Response to IEEE paper that characterizes P2P as undesirable and illegal
Consider, for a moment, the issue most often cited for "traffic shaping", the practice of filtering a users traffic based on the type and source: legality of content. While there is an abundance of content with questionable copyright origins based on the current interpretations of the DMCA (in America), there is also a sea of legal content being acquired by the same means: Nine Inch Nails, Radiohead, and a number of other musical artists have experimented with a freely available online distribution method, in addition to countless young movie producers that are only interested in their content being available and seen.I've never understood the ISP/admin approach to P2P that says, "We've provided you with a pipe so you can access the Internet, but stop accessing the Internet so much!" If users want P2P, then P2P is what makes paying for an ISP valuable, so why would ISPs want to reduce its availability? That's like a phone company that discovers that teenagers use phones to send a lot of texts to one another, overwhelming their capacity (based on assumptions about how much text users will want to send) who then throttles text-sending rather than changing their assumptions about use-patterns.How can network monitoring practices differentiate between "legal" and "illegal" P2P traffic? Filtering by content source, such as a band's official website vs. IsoHunt, is impractical - the content available via the official source is likely licensed for free distribution and sharing by other means. Filtering by traffic size, as in number of bytes transferred, is a gray area at best - setting an arbitrary size for acceptable P2P traffic, or any type of traffic, creates artificial pricing levels, not to mention potentially endorsing the acquisition of questionably sourced content. There is really only one option left, and it is what most ISPs choose in such cases: filter by traffic type.
Incorrect Base Assumptions About Network Management (Thanks, Kyle!)
the latest
latest episodes
This issue isn't really related to the legality or desirability of P2P. It's an endemic group-think between network managers who see usage patterns blowing through their capacity plans and business managers who don't have the budget to expand capacity faster than planned. The same thing happens within corporate networks: "We need to stop the users from using the network."
P2P does make a convenient scape-goat when this situation applies to the internet, though. The ISPs don't need to come up with new justifications for curtailing traffic, but can just parrot the industry "knowledge" that is already out there.
Obviously there needs to be an RFC put forward that defines a Legality Bit, and those who use P2P should be on their honor to set it appropriately.
The reason why there is so much objection on the ISP front is because P2P content delivery shifts the cost of delivery to the ISP and, at the same time, greatly increases the aggregate cost of content delivery.
Unless there is caching built into the P2P infrastructure, the basic P2P design acts to magnify costs.
@2: You laugh, but the FCC actually adopted this with the Broadcast Flag and only an EFF lawsuit killed it.
Of course, they deny the most useful application of P2P software: to prevent network congestion by spreading popular data far and wide, such as how it is used by Blizzard to distribute updates to World of Warcraft.
P2P actually makes the network more efficient, but you'll never hear the ISPs and telcos admit that, because they don't care about efficiency, they care about profit, and increasing the capacity of their existing systems without adding additional users gives them no excuse to extract more money. To them, using the network at an efficiency greater than they assumed in their initial models is stealing. And they won't update their models to represent reality, because to do so is to admit that they are nothing more than a seller of commodity bandwidth, indistinguishable from each other and subject to the laws of supply and demand.
the main thing that the ISPs don’t want to admit is that they have oversold there bandwidth so when people complain "I am not getting the speed you promised" its easer to blame "kids" and there p2p traffic many programs are updated using p2p now and to start blocking it is just short sighted and dumb
@2, @3:
See also the humourous April Fool's Day RFC from 2003:
RFC 3514, "The Security Flag in the IPv4 Header"
http://www.faqs.org/rfcs/rfc3514.html
OMG!
IEEE is supposed to be a reasonbly reputable organisation, this is more like something spun by a Ruper Murdoch powered tabloid....
ISP's are profit sluts, no better then tobacco companies, banks, Mobile Phone companies ( with brain frying RF thats just soo totally safe ! ) & big oil companies
Maybe the guy was simply payed to write the nonsense?
From reading the IEEE article, it describes what appears to be a useful optimization in the design of IP routers: by caching the results of routing decisions, you can build more efficient routers.
More than that, you then have a useful set of data structures which help you design routers that make better packet-drop decisions that take into account the class of data being transferred.
This seems entirely reasonable, even interesting.
The complaints about P2P simply echo the concerns of network managers: demand for bandwidth is growing faster than expectations.
The 'problem' with P2P is not technological, it is economic: ISPs have misjudged how much it will cost them to provide the services they've promised to their customers.
(UK ISPs that use BT's backbone network are charged a metered rate by BT -- but only charge a flat rate to their customers.
