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
February 02, 2007 -- WYSIWYB
Status: [CLOSED]

IMO it will require too much far sighted thinking for most ISP's to voluntarily do this.

I suspect they will put up with degraded service until their customers bail for a better ISP. One using this technology, then widespread adoption will follow.

Although the NAT workaround slowed the adoption of IPV6 perhaps a workaround will be found for this!

Jimbo | Feb 02, 2007 | 7:34PM

Bob - One of your best articles in some time.

There are two things I don't understand (at this moment): Firstly, what does CISCO and the other big network router people have to say about this. Do they see a problem and do they have in mind a solution. Second, I was under the impression that the backbone is really ATM with TCP/IP layered over it. Does using ATM with its quality of service facilities built-in make this easier or more difficult?

Steve | Feb 02, 2007 | 8:41PM

This sounds like a way for large telco ISPs to make their solution (of offering movie downloads themselves) even more attractive, thus killing off the smaller ISPs and other content providers, solidifying their position as a monopoly.

A prime example is BT, who owns the telephone exchanges and the fiber that connects them, and so their bandwidth costs are much lower than other ISPs who buy bandwidth wholesale from them. Their movie download service has an unfair advantage, and frankly the regulations preventing them from entering the IPTV market should never have been lifted.

Davoud | Feb 02, 2007 | 8:46PM

why isn't it ISTP, not ITSP?

dan | Feb 02, 2007 | 9:33PM

Maybe NIST needs to get involved to ensure the bandwidth consumer is getting what is sold. This would force ISPs to compete.

We don't need to guess whether we are being shorted when we buy a gallon of milk; why should we have to guess about bandwidth?

Honest competition couldn't hurt the consumer.

Doug Whitehead | Feb 02, 2007 | 9:58PM

The rare typo! Your proofreader must have missed their morning java - the acronym for InterStream Transport Protocol (according to NuMetra's website) is ISTP.

However really intesresting technology, I agree though that NuMetra has a tough sell to the ISPs for adopting a technology they don't fully control and could expose the flaws and poor performance of their networks that might in turn force them to invest more in their infrastructure. I'll be looking for the first taker...

Buckdakat | Feb 02, 2007 | 10:52PM

Actually the ISP's just need to monitor their networks and add bandwidth as needed. It would be wasteful to provide everyone full bandwidth when everyone does not use full bandwidth.

Also, if the video is not streaming, then it does not matter if there is a hiccup. Basically what we need is an Internet PVR that collects video and stores it for us to watch locally when we are ready.

Chris Nystrom | Feb 02, 2007 | 10:58PM

So where does Joost, the video technology from the folks that brought us Skype, fit into all these scenarios?

Page Hite | Feb 03, 2007 | 1:11AM

You are right about Google taking over the internet traffic, and to know how right you are, go check:

What does this small piece of software do for you, you wonder? Why, it just reroutes all your internet traffic through Google. For real.

(sorry for the reposting, but i thought it would be more visible here)

Vahid | Feb 03, 2007 | 1:48AM

Bob, speaking as a DSL ISP, you are missing quite a bit in the cost-side to the ISP providers. You have forgotten the toll that the telcos add to the circuit. In a typical $40 3Mbps DSL package, $28 goes to the telco, basic support costs are about $3, and 3Mbps of bandwidth would cost $45. So assuming zero profit for the ISP, you could provision at best 600 Kbps; and still go out of business.

Milton | Feb 03, 2007 | 2:02AM

Couple of comments Bob, and thanks for the articles - you've been on a role these last few weeks.

1. Not all of this content needs to be real-time. I consume a lot of internet sourced content (podcasts) and my iTunes just pulls it down then synchs to my iPod - I don't need to worry about it until - hey presto! - there's some new content. We only have the music store here in NZ but for you in the US you now have this model on your TV with Apple TV - seems great! Does it really matter how long it take for the content to trickle down?

