-
Website
http://mashable.com/ -
Original page
http://mashable.com/2007/11/23/mark-cuban-stop-p2p-traffic/ -
Subscribe
All Comments -
Community
-
Top Commenters
-
Robert Basil
142 comments · 8 points
-
Jennifer Van Grove
149 comments · 23 points
-
r0cketman22
317 comments · 52 points
-
rajagiri4
160 comments · 2 points
-
barringtonarch
150 comments · 4 points
-
-
Popular Threads
-
Enter the Zappos Sharing Happiness $3,000 Shopping Spree Giveaway Contest
4 hours ago · 92 comments
-
iPhone App Offers Instant Speech-to-Text Transcription
3 hours ago · 16 comments
-
Your Next Car Radio Might Be Pandora
4 hours ago · 21 comments
-
Google Launches Chrome for Mac
6 hours ago · 28 comments
-
BREAKING: Google Launches Real-Time Search Results
1 day ago · 96 comments
-
Enter the Zappos Sharing Happiness $3,000 Shopping Spree Giveaway Contest
I said Google Video should replace P2P..no. I said it was better for audio and video. Understand the difference ?
And your example on efficiency only applies in the rare cases where 100pct of the seeds are on the same network segment.
and there are far better ways to do file distribution than p2p. Look up multicast. I guess i should have expected your post and done a "Turn Off P2P, replace it with multicast and...." post first. Would that have helped you ?
the redunancy and load balancing of P2P ? I would love to see that paper confirming that. The issue with P2P is that you dont know who your seeds are or when they are available and exactly how much bandwidth will be available. if you did, it wouldnt be a P2P network, it would be a CDN.
Im all for applying P2P principles on a CDN. A CDN has resources thats are clearly definable and for the most part deterministic. Something that an open internet P2P application doesnt have.
More importantly, CDNs pay to host their servers and for the bandwidth they use.
Open Internet P2P is a leach network. Its theft by stealing bandwidth a few bits at a time.
As bad as an idea as it would be, it would be fun to watch ISPs go to a pay per bit model rather than the disingenuous all you can eat model they currently have.
If not a pay per bit, how about an all you can download for free, but pay per bit for upload model ?
Want to see how inefficient a protocol P2P is ? Watch how quickly users would unload their clients when they got a bandwidth bill for their upstream bandwidth use.
As far as CDNs go, aren't they unsuitable for transferring a lot of very large files? A combination of CDN and P2P, maybe, for QOS reasons; but CDN replacing P2P, I don't see it happening.
About load balancing and redundancy, well there are plenty of papers on P2P out there; I don't mean to say that P2P is perfect, but it sure as hell beats simply everyone downloading from the same source, which is the common alternative.
About paying for upload, well, that would be another step back into the dark ages of the Internet and it's simply not going to happen. You said it yourself: it's a bad idea.
Bottom line: P2P is here, and it solved a lot of problems for the users. Has it created some new ones for the ISPs? Maybe. But we're paying them. They need to solve it to our satisfaction.
This doesnt mean that Joost isnt a valid application. It is. If they were to put P2P on a CDN and pay for the bandwidth they consume, I would be all for it.
To be clear, Im not against P2P across the board, just in instances where its used as a bandwidth leach.
As far as CDN vs P2P, i would guess that for non pirated bits, that CDNs consume more bandwidth than P2P does. Again, Im fine with this because it requires companies to pay for the bandwidth they use.
And P2P beats downloading from a single source in many cases, but who downloads from a single source these days ? There are untold non P2P options, CDNs and multiple file download hosts that require the distributor to pay for the bandwidth they consume and distribute the bits to end users in an very efficient manner.
The Google option is for file distribution rather than emulating a video distribution option. Google is preferable because they have what is arguably the best network for content distribution in the world and they have made a business decision to subsidize the cost of anyone wants to use it.
And I think its a bad idea because of the confusion it would cause, not because its not a viable business model. I havent done a complete analysis, but from a cursory glance it seems like it could lead to the opportunity for lower pricing alternatives (if ISPs were smart enough to avail themselves of this alternative)
Bottom line is that P2P hasnt solved a single problem for recipients of files. It has solved a cost problem for distribution of content/files/other bits. It has moved the cost of distribution from the source to the destination. I think thats wrong.
obviously , not everyone agrees.
If we want to push for ISPs to optimize our experience and minimize our costs, then lets push for multicast. Lets push for better accounting of usage and more realism in advertising of just want we get for our money and what we dont get.
I can think of 100 different ways to improve the ISP experience.
I can also think of 100 internet sacred cows that need to be challenged. Challenging them in blog posts is a lot more fun
m
Wouldn't it be something if this guy won the trolling contest? Mark Cuban has my official nomination.
I totally agree with your idea of a Centrally Planned and Controlled People's Internet, Mr. Cuban. I think there should be heavy restrictions on on what people can and can't do with the bandwidth they pay for.
And you are so right about P2P. It's ridiculously inefficient to have each client acting as a server, taking the load off the main site. So much bandwidth wasted!
I was on Ares last night and it brought down my ISP, because I was trying to download 8 songs at the same time. So that's proof right there that P2P is a failed technology.
