YAPD (yet another P2P diagram)

OK, these are a couple of diagrams that try to do simple description of the magic of peer-to-peer, in the 1st image, the ISP ignores the p2p traffic, in the 2nd, it implements a layer of caching near the edge/access network (CMTS/UMTS/ADSL/FTTH):

(1) - end-user accesses metadata repository to find content of interest;
(2) - then complete (if any) access lifecycle (purchase, drm key generation, etc.);
(3) - peer-to-peer protocol then kicks in, trying to discover peers and find appropriate "segments" amongst near ones (moving across to peers further and further away)
(4) - Peers start deliverying segments to requestor (until enough segments have been collected to create a buffer and 'play' the content)
(5) - If peers don't have the segments, it is necessary to access the peer-2-peer data centre.



In this second diagram, pretty much steps 1 to 4 remain the same, difference is now:

(5) - ISP has a caching layer near the edge/access that allows not only to reduce the amount of times:

• segments have to be retrieved from remote peers
• content/segments have to be retrieved from the p2p data centre

(6) - If peers don't have the content and/or segments are not found, it is necessary to access the peer-2-peer data centre.

It is a very high level description - but as common sense would tell you: the closest the content is to the edge, the faster the access from the end-user's perspective.

There are also benefits for the ISP (i.e. ability to reduce transit bandwidth, mediate if there is profit share involved, etc).