Date Added: Apr 2011
Byzantine fault-tolerance is a fundamental requirement for many contemporary services. Shorter development cycles, bigger server exposure to attacks, and an inherently hostile network have demonstrated the weaknesses of replication mechanisms designed for benign failures. Unfortunately, Byzantine Fault-Tolerant (BFT) services usually have increased latency, when compared to simple client-server interactions, and limited scalability, in the sense that adding replicas does not translate into higher throughput. Except for a few exceptions discussed later, most recent contributions in the area of BFT protocols have sought to address the latency problem.