|
June 26, 2011, 12:37:04 AM |
|
Bumping this thread.
I am interested in this kind of system. Need to outline some goals here.
Want to blend in bitcoin's aspects into the mesh system more deeply. Not just as a payment system incentive, but as a stand-in for trust.
The forced-work-as-standin-for-trust system is a useful solution to the Byzantine Generals problem, but we need to outline what trust we're trying to mitigate.
Right now centralized internet access is the norm, in a trickle up system. We must trust that the internet we receive from upstream nodes will a) work properly (ie not be shut off) b) not be filtered in some way.
To avoid the shutting off, we go to an ad hoc model where trust in a particular mesh node is not required. This is done by supplanting work required on part of the node for the quality (uptime? bw rate? lack of filtering 'canary in a coalmine' content?) of the connection. However, in bitcoins the incentive for doign the work is to realise the value of the bitcoins you mine. In this case we have to realise some benefit to the node that does the work. Perhaps in this case then instead of like namecoin which is its own advantage as an alternate system to bitcoin, we actually use bitcoins to generate a payment for the node.
However we must be careful in that very very good nodes dont dominate the local (or larger) tree, or they'll become attackable as single points of failure/access. Not sure what can be done here - if we make an incentive to provide connectivity too large, nodes will dominate. If we make a timeout situation, then we'll be punishing reliable nodes and encouraging large organizations with wider resources to setup multiple nodes in cahoots and dominate. This is a tricky problem for local access - not the same issue as bitcoin dominateing > 50% of computation resources as connectivity may locally expose nodes to >50% (or more) of all connectivity travelling via just a few nodes, making a local attack on the leaf nodes QoS possible.
Throwing some ideas around here. Need input.
|