC (digital over-the-air) signals and not the QAM standard that is typically used by cable companies. And the program can be used to purchase only the most basic set-top boxes (no high-def, no DVR capability). So after you get your voucher from the federal government, you get to pay $40 of your own money for an $80 set-top box that provides basically the same TV that you get right now. Rest assured that you will have hundreds of thousands of angry citizens demanding an accounting for the resulting fiasco.

ploeg | Dec 22, 2007 | 1:22AM

Okay, it's a nit, but you and I have always been into picking them so I'll pick this one. There weren't any "x386" computers, the "x86" terminology didn't come out until the Pentium was running out of legs with 32 bits and Intel built the Itanium. The x86 designation was created for AMD's more compatible response to Itanium. You're thinking of the i386, which fit between the 80286 and the 486. (We still used the full model number for the '286, and I don't recall their being a letter prefixed to the 486 very often.)


G. Armour Van Horn | Dec 22, 2007 | 1:23AM

Van: to pick on a double-nit, what you're thinking of as i386, is actually the 80386 processor. i386 is the name for a computer architecture (a "platform") that is based on Intel 80386-compatible instruction set of the CPU and is supported by successors, including the most modern Intel Pentium and AMD CPUs. As such, you can hear operating system programmers referring to supporting the "i386" platform, although some of the systems (for example, some flavors of Linux) no longer support the 80386 processor that gave the architecture its common name.

MaxVT | Dec 22, 2007 | 2:17AM

This is not just a technical issue. Show me the money. Here in Australia I have not seen a viable business plan for ISP's to make money from the service, rather than it being an added cost to provide a fragile service. All the indicators I have seen is that it is a solution looking for a problem. Unless it increses profit by a healthy percentage it wont happen.

I agree with Luis, I believe download video will become a defacto standard

Peter Bolton | Dec 22, 2007 | 7:44AM

So is the lack of multicast holding back both NerdTV and the new book? Eh? Eh? ;) Not a peep about either... humph!

Derek | Dec 22, 2007 | 2:58PM

There are multiple technical views of IP Multicasting and related products. Could someone really nail it down for us?

Feb 09! Here in Hong Kong, digital TV (I think not just air) is the law begining Jan 08! And, there is no voucher for digital to analog set-top box. You are on your own, man.

Tony | Dec 22, 2007 | 3:50PM

It seems that co's the size of Comcast are likely to gin up a solution somewhat more complicated than can be discussed in a 1600 word essay.

E.g., how complicated would it be to push the HD-MPEG4-to-other (analog, MPEG2, etc) converters out to the last mile, so that the backbone only tied up a single, low-density format? ... or to give each household a converter box?

Somewhat ironic for an article about Hi-Def to be limited by big-picture, skip the details, thinking, no?

Walt French | Dec 22, 2007 | 6:53PM

Would it be possible to implement a p2p application that uses a multicast delivery system?

Steve Pelletier | Dec 22, 2007 | 6:55PM

Well you can multicast this...............

Have a Happy New Year.

Fast Fred | Dec 22, 2007 | 7:40PM

Hard to be sure, but one way streaming video - multicast or not - is going to be forced onto the ISPs is if entertainment artists realize that file sharing is here to stay and the ONLY way to monetize their output is LIVE (or pre-recorded live, like "The Tonight Show") is broadcast over the Net.

Downloads won't work here - the sole appeal is that it's LIVE. After the performance is done, THEN you can also try to sell the downloads of the performance. And that will work most of the time since why would anyone pirate what they can get for low cost direct from the source? Of course, some people will, but since the bulk of the revenue is from subscriptions, the smart artists won't care and will see the pirated downloads as what they really are - free marketing for the subscription to the live broadcasts.

If this model for the music industry takes off, the ISPs are going to have to stream that video somehow.

Richard Steven Hack | Dec 23, 2007 | 3:38AM

Has anyone patented multicast downloads of applications and live-CDs? I would expect MS, software vendors, and even Linux live-CD distros to let ISPs stream their ISO files out for general release. Not the authorization codes, of course. One channel could probably deliver all the software ever written every day. Users could select ahead of time from a menu to choose what to grab and store. Given that all the set-top boxes have USB ports and PVRs these days, temporary storage for stuff until transfer to a USB flash shouldn't be a problem. Will it happen?

Bob Carney | Dec 23, 2007 | 3:19PM

Here in the UK the BBC has been trailing multicast. At the University where I work we can receive certain trial BBC multicast streams and it's interesting because on Linux the firewalls are set to block multicast by default. On Windows XP you seemingly need administrator privileges to see multicast streams because you need to poke a listening hole in the default firewall (I haven't tested Vista so that could be even simpler). I couldn't get any of the sample streams to work on OSX.

