Self-stabilizing Connected Components

Conference · Fri Nov 01 00:00:00 EDT 2019

DOI:https://doi.org/10.1109/FTXS49593.2019.00011· OSTI ID:1649445

Sao, Piyush  ^[1]; Engelmann, Christian  ^[1]; Eswar, Srinivas ^[2]; Green, Oded ^[2]; Vuduc, Richard ^[3]

---

1. ORNL
2. Georgia Institute of Technology
3. Georgia Institute of Technology, Atlanta

For the problem of computing the connected components of a graph, this paper considers the design of algorithms that are resilient to transient hardware faults, like bit Hips. More specifically, it applies the technique of self-stabilization. A system is self-stabilizing if, when starting from a valid or invalid state, it is guaranteed to reach a valid state after a finite number of steps. Therefore on a machine subject to a transient fault, a self-stabilizing algorithm could recover if that fault caused the system to enter an invalid state. We give a comprehensive analysis of the valid and invalid states during label propagation and derive algorithms to verify and correct the invalid state. The self-stabilizing label-propagation algorithm performs O (V log V) additional computation and requires O (V) additional storage over its conventional counterpart (and, as such, does not increase asymptotic complexity over conventional label propagation). When run against a battery of simulated fault injection tests, the self-stabilizing label propagation algorithm exhibits more resilient behavior than a triple modular redundancy (TMR) based fault-tolerant algorithm in 80% of cases. From a performance perspective, it also outperforms TMR as it requires fewer iterations in total. Beyond the fault tolerance properties of self-stabilizing label-propagation, we believe, they are useful from the theoretical perspective; and may have other use-cases.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
• Sponsoring Organization: USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
• DOE Contract Number: AC05-00OR22725
• OSTI ID: 1649445
• Resource Relation: Conference: The 9th Workshop on Fault Tolerance for HPC at eXtreme Scale (FTXS) 2019 - Denver, Colorado, United States of America - 11/22/2019 5:00:00 AM-
• Country of Publication: United States
• Language: English