2. I have just started to see products being presented to ISP's and telco's that allow them to insert their own advertising into content streams - so not the source websites Google ads, but the ISP's ads based on their commercial relationships. The question is - who owns the user's internet experience? But ISP's/telco's need to do something like this to recoup costs.

3. At the end of the day the last mile access is a utility product - it is very hard to see what value the ISP really ads above and beyond the telco/cable provider. Wouldn't Google be best trying to forge relationships with the last mile provider and forget the ISP's altogether?

Tim | Feb 03, 2007 | 3:33AM

Just re-read the endless predictions for the last 10 years that the internet was about to crash for lack of bandwidth. How many times are idiots like this going to "cry wolf"?

Take everything this guy says with a pallet of salt. He has nothing new, nothing smart, nothing interesting. Use your common sense.

Joe | Feb 03, 2007 | 3:54AM

I echo Tim's comments... I'm willing to wait for a download.

One question - the other day I downloaded Windows Live Messenger 8.1. I was getting 500 KB/s bandwidth on a cable system. How can I get bandwidth an order of magnitude more than what my ISP provisions?

Graham Fair | Feb 03, 2007 | 4:58AM


A good article: very interesting.

Your P2P comparison makes me wonder about the one thing that I haven't seen analysed: the impact of using ADSL rather than DSL.

P2P seems to imply a fairly large upload bandwidth. It obviously works well on DSL systems, but here in the UK we're largely on ADSL and typically the download:upload ratio is 10:1.
It would be interesting to know what impact this has on network performance and especially on P2P.

Martin | Feb 03, 2007 | 6:26AM

Bob I seem to recall reading about several large companies going broke around the turn of the century because they laid so much dark fiber and no one wanted it. Hell in this little burg I live in there are two big fibers running down the railroad track and only one in use. How could all of this dark fiber be used up in 6 years. Keep writing, I only understand about a fifth of it but it keeps me thinking I'm in the know. Thanks

Art | Feb 03, 2007 | 7:58AM

So which is it? ITSP or ISTP? and WYSIWYB or WYBIWYG? Picayune, I know.

Steve | Feb 03, 2007 | 7:59AM

Here we go again. "It's quite simple really... all you have to do is install our protocol on all your routers, switches, DSLAMs, CMTSs, and clients (and their routers, switches, VPN devices, etc), and you'll speed up the Internet!" We can't even get any major player to implement IPv6, and it is already supported by most devices out there, just turn it on.

If only we'd all got ISDN (backed up with ATM), we'd not be in this mess now. I say we tear the TCP/IP stuff out and go back... to the future!

Eric | Feb 03, 2007 | 9:13AM

Ummm... first it was images in HTML that were going to kill the Internet. Then it was the verbosity of XML. Now it is video.

Maybe the Web should just die and get it over with so we can all go watch Toyotas race at NASCAR. :-)

len | Feb 03, 2007 | 11:00AM

Joe - you were able to attain that speed because while you were downloading, enough of your neighbors were idle to free up the bandwidth required to do so. Web surfing typically generates 'bursty' traffic and that's why the 10:1 ratio that Bob talks about has worked until now. Bob is saying that watching Internet video generates constant, high demand, while ISP's have based their business models on satisfying sporadic requests.

Try re-reading the article - it should help you understand better.

allen | Feb 03, 2007 | 12:50PM

It isn't that P2P is really so much cheaper, it's just that a lot of the costs are being borne by the ISP without specific compensation.

I think that's BS - the "P"'s in P2P are paying for their connections. There isn't any unbilled traffic on the ISP's network, everyone pays for their pipe and they're using it. Why should the ISP have "specific compensation" for different types of traffic? They don't bill separately for web traffic vs email etc, just for the connection and a monthly bandwidth ceiling (depending on the package).

Sri | Feb 03, 2007 | 12:57PM

The reason we keep hearing about the death that never happens is that market forces prevent it from ever happening.

When the net slows down, the ISPs step up and do something about it, for more money. That is why I am now paying $60 per month for Internet access, when I used to pay $14.95 for a dedicated phone line and 60kbit access.

