There needs to be a real distinction made between “P2P technology” and “P2P file-sharing” and the impact they each have on ISP network traffic. There are many different applications of P2P that are often painted with the same broad brush. Commercial P2P traffic is still very low compared to the traffic generated from P2P file-sharing applications. These things are evolving separately, driven by companies with very different objectives and desired outcomes.

Since Nov 2007 (the date of your post) a lot has happened on the commercial P2P technology front that is moving towards it playing a larger, more contributing role in content distribution. For starters, take a look at the work of the P4P Working Group which accomplished quite a bit in 2008. P4PWG is based on research at Yale that concluded that ISPs and consumers could benefit if the ISPs provided guidance to P2P companies to make better peer disclosures. Verizon, Comcast, AT&T, Microsoft and a host of other companies are investing resources in this effort and a P4P standard has recently been submitted to the IETF. ISPs now have a way to impact commercial P2P traffic in a way that reduces their costs by encouraging peer communications to remain “on network” and reducing the number of costly off-network peer connections. And the data provided by ISPs can be dynamically updated based on the current traffic conditions on its network. What a difference a year can make – now in fact P2P has a way to of offering a more efficient way to deliver to the last mile! Field trials have shown that using P4P can reduce ISP transit costs by 80% while improving transfer speeds by 200%. The advent of P4P has brought a new method of improving network efficiency that holds the potential of becoming competitive with other efficiency alternatives such as ISP caching, direct network peering and IP multicasting. Every delivery method has its benefits and its shortcomings. P4P has the distinct advantages of being easy to implement, dynamically adjustable and most importantly, requiring relatively little in CAPEX.

To your point about consumers subsidizing content distributors with bandwidth – why not? Everybody wants content but few want to pay for it. But delivery costs publishers real money. With online ad rates dismally low, subsidizing the cost of delivery by contributing a resource that consumers have already paid for may be a good compromise for some publishers who have consumers who want their content but can’t find an effective way to cover the variable costs of delivery.

That said, it is my view is that content distributors who use P2P but require consumers to “seed” content for some period of time after the consumer interaction is over are being exploitive and that this practice is bad for the industry. Even worse is when there is inconspicuous or even no disclosure as to how and when the content will subsequently be shared. This will inevitably hurt consumer adoption as many users will ultimately have a real problem with this and reject this exchange as unfair or even deceptive. Unfortunately, I think too few P2P companies share this view.

I have long been a proponent that users should be contributing only when they are actively engaged with the content publisher – either by downloading content, playing a game, viewing video on a website, etc. When consumers disengage (leave a site, stop playing a game, etc) then they should no longer be contributors to the network. Providing users transparent and intuitive control of their participation in P2P will promote trust and adoption of these types of services. And the upshot of this approach, for the ISPs and its users, is that in the aggregate *the amount of data transferred on the last mile networks will not be significantly more than it would be if everyone got the content from a web server or CDN.* The amount of data transferred is roughly the same except that the traffic load is distributed across the network. Yes, there will be bytes retransmitted when there are broken connections, but this does not result in 50% less efficiency as you have described, not even close. And with P4P, it can be done in a way that works cost effectively for the ISPs, perhaps even better than relying on Internet transit to deliver all the content. The real limitation right now is in ensuring QoS, which if necessary can be addressed by utilizing the traditional delivery methods such as webservers or CDN. Finally, it is worth noting that employing P2P technology in this manner does nothing to contribute to or promote illegal file-sharing. It does not require the adoption or promotion of P2P applications that can be misused for piracy, illicit file-sharing or downloading of content illegally.

File-sharing applications aside, I believe that the future for commercial P2P technology is in making network utilization more efficient by distributing the traffic load based on resource availability, which will include input from the ISPs. Users that request content will temporarily connect with other users within proximity, either on the same network or well connected networks that are low cost links for the ISP. The objective will be relying on fewer connections each with high transfer rates so that the transfer is completed quickly while avoiding network congestion.

Getting people to understand the distinction between P2P technology and the P2P applications used for file-sharing has and will continue to be a long row to hoe. I know that you have been down on P2P in general for a long time and have repeatedly condemned it in the past as never to become an effective and efficient means of commercial content distribution. However, I am hopeful that you will be open to the idea that P2P technology, at least when used in the manner I have described, has the potential to be beneficial for the last mile networks, and need not be a freeloading activity nor a means of exploiting consumers resources. If so, please take this into consideration before using broad statements in any future diatribes against P2P file-sharing that do not rightfully distinguish it from those commercial efforts to deploy P2P technology in an efficient and responsible manner.

