[p2p-hackers] Byzantine Quorum Systems

[Wes Felter]

Brad Neuberg wrote:

> My understanding of byzantine quorum algorithms are that they break down 
> under the kinds of conditions found in large-scale, P2P systems, which 
> have very high node churn and high latency, described above.  They seem 
> to be focused on very stable or LAN type networks.  Is this a correct 
> assumption?  If it is, it seems that byzantine quorum algorithms need to 
> be refocused on the kinds of networks that we are dealing with today, 
> rather than LAN centric networks or networks of very stable servers on 
> the public Internet.

OceanStore solves this problem by using a byzantine quorum protocol only 
between supernodes. IIRC, performance of this protocol over the Internet 
was not too bad. Unfortunately, I didn't understand Nick's recent 
messages (and thus the context for this discussion) at all, so I don't 
know if this is relevant.

Wes Felter - wesley at felter.org