Recovering from Distributable Thread Failures with Assured Timeliness in Real-Time Distributed Systems

We consider the problem of recovering from failures of distributable threads with assured timeliness. When a node hosting a portion of a distributable thread fails, it causes orphans — i.e., segments of distributable threads that are disconnected from the thread’s root. We consider a termination model for recovering from such failures, where the orphans must be detected and aborted, resources held by them must be released and rolled back to safe states, and exceptions must be delivered to farthest, contiguous surviving thread segment from where execution can be resumed. Since distributable threads are subject to time constraints in real-time distributed systems, such recovery must be conducted with assured timeliness. Toward this, we present 1) a real-time scheduling algorithm called AUA, and 2) a distributable thread integrity protocol called TP-TR. We show that AUA and TP-TR bound the orphan cleanup and recovery time (thereby bounding thread starvation durations), maximize total...

• Type: Conference
• Year: 2006
• Where: SRDS
• Authors: Edward Curley, Jonathan Stephen Anderson, Binoy Ravindran, E. Douglas Jensen
• Keywords: Assured Timeliness, Distributable Thread Integrity, Distributable Threads, SRDS 2006