Also, while I have you here, could you sign my petition to ban electric lightbulbs? They're totally inefficient and they waste electricity. Once we make them illegal, we can go back to using gas lamps.
Bandwidth is a measure of throughput, so 100Mbps is a measure of bandwidth, 100MB (or 10GB) is just a file size.
In general, the issue with p2p for ISPs isn't so much that there is more or less data being used up, they know full well that the bandwidth will still be consumed, but that the number of simultaneous connections between homes is far higher than they built the networks to support. One movie could well be downloaded using 100s of separate download streams using p2p when 1 stream would be used from a central ftp site.
This more predictable behaviour of centralised networking makes it much easier for ISPs to shape traffic sensibly so peoples HTTP requests and VOIP are transmitted quickly and consistently while keeping the ftp transfer going at a reasonable rate, but instead they end up resorting to just crippling p2p entirely to try and reduce the load.
In the long-term I think ISPs will end up rebuilding their networks around the p2p model to support higher numbers of seemingly random connections, and the problems currently experienced will largely go away.
You bring some good points though, I partly addressed your comment in my reply to Mark up there. The thing is, I think Mark is talking about a lot of different things, and in doing so he misses the overall point. You can't say: I'd have more bandwidth if it weren't for P2P, so let's ban P2P. It's a gross oversimplification on several levels.
It is the only manner of bandwidth consumption that commercial operations use in order to avoid paying for bandwidth. A distinction that I point out multiple times
m
"Let’s talk elementary math here. 100 users want to download a 100 MB file. If they all download it from a single source, HTTP or FTP style, the overall bandwidth toll will be approx. 10 GB. Now, if they all do it by using P2P, downloading the file off each other, how much bandwidth will they spend in the end? You guessed it - approx. 10 GB"
so yes you did say that 100MB was a measure of bandwidth, in terms of 10GB being a measure of 100 users downloading 100MB each, in fact you said it twice.
Oh, and I'll be expecting my troll of the week prize in the post, I can't imagine many entries getting a reply from yourselves :)
The word of the day is................ *DING DING DING* Bandwidth!
this is truly funny !
I pay them for unlimited bandwidth, so what's wrong with me using what I've paid for? What if I use HTTP? What's wrong if it's P2P? Hey Mr Stupid Mark, are you sure you want to pay for unlimited bandwidth and use limited bandwidth?
P2P is a great technology, with the sharing in mind. It helps downloading from busy servers easier, take loads of servers, files live longer. It's just that servers pay for their limited bandwidth, while I have unlimited bandwidth. P2P increase bandwidth at end user side and significantly reduce server load. It's not much different in total amount of bandwidth used. But it's that ISPs have to pay more for their network.
ISPs oversells all the time, they lie about network speed, they charge me for unlimited and want to limit my bandwidth, they're f*cking bitches dirty.
They simply give higher priority to http traffic and do some simple tricks to show their network speed, which is absolutely stupid under an network administrator's eye.
They blame P2P for the slow speed that instead they have for pay for an upgrade.
Have a nice day, Mr Stupid Mark, aka. the most stupid puppet of ISPs.
however, i have a shitty isp and i realise that, so thats the problem, not p2p.
Obviously, there are technical and efficiency problems with existing networks; shared connections mean that broadband isn't really comparable to the processing power of my computer.
But the way broadband is currently sold; I am buying an always on service if I choose to, while the companies selling the broadband are operating on the same model as airlines over selling flights. I choose to use as much as possible the capacity I have been _sold_, and I think the fault lies with network providers (and awful advertising and regulation).
1. i have never known anything on p2p networks to be legal, for the most part the p2p is an underground leach that allows the smuggling of illegal audio, video and software.
2. Servers pay for the bandwidth to upload to the user files, If everyone that wasn't useing p2p networks, started to we would all see our much smaller pipe lines at our homes getting clogged and IF... p2p networks were emliminated then our speeds at home would be greater because our consumer lines would not be clogged with data that should be going through a much wider server line.
3. Bandwidth is measured in GB's in the end, you CAN use 10 GB of bandwidth, because those that pay for it don't pay for the speed, but rather pay for the total pass through in a given amount of time. Bandwidth can be both a speed AND when quantified, a measurement of the actual data.
4. Another factor is that videos that are out on p2p networks are poorly encoded and compressed. If video was streamline through a server that could compress and encode all videos to a standard, it would reduce the amount of bandwidth that would have to be used, by reducing the size of the files.
Adobe needs to hurry and get there new flash player out so we can all switch over to the H.264 format.
get off his back..
"in the land of the blind the one eyed man is stoned"
The fact is that Energy delivery per unit time (Power) is also finite. Brownouts during the summer, because everyone has their AC cranked, are exactly analogous to when your connection slows because your neighbors are running a bunch of torrents.
What Mark Cuban seems to be suggesting is "bandwidth-conservation." To me this is basically the equivalent of Rockefeller-esque Industrialist recommending that we use less energy in the middle of an Industrial Revolution. Conservation may have it's merits, but certainly not yet. ISPs should be focused on one thing -- building infrastructure to fuel the IT revolution.
If bandwidth-conservation were to ever become a real issue (doubtful), the last person I'd want to hear about it from is a dot-com billionaire.