If video slows the net, ISPs will find a way to guarantee reasonable service at a higher price; as a competitive move to gain the unhappy customers of their competitors. But the key is, this isn't going to happen a priori; you need a critical mass of unhappy customers to capture, and they aren't there at the moment.

So wait until video DOES overwhelm the net, and about the time we find it irritating someone will start providing noticeably greater bandwidth at an unconscionably high margin, and those that care more about convenience and performance than price will shift over.

Tony Castaldo | Feb 03, 2007 | 1:05PM

Joe - I assume you're using a mozilla based product to download. While you were looking at the 'save as' box, it was already downloading in the background, but didn't start timing the download until you hit yes. at that point, the download speed was the ratio of the amount already downloaded by 0 seconds. I would wager that the 500Kb was either not the highest you saw, or that it dropped in speed rapidly.

robert | Feb 03, 2007 | 1:35PM

You wrote: "Nobody really understands the math of TCP/IP well enough to predict traffic glitches or know how to avoid them, but the folks at NuMetra know some about how to exploit them. NuMetra's TCP Mustang code can be used to effectively game TCP, creating a clear path for packets to follow. The problem is that TCP Mustang's success comes at the expense of all other traffic, which can be severely detained as a result. Slow things down enough and packets have to be retransmitted, grinding the net to a halt yet again."
Just one question. Is this what has happened to Season 2 of NerdTV? It's being held up in TCP/IP limbo land by TCP Mustang?

drewby | Feb 03, 2007 | 2:27PM


"time users will be bale to compare ISPs based on performance, not claims or specifications."

bale -> able

Lucks | Feb 03, 2007 | 4:45PM

To summarize:

(1) I, Cringely, don't really have a clue how the internet works.
(2) I, Cringely, cannot see that none of this is a real problem if you are *downloading* material as opposed to BS concepts like streaming material.
(3) The above two of which mean that I am a perfect shill for people selling stupid ideas like QOS, an idea the internet has not needed for thirty years and does not need today.

Ten years ago parasites like real got rich by selling companies on streaming "technology", a solution in search of a problem that made the lives of the server end worse (they had to pay real and needed expensive servers), and the lives of the public worse (they couldn't just download a file and use it on, say, a portable device; and the viewing/listening experience was awful if the server or internet connection were overloaded).

At much the same time, we had idiots from the telecom world trying to get us all to run the internet using ATM, virtual circuits, and similar stupid ideas from the world of 1900.

Now we have a new generation of parasites selling us a new version of this BS - an expensive solution to a non-existent problem that will make us basically worse off. And Cringely is happily telling us what fantastic people they are, justified by "explanations" of the internet that make no sense whatsoever.

On reading this material the impression I get is that this "crisis" he talks about has to do with the pricing structure of the agreements between ISPs; a purely social issue that can be modified by human beings. But the material is presented as though this "crisis" is technical, using a mishmash of vocabulary and ideas from the technology of networks. The result is incomprehensible.

Maynard Handley | Feb 03, 2007 | 8:30PM

Bob, NuMetra's got some interesting ideas and a real steep hill to climb when it comes to getting cooperation at all the critical points in the stream's path. But what NuMetra really needs to worry about is linking to the web pages of their strategic funding plan's source, VentureBank LLC, and their angel funding source, Pivotol Ventures, LLC from their investor's page. The latter's web site is a typical empty brochure page. But the former's web site looks like something I would expect to see on MySpace with all the pictures of parties and horses and ponies.

yDNA | Feb 03, 2007 | 8:37PM


I've been reading bob for over (I think) 10 years now, and wrote about the google factor (See above)
or here my synthesis of Bob's GOOGLE stories!

And this is from last year!