Multicast seems to do little caching (in the sense of saving data for hours later and serving it up again in the way an HTTP cache might). Instead it seems to be geared towards live streams where after the setup is done only one packet is sent to the University before it is split up and sent to everyone interested in it. I think Wikipedia had a good multicast description (although that may have changed by now).

At home two problems turned up. Most UK ISPs are not multicast capable - their routers are simply not set/capable of dealing with it. Secondly, even if your ISP is multicast capable you may find that your off the shelf cable/ASDL modem presently is not.

Obviously content providers would like multicast but I think the bad news is that it's a good technology that is facing an uphill climb in a similar fashion to IPV6 - a real shame.

(By the way some imaging software like Symantec Ghost can make use of multicast if your network is up to it)

Anon | Dec 23, 2007 | 4:46PM

If multi-casting is officially implemented, won't this be one of the greatest things to happen to Internet Radio stations? Suddenly, the cost associated with paying a lot of money for upload bandwidth in order serve each individual listener goes away.

If I understand this correctly there would be no need for the ISP to cache anything since the streams are always "live". If I tap into a stream the ISP just sends it on to me. If my neighbor taps into the same stream, then the router at the ISP kicks in and takes the single stream coming in from the Radio station and splits it into 2 identical streams going out. One to me and one to my neighbor. Since the multiple streams are contained within the ISPs Intranet, the bandwidth issue is "cheaper" and easier to manage.

Will mulit-casting be available to a small individual "hobbiest provider"?

drewby | Dec 23, 2007 | 5:36PM


I've built, deployed and maintained dozens of 'triple play' networks in Europe and the USA. 'Triple play' is an industry term that means a network that delivers voice, video and internet on the same infrastructure. There is a convergence going on between ISPs, power companies and traditional cable companies. That much I agree with. And every one of these companies is looking to multicast for video delivery at the last mile.

But no one, really no one is looking at using multicast across the public internet as a means of video distribution. The simple reason is that it's a security and administrative nightmare.

Here's the breakdown:
Once an IGMP packet is received at a subscriber's L3 boundary the request for a specific multicast stream needs to be handled by some multicast routing protocol. This is typically PIM sparse(PIM-SM) mode in local networks. But PIM-SM and other multicast routing protocols needs to administered locally because it requires trust among all routers not to poison the multicast cache of other routers. So PIM-SM and other multicast routing protocols cannot be run on the public internet.

So what about MBGP you ask? MBGP stands for Multiprotocol BGP and provides a mechanism for advertising unicast nlri's and unicast next-hops. It is not a multicast distribution protocol like PIM-SM. It is a means to distribute unicast reverse-path-forwarding(RPF) routes across administrative domains. But that's it. I'm not going to go into what RPF is about.
I've setup a couple of ISPs to use MBGP as a means to share multicast routing information between one another. But the level of trust that must exist between them is high. At this point the full security exposure of multicast routing protocols is a very large unknown. Simply because no ISP in their right mind would expose their multicast routers to the outside world. And it's not a question of buffer overflows that are not yet discovered or tricky coding. The protocols as they exist today are not designed to be that secure because they were never intended to be used on the public internet.

People have been talking about multicast in the public internet about as long as people have been talking about IPv6. It's just not going to happen until some major hurdles are overcome. IPv6 alleviates some of these problems but who knows what kind of security nightmares a full switch to IPv6 will bring. What we might start getting is something where a content distributor offers up a stream via multicast and people start thinking they are receiving a multicast stream when instead they are receiving a directed-unicast stream with a multicast destination IP address. This only requires that the user's ISP and content distributor communicate with one another somehow. But to get real multicast routing in the public internet like we have now with unicast routing is not going to happen anytime soon.

I can always be wrong and I invite others to comment on this. Please point out any nonsense I tried to pass as fact.

As an aside the multicast address you suggest is flawed. The IETF says you should use 239/8 for things like public video streams. If you start using addresses in the 224/8 range you run the risk of using a well known multicast address or one that might be reserved later by the IETF. For instance is all hosts on the local subnet.

Andrew McConachie | Dec 24, 2007 | 6:04AM

You're wrong about the preservation of analog cable channels. Comcast has been cutting them out one by one - at this rate, it will take a decade to pare us back to 4 channels, but if the weather channel and the radar channel and the guide channel and one of the Spanish channels can be pared back, I expect it's not long for the local college channel or the local school channel or any of the other extremely-low-viewership stuff.

