What Intel Distribution for Apache Hadoop software means for PowerEdge customers

Dell is committed to enabling customer workloads. We know that many of our customers are running applications leveraging Hadoop both in test / dev environments as well as within their IT environments. Our award-winning PowerEdge R720xd  server and cloud enabled PowerEdge C8000 series servers are particularly well suited for these customer environments because Hadoop deployments often call for a 1:1 spindle to core ratio for optimized performance. The PowerEdge R720xd and the PowerEdge C8000 both have high spindle to core counts, with plenty of options to help prevent or eliminate I/O bottlenecks.

The PowerEdge R720xd offers twenty-four 2.5-inch disks or twelve 3.5-inch disks (plus two 2.5-inch disks in the back of the server), up to 768GB of memory (24 DIMMs) and integrated PCI Express 3.0-capable expansion slots.

With the PowerEdge C8000 series, big data applications can get the right mix of compute, storage and coprocessor resources in the same chassis. Choices include up to eight single-wide C8220 compute sleds with up to 16 cores each; four C8220X double-wide compute/coprocessor sleds; four C8000XD storage sleds with up to up to 36TB of SAS, SATA or SSD storage; and a variety of combinations.

Introduced today, Intel Distribution of Apache Hadoop software (Intel Distribution) is open source, leveraging the latest Intel innovation and server optimization with Dell PowerEdge R720xd and C8000 series servers.  Intel’s engineering team seized an opportunity to move Intel hardware innovation and optimization up the software stack – leveraging performance, security and manageability features into Hadoop natively.

To increase performance, the Intel Distributions has:

  • Optimized HIVE to run real-time queries directly on HBase
  • An AES-NI instruction set to speed file encryption and decryption, with up to 20x performance improvement with Intel Distribution on Intel-processor-based servers
  • Automated tuning by the Intel Active Tune included in the Intel Manager for Apache Hadoop

We look forward to extending the value of the Intel Distribution to more Dell PowerEdge customers!

About the Author: Joseph B. George