Regards,

Rick Buonincontri
CEO, Solid State Networks

Dynamic data such as chat, gaming, blogs, message boards, etc are miniscule in comparison.

So why aren\’t ISP\’s caching this content instead of letting the same data travel through many networks over and over again? Is their bandwidth cost less than supporting a distributed content caching solution?

http://www.msnbc.msn.com/id/22707271/

We\’re a small fledgling business struggling to pay our phone bill as it is, and use uploads to update our website, or to send (as well as receive) large files like DVD brochures from suppliers on yousendit.com, as well as frequently do backups online. I\’m also unsure about how VOIP or video conferencing works, but I\’m pretty sure my voice or image would be uploaded using heavy bandwidth?? What about my current reliance on online office apps to allow for online collaboration?

So Mark, would you have small businesses like ours who use yousendit, or google apps, or zoho or a million other innovative and cheap or free solutions (TRUE market solutions employing innovation and genius business models) revert to the old way of doing business? I\’m sure FEDEX, Microsoft, and the phone company would LOVE that.

Please Mark, you don\’t suck, but don\’t become so wrapped up in your own perspective on these technologies that you forget how invaluable the availability of the internet has been to the little guy with big ideas and no budget. Your analysis smacks a bit of \”Let them eat cake.\”

–Clark

I live in a country called New Zealand, not sure if you have ever heard of it, but if you haven\’t come for a holiday some time, the scenery is breath taking. I remember they unrolled unlimited broadband usage, meaning no caps and unlimited speeds. Do you know what happened Mark, I am pretty sure you do. The massive p2p downloading, etc that took place crippled our small countries network. It was like being back on dial up again, because people were using and abusing the bandwidth by downloading large amounts of things in peak service. They canceled that service until they can sort out the problems. But essentially people are blind to the idea that there is a limited amount of bandwidth available, and if they are 24 7 downloading and uploading they are affecting others experience.

Sports games: you can go to actual game and tape it too, watch it on TV and record it, no way to stop copying!
Movies: you can go to theater and record it, copy VCR/DVD/HDTV and more, and you cant stop it!
Music: go to concert and record it, off TV, from CD/DVD/Tape etc., and cant be stopped!
Patents: go to patent office or online, get copy, build it, cant top this either!

Example: Just try to tell all students to stop copying from teachers backboard, HAHAHA.

American DOD rejected Internet for its totally insecure structure. It was given away for the purpose of FREE sharing of information for all [all on this planet]. Distributed networking guarantees this freedom and CAN NOT be shut down no matter what.
Result: file/information sharing will continue as long as Internet is up.

These days almost all users have Server capable Operating Systems. This means they become a Server as soon as they log on. It also means they can share any and all information they want.
[End of refresher coarse]

We already pay a levy for downloading content so why are we penalized again?

RIAA, CRIA and rest of Spoiled Little Brats trying to stop users from sharing can go piss up a rope. They are wasting Billions of our money to promote their drivel so they can sit on their Asses and get a free ride. Yet they have declared WAR on the general public. Are we going to put up with it?

If you notice, the IPs attacking any flow of FREE information are the governments, big business and their related SPAM factories.

I propose a Tarpit/Honeypot be coded to trap all offending IPs. You know, the ones that attack file sharing and any user communications. This would solve any infraction against any user.

Theres no call what-so-ever to throttle any connection in Canada for bandwidth use. All ISPs that throttle bellow what user pays for [example: 5 Mbps dwld and 3 Mbps upld] should be charged for Fraud or charged for ripping off customers by whatever they take away from customer. You might include your wasted time on phone, loss of work, business etc. on invoice.

Also, as Michael pointed out, complain to CRTC and Competition Bureau and dont stop till they change it back to acceptable levels. Contact all Premiers of all provinces with your complaints too. Many of us use P2P for beta testing software, files, sending / receiving security or home video and much more which is none of ISPs or anyone elses business.

.tudmax,
Senior Network and Systems Analyst