Archive for January, 2007

The Four P2P Models

Every time I pull up my Google News Alerts on ‘P2P’, there’s yet another P2P start-up getting funded, and yet another big company making some kind of partnership in P2P, and of course some kind of new-fangled P2P killer-app that will “change the world”. Isn’t like every cool app that comes out now powered by P2P? Isn’t Bittorrent the company taking over?. .. oh, maybe that’s bittorrent the technology. . what’s the difference again?? I saw one article that said P2P filesharing was going to help find the cure for cancer ,. . .oh wait, that was me 6½ years ago ;)

There are a few major things going on in the P2P industry, but man, there is A TON of vapor in the P2P industry press, and with it even more confusion. Here’s my attempt at breaking the P2P tech space down into its parts so we all can be a little more informed, and maybe help the pundits and journalists more easily cover the chaos.

P2P Business Model #1 – Illicit Distribution

To date, the only P2P companies that have become popular got their millions of clients distributed (or their 10’s of millions of daily page views) by powering mass piracy. There’s no better way to get your client out there or to get your site in the top 1000 than to help people get their favorite TV, movies, video, music for free. Think Napster, Scour, Kazaa, Azureus, uTorrent,, These guys are easy to recognize, they are providing you with the access and downloads to content you want for free. Almost exclusively, these applications and sites are powered by ads and adware.

P2P Business Model #2 – Content Portal

Conversation amongst two Silicon Valley Entrepreneurs:

Entrepreneur #1: Hey, YouTube is pretty cool
Entrepreneur #2: But the quality sux
Entrepreneur #1: True, but I can get anything I want with a click of a button
Entrepreneur #2: Maybe, but all the clips are super short, like 3 minutes long
Entrepreneur #1: True
Entrepreneur #2: Hey, bittorrent is cool
Entrepreneur #1: Yeah, but the downloads are slow, unreliable, and how do you make any money delivering illegal stuff for free
Entrepreneur #2: Yeah, but iTunes is pretty cool, pretty decent quality and they sell stuff for $$
Entrepreneur #1: Maybe, but the bandwidth bills must be killing them, there’s no way they can scale that for video
Entrepreneur #2: Hey, what about YouTube meets iTunes, with a little high-performance bittorrent underneath the hood??
Entrepreneur #1: Muahahahah. . . .we’re going to be RICH!!

The P2P content portal is the flavor of the month. Bittorrent Inc. is pitching their own content portal where they are PAYING the studios for licenses to their content . Look at Joost . . . the Kazaa/Skype guys are at it again, redefining TV by trying to create an online Cable Co. of sorts powered by P2P. . .and again, Joost will need to pay content owners for access to licenses and content, and then will create a branded channel/player where they sell the content and ads around it. Look at Azureus , Inc. (primarily an illicit distribution background) with their legit site Zudeo paying BBC for access to content. . .

The keys to success are going to be licensing a broad library of content while marketing to and acquiring a consumer base. They’re going to have to compete head to head with Yahoo!, Google, AOL, MSN, and iTunes to do this. That’s a tough fight. In addition Azureus, and Bittorrent Inc. are going to have to turn illicit piracy products into legitimate commerce brands. Didn’t work so well for the folks that have tried before: Napster, and Scour come to mind. They’re going to need to deliver the content for cheap, and P2P is necessary, but P2P won’t be a strategic part of the business if/when they’re successful getting the marketing and licensing right (P2P won’t matter until they have lots of users).

They’ll ultimately compete with the other portals, and ultimately want the best technology out there to make the economics work, and it won’t matter whether it’s their P2P tech or not, the same way most Internet co’s lease bandwidth instead of laying their own fiber (who wants build the power plant?). Regardless, there have been couple of big gambles with 10’s of millions of dollars put into this space, and only time will tell if they play out.

P2P Business Model #3 — Infrastructure

Websites pay BILLIONS for bandwidth every year. We’re entering the video era of the Net, television and video will be delivered over IP networks. . . bandwidth expenses are going quickly to 10x what they are now. Why should people pay for bandwidth?? Can’t data transfers of popular content be localized to ISPs? Commercial-grade P2P infrastructure definitely makes better use of scarce existing bandwidth on the Internet.

