Slashdot Mirror
P2P Set-top Boxes To Revolutionize Internet
An anonymous reader writes "The European Commissions 7th Framework Program (FP7) is working on a project called Nano Data Centers (NADA) as part of the its future Internet initiative. NADA will seek to build an Internet architecture that delivers data from the edge of the Internet using set top boxes and Peer-to-Peer technology, instead of the network-centric architecture that stores and delivers content from data centers via Internet backbones. NADA is proposing a network of hundreds of thousands of set top boxes, hugely popular in Europe, to be essentially split into two — one side is the user interface side, the other a virtualised Peer-to-Peer storage client that stores and sends media in the same way a data center would. Ideally there would be millions of these boxes each acting as a mini data center — hence the Nano Data Center moniker.
The NADA project is convincing enough to have attracted some of Europe's largest telecommunications companies. Set top box manufacturer, Thomson SA, and European ISP, Telefonica, are among nine contributing partners to the NADA project.
NADA could see a dramatic reduction in the size and frequency of data centers that serve all kinds of media over the Internet."
I gather Cable + movies (maybe eventually games) is what they are after here. It seems as though the idea is to be able to deliver content faster and with less stress on a centralized data center than we have now for things like digital cable et al.
The thing I am wondering though is how would they maintain quality with such an uncontrollable system. Basically it seems that it will, of course, benefit the content delivery company in reducing bandwidth overhead. But where is the benefit to the user? What happens when a particular "torrent" is less popular? Will it be able to stream fast enough for the end user to see the video in reasonably close to real-time? Or, would they be distributing every file equally? essentially consuming the user's bandwidth and hard drive space for files they don't use/need/watch?
meep
YEAH....you know...I could go quite a ways with this one...but we'll leave it at "the guy who made this comment is full of shit".
It's in the summary, and again in the article:
"Telefonica is a key partner"
Either this scheme's proponents are hoping to sneak a lot of their costs onto other people's plates, or they have a seriously dubious grasp of the economics of IT.
In effect, bittorrent's success is not based on its efficiency as a file transfer mechanism(which is actually quite lousy); but on its effectiveness as a download micropayment system(which isn't fantastic; but is better than anything else we have). Bittorrent reduces the cost, to the distributor, of distributing a file by making it easy for downloaders to contribute their own bandwidth. Even more conveniently, for anybody with a fixed-price internet connection, the marginal cost of their bandwidth contribution is near zero. Unfortunately, the total cost of distribution is actually fairly high, since bittorrent uses a lot of "last mile" bandwidth(particularly upstream last mile bandwidth) which is quite limited and expensive compared to bulk datacenter bandwidth. If micropayment were possible, and if individuals paid for bandwidth per-unit-use, rather than fixed rate, it would be cheaper for them to just pay the file distributor's upload costs directly, at bulk rate, rather than "in kind" at retail rates. The exact same argument applies for electricity and disk space. Bittorrent is great because it is an efficient method of aggregating the limited amounts available at zero(ish) marginal cost, not because it is actually efficient per unit.
Given this, I find it hard to judge TFA's scheme kindly. Either it is based on a frankly delusional understanding of relative costs, or it is essentially a cynical attempt to shift costs onto end users. This will only get worse if the ISP pressure toward caps and overage fees gets stronger, since the amount of "free" bandwidth will decline, and the impact of shifted costs will become much more direct.