Quobyte vervollständigt paralleles Dateisystem mit Version 1.3 um Erasure Coding

Berlin, Starnberg, 21. Sept. 2016 - Durch den Einsatz von Erasure Coding können mit dem Update 1.3 jetzt auch Medien- und Archiv-Workloads verarbeitet werden…

Zum Hintergrund: Skalierbare Storage Filesysteme gewinnen auf Grund der zunehmenden Menge an zu verarbeitenden Daten an Bedeutung. Mit seinem Produkt-Update in v.1.3 erweitert Quobyte das Konzept des parallelen Dateisystems, das bisher auf HPC Anwendungen zugeschnitten war, zum skalierbaren Data Center File System, das File-/NAS-, Block- und Objekt-Storage-Workloads verarbeitet. Die Software Defined Storage - Lösung benötigt laut Entwickler als Hardware-Grundlage ausschließlich Standard-Server, da alle Redundanz in der Software abgebildet wird. Durch den Einsatz von Erasure Coding sollen mit dem Update 1.3 jetzt auch Medien- und Archiv-Workloads effizienter und schneller verarbeitet werden können. Weitere technische Informationen hierzu finden sich im Blog zu Erasure Coding von Felix Hupfeld (QuoByte Co-Founder & CTO).

Nachfolgend ein (gekürzter) Auszug aus der Originalmeldung des Herstellers zur Ankündigung:

  • "Quobyte’s software storage system combines a high-performance parallel file system core with native data redundancy and safety features such as replication for data and metadata. The 1.3 release marks another milestone in Quobyte’s value for a wide range of HPC and scale-out NAS workloads. It improves performance significantly in terms of IOPS and throughput, delivers a sub-millisecond latency for direct I/O, and fully parallelizes metadata operations. The release also extends Quobyte’s data management capabilities for organizations that have a large number of users and file systems through cross-interface ACLs, integrated multi-tenancy and hierarchical quota support.

  • Core of the 1.3 update is erasure coding. For sequential workloads, erasure coding combines high storage efficiency and resilience without compromising performance. Just like Quobyte’s end-to-end checksums and quorum-replicated metadata, erasure coding is a mechanism implemented purely in software and is therefore able to turn standard server hardware into a fault-tolerant storage system. Compared to hardware redundancy like RAID it is not only more economical, but also makes the system much more resilient and easy to manage.

  • Quobyte’s roots lie within the academic HPC community. Before creating Quobyte, the two founders, Björn Kolbeck and Felix Hupfeld, together with their team developed the open source parallel file system XtreemFS at the Zuse-Institute Berlin, one of Germany’s supercomputing centers. Taking the best parts of their previous work and harnessing the founders’ experience with large-scale storage systems at Google, the Quobyte team built a brand new POSIX-compatible file system with Quobyte, adding cutting-edge technologies to make it suitable for enterprise use. The resulting storage system runs any application and any workload on any kind of x86 server.


About Quobyte: Quobyte is software that turns commodity servers into a reliable and highly automated data center file system. It combines latest distributed systems research with proven concepts for large-scale infrastructure. Quobyte can handle all workloads in a single consolidated deployment: from virtual machines over shared file storage to Big Data and high-performance computing. It comes with all the benefits of a data center file system: full fault tolerance, resilience, scalability and self-management – in short, it greatly simplifies operations.

Quobyte was founded in 2013 with VC funding and is based in Berlin, Germany. Despite its young age, the company draws from nearly a decade of research and experience with the open-source distributed file system XtreemFS and from working on Google’s infrastructure. Quobyte’s customers profit from the system’s capabilities in such diverse areas as container infrastructures, public OpenStack clouds, supercomputing, and large-scale web services.


Abb. 1: Bildquelle Quobyte / Erasure Coding Screenshot, Sept. 2016