2026-07-25T00:41:41Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:17216 2024-03-06T10:59:15Z ddc:004 open_access

Fragmented ARES: Dynamic Storage for Large Objects Georgiou, Chryssis Nicolaou, Nicolas Trigeorgi, Andria Distributed storage Large objects Strong consistency High access concurrency Erasure code Reconfiguration Data availability is one of the most important features in distributed storage systems, made possible by data replication. Nowadays data are generated rapidly and developing efficient, scalable and reliable storage systems has become one of the major challenges for high performance computing. In this work, we develop and prove correct a dynamic, robust and strongly consistent distributed shared memory suitable for handling large objects (such as files) and utilizing erasure coding. We do so by integrating an Adaptive, Reconfigurable, Atomic memory framework, called Ares, with the CoBFS framework, which relies on a block fragmentation technique to handle large objects. With the addition of Ares, we also enable the use of an erasure-coded algorithm to further split the data and to potentially improve storage efficiency at the replica servers and operation latency. Our development is complemented with an in-depth experimental evaluation on the Emulab and AWS EC2 testbeds, illustrating the benefits of our approach, as well as interesting tradeoffs. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Chryssis Georgiou and Nicolas Nicolaou and Andria Trigeorgi 2022 Is Part Of LIPIcs, Volume 246, 36th International Symposium on Distributed Computing (DISC 2022) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.DISC.2022.25 urn:nbn:de:0030-drops-172161 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.DISC.2022.25 eng https://creativecommons.org/licenses/by/4.0/legalcode