I don't think things have changed, and short of the Verizon FIOS to my neighborhood, I think I have very few LONG TERM options.


mark | Feb 03, 2007 | 9:55PM

Well im in good company! Actually my employer, Omneon, makes video servers...web enabled... and has been suppling broadcasters with servers, storage, software. In fact our customers are world wide. All this talk has yet to mention the need of HD MPEG4 encoding/decoding. CISCO and many others dont even come to handling video.

Ace | Feb 03, 2007 | 10:52PM

Well im in good company! Actually my employer, Omneon, makes video servers...web enabled... and has been suppling broadcasters with servers, storage, software. In fact our customers are world wide. All this talk has yet to mention the need of HD MPEG4 encoding/decoding. CISCO and many others dont even come close to the need of hardware handling video compression on a multi-channel studio production.

Ace | Feb 03, 2007 | 10:54PM

Forgot to mention. The Europeans and Asians are already far ahead of US companies on this. Its true folks! Ever go to Internet Broadcasters Convention in Amsterdam?

Ace | Feb 03, 2007 | 10:57PM

You know what's the scariest part of this story? The idea that people have the time resource to download & watch all these movies. Maybe the children of the six/seven figure parents do have that time, but the parents do not. They are too busy earning the money! (so their children can waste it away on videos they do not need). Secondly, the worse idea ever is to say that we need QOS. Bah! What Internet users need is a way to rid their inboxes of spam & forego all the advertisements. With the time/bandwith savings gained from getting rid of all the spam and ads we never read, we will be more than glad to pay for and download a few videos.
Not many, but a few.
So don't build those pipes too big! Free media comes with a price. Luckily the air we breathe and the water we drink doesn't just yet. I'm sure some megasuperhugecorporation has a solution for THAT as well!

Josh Gumbert | Feb 04, 2007 | 4:40AM

I'm sorry that the responses on this amazing technology gets off course to focus on Venture Bank and my addiction to horses..Just think " Mustang TCP " !! I do have a My Space page because Numetra supports the vastness of broadband needs their audience demands. I hope this comment isn't from someone I didn't want to date but I support Numetra and have been down this path before in emerging broadband technologies like Microcast Network. Giddy Up...Cheray

Cheray | Feb 04, 2007 | 2:57PM

Here fans is my My Space page so you can understand how streaming media is integrated with the masses and advertising. I worry about the investors who don't understand these portals.. I love my Mustangs..Giddy Up !!

Cheray Unman | Feb 04, 2007 | 3:01PM

This is a great idea, but because it requires a wholesale change to everything, I can't see it happening. If North American ISPs won't upgrade to IPV6, which addresses many of these issues, what possible viable incentive could they have to adopt this other technology?

Too many of these technologies require end-to-end buy in, and so the benefits don't make it to the end user until everything is replaced. Call me cynical, but I just don't see it happening.

Scott | Feb 04, 2007 | 5:03PM

Your discussion appears to miss a few points on how ISPs (particularly, the DSL services of the large telcos) work. The bandwidth constraint at the moment is not at the interconnection with the broader Internet, but within their own networks. It is this "internal" bandwidth demand that causes them problems, not because they need infrastructure (they already have the fiber in place), but because most of the existing DSL equipment was built on ATM (with IP merely as an overlay). It is complex and inefficient and expensive to upgrade.

The dedicated line (where they get there marketed speeds from) is a pipe to nowhere -- it is the speed between the modem in your home and the DSLAM in the central office (or a neighborhood node for FTTN services like AT&T's). From that point on, you are being sold a shared, oversubscribed connection (and yes, assuming it is 20:1 is being generous in most cases). This probably, but not necessarily, applies to the service MegaPath is buying from AT&T and reselling to you.

Going to 1:1 provisioning would be excessive because there are no circumstances where people need constant access to such speeds (as people have pointed out, trying to stream mass media is just foolish when download models are much more efficient and better for the user in every instance but a small set of live events).

In addition, the great demon of bandwidth usage -- bittorrent -- actually makes the network more efficient. It shares the files across multiple peers, reducing the demand on any one link, and is designed to run in the background when machines re otherwise unused. While the total usage is very high, it serves in many cases to smooth usage throughout the day and therefore more efficiently use the available capacity.

