Orphan problems and remedies in distributed systems

In large distributed systems, site and communication failures can be relatively frequently occurring events. Such faults may allow unwanted remote procedure calls (RPC's) to execute, leading to a problem known as an orphan. In this paper, the impact of orphans on reliability aspects of t tPC's in general and the entire system in particular is examined. The paper is organized in such a way that how orphans are created is presented first. Then a survey of three different algorithms for handling orphan problems in distributed systems will be discussed. The implementation of these algorithms in Rajdoot RPC will be discussed next. Finally, we conclude with a critical assessment of the three algorithms.