Those ISPs would be fine if their customer's bandwidth use had remained stable, but it hasn't.)
P2P was only the beginning: other high-bandwidth services, such as the BBC iPlayer, YouTube, and others are now causing major headaches for ISPs that are unable to afford the promised made in their contracts.
However, unlike P2P, the services provided by these sites are clearly legitimate -- so they've had to take a different tack: they're now trying to charge the upstream service providers for carrying their traffic across their network.
But that's unlikely to work; Google and the BBC have already paid for their bandwidth, too.
I manage bandwidth, not for an ISP, but for a small college. Our purchased b/w has about doubled every 18 months since I've been here. And it's never been enough. One difference between P2P and, say, web browsing is the latter is something people wait on. Click on link, wait for the page to load, wash, rinse, repeat. It should be fast and responsive. P2P apps either d/l in the background or could queue up downloads and get the lot while the user does something else. We don't block P2P traffic, but we try hard to keep human wait time low, and if that means music files take all night to download, ah, well...
We could hire another professor or two for what we spend on Internet bandwidth, and we get the best rates available as a non-profit .edu. We work with what we have.
Apparently it is cheaper to re-engineer their customers expectations than to meet them.
Hey, it worked for Detroit!
It works for boingboing, too!
laughs diabolically
love you guys!
Those ISPs would be fine if their customer's bandwidth use had remained stable, but it hasn't.
I'm not sure what you mean by fine.
Why should any capitalist business that failed to anticipate the markets demands be fine?
I want personal automatic traffic shaping so my torrents don't flood out my voip calls and web browsing.
This is sorta like the post office finding out that contraband is mailed in parcels, so in order to cut down on people sending illegal things, they begin throwing out all parcels.
Over the course of a week last month I downloaded every linux distribution I could find and burnt them all to discs. My ISP cut me off, and I ended up having a very surreal conversation with their billing department, convincing them that not only is linux distributed by p2p, but it's also legal.
What happened here is Roberts started a company (Anagran) and built his cool router, but nobody bought it. Then he realized that his router could be easily modified to throttle P2P and thus ISPs would buy it. I doubt Roberts personally cares about P2P one way or the other; he's just whoring out to whoever's paying. It's unfortunate that IEEE chose to uncritically print his anti-P2P views, though.
@9,
"ISPs have misjudged how much it will cost them to provide the services they've promised to their customers."
Maybe. I *did* enjoy your thoughtful post.
But my ISP complains about how the volume of spam has increased to the point that he has to continually upgrade his capacity. And there is certainly a lot of spam out there. I can't entirely blame ISPs for misjudging the cost of the services they've promised to provide. If you over-build the data center, you go broke. If you under-build, customers are unhappy.
@ #10: We have the same problem in our library. Due to budget cuts, a lot of our upgrade projects are on hold, including increasing bandwidth and/or making all our libraries wifi hotspots. Streaming video, games and chat had slowed Internet access to dead slow, so we sadly had to block access for those functions, at least until we can get better bandwidth. Most patrons understand about the blocks, but making excuses for it is grating, especially when YouTube and other social networking site links are so ubiquitous.
@14:
Check out DD-WRT and other firware for routers, they allow QOS like that.
P2P is already an underutilized protocol. There's a http(p2p) protocol which has never been implemented (to my knowledge) in browsers, but in many cases could solve the effect of boingboing (or digg, fark, or other high traffic sites) linking to tiny sites.
If you can't get static web content from the original source, you get it from a p2p network. A combination of p2p and ordinary web protocols would even be able to take some of the burden off servers running dynamic content (allowing images or other static portions of the site to be retrieved via p2p).
It's worth mentioning that I had a conversation with Dr. Roberts and he clarified things quite a bit - it turns out some of my criticisms may not be on him, but on the editors.
I posted the update at the bottom of the article on my site.
--Kyle
Dijjer?
Perhaps they should alter their expectations then?
Attention telecom CEOs: Expect Moore's Law to apply to bandwidth demands, ok?!
Maybe your .edu should consider retiring / laying off some professors and buying more bandwidth instead?
Pitch the idea as "telelearning" or contributing to Open Courseware, if that helps.
Precisely! Business exist solely to satisfy customer demands; not vice-versa.
Businesses that fail to anticipate customer demands need to go bankrupt and liquidate to free up those resources to new businesses who will make a better attempt at serving their customers.
Businesses aren't (really) people, it's ok for companies to "die".
Overprovisioning is actually the least expensive option.
c.f. Assuring Scarcity, When Bandwidth is Free
Also: dumb network, end-to-end principle