Furthermore, the only difference between a webpage with a bunch of links to content, and a P2P app is that the P2P app delivers the same links with no bandwidth costs. Bottom line is that the only unique technological benefit to P2P is the substantial efficiencies of data transfers. Sounds like a no-brainer to become the new architecture and infrastructure for the Internet?? The main challenge for P2P Infrastructure companies (Red Swoosh included) is their ability to get their clients distributed in a massive fashion. Once the clients are out there in a major way, the rest will be a slam dunk. If 20% of the users have the Red Swoosh client, then any website can save 20% on bandwidth using Swoosh p2p infrastructure.

P2P Business Model #4 – Device licensing

Any of you guys seen DivX’s business model? DivX has been the video codec of choice for p2p pirates since the last decade. Because pirates want to play their “acquired” content in their living rooms on their big tv’s and with their great sound system, DivX is making brisk business by getting paid $1 for every box deployed that is “certified” DivX.

Bittorrent is now trying the same thing, except they are trying to license the Bittorrent name to boxes that want do download torrent files . Granted bittorrent was initially open-sourced under the MIT License , so it’s unclear if it’s even possible to enforce that trademark or if it’s the new Kleenex, but we’ll have to watch and see. Regardless, there’s lots of content going to boxes, and P2P will most definitely make it there.but we’ll have to watch and see. There’s lots of content going to boxes, and P2P will most definitely make it there.


So there’s a lot of activity in the P2P space, but in the chaos there is so much more more hype. Good news is that it can all be broken down into these four basic business models. Journalists should use these to parse through noise and find the goods. Who is blowing smoke, and who's got something real that they bring to the table.
There's still a lot of work before the land is grabbed, and the winners chosen. This is the map for the P2P business battles to come.

The Web Responds to Our Steve Jobs Letter

Ever since we posted the letter Johnny Cakes wrote to Steve Jobs, we've gotten some great perspectives coming from around the blogosphere.

One well-thought post written by Mike from Buzzword Compliant discusses some of our major challenges.

Some of them really hit home with us - we need to step our game up! But at the same time, we want to clear up some misperceptions about how the swoosh service works.

For instance,

"Hurdles to content drive away users. As anyone who has designed online systems knows, every single additional page load you force users to through will drive some of them away. Why do you think Amazon decided to patent the “1-Click checkout” and not the “12-Step Checkout”

We're working on improving our usability right now for both downloaders and distributors. But truth be told, swooshing is the easiest media delivery system we've ever used, and we don't just say that because we made it.

Don't believe us? Check out this tutorial from the Learn More section of our site.

Swooshing is really that easy, and it's getting even easier as we're beginning to integrate directly into popular file-hosting services. Our goal is to be as transparent as possible. We want to keep the focus on the content while increasing download speeds and lowering bandwidth costs.

If you're an end-user and bandwidth costs aren't a concern to you, remember that with up to 99% percent less bandwidth costs, your download servers won't have to overwhelm you with pop-up ads, donation requests, and other ways of paying the (avoidable) bills.

"Metcalfe’s law: networks increase in value proportional to the number of users in the network. I’ve never come across a link that’s asked me to load up RedSwoosh. "

Redswoosh starts really heating up with as few as ten users. It's easy to compare RS to all other P2P apps, but its our sheer efficiency, even in small numbers, which makes us distinct.

"It doesn’t work on the Mac. I’m hoping that they’re doing browser sniffing of some sort and redirecting links from Mac users so they aren’t totally out of luck."

Coming soon! Really really.

"RedSwoosh is positioning themselves against Akamai, which is most likely about the most expensive hosting on the planet, but I think they’re missing the true game changer in the field which is Amazon’s S3. While S3 is mostly referred to as a way to backup your files, it’s actually a massive file hosting operation as well. It’s debut has forced up the amount of bandwidth associated with many other web hosting services; blunting RedSwoosh’s value proposition. Nevermind the fact you can take advantage of Amazon’s huge uptime and bandwidth to directly distribute files for 15 cents a Gigabyte."

We're with you on that one. We're glad that Amazon is opening up S3 to developers, just like we do. We're hoping for some innotative crossovers that combine the best of our P2P with massive utility computing offered by S3.

