Reputation Consolidation Concept




A breif outline of a theoretical mechanism to allow for the unmasking of bad actors in a decentralised p2p environment


This document describes a theoretical mechanism that can be employed in any of the major two-tiered decentralised peer-2-peer networks that use the flood broadcasting method or a variation of the same, its efficiency relies on the number of genuine peers outnumbering those controlled by an attacker who will be attempting to interdict and inject false network traffic into the high level network group.


The reasons for proposing this mechanism and its suggested method of usage


The ability to follow a data trail/track back to the originating node is currently ones thats only possible misusing selected packets and is not an available option in the normal release of all clients using this type of broadcast method and thus its non-existence in all of the upper tier clients on the network is one of the major roadblocks to undertaking such auditing of packet routing data. Even misuing a current packet to try to do this involves communicating with anything from ten to hundreds of nodes, cross checking the data they provide to pick out the rogue nodes in the network in this way is far from ideal and pretty much impractical so we need to look at the addition of a packet that delivers this data to each node along the data routing path and allow for an automatic challenge that does a simple handshake of a randomly generated number directed only at the IP handing the data along the route, thus a second node will confirm the IP of the originating node and the 3rd will confirm the IP of the 2nd and in this fashion the scope for rogue broadcasts that are injected into the network to remain undetected diminishes with each spoof packet they generate. By also only challenging the node prior to itself and not by attempting to try all of the nodes back down the routing chain, this mechnism virtually eliminates the vast overhead normally necessary to undertake this same action. Thus a chain of "reputation" exists for each relay request across the decentralised network, bad actors can only spoof/interdict data pror to or from themselves but to make of its disruption value are strictly required to pass the data along the chain and thus the trail will in enough cases always lead back to them and allow for them to be placed on a network specific blacklist.

 















A breif outline of a theoretical mechanism to allow for the unmasking of bad actors in a decentralised p2p environment

©2005-2024 WinMXWorld.com. All rights reserved. Page last updated Tue Dec 17 2013