Fearmongering aside, there have been predictions of the Internet collapsing for a decade. Bob Metcalfe predicted the Internet's collapse in 1996 (and famously ate his words -- literally -- when it didn't occur).

If you want to get NerdTV out there, seed a bittorrent and let it run. And, by the way, this is not secretly subsidized by the ISPs. It is paid for by their customers.

Bruce | Feb 04, 2007 | 5:28PM

I see two problems.

First, the idea that a bunch of geeks are going to convince every ISP and device and software manufacturer to switch from TCP/IP to their new thingy flies in the face of human nature. It simply isn't going to happen.

Second, as a result of the first, it isn't going to happen soon enough. It will be ten or fifteen years before this new tech can replace existing tech. Meanwhile, the bandwidth crunch continues unabated.

If your Google scheme is for real, Google will beat that approach pretty quickly. If fact, I wouldn't be surprised if Google bought the company and tried to use the tech inside its network of data centers and fat pipes, if that's feasible at all.

So my suspicion is: fergeddaboutit.

As for ISPs oversubscribing, it's worse than that. SBC (now AT&T) has been deliberately selling people 3Mbps DSL access that they KNOW the customer CANNOT USE because the customer is too far away from the CO to get that speed. This happened to me last year - I upgraded to 3Mbps because it would have cost me $20 less a month than the original 1.5Mbps contract. After a bit, I did a speed test and discovered I was only getting ten percent more speed than at 1.5. I called SBC, got referred to their provisioner, where the techs said there was no way I could get 3Mbps without the risk of my line progressively getting worse until it went down and stayed down. I was 12,000 feet from the CO, and 3Mbps is only available at 10,000 feet or less.

This is a class action lawsuit waiting to happen, because essentially AT&T is selling a fraudulent service - a 3Mbps service they CANNOT deliver. And they know it based on your distance from the CO.

Consciously fraudulent.

Richard Steven Hack | Feb 04, 2007 | 6:12PM

" - 8k - Cached - Similar pages"

Bob, I think you're half-right but not all the way there. Google may be slapping datacentres down all over the USA but I'd wager that the breakthrough they're actually in the process of deploying is that damn fine cache coupled with (speculation alert!) a way of making streams cacheable. Put that as close to the edge as you can, connect it to the ISP's datacentres (for a small fee) and you've got a great service.

Angus Wood | Feb 05, 2007 | 9:23AM

My question is why IPTV at all? Yes, narrowcasting has some value along with the potential for another order of magitude change in viewer choice. But it's a lousy solution for the majority of the viewer market.

Let's look at it another way. For traditional broadcast methods (OTA, satellite, cable) the incremental cost of each viewer is basically zero. But for IPTV there's both a head-end and a viewer bandwidth cost since traditional IP is point to point. Peer-to-peer can shift that around somewhat but P2P doesn't fit the narrowcasting hyper-choice model since that assumes (especially for an ISP-first scheme) I have the same interests as my neighbour.

EricBall | Feb 05, 2007 | 11:21AM

This is just the type of approach by ISP's and telcos that is going to give broadband over powerline a leg up, especially in that critical last mile to the home.

DwayneH | Feb 05, 2007 | 12:02PM

This is just the type of approach by ISP's and telcos that is going to give broadband over powerline a leg up, especially in that critical last mile to the home.

DH | Feb 05, 2007 | 12:02PM


Ashley | Feb 05, 2007 | 1:33PM

If you look at the tivo experience, basically downloading programs in advance of watching them, I don't see the bandwidth problem.

Other than breaking news, I do not see a problem with watching "delayed" programming. If we had a smart tvio type of box that was able to share programming p2p amongst other isp clients I think most of the issues you are bringing up would disappear.

DanB | Feb 05, 2007 | 2:39PM

A national fttp rollout is whats actually needed, the rest is just a waste of time if we still deal with copper on the last mile.

