Blog Post

Flash P2P: Now That's Disruptive

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Don’t blame me for getting caught up in the whole hoopla around media-buying-media…we media types are known for being narcissists. Blame me for not being able to blog about the new beta of Adobe Flash Player 10, which has built-in P2P features and is able to save files to the local drives. I was reminded by Hank Williams about the new release, and its big impact on the world of video in particular and other web apps in general.As some of you might remember, I wrote about Adobe’s P2P ambitions that revolved around buying a company called Amicima.

Through LinkedIn, we were able to find that amicima co-founder Mathew Kaufman has been working as Senior Computer Scientist for Adobe since October 2006. His co-founder, Michael Thornburgh, is also said to be at Adobe. Both of them have vast experience in networking and P2P technologies. The two of them worked at Tycho Networks, and later at, after that company acquired Tycho.

I have been following this closely, and my sources say that this is a solid technology with the potential to seriously disrupt the CDN market, especially those companies that rely on clients. I wonder, for example, what will happen to RedSwoosh, which is owned by Akamai, or to other, similar providers of P2P-based client services. I think one shouldn’t get caught up in the CDN-killer aspect of this technology.

From what I have learnt, there are some elements of this technology that make it necessary to have a server infrastructure for situations where traversing NAT’s/Firewalls isn’t possible. It also needs a centralized registrar is also needed that maintains the ID’s of all the P2P clients (nodes) connected to a service. In other words, a CDN operator work with Adobe, charging for traffic that goes through their proxies as and when needed by the Flash 10. By the way Adobe has an arrangement with Kontiki, a CDN operator of sorts.

Williams’ post digs deeper into this in a thoughtful, intelligent way. “[I]s the innovation that will be unleashed by making P2P technology an assumed part of the web protocol stack?” he wonders. (I think that’s why it’s important that we start harping about upload speeds on our broadband connections.)

The reason we should pay attention to this product is Adobe’s distribution strength. The company can easily upgrade its Flash clients and instantly become owner of one of the largest P2P services. What that means is that now anyone can contemplate a Joost-like service that works within a browser. Using AIR to extend those P2P abilities to the desktop would be fairly easy as well. Ironically, both Joost and Jaman have spent considerable time, money and attention doing this.

The early version of Flash is rather simple, but it does offer a way to lower bandwidth costs while still delivering high-quality video. In addition, companies like Tokbox (our story) and Woome (NTV story) can add more functionality, such as cheaper, live video-voice service, without spending too much money.

It’s clear that Adobe is not going to become a huge P2P service overnight. But this release does portend to an interesting future.

PS: If anyone wants to share their thoughts, please leave a comment or drop me an email.

27 Responses to “Flash P2P: Now That's Disruptive”

  1. Reading between the lines, the protocol implementing this and the server back-end are going to
    be proprietary. Adobe presumably hope to sell lots of media server licenses.
    It will be interesting to see which of the open source servers re-implements it first.
    In theory it should be Red5, but my money would be on Freeswitch.

  2. I do deeply believe in the power of p2p delivered video – but the problem here seems to be the protocol: rtmfp seems to be a cool streaming implementation, but it’s also restricted to the use of Flash with FMS. No doubt, this team is playing big, but it’s a wallet garden again. We at emundoo try to go the open way, by implementing rfc confomrm rtsp, which in the end will be able to play to any device (we currently support qt and vlc for playing vids in your browser). Another aspect for the success of p2p is the ability to charge viewers for their content – we are talking about security and chargeability here: how do i know, as a content owner, which of my viewing peers gave away how much upload capacity – that’s a point Adobe should try to implement in order to shift their flash player away from “funny-video-youtube-alikes” to serious content delivery. Ironically we thought about that, here at emundoo, some time ago… ;)

  3. Andy Abramson


    As I mentioned to you at breakfast, I installed Tokbox and Meebo. My Mac Book Air lost it’s hearing and its voice with other telephony apps. I ended up having to do a complete hard disk wipe and OS reinstall.

    The same thing happened with a prior Flash based install a few years back and was more easily fixed, but that was Pre Intel chips in Macs.

    So, if I say, I’m not impressed its because of how much time was wasted trying to fix it, then finally fixing the Mac.

  4. @Me
    Quick correction for everyone to my first post on this thread about Henry Sinnreich and P2P SIP. I talked to Henry and did my own outside research, and the SIP work apparently is all going in to Adobe Pacifica.

    All the p2p in Flash player is the Amicima work from Matthew Kaufman and Michael Thornburgh and their very cool RTMFP protocol. Matthew is a longtime participant on the p2p hackers list and on various IETF lists (he was thankfully on my “side” of the infamous “what’s wrong with DNS?” thread linked above — I’m still unclear on how that became the ultimate flame war so quickly). If you have any interest in protocol design, their work is frankly just sick. The Adobe stuff appears to be the latest iteration of their original Secure Media Flow Protocol (SMP) at

    I go into much more detail on all of this at:

    If you’re in the mood for some totally geeked out technical drama, check out the DNS discussion:

  5. Cut from my blog but on this topic..

    Adobe, the Apple killer.

    If you follow my blog, you will notice that I take a lot of effort in looking at Adobe and what they are doing with their flash technology. To help position this post, let me first refer to previous posts in reverse chronological order:

    Flash DRM, a disappointing implementation: where I discuss pitfalls in Adobe’s plans, but for which recent developments appear to have fixed.

    Adobe’s plan for world domination: where I discuss how Adobe may leap over the current trends in digital media like, for example, what Apple/Itunes is doing.

