Here’s a question for you: when you use peer assisted delivery, who pays for the data?
Ok. So here’s the deal. What if I told you there’s a way to reduce your costs as a service provider by pushing some of the “serving” part of your work to your customers?
Let’s say you have a website (think this blog, just… with a bit more viewers), and it costs you money to have it running smoothly and fast. And someone comes to you, with this P2P magic and enables you to reduce your bandwidth costs while improving the service you give. Would you use it?
Not sure about you, but there are a few WebRTC vendors who are doing just that with the data channel.
During the meetup in Paris of the WebRTC guys there, Streamroot gave a presentation of what they do and how. Essentially – it is what I explained above. A question was raised – who pays for it?
If I go to your website, and the resources on that page are the size of 1Mb, and I am used as a peer within that network, then the bandwidth I “spend” might be 3-4 times more than that – just because I’ve been a “good neighbor”.
And if that goes on a wireless bill, where data caps are the norm then accessing your service actually cost me more.
In a way, the one providing the service here has effectively diverted some of his bandwidth costs to his users – he didn’t really eliminate them – just shifted them around.
–
Does this make such solutions bad? Does using the data channel to add a layer of anonymity and privacy over the web a bad thing if it eats up more data? And how is it different from Skype’s super nodes who slave your machines, making them relay points to calls of others?
There are no free meals. Not even in reading this blog…
There are some cases where it’s relatively safe to use the bandwidth without harm, typically with host candidates that are from the RFC 1918 local address space.
Without STUN, this will only work on a LAN.
Whether WebRTC will add user consent, as flash required, before getting beyond a limited traffic/bandwidth cap remains to be seen.