Regards from the UK

zemadeiran | Feb 05, 2007 | 2:45PM

WYBIWYG: That kind of flies in the face of HillaryCare, doesn't it?

Jeff H | Feb 05, 2007 | 3:14PM

I'm going charge right in here, without reading any other comments, and challenge some of your assumptions.

First, you wrote:
"It isn't that P2P is really so much cheaper, it's just that a lot of the costs are being borne by the ISP without specific compensation."

Um, not to quibble (well, yeah, to quibble) but the ISP does in fact get specific compensation; subscriber fees.

I pay my ISP specifically to gain access to CONTENT THAT I WANT. Now, I'm beginning to join the rush to Internet 'video', and I regularly want content such as DVD images of various legal software packages. So my habits are changing. I'm demanding larger amounts of data, and more timely (faster) delivery.

My ISP is going to have to accomodate that, or I'll move to another, or a new ISP will find a way to deliver. How would an ISP deliver greater amounts of data faster?


Examples of how caching might work? I get Linux distributions as DVD images now. How about caching those locally. Hey, there are only so many distros out there, and storage is realtively cheap. Then again, maybe this is being done.

YouTube? Well, I bet caching doesn't solve the ad problem. And I wonder, if YouTube-like video is the problem, we are nto far away from ads being as 'big' as the 'content'. Darn, caching is out.

But another thing bothers me. Point #2:

Internet video is not much different than cable TV. Much of it is so narrowcasted that it's like those wacky channels you can find if you look hard enough.

Imagine how my wife will react when I try to explain how our Internet bill is going up $20-$40 per month because of video. Why do we have Internet and Cable?

Good question.

Sooner or later, as video as a service of the ISP gets more important, I suspect ISPs will not care about bandwidth. I get DLS now, but sooner or later my ISP will want to sell me TV over DSP. Why would they care about me getting true Internet TV? In fact, they already sell 'digital' telephone service over my DSL line. Why would they care about Vonage or Skype? Just competitors. If I go to cable for my ISP, do they care about 'Internet TV'. Not if it isn't theirs...

And if I whine, I wonder. Do they listen? Where do I go?

But I'm already paying them for access to the content I WANT. They are in fact getting paid. It's just that I want a lot more than they thought I wanted. Kinda like my landlord looking at me, and thinking, "Sheesh. He rents a two-bedroom unit, and he wants a courtyard, garden, and outdoor theatre too? For the same money? Ha! HA HA!

So charge me more. I can get DSL for $25-$50/month. Cable for $40-$90. Would I pay $75/mo for functional Internet that delivered my video and DVD images?

Oh, well you did make the ultimate point. My ISP says I get 7MB download, the maximum speed I can get. I can't get it to reach that speed by any measure. Time to change.

Will the other ISP I want do any better? Good question. Worse? Doubtful. But, importantly, will any ISP do better? Nope.

In the end, this is a lot about the technology. I suspect there is no good solution, save for someone out there to make a major change, an breakthrough. Google might do it. Maybe NuMetra.

Maybe not.

But don't tell me the ISP isn't being specifically compensated. They are just priced wrong.


Rick | Feb 05, 2007 | 3:22PM

Maybe I just don't understand the intricacies of networking "solutions", but the nuMetra site seems completely devoid of content. If I hadn't gotten the link from I,Cringely I wouldn't have given nuMetra a second thought, but even with the reference I'm suspicious. They claim to use an open source model, but I don't see one link to their source trees. In fact I don't see anything on the page but jargon and flash animation; the page reads like a Dilbert comic. The credentials they list on their site are intimidating, but until they give some details on how their model works(which should be forthcoming if their product is open-source), there doesn't seem to be any reason to think this is anything other than silicon valley snake oil.

Patrick | Feb 05, 2007 | 4:45PM

"ISTP protocol, which is compatible with TCP/IP but must be supported by devices end to end". If end to end supported is needed then I see almost no way it will be successful, at least not within 10 years. Any technologies that need upgrade across company border failed. For example, ATM, MPLS, both never be able cross company boundary although they might be successful within one organization.