And keep in mind that S3 is hyped for its disruptive capability in storage - that's not our bag baby. Amazon's bandwidth still costs twenty-cents a gigabyte, and for popular bandwidth-intensive services that use S3, the cost will make a difference.

With even a few users, Redswoosh brings down bandwidth costs to nothing - "free bucks and thrifty-free cents", as Johnny Cakes likes to say.
Mike also suggests that we "license RS technology to power individual applications. For example, WoW distributes their massive patches and updates via P2P."

That's a great idea, Mike. We'll be keeping an eye on your blog in case you have any others.

Until then, it's time to start rolling out more changes. If you have any suggestions for how we can improve our service, drop them in the comments.

We'll be watching, and featuring good suggestions regularly on the blog.


iWaste: How Apple Can Save 15 Million Dollars

Johnny Cakes      
Dear Steve,
        Fifteen million dollars is a lot of money.   It looks like this:   15 million dollars     You make fifteen million dollars every time
  • 100,000 iPhones are sold.
  • 50,000 fourth-generation Ipods are sold.
  • You, with your personal networth of 4.9 billion dollars, sit on your ass, eating seaweed salad and planning out the rest of our lives.
With fifteen million dollars, you could buy:
  • Sealand.
  • 3,750 more Segways for Woz, and a lifetime supply of Nathan’s famous hot dogs.
  • Exactly one talking car for the Pixar campus parking lot.
  • (True, only one. But then again, it *is* a talking car, and they’re even more rare than electric cars!)
Fiften million dollars also buys about 200 terabytes of bandwidth from Akamai. And because of iWaste, those terabytes of bandwidth are being purchased with money that could buy iGet iPhones for 100,000 iPeople. But, wait! You need to spend lots of money on bandwidth, for “peak traffic”. What if millions across the globe decide to download Mean Girls as a treatise, a pledge in support of Lohanesque pacifism? What if??? So you keep the server farms iWasting kilowatts like candy. Well, allow me to iRetort. Time to bust out the abacus, Mr. Jobs. Let’s go over some iTunes Store math: According to the BBC, Apple has sold “1.5 billion songs, and tens of millions of TV shows and movies.” The average film is about two gigabytes. The average TV show is about 700 megabytes, or 70% of a gigabyte. The average song is about 5 megabytes. So let’s go over how many terabytes are needed to transfer iTunes material.
  • Music: one and a half billion songs x 5 MB = 7,500 Terabytes
  • TV Shows: 25 million TV shows x .7 GB = 17,500 Terabytes
  • Movies: 25 million movies x 2 GB = 50 million GB, or 50,000 Terabytes
  • 75,000 terabytes = A #*#&[email protected] TON OF BANDWIDTH. Apple’s bandwidth, using pricely Akamai servers, costs about $200 per terabyte. A lot of that money is wasted, because of the need to always be ready for the mythic International Downloading Sesh. Anyways,
    • Seventy-five thousand terabytes x $200 = 15,000,000
    Fifteen million dollars. For fifteen million dollars, you could buy the world a Coke. Or at the very least, a Shasta grape-flavored soda. So how can Apple’s fifteen million dollars be saved? Or better yet, how can a hundred thousand iPeeps get their new iPhones for free bucks? Using P2P distribution, of course. The iTunes store now uses a client-server setup. The client (your computer) asks the server for the new episode of Battlestar Galactica (Not Starbuck! Anyone but Starbuck!!!) or the new Amerie song or whatever whatever. And everyone downloads the same files from the same electricity-hungry servers. It’s not just boring. It’s wasteful. With only one person in the P2P network, downloads would be as fast as the client-server setup. In fact, the client-server setup is just a really weaksauce P2P network with you, yourself, and you. As a P2P network grows in size, the download speed increases by magnitudes, and bandwidth is used as efficiently as possible. And some of the newer P2P systems like the Swoosh have all kinds of crazy as-yet unexplored features like prediscovery and a versatile API. I insist, Mr. Jobs, that you don’t sleep on it. Save your cash. Save our strained electricity grids. Lindsay would want it that way. -Johnny Cakes Johnny Cakes is the newest member of the Swoosh entourage. He is our Downloader in Residence, and Ranter in Chief.