Detecting Arbitrary Stable Properties Using Efficient Snapshots

Abstract

Abstract-A stable property continues to hold in an execution once it becomes true. Detecting arbitrary stable properties efficiently in distributed executions is still an open problem. The known algorithms for detecting arbitrary stable properties and snapshot algorithms used to detect such stable properties suffer from drawbacks such as the following: They incur the overhead of a large number of messages per global snapshot, or alter application message headers, or use inhibition, or use the execution history, or assume a strong property such as causal delivery of messages in the system. We solve the problem of detecting an arbitrary stable property efficiently under the following assumptions: P1) The application messages should not be modified, not even by timestamps or message coloring. P2) No inhibition is allowed. P3) The algorithm should not use the message history. P4) Any process can initiate the algorithm. This paper proposes a family of nonintrusive algorithms requiring 6ðn À 1Þ control messages, where n is the number of processes. A three-phase strategy of uncoordinated observation of local states is used to give a consistent snapshot from which any stable property can be detected. A key feature of our algorithms is that they do not rely on the processes continually and pessimistically reporting their activity. Only the relevant activity that occurs in the thin slice during the algorithm execution needs to be examined.

A Lightweight Heuristic-based Mechanism for Collecting Committed Consistent Global States in Optimistic Simulation

“…Instead, we exploit the result of the last GVT calculation (anyway required for memory recovery purposes) in order to heuristically determine where to realign the local states of the involved LPs while guaranteing global snapshot consistency. Similar considerations can be made when comparing our proposal to the global snapshot collection protocols in [6,10,13,14], some of which have been proposed just in the context of detection of global predicates in distributed systems. These solutions require an explicit distributed algorithm to be executed among the processes participating in the computation for the determination of the consistent global snapshot, which is avoided in our proposal thanks to the exploitation of the results of the GVT protocol.…”

Fully-Distributed Debugging and Visualization of Distributed Systems in Anonymous Networks

“…They assume that processes have unique identifiers and/or there is exactly one initiator. Many papers give also specific algorithms to detect some specific predicates that hold in a system such as termination or deadlock (Mattern, 1987;Marzullo and Sabel, 1994;Kshemkalyani and Wu, 2007;Kshemkalyani, 2010). Among well-known global predicates of distributed systems detected with snapshots, one can also consider loss of tokens and garbage collection (see (Tel, 2000;Santoro, 2007;Kshemkalyani and Singhal, 2008)).…”

Detecting Locally Distributed Predicates

“…We compare our algorithms to the classic Chandy-Lamport distributed snapshot algorithm. While there are more efficient snapshot algorithms [Kshemkalyani and Wu 2007;Venkatesan 1989], the Chandy-Lamport algorithm is both widely known and very succinct.…”