JeffreyH | Feb 05, 2007 | 7:10PM

I think that that the NuMetra approach doesn't sound like it is a whole lot different, in the end, what the non-Net Neutrality approach that some ISP's have thrown out. You apply QoS, by some mechanism, and then the priority traffic gets where it needs to go and the non-priority traffic suffers. If the people applying QoS watch the statistics, they know when they need to add bandwidth to feed either queue. People pay more to access the priority queue, but then I feel like the slow lane suffers in a push to move people onto the toll road. It all ends up sounding like the push to extract the type of per-minute revenues that they used to get so much of before.

Bill F | Feb 05, 2007 | 7:20PM

Something sounds fishy. Not saying it can't be true but something sounds off. If this requires system wide adoption, how does this system look when compared to IPV6?

rhodope | Feb 05, 2007 | 10:19PM

“If only everyone would adopt it”

Somewhere somebody has gotta be tracking all of these proposals to replace the old plumbing with new plumbing, ‘cause it’ll solve some specific current need or another. Of course it never does because the adoption cost is just too high and each time clever folks have figured out how wring enough of the benefit out of good old UDP/TCP-IP to keep ‘em.

So interesting stuff from NuMetra but they’ve got a snowball’s chance. We’ll see ISP-endorsed multicast, end-to-end supported QOS, and IPv6 broadly implemented in North America before we’ll see a it’s-like-TCP/IP-but-breaks-it NuMetra-packet-technology adopted.

maggard | Feb 06, 2007 | 2:14AM

Maybe Mediacom could be that national ISP o adopt this and clean up. My service cost more than DSL, but it has been quite good over the 4 years that I've had it.

As for internet video, though, I'm not impressed. YouTube, for example, contains nothing of value that I could find. Sure, some of the things are humorous (Old Lady and Slingshot), but usually at the expense of someone else. The site should be called for "Lowest Common Denominator."

Speaking of which, maybe NuMetra should approach the pornsite operators, the industry known for adopting new internet technology first.

Tom | Feb 06, 2007 | 12:08PM

Get a company like Tivo to build a IPV6 P2P appliance which is essentially a 1 Terabyte File Server. Operate it over cable or FIOS networks, charge $400 for the box and $50 per year subscription. Make an agreement with all content providers for measured cuts of the files and give the user unlimited peer sharing.

Cobb | Feb 06, 2007 | 12:26PM

Something seems strange about Venture Bank. Lots of jargon and buzzwords.

And this: "The second spin off will be Venture Bank CHina headed by Dr. Charlie Chan."

Um, riiight.

BarnabyWilde | Feb 06, 2007 | 6:46PM

NuMetra's solution sounds like a form of "quality of service" which prioritizes streaming video. If you are going to stream video you also need to consider peak load times. The ISP and backbone provider would need to make sure that they have sufficient bandwidth to the source or possibly the closest Google site at specific times of the day.
If on the other hand, you plan to distribute video with BitTorrent my guess is that your customer will be happy with next day delivery for high quality video. Blockbuster and Netflix take longer than that. The other thing to consider is that "quality of service" solutions decrease the amount of bandwidth delivered to everybody except the one or two preferred services. You could provide good service for streaming customers at the price of slowing down BitTorrent.

Allen Cole | Feb 07, 2007 | 8:46AM

If the ISPs are all worried about internet video bandwidth problems, they should all have turned multicast on by now. They haven't.

leeg | Feb 07, 2007 | 12:20PM

The battle over net neutrality assumes a model where broadband providers could charge fees that segregated deep-pocket corporations from the Little Guy, thus destroying the egalitarian nature of the internet as we know it. But what about segregation not by arbitrary ability to pay, but by use? If video downloading threatens to make ISPs unprofitable, why not enable them to filter video content over a set file size, and charge extra for it? I can't park my semi truck in the garage on the corner, but that doesn't mean the guy running it is disriminating against me in favor of my richer neighbor.

