More P2P Fun with Mark
Reusing headers. How gauche.
While they help to clarify his position regarding his first post (which I wrote about below), they make his overall viewpoint (in my opinion) even more skewed.
In one breath he argues that consumers really can’t tailor their individual ISP accounts to their personal ideals, and in the next he advocates a system where outbound traffic is charged a per kb rate independently of all-you-can-eat inbound traffic, because that makes the most sense to him personally:
For most internet users, like probably 99pct of us, it wouldn’t make a bit of difference in our bills or consumption. In fact, many of us could opt for cheaper plans because beyond the family photos or videos we may upload every now and then, or the rare backup of our hard drives, most people don’t consume much outbound bandwidth at all.
which is true if 99pct of all internet users don’t play on-line games, or have on-line connected devices (especially game consoles), Or use Skype (or any other VOIP-based telephony), or webcam. or iChat, Or VPN, or FTP, or upload to their own websites, or participate in distributed computing projects…
Again, I’m not hating just to be a hater here. Bandwidth has to come from *somewhere*, and it has a finite service limit, and a tangible cost. If performance is truly such an issue across the network that ISP’s need to move to a “per kb” pricing system, so be it (but that would also open the ISP’s up to admitting they’re utilities… and increase the pressure on them for the requisite oversight and disclosure that telcos face – which I suspect they’d rather avoid).
What really sticks in my craw about Cuban’s position is that he clearly sees a fundamental difference between being a “commercial provider” (like Google, or Flickr, or HD Net) and a “consumer”… to really beat it with a simplistic stick – that “downloading” is a different product than “uploading”. This is a view of the Internet that treats it as a provider/consumer relationship. Which isn’t incorrect from a functional point of view (I could never host my podcast series on my home internet connection, because the connectivity would be atrocious… most ISP’s configure their network to greatly prioritize downloading) but is a much different interpretation of “the internet” than many would share. The Internet is a utility not a content channel (with a mythical “pay-per-upload” system tacked-on). The fundamental difference between (ugh) “new-media” and “old-media” is that “old-media” is a one way communication and “new media” is a two way collaboration. If I want to run my phones, my television, my home-office via the utility that is the Internet that is as valid a use as reading websites, or watching Google Videos. To split bandwidth into “up” and “down” streams somehow weights categories of useage as being more or less “correct” useages when really they are all pieces of the same whole.
I’m even wary of Cuban’s baseline comment about last-mile P2P (in)efficiency and it’s effect on end-user overall connectivity. Certainly P2P traffic generates more upstream traffic than just downloading a file from a centralized server – but as I mentioned ISP’s usually strictly QoS upstream and downstream, so even if Mark’s neighbours are Level-50 Night Elf Mohawks with a nasty P2P habit on the side, the amount that their *upload* stream is going to degrade his connectivity experience is negligible. Especially since Mark counts himself among the “99%” who only use the outbound stream for occasional family photos.
After my initial post, it was pointed out to me that this isn’t the first time Cuban’s taken a swipe against BitTorrent in particular, and that he is/was also involved with financing a commercial file-sharing competitor to the ubiquitous P2P protocol. Perhaps in the end these series of posts are really less about issues facing his personal internet connection and more about how difficult it is for commercial concerns to compete against open source models when the actual commodities involved (time and bandwidth) are either given away by ISP’s, or donated by the Open Source model. This at least would be an argument I could understand, but it’s a very different concern and debate than the one he initially put forward.