When it becomes more financially viable for them to reclaim all their cable at the cost of giving a free set top box or two to all their subscribers, they will take that route - Comcast here has been offering free HBO/showtime for a year to existing subscribers if you agree to switch to digital for free. Analog is something they're going to gradually phase out - I don't know if it will survive to see the bump from OTA subscribers switching, but it won't survive for many years after that.

anonymo | Dec 24, 2007 | 6:44AM

I want to know when the cable company will allow me to see how many other people in my neigborhood are watching the same show as me.

Neighborhood: 7 out of 127 active settops
Region: 4,020 out of 120,489 active settops
Nation: 942,004 out of 8,994,245 active settops

Once the boxtops are digitial they have the info, they should share it back with us.

Mark | Dec 24, 2007 | 8:41AM

I want to know when the cable company will allow me to see how many other people in my neigborhood are watching the same show as me.

Neighborhood: 7 out of 127 active settops
Region: 4,020 out of 120,489 active settops
Nation: 942,004 out of 8,994,245 active settops

Once the boxtops are digitial they have the info, they should share it back with us.

Mark | Dec 24, 2007 | 8:41AM

I suppose it's a stupid question, but why are the cable TV companies still bothering with cable TV?

Is it because the physical cable lines don't really belong to them and can be turned over to another cable TV company if they stop showing television? Because they have contracts with local municipalities that mandate television be provided? What?

Think of all the bandwidth that would be freed up if they just told people, "Dude, call DirecTV or Dish Network" when people called to order cable television.

Yes, the "triple play" becomes a "double play" but a heck of a lot of bandwidth would be freed up and they could get on with life. Start selling truly high speed internet connections to the Uber-Geeks of the world.

They wouldn't have to hassle with the local NBA owner who decides to create his own TV station that carries the team's games exclusively, HBO wanting more money even though the Sopranos is over, etc.

Just my $0.02.

Winston Smith | Dec 24, 2007 | 6:34PM

It's still a problem of lack of competition in broadband internet. Most areas have a choice of the monopoly cable company, the monopoly phone company, or maybe both if they're lucky - hardly ever any other alternatives. If there were competition in broadband internet, ISPs couldn't get away with the interferences with internet use. Not only would they have to avoid abusing net users, they'd have to actually improve infrastructure.

swhx7 | Dec 26, 2007 | 1:22AM

And no comment at all about how AT&T is taking IP Multicast and turning it into a complete alternative to "cable" or satellite TV. You're slipping, Bob.

But seriously... IP Multicast does take up a LOT of bandwidth. However, a separate network dedicated to IP Multicast works much more efficiently than one where IP Multicast is tossed onto existing infrastructure.

Plan big; you're looking at a LOT of servers, not just for your stations, but also for your subscriber administration. Your routers and customer-endpoint switches have to be extremely robust, and the cable plant connecting it all needs to be top grade.

But, once it's in place... look out.

George E | Dec 26, 2007 | 8:30AM

America: finding new and innovative ways to feed off the glass teat since its inception.

Troy | Dec 26, 2007 | 10:11AM

America: finding new and innovative ways to suck off the glass teat since its inception.

Troy | Dec 26, 2007 | 10:11AM

Google as a replacement for TV Programming Aggregators and Satellite delivery of Programming.
Seems to me Google is developing and deploying a TV Programming delivery network as we speak in the way of their Fiber linked Nationwide Data/Content Centers.
When the local MSO provide a Direct fiber feed into their local Google Data Centers they have effectively emulated a Video Distribution Network that will replace most of their Headend investment, gain them control of their bandwidth (IP MultiCasting more efficient) and allow them to eliminate the ludicrous Fees the Aggregators (like TVN) charge them (Revenue share). All Google needs do here is negotiate direct feeds from Video Content providers, provide Fiber links to each, package the video content (add Meta Data) and broadcast Nationwide, while gaining control of the ad revenues from all TV programming which they can share with Content Developers and local Service Providers.
When these local MSO are also ISP and ILECs they gain added value from the direct Fiber links allowing them access to and use of the Google CLoud Computing services as well as a Premium Internet replacement service.
Jim (aka Jacomo)

Jim A | Dec 26, 2007 | 3:27PM

If there were competition in broadband internet, ISPs couldn't get away with the interferences with internet use. Not only would they have to avoid abusing net users, they'd have to actually improve infrastructure.

swhx7 | Dec 26, 2007 | 1:22AM

No, they wouldn't. I have three service providers in my area, and all of them suck in different ways. Service #1 throttles just about everything - if you can pull a download at +50kbps, it's because it's 3:00am and you're the only one up. Channels have started disappearing from the analog basic service, replaced with "This channel now on digital channel #232" banners. Service #2 is okay, but blocks anything but HTTP, POP3 and SMTP unless you buy the 'business' package - fine for Grandma Smith to surf and check her mail, but bad for Mr Contract Programmer who needs VPN or likes to play networked games with his comrades. Service #3 offers both cable and DSL, but you must rent their equipment, and the basic subscription is *really* basic ... unless you also buy the VoIP service package or pay up for "Basic +".

