The cascade fault tolerance message system
H. Al-Jabar, Shmuel Rotenstreich · 1989
This paper addresses the problem of constructing a highly reliable message delivery system in a distributed environment. It presents a fault tolerance algorithm that guarantees the delivery of a message to its destination despite faults in one or more nodes in the system. The algorithm is distinguished by not using extra hardware or checkpoint facilities that are common to many algorithm of its type. Instead, it maintains an appropriate number of cascading copies of the message in nodes where the message passes. In case of fault, the algorithm locates a copy of the message closest to the destination, and resumes delivery of the message from this location. This special mechanism makes the algorithm suitable to systems without disks, real-time systems, and systems where message transmission time must be minimal.