    Just a few days ago, Adobe released FlashPlayer10 BETA preview. Among many improvements, the main ones of note are:

    1. built in P2P networking which could turn Adobe into one of the biggest controlled P2P networks
    in the world over night.
    2. Capability for Flash Player to write files to your Hard drive.

    These two developments are exactly where I explained Adobe should be going with flash in my post “Flash DRM, a disappointing implementation”. Let me explain.

    Adobe Media Player is all well and good, but the requirements to install it, even being so easy, is a hindrance. It is also designed as a portal with ultimate control going to Adobe. (Similar to Itunes and Podcasts).

    The problem I see with all this is that the idea behind Apple and Itunes and similar business models, is that they intends to be a totally dominant in the distribution chain. Trying its best to stop any direct producer to consumer transactions.

    This plan has been working well, so well in fact that the incumbent distribution gatekeepers have been doing there best to erode Apple/Itunes market lead. For example, DRM free music available on Amazon and not on Itunes. NBC shows available on hulu, and other non Itunes services and ignoring/forfeiting some short turn profitability on the content, if on Itunes, in the hope to break up the market into competitive players.

    During this fierce battle for digital media dominance, Adobe has steadily been building tools that will make it all mute. I may not agree with all the pricing structures Adobe has but there is no dough that Adobe has built a platform in which any content producer could build an Itunes like experience either on their cross platform AIR technology, or with this new FlashPlayer10, simply in any Flash capable browser.

    This leads into the likelihood that Adobe is likely to soon be distributed in SetTopBoxes, or embedded in all TV sets sold. Imagine a TV set with a H.264 engine, great interactive capabilities, secure content (DRM capable) and secure P2P capabilities. And top it all of, millions of Web developer/content producers that are trained up in the technology and cheap.

    This is why, in my opinion, Steve refuses to let Flash on the IPhone. It is also why there has been rumors of Apple purchasing Adobe. These titans are headed for a showdown.

    Let me paint the picture here. Imagine you are the owner of a successful content production chain, for example, the CSI franchise or the Law and Order franchise. You most likely are connected to a large traditional content gatekeeper like NBC etc. NBC is now controlling how this content is distributed on TV and Web. Until these tools where available, there was no choice in the matter. You had to use a gatekeeper with connections to online distribution technologies. This has predominantly been Apple. Microsoft is also in the mix but I tend to think they are a bit clueless as they listen to their major clients too much, the big incumbent companies. And we know they have no clue.

    As NBC really controls your content, they ultimately have a lot of power over you. They can, to a degree, make or break you. Apple/Itunes, with its dominance in digital media now has the edge over the traditional gatekeepers. As they have used this power over their content producers, they can see how Apple could use it over them. This is why they are so paranoid. (No one is more paranoid then a thief.)

    Considering all this, I feel that once the tools to allow content producers to distribute to consumers are available at reasonable prices, the dominance of Itunes and similar gatekeeper type models, (Hulu, etc) are likely to become less relevant. What will become more relevant will be social network models that help people discover these islands of content. And to a degree, this is where the traditional players are likely to fall into, as its that or nothing.

  6. @Wes

    There’s no standard definitely of what a “P2P CDN” is. A P2P CDN could be implemented on top of P2P SIP. It could also not. There are many, many ways to implement a P2P CDN.

  7. @Om
    I’ve been a silent reader of your postings on this topic for some time, Om. You’re the one person I’ve seen following it, and I agree it would be an unbelievably disruptive shift if they did it right. Henry is a powerhouse in the SIP community and a very smart guy, so they’ve got the personnel to pull it off. The biggest obstacles are the security and the privacy concerns with allowing incoming connections to the Flash player — turning Flash into a server.

    Henry and I have certainly had our differences at times (see the dizzying “What’s wrong with DNS?” thread at, but he’s a top-notch engineer.

    I’ll let you know if I hear anything more from him or another friend on the Flash team, and I’ll certainly send an e-mail.

    -Adam Fisk

  8. @ Adam Fisk

    I have been on the same trail for a long time, though I am wondering if they will ever get their SIP/VoIP act together. Any updates you can provide are handy. Drop me an email if you want to stay in touch.

  9. Henry Sinnreich from Adobe has been active on the P2P-related IETF working groups for some time. He could never talk about what he was working on, but this certainly seemed like the most likely candidate. Given his participation in the groups, it’s likely this is SIP-based and possibly P2P-SIP based. If it’s using the P2P SIP stuff, then it’s almost entirely decentralized, even using distributed registrars. It’s unlikely Abobe would build that in to Flash given the security concerns, however, so their architecture remains an open question. As you say, NAT/firewall traversal requires an intermediary with a public address, most easily accomplished with central servers. It’s possible Adobe is firing up a server farm of SIP proxies and registrars for this purpose, although that would be at non-negligible expense.

    I e-mailed Henry this morning to see if he can clear any of this up further, so hopefully I’ll know more soon. My work at uses a very similar architecture (you should definitely check it out), and I’ll be blogging about this at some point at

    -Adam Fisk

  10. Thanks Om, once again I get the best info from GigaOm. Other sources that reviewed the new Flash player did not even mention this P2P feature.

  11. prefer not to say

    after reading about flash player 10 P2P technology 2 days ago it kept my mind thinking about how would this affect application development and how it would affect Microsoft’s Silverlight! i think with a built in CDN companies might want to build their applications in flash since now the cost on the cloud could be reduced and companies could offer a richer experience using flash, another thing is that it might push companies to offer a free plan (as an entry plan) for using their applications because having a bigger user base will help those companies pay less for the cloud, now what about Silverlight? does microsoft need to take a similar step and build P2P into Silverlight?!! think yes.
    Adobe really took a smart step doing this.