The new ETERNUS CD10000 S2 includes many enhancements to reduce storage costs and space requirements in the data center. New types of storage nodes can host up to 60 TB of data in a single rack unit, further reducing space requirements. New hard- and software functions and technologies also reduce the cost per TB by a factor of up to ten.(2)
New design storage nodes called ‘Flex Node’ and ‘Density Node’ enable higher granularity in adding capacity and performance. Flex Nodes support storage performance-oriented usage scenarios while providing good scalability, while Density Nodes are optimum for storing large amounts of data in a minimum amount of space and at the lowest possible costs per terabyte (as low as between EUR 0.007 to 0.013 per GB per month). Simply by adding system nodes, the ETERNUS CD10000 S2 can cover capacity up to 50 PB and beyond, supporting scenarios of unpredictable data growth.
Since the ETERNUS CD10000 S2 internal system management is based on Ceph software, older hardware nodes can be replaced by fully compatible newer generations without system downtime, therefore extending system lifecycle by 10 to 15 years in comparison with traditional RAID storage systems. This also drastically reduces system migration requirements.
One of Fujitsu’s latest contributions to the Ceph open source community has been an optimized Erasure Coding, providing data protection through redundancy while reducing disk capacity requirements and costs.
A further enhancement to the ETERNUS CD10000 S2 is the support for more disaster resilience options, achieved by either stretching one storage node cluster across two sites, or by implementing two discrete system clusters. As traditional backup solutions reach their limits in the multi-petabyte dimensions, Fujitsu has introduced an innovative backup-in-a-box concept to provide large-scale data protection. This adds flexibility and the ability to deal with complete site failure, while ensuring fast data recovery, even on a massive scale.
Based on existing customer scenarios, Fujitsu has developed validated solutions enabling a fast implementation of software-defined storage. These solutions cover Enterprise File Sync & Share, online archives and cloud back-up. For OpenStack projects where Ceph-based storage is the most suitable storage architecture, Fujitsu recently announced PRIMEFLEX for OpenStack
, a complete solution that integrates ETERNUS CD10000 with Fujitsu’s PRIMERGY servers and OpenStack cloud software. Fujitsu is an active contributor to the OpenStack Initiative and recently attained Gold Membership
Jack Rondoni, Vice President, Storage Networking at Brocade says: ”Clustered storage platforms like the ETERNUS CD10000 S2 require resilient, low-latency networks as a foundation to deliver high-quality data services. We are partnering with Fujitsu to deliver high-performance storage fabrics as part of an overall solution that improves application availability and helps ensure non-disruptive scale-out to support a variety of high-performance workloads and data protection initiatives.”
Ben Cherian, Director of Partners, Channel and Alliances, Storage and Big Data at Red Hat says: ““Fujitsu was the first hardware provider to embed Ceph technology, and, today, that commitment is reaffirmed with the announcement of the Fujitsu ETERNUS CD10000 S2 storage system. By combining Red Hat Ceph Storage with the new features in the Fujitsu ETERNUS CD10000 S2 storage system, customers may more quickly and cost-effectively deploy a software-defined storage solution in their next-generation hyperscale environments.”
The FUJITSU Storage ETERNUS CD10000 S2 is now available in EMEIA and in the Americas from Fujitsu and Fujitsu SELECT Experts that offer Advanced Server Solutions. Prices vary by specification and region. Other regions will announce local availability later in 2015.
1. According to Fujitsu’s own calculations. Compared to the the predecessor based on a of 4-node minimum configuration.
2. According to Fujitsu’s own calculations. With introduction of the Density Nodes compared to the predecessor based on a of 4-node minimum configuration.