The customer service at all three is non-existent, unless you count automated email responses or hours of hold music as 'service'.

We're to the point that at the end of the school year, we dump our current basic+internet package, switch to a slightly higher-priced competing internet package, and dump the cable tv service entirely. The nickel-and-dime increases over the years have priced me out of the cable market for good, and everything of interest is now at Blockbuster, NetFlix or on line anyway.

GuyFromOhio | Dec 26, 2007 | 4:07PM

Hi guys,

I hope you all take a look at TXP Corporation. It's my investors hope they will do extreemly well within the FTTH roll-out. Fiber is the future, baby!!

steve schiets | Dec 26, 2007 | 4:53PM

Whatever happened to Ultra-Wide band over cable? UWB always seemed like the ideal solution to me. Didn't Cringley do an article about it a few years back.

bill | Dec 27, 2007 | 1:46AM

But do they watch the *same* 10% of available channels, or does each customer have a different 10% that they watch? This was always the problem with the old "80% of people use 20% of word processor features" - each used a different 20%

tom | Dec 27, 2007 | 9:28AM

Is it because the physical cable lines don't really belong to them and can be turned over to another cable TV company if they stop showing television?

Sort of.

Because they have contracts with local municipalities that mandate television be provided?

Bingo. Most cable contracts require a cheap basic package with the local broadcast channels, the local access and government channel - all of which must be analog.

Think of all the bandwidth that would be freed up if they just told people, "Dude, call DirecTV or Dish Network" when people called to order cable television.

You must not live in an area with tall trees and/or frequent thunderstorms. I had DirecTV for a couple of years and signal fade during rain was a major drawback. And it rains a lot here in Houston.

Also, cable TV is their foot in the door to sell you phone, ISP and security service. It's the loss leader that "gets you in the store" like the cheap steaks on the front of the grocery store flyer.

ech | Dec 28, 2007 | 10:04AM

The February 2009 deadline is very important for the cable companies too! It singles the irreversible death of analog TV. From now on, all programming will be digital. All TVs sold will be digital, and customers are going to demand digital programming from their entertainment providers

Sure, it's not a hard deadline as it is for broadcasters. Analog cable will be around for another five years after that. But, cable and other providers of video entertainment will soon put HD offerings on their main tiers instead of being on a separate tier as it is now.

With my provider, everyone will have the IO digital cable package as their base package. Local franchise agreements will be renegotiated. Analog channels will be dropped one-by-one. Sports networks will be the first to go digital only. Then movie channels, and finally the broadcast channels.

Customers who insist on analog cable will sooner or later find themselves in the same position as AT&T's analog customers: The company will say to their customers, it was nice knowing you. Now scram. It won't happen in 2009, but by the time 2014 comes around, analog cable will be dead.

David W. | Dec 28, 2007 | 1:30PM

Until 4 years back, 'Multicast is a research only area' when I took a course on networking. It looks like things have come a long way since then.

listen_to_blogs | Dec 30, 2007 | 1:59AM

Some form of multicast is the obvious way to replace
broadcast media transmissions the current multicast implementations are not it. If you don't use some form of source specific multicast protocol then anybody can pump out packets onto a broadcast group
to either DOS everyone watching that channel or to
subvert it to some other purpose (multicast file sharing anyone).

stephen booth | Dec 30, 2007 | 10:09AM


Can you please elaborate on your comment about Verizon FIOS:

"Even Verizon's fiber-to-the-home service, FiOS, is moving to multicast because it was architected in a dumb way that sorely limits what should be a lot of throughput."

Perhaps this is addressed in a previous column that I missed?

David Strom | Jan 02, 2008 | 9:58AM

And another question -- if the cablecos move to IP Multicasting for TV delivery, what happens to CableCard support? FCC requires cablecos to support CableCard, I believe, and I doubt that there is support for IPTV by CableCards.

David Strom | Jan 02, 2008 | 10:36AM

Having noticed this late because of the holidays, this is a late correction/comment.

Having conferred with some of my colleagues here at TIBCO, I can confirm that multicast occurs all over the place - in local networks, and not just in our products. My colleague mentioned Apple's Bonjour, UPnP, in addition to our own products.

My colleague does concur that routing of multicast is difficult.

Eric Johnson | Jan 03, 2008 | 1:58PM