SGI^® Hadoop Clusters

Having moved beyond its origins in search and Web indexing, Hadoop is fast becoming the de facto standard for large-scale, data storage and analytics applications. Because Hadoop deployments can have very large infrastructure requirements, hardware and software choices made at design time have significant impact in terms of performance and total cost of ownership.

Solutions Brief   

SGI Hadoop Solutions

Solutions for Data Analysis on Large Clusters
(Registration Required)

SGI is one of the leading commercial providers of Hadoop clusters to the industry's most prominent Hadoop users. SGI offers complete, fully configured Hadoop clusters built on the highly successful SGI® Rackable™ and CloudRack® server lines. A Cloudera Connect Partner, SGI distributes Cloudera software pre-installed on SGI Hadoop Clusters. SGI is a performance leader, having recently announced world record Terasort benchmark results using SGI Rackable servers and running the Cloudera distribution of Apache Hadoop.

SGI Hadoop Clusters are purpose built to match customer requirements for performance, power optimization, density, and price performance. SGI can build and deliver fully integrated Hadoop clusters based on our reference implementations, or can help you design an optimized Hadoop cluster based on your application requirements, optimization points, and facility/operations constraints.

SGI Hadoop Clusters are built on:

• SGI Rackable Servers: Rackmount servers that can be deployed in either back-to-back or flow-through configurations, using either traditional AC or rack-level DC power options. SGI Rackable servers provide industry-leading density, power-savings and flexibility. With the ability to draw from a wide variety of commodity components, the Rackable line allows customers multiple configuration design points.

• SGI CloudRack Servers: CloudRack extends the level of flexibility and power savings available in the Rackable line by implementing a tray design that creates a 'breadboard' over which different combinations and layouts of motherboards and disk drives can be arranged. Hadoop configurations based on CloudRack allow for the largest disk to motherboard ratios available in the industry today. Cooling is provided at the rack level with large, efficient fans. DC variants provides highly efficient rack level rectification, while the AC version provides power supplies at the tray level.

SGI Hadoop Clusters are factory integrated, installed, and serviced by SGI. SGI Management Center, included with your system, provides complete remote monitoring and control of your Hadoop cluster. SGI Hadoop Clusters along with Cloudera's software, services and support form a complete, end-to-end solution for enterprises deploying Apache Hadoop in production or performance-intensive environments. SGI Professional Services provides complete services from conception through implementation, including design optimization, system integration with existing infrastructure, tuning for optimal performance, and more.

Additional Information

• Press Release: SGI Establishes New World Record Apache Hadoop Benchmark
• Press Release: Cloudera and SGI Partner to Take High Performance Computing on Apache Hadoop to the Next Level
• Solutions Brief: SGI Hadoop Solutions - Solutions for Data Analysis on Large Clusters - registration required
• White Paper: Increased Hadoop™ Optimization Using Intel®-based SGI® Rackable™ Solutions (PDF 2.0M)