PADLL, a storage middleware that enables system administrators to proactively control and ensure QoS over metadata workflows in HPC storage systems, was presented by our researcher from INESC TEC, Mariana Miranda, at the 2nd Workshop on Re-envisioning Extreme-Scale I/O for Emerging Hybrid HPC Workloads, on the IEEE Cluster Conference 2022 and within the scope of the BigHPC project.
The results of the paper “Protecting Metadata Severs From Harm Through Application-level I/O Control” show that PADLL can dynamically control metadata-aggressive workloads, prevent I/O burstiness, and ensure I/O fairness and prioritization. With it, system administrators can holistically control the I/O rate of all running jobs, and thus, prevent metadata-aggressive ones from harming the PFS, as well as other jobs in execution.
The paper was written by Ricardo Macedo, Mariana Miranda, and João Paulo, from INESC TEC and University of Minho, Yusuke Tanimura and Jason Haga, from AIST, Amit Ruhela and Stephen Lien Harrell, from TACC and UT Austin, and Richard Todd Evans, from Intel.
The IEEE Cluster Conference 2022 took place between September 6-9, 2022, in Heidelberg, Germany. This Conference serves a major international forum for presenting and sharing recent accomplishments and technological developments in the field of cluster computing as well as the use of cluster systems for scientific and commercial applications.
The paper will be available soon.