Power Systems™ offers a range of solutions for big data, from entry level for dedicated Hadoop infrastructures with the IBM Solution for Hadoop Power Systems Edition to enterprise solutions with the IBM GPFS™ Storage Server (Elastic Storage Server), which is capable of hosting multiple Hadoop tenants along with traditional analytics workloads. The IBM Elastic Storage Server is included in the IBM Data Engine for Analytics or can be purchased independently. It is a building block for big data. For further information on the IBM Data Engine for Analytics, refer to Hardware Announcement 114-176, dated October 06, 2014.
The IBM Elastic Storage Server is a big data storage system that combines Power servers, storage enclosures, and disks along with GPFS and its Native RAID technology, providing analytic and technical computing storage and data services for elastic storage workloads.
The Elastic Storage Server, built on POWER8™ server technology, offers the following capabilities and features:
- Scalability from 40 TB to hundreds of petabytes, supporting 10 Gigabit Ethernet (10 GbE), 40 Gigabit Ethernet (40 GbE), and FDR InfiniBand.
- Support for GPFS V4.1 capabilities in complete, ready-to-deploy models with enhanced server-based RAID technologies supporting GPFS NSD and NFS protocol access.
- ESS provides a common repository, infrastructure, and name space for modern workflow such as analytic, cloud, and technical computing.
- Big data requires easy storage growth. The Elastic Storage Server scales in a building block approach, where adding additional Storage Servers adds to the overall capacity, bandwidth, and performance all with a single name space.
- Big data requires price/performance. By leveraging JBODs and Power Servers, ESS provides low-cost storage for the analytics, technical computing, and cloud computing environments.
- Big data requires big data performance. The Elastic Storage Server provides industry-leading streaming data performance for modern workflows, up to 15 GB/second from a single system, with near linear performance improvement as systems scale.
- Big data requires predictable performance and data protection. The ESS modern declustered RAID technology is designed to recover from multiple disk failures in minutes, versus hours and days in older technology. 8+2 and 8+3 RAID protection and platter-to-client data protection are included with ESS.