Paper #: 4.3.12
Title:  Secure routing for structured peer-to-peer overlay networks

1. Problems
-P2P overlay networks are notoriously unsecure. Large scale decentralized networks promote privacy and anonimity, which also provokes identity 
misrepresentation and a higher difficulty of catching malicious use.
-I should note that I took a programming class from one of the authors, Dan Wallach. He was a royal ass, and a former Netscape security guy who
 bashed Microsoft on a regular basis. It was a little satisfying to see him participate in a paper testing on Pastry, Microsoft's p2p structure.

2. New Idea and Strengths
-As mentioned in the previous paper I read, the authors admit that without a centralized authority the Sybil attack is indefensible. It's 
effects can be minimized through small secure systems, or extremely large systems where malicious users would require significant resources to 
have negative impacts. A monetary nodeId registration would also curb unauthorized identities.
-By approaching this by getting high probablities of success instead of solving the security problem, an effective solution is created that is 
closer to human interaction than I have seen. Although message reduncy is necessary, the performance slowdown is acceptable.

3. Weaknesses and Extensions
-The monetary registration feels like a cop-out. The ultimate answer is the trusted authority for registration. I feel that a centralized system 
in conjuction with a decentralized network that acts as a backup system would be ideal. In the event of a crash of the centralized system due to 
an attack, the backup network would become the primary one. By maintaining the base structure of the decentralized network, the motiviation for 
attacking the centralized system is significantly decreased, as DoS attacks would only harm the centralized system, not the entire network, as basic 
users would hardly notice a difference, just increased trust issues...