From SCP's white paper, it states
Our definition of liveness is weak in that it says a node can externalize new values, not that it will. Hence, it admits a state of perpetual preemption in which consensus remains forever possible, yet the network continually thwarts it by delaying or reordering critical messages in just the wrong way. Perpetual preemption is inevitable in a purely asynchronous, deterministic system that survives node failure [Fischer et al. 1985]. Fortunately, preemption is transient. It does not indicate node failure, because the system can recover at any time. Protocols can mitigate the problem through randomness [Ben-Or 1983; Bracha and Toueg 1985] or through realistic assumptions about message latency [Dwork et al. 1988]. Latency assumptions are more practical when one would like to limit execution time or avoid the trusted dealers often required by more efficient Randomized algorithms [?]. Of course, only termination and not safety should depend upon message timing.
Theorem 16 assures us there are no dead-end states in SCP. However, a set of illbehaved nodes with very good timing could perpetually preempt an SCP system by delaying messages so that some fraction of intact nodes update ℎ right before timers fire and the remaining update it after, preventing intact nodes from converging on the next ballot. Nodes can recover from such an attack by removing ill-behaved nodes from their slices.
Can some elaborate on this? I'm not understanding exactly how this attack happens and what the severity of this is.
Does this mean that an extremely calculative adversary can create a DoS attack?