Checkpointing Models for Tasks with Widely Different Processing Times

A server subject to random breakdowns and repairs offers services incoming jobs whose lengths are highly variable. checkpointing policy aiming protect against possibly lengthy recovery periods is in operation. The problem of how choose a interval order optimize performance addressed by analysing general queueing model which includes breakdowns, repairs, back-ups recoveries. Exact solutions obtained under both Markovian non-Markovian assumptions. Numerical experiments illustrate the conditions where checkpoints useful they not, former case, quantify achievable benefits.