News Article

  • 15 December 2015

    Mediaflux® Cluster Edition Adds Scale-Out Capability to Parallelize I/O

    Arcitecta has added a significant enhancement to the Mediaflux data management platform with the release of Mediaflux Cluster Edition, a scale-out capability that enables multiple Mediaflux servers to work in parallel to accommodate high-throughput requirements for large data environments.

    The addition of clustering capabilities to Mediaflux software means that as data sets grow and throughput requirements increase over time, customers can add additional Mediaflux Cluster Nodes as needed to increase the I/O (input/output) capacity of their existing Mediaflux system.

    In this way, IT administrators can future-proof their system, increasing throughput incrementally as needed without impacting existing systems or user workflows.

    Mediaflux is being used in data environments where rapidly growing volumes of data require the ability of the system to increase throughput to keep ahead of the growth.

    With Mediaflux Cluster Edition, the throughput of a customer’s existing system can be increased by simply adding as many Mediaflux Cluster Nodes as required.

    This ability to incrementally expand an existing system protects Mediaflux customers’ investment, and enables them to proactively accommodate future requirements without disruption to users.

    Mediaflux software has long been used by environments that depend on its ability to provide global management of data across disparate data types, across otherwise incompatible storage environments, and across multiple locations.

    Leveraging the power of metadata, Mediaflux enables users to find and act upon data in ways that would otherwise be difficult or impossible in such heterogeneous environments using conventional approaches.

    The addition of Mediaflux Cluster Edition enhances this capability by adding virtually unlimited scale-out capacity to the system. Existing Mediaflux customers, for example, can expand I/O on current systems by simply adding as many Mediaflux Cluster Nodes as needed to keep ahead of evolving requirements.

    As with any Mediaflux installation, the cluster node software runs on almost any computer, small or large, and on standard operating systems. Hardware choices may therefore be tuned to the specific I/O load and customer’s preference.

    Cluster nodes may be added or removed at any time, using any mixture of hardware as required to achieve the throughput requirements. This flexibility means that customers can repurpose existing servers, or minimize the need for new servers, to dramatically increase I/O throughput with the existing storage infrastructure.

    Originally released to selected customers one year ago, Mediaflux Cluster Edition has since been refined to deliver a multi-fold increase in performance across all data types and networks. Arcitecta has deployments where throughputs of 10s of TB per hour are now possible with a relatively small cluster. When these workflows require additional throughput, more Cluster Nodes can be added at any time to keep ahead of requirements.

    Journaling and other improvements to the underlying Mediaflux XODB NoSQL database that dramatically increase transaction rates also contribute, along with Clustering, to expanding Mediaflux’s ability to manage many billions of digital assets in a live system.

    “Democratizing access to data is a key tenet of Mediaflux,” said Mr. Lohrey. “The scale-out capability of Mediaflux Cluster Edition enhances our ability to deliver on this promise, by enabling much greater flexibility and scalability to support increased customer workflows.”