Recent development on distributed systems has shown that a variety of fairness constraints （some of which are only recently defined） play vital roles in designing self- stabilizing population protocols. Existing mod...Recent development on distributed systems has shown that a variety of fairness constraints （some of which are only recently defined） play vital roles in designing self- stabilizing population protocols. Existing model checkers are deficient in verifying the systems as only limited kinds of fair- ness are supported with limited verification efficiency. In this work, we support model checking of distributed systems in the toolkit PAT （process analysis toolkit）, with a variety of fairness constraints （e.g., process-level weak/strong fairness, event-level weak/strong fairness, strong global fairness）. It performs on-the-fly verification against linear temporal prop- erties. We show through empirical evaluation （on recent pop- ulation protocols as well as benchmark systems） that PAT has advantage in model checking with fairness. Previously un- known bugs have been revealed against systems which are designed to function only with strong global fairness.展开更多
文摘Recent development on distributed systems has shown that a variety of fairness constraints （some of which are only recently defined） play vital roles in designing self- stabilizing population protocols. Existing model checkers are deficient in verifying the systems as only limited kinds of fair- ness are supported with limited verification efficiency. In this work, we support model checking of distributed systems in the toolkit PAT （process analysis toolkit）, with a variety of fairness constraints （e.g., process-level weak/strong fairness, event-level weak/strong fairness, strong global fairness）. It performs on-the-fly verification against linear temporal prop- erties. We show through empirical evaluation （on recent pop- ulation protocols as well as benchmark systems） that PAT has advantage in model checking with fairness. Previously un- known bugs have been revealed against systems which are designed to function only with strong global fairness.
