High-Quality Fault Resiliency in Fat Trees

John Gliksberg (1, 2, 3) , Antoine Capra (3) , Alexandre Louvet (3) , Pedro Javier Garcia (2) , Devan Sohier (1)

1 LI-PaRAD - Laboratoire d'Informatique Parallélisme Réseaux Algorithmes Distribués
2 UCLM - Universidad de Castilla-La Mancha = University of Castilla-La Mancha
3 Atos Bull

Coupling regular topologies with optimised routing algorithms is key in pushing the performance of interconnection networks of supercomputers. In this paper we present Dmodc, a fast deterministic routing algorithm for Parallel Generalised Fat-Trees (PGFTs) which minimises congestion risk even under massive network degradation caused by equipment failure. Dmodc computes forwarding tables with a closed-form arithmetic formula by relying on a fast preprocessing phase. This allows complete re-routing of networks with tens of thousands of nodes in less than a second. In turn, this greatly helps centralised fabric management react to faults with high-quality routing tables and no impact to running applications in current and future very large-scale HPC clusters.

Domaines: Calcul parallèle, distribué et partagé [cs.DC]; Réseaux et télécommunications [cs.NI]

John Gliksberg, Antoine Capra, Alexandre Louvet, Pedro Javier Garcia, Devan Sohier. High-Quality Fault Resiliency in Fat Trees. IEEE Micro, 2020, 40 (1), pp.44-49. ⟨10.1109/MM.2019.2949978⟩. ⟨hal-03861715⟩