Time, Clocks, and the Ordering of Events in a Distributed System.

Time, Clocks, and the Ordering of Events in a Distributed System.

by Leslie Lamport

My thoughts on the paper -

Important Points ---
* Provide an effective simple approach, that seems to be easy to implement, that resolves an very important synchronization problem.

* They effectively present their approach in the form of a set of 2 conditions that if satisfied would provide total ordering of system events. They provide an efffective algorithm to involving multiple processes that uses message acknowledgement protocols and message numbers.

* Paves the way for future research in areas like failure handling , i.e. when a process fails to respond.

Deficiencies ---
* They also argue that this problem applies to any multiprocessing system, even if all the processes are on the same machine. I think there could more simplistic ways, if the processes are on the same machine, it was not clear from the paper what was the synchronization problem when the processes were on the same machine.

* The algorithm given to synchronize access to a single resource from multiple processes seems to have a lot of overhead as for every resource request messages are sent to and recived from from all other processes.

Conclusion ---
This paved a way for a solution to a very important problem is distributed systems, and is still commonly used paradigm in distributed systems. Nowdays, Network TIme Servers are used quite a lot for synchronization, which extends the same paradigm.