The Aureum architecture scales out organically. It decouples the namespace from the data stores, and distributes the namespace across servers, making it fully expandable with parallel access to data while eliminating performance bottlenecks, even for small files. This unique architecture allows customers to grow their storage infrastructure as needs dictate, even to exabyte scale with linear performance.
To ensure high availability, the architecture is self-healing. It features remote and local replication to ensure that data is always available. Replication of up to four copies of both data and namespace is supported across servers to ensure data survives even when you lose entire servers (depending on the level of redundancy used).
The system also ensures data integrity through the use of Cyclic Redundancy Checks (CRCs) for both metadata and data, and can self-heal by retrieving any affected data from a replica server. Use the Asynchronous Replication feature for disaster recovery services through remote replication. Asynchronous Replication is also used to replicate data between sites working at exabyte scale.
Aureum is as easy to use as it is powerful. Virtually anyone can configure the system, keeping information management as close as possible to those who use the data.
Combining user-defined storage classes with data management policies, Aureum ensures that data resides on the hardware that best fits the requirements. A comprehensive dashboard enables easy configuration by users with appropriate access. This allows enterprises to easily set organizational data policies and storage classes across all data managed by Aureum.
Peaxy® Threaded Find helps categorize, catalog, parse and thread data on the Aureum platform instantly. The immutable pathname ensures that as the system is scaled out, data is never lost and can be stored for the long term across technology refreshes.
Threaded Find capabilities are fully integrated with other Aureum data management functions. Aureum's data classes aggregate data sets by project or data type regardless of the file pathname, group or user. Find can utilize data classes to create more useful data catalogs and digital threads. The system indexes data whenever it is created or modified, eliminating the need to periodically crawl the entire infrastructure and improving overall performance and speed. Our technology also enables rich metadata tagging for speedy discovery and retrieval of data.
A three-level security architecture provides strong authentication. Security levels include authentication only, message integrity checking, and encryption.
The first level uses Kerberos®, X.509 certificates, and HTTPS to ensure that only authorized machines and users can connect to Aureum.
The second level adds Poly 1305-AES cryptographic message authentication code (MAC) on each data packet to ensure that the contents of a message have not been tampered with or altered.
The third level provides full 256-bit authenticated encryption on the wire, to protect data across locations.
Today’s enterprises are aggregating, mining, and analyzing data to gain insights using the Hadoop open source framework for distributed processing of large datasets. Ingesting and storing data in Aureum provisions the data set with robust data services, including management and availability, search, and security.
Support for POSIX® and HDFS guarantees a robust, stable, highly available, and high-performance foundation to ingest and store data in Aureum, and to perform computations on it with Hadoop or Spark.
A fully clustered SAMBA server is integrated with Aureum to allow efficient interoperability with Windows-based clients. This approach provides scalable performance and resiliency against machine failures by distributing the load across virtual machines and providing failover capabilities.
Aureum is hardware-agnostic and runs on industry-standard commodity hardware to reduce cost of ownership. It is built to run across heterogeneous hardware, so enterprises can mix and match vendors, form factors, hardware generations, and types of storage devices while still offering a single namespace that aggregates all devices.