India to the World

Computing, IT & AI

Chandy-Lamport distributed snapshot

1985 CEK. Mani Chandy and Leslie LamportDistributed snapshots: determining global states of distributed systems (ACM TOCS, 1985)Diaspora (work abroad)

1985

An algorithm for recording a consistent picture of a distributed system without stopping it.

Dated
1985
As scholars put it
1985 CE
Originator
K. Mani Chandy and Leslie Lamport
Source text
Distributed snapshots: determining global states of distributed systems (ACM TOCS, 1985)
Region
Diaspora (work abroad)
Discipline
Computing, IT & AI
Evidence
Documented
Confidence
High
Verification
Verified-web

Background

In a system of many computers exchanging messages there is no global clock, so it is hard to say what state the whole system is in at any moment. K. Mani Chandy, an IIT Madras graduate now at Caltech, and Leslie Lamport showed how marker messages let every process record its state, together with the messages in transit, consistently.

Significance

A foundational result of distributed computing, used for checkpointing, debugging and detecting deadlock or termination.

Path to modern innovation

Stream-processing engines such as Apache Flink checkpoint with variants of it.

Global impact

  • Fault-tolerant checkpointing in large data-processing systems.

References

  • 3Distributed snapshots: determining global states of distributed systems K. M. Chandy and L. Lamport, 1985

Further reading

Corrections and additional sources: edit IND-391 on GitHub.