Hortonworks on Ryan Orban

Nutanix Joins Hortonworks Certified Technology Partner Program

me@ryanorban.com (Ryan Orban) — Wed, 26 Jun 2013 00:00:00 +0000

Summary

Nutanix joined the Hortonworks Certified Technology Partner Program in June 2013, a business development milestone that gave enterprise customers formal validation that HDP (Hortonworks Data Platform) workloads would run on Nutanix hyper-converged infrastructure. Ryan’s bookmark note (BOOM!) captures the enthusiasm inside Nutanix — this was an important step in the company’s strategy to expand beyond pure VMware virtualization into Hadoop analytics workloads.

Get Started: Ambari for Provisioning, Managing and Monitoring Hadoop

me@ryanorban.com (Ryan Orban) — Sat, 04 May 2013 00:00:00 +0000

Summary

Apache Ambari was Hortonworks’ answer to Hadoop’s operational complexity problem. In 2012-2013, setting up and managing a Hadoop cluster required deep knowledge of XML configuration files, SSH scripting, and per-service configuration for HDFS, MapReduce, Hive, HBase, Pig, and a dozen other components. Ambari wrapped all of this in a web UI with guided install wizards, centralized configuration management, and service health dashboards.

Best Practices for Virtualizing Hadoop

me@ryanorban.com (Ryan Orban) — Wed, 01 May 2013 00:00:00 +0000

Summary

This Hadoop Summit presentation covered running Apache Hadoop workloads on VMware infrastructure using Hadoop Virtualization Extensions (HVE) with Hortonworks Data Platform (HDP). In 2013 this was a contested topic: the conventional wisdom was that Hadoop should run on bare metal because virtualization overhead degraded the I/O-intensive HDFS and MapReduce workloads. HVE was VMware’s answer — a set of extensions specifically designed to preserve data locality when running Hadoop in VMware vSphere VMs.

Big Data: Hadoop Distributions Compared

me@ryanorban.com (Ryan Orban) — Sun, 17 Mar 2013 00:00:00 +0000

Summary

By early 2013, the commercial Hadoop ecosystem had consolidated around three main distributions: Cloudera CDH (Cloudera Distribution of Hadoop), Hortonworks HDP (Hortonworks Data Platform), and MapR. Each took a different approach to what enterprise Hadoop meant, and choosing between them was a real procurement decision for organizations investing in big data infrastructure.

Proprietary Hadoop Is a Losing Strategy

me@ryanorban.com (Ryan Orban) — Tue, 12 Mar 2013 00:00:00 +0000

Summary

The core argument in this 2013 ReadWrite piece: the Hadoop ecosystem was too large and too community-driven for any single vendor to win by locking customers into proprietary extensions. Any feature Cloudera or another vendor added that wasn’t upstreamed to Apache would eventually be replicated in the open-source project or competed away by rivals. Therefore, vendors should win on service quality, support, integration, and ease of use — not on proprietary code that created lock-in.

Hortonworks Joins