tankman1989
Active Member
- Reaction score
- 5
Ok, so this is a combination of question and idea/proposal. I was using Netflix the other day and I was having major bandwidth issues and the movie was interrupted every few minutes until it got to the point that it wasn't watchable. The bandwidth problem didn't seem to be on my end so I figured that it was on the server side. I was also listening to a podcast about how Verizon is throttling the top 5% of their wireless data users and it is also utilizing a very interesting technology that is going to free up more bandwidth for their network (all this because of the Iphone release on the network). Basically what they are doing is re-encoding video at higher compression rates and using a cache system to store popular videos at regional distribution centers instead of only being available at one source (utilizing mirrors basically).
What I am thinking here is creating a content distribution network which utilizes P2P connectivity to distribute revenue generating content. I have been following the availability of Netflix videos and it seems that they only have so many videos available at once. IDK if this is due to limited capacity w/n the system or due to licensing agreements limiting the amount of content they can provide. If it is the former, then this distributed network would be very beneficial and allow for very rapid expansion of the network.
What I envision is the client/customer downloading a client program which is similar to uTorrent or other torrent clients. Let's use Netflix or any movie serving company as an example. Let's say you have a FiOS connection with ridiculous upload speeds. This P2P network would take advantage of the idle network capacity much like the distributed F@H makes use of unused CPU cycles.
What would happen is everytime you watch a movie it is downloaded onto your machine where it is stored until someone on a network close to you needs to watch the movie. When someone else ques the movie, the server looks for clients with the movie in their "stored movie list" and then starts the "silent" transfer from the server. There would have to be extensive test run to determine network speed ATM of the request and maybe look at past trends of network usage for that time and day in order to determine availability for the length of the movie. Like P2P there may be a number of clients serving the movie to you, each on uploading a specific part where it is later re-assembled.
So why would anyone do this? The content provider would provide compensation of some kind to those involved, maybe they also get content before others have the ability to watch it, this would make a lot of sense as they would distribute the content to the participating clients before the rest of the customers are able to see it. Once it is d/l'd onto the client machines there is much less bandwidth used in the further distribution of the content once fully released to the entire customer base.
I've come up with a lot of ways of maximizing this setup and see that it has great potential, especially when distributing unique content.
So, if you have followed my tangent, what do you think? I would think that movies would be the most widely distributed content but can anyone else think of anything else that is distributed in such large capacity?
So basically you could have something like 5,000 customer/clients, each with a very high speed connection distributing content "locally" - for compensation maybe they get the service for free or something, this isn't the area on which I have been focusing..
What I am thinking here is creating a content distribution network which utilizes P2P connectivity to distribute revenue generating content. I have been following the availability of Netflix videos and it seems that they only have so many videos available at once. IDK if this is due to limited capacity w/n the system or due to licensing agreements limiting the amount of content they can provide. If it is the former, then this distributed network would be very beneficial and allow for very rapid expansion of the network.
What I envision is the client/customer downloading a client program which is similar to uTorrent or other torrent clients. Let's use Netflix or any movie serving company as an example. Let's say you have a FiOS connection with ridiculous upload speeds. This P2P network would take advantage of the idle network capacity much like the distributed F@H makes use of unused CPU cycles.
What would happen is everytime you watch a movie it is downloaded onto your machine where it is stored until someone on a network close to you needs to watch the movie. When someone else ques the movie, the server looks for clients with the movie in their "stored movie list" and then starts the "silent" transfer from the server. There would have to be extensive test run to determine network speed ATM of the request and maybe look at past trends of network usage for that time and day in order to determine availability for the length of the movie. Like P2P there may be a number of clients serving the movie to you, each on uploading a specific part where it is later re-assembled.
So why would anyone do this? The content provider would provide compensation of some kind to those involved, maybe they also get content before others have the ability to watch it, this would make a lot of sense as they would distribute the content to the participating clients before the rest of the customers are able to see it. Once it is d/l'd onto the client machines there is much less bandwidth used in the further distribution of the content once fully released to the entire customer base.
I've come up with a lot of ways of maximizing this setup and see that it has great potential, especially when distributing unique content.
So, if you have followed my tangent, what do you think? I would think that movies would be the most widely distributed content but can anyone else think of anything else that is distributed in such large capacity?
So basically you could have something like 5,000 customer/clients, each with a very high speed connection distributing content "locally" - for compensation maybe they get the service for free or something, this isn't the area on which I have been focusing..
Last edited:
