Restart Behaviour of Distributed Status Messages

The project compares restart rules that make outdated status messages recognisable after a connection loss. Completed in the scenario; outputs are illustrative.

P01 · Completed · Networked Systems

Problem

After a connection loss, old messages can look like current states. Anyone who accepts them unchecked works with a state that has long been superseded. The project treats this confusion as a problem of message ordering and restart rules.

Research question

Which combination of sequence number, timestamp and restart rule makes outdated states recognisable? The study examines how the three elements complement one another and where each falls short on its own.

Method

Three state machines are compared using synthetic messages and deliberately interrupted connections. The experiments run on the network test bench with controlled fault scenarios.

Evaluation criteria

The criteria are the share of outdated messages correctly discarded, the time until a consistent state is reached, and the communication overhead under defined loss patterns. These criteria frame the comparison; no measured values are reported here.

Expected outcome

Completed in the scenario. The illustrative outputs are comparison logic and example traces that make the behaviour of the three rules easy to follow. They are not measured performance figures and not a performance guarantee.

Limits

Only modelled message paths are considered. The work makes no claim to certified control-room or real-time functionality.

Publications