DC | Feb 07, 2007 | 1:12PM

Y'all have to stop thinking we're still in the 20th century. The way I see the future is Wi-Max mesh networking with distributed serving. There will be no last mile wire, and most of the rest of the wire will go away. Hardware is too expensive to install and maintain, so you eliminate as much as possible, and/or get your customer to do it for you.

The distributed serving is the easy part, just think about everyone having a bit torrent client installed & sharing their video (TV & movies) & audio (radio) downloads, plus cashed web pages and anything else they want.

Wi-Max mesh networking: Take my own metro area as an example, Seattle. Urban area around 600 square miles. I could set up a Wi-Max network by putting up eight to ten antennas. Sell the customers a network adapter & software that lets them connect to the network & stores all their downloads on an area of their hard disk. When someone close to them wants to download the same content, instead of downloading it from the provider, they download it from the first user. Data is only being sent over the fiber once, afetr that it's provided localy.

If the customer is to far from the access point for a good signal, it’s relayed through other customers network adapters. This could also be accomplished by selling a set-top network adapter with a hard disk inside. Always on for the network to use even when the customer’s computer is off. Plus this would eliminate some security concerns for customers.

Lots of content providers and lawyers will be crying fowl, but the days of making large amounts of money from selling data are almost over. Small profits, sure, just not money to keep the stockholders happy.

There are already wireless networks in rural areas where non-commercial use is free. The days of large ISPs making money are almost over.

But wait, there’s more! What happens when people realize they can upload their own created content? The producer, distributor and user all become one. Large media companies will have a tough time competing with free or nearly free content.

Wi-Max phones will put cell phone companies out of business. Also, imagine shooting video with your phone anywhere in the world, and instead of storing in on the phone, it’s sent to your home server as you shoot.

This is the quick and dirty version, I’ve left out things like: How do content providers get paid at all? Short answer: It’s possible, but you will be competing with millions of other content providers, so you won’t be making much.

Dave K | Feb 07, 2007 | 6:34PM

TCP/IP via ATM is so 1990's...

Today data networks are large large frame size ethernet based. And with voice going VOIP, ATM is a dinosaur, although far from extinct (yet).

Tom Hedges | Feb 08, 2007 | 4:09AM

Somebody is going to have to explain how exactly this is different from in another comment or something.

Devang | Feb 09, 2007 | 8:12AM

The idea sounds interesting in principle but to get all of these ISPs to adopt this is simply never going to happen. What is really needed is for the standards bodies and equipment manufactures to implement a viable next generation QoS scheme that fits with the current rich media reality of the Internet. An ATM type solution for the IP world anybody?

dogsta | Feb 13, 2007 | 2:29AM

I'm shocked, I mean *shocked* that Bob has joined the 'ISPs need compensation' bandwagon. Why am I paying then $35 a month if I can't use the pipe any way I see fit? A couple of people have already commented on that, but it is worth repeating.

Abhijeet | Feb 17, 2007 | 9:20PM

I would love to see the papers on some of these technologies. I can't fine any references to TCP Mustang or ISTP at IETF, and NuMetra's site is shockingly sparse on specific information about how these technologies actually work.

Eugene | Feb 18, 2007 | 12:06PM

Great insight...

Although, if you look carefully, nuMetra isn't really that new an idea. The academic community has been focused on fixing TCP for the better part of the past several years (see TCP Fast (CalTech), and TCP Africa (Rice)). The real innovation here is in their economics. If you look at their association,, it looks like the real objective is to get the "old Big Media" to partner with the ISPs to only stream authorized content over their "diamond lane" service. Bill Smith, the CTO at BellSouth, proposed something similar about a year ago. nuMetra's approach appears to be one heck of a lot more elegant and scalable...
It looks like old Media needs to get a clue and partner with them or perish to the likes of Google.

Jack | Feb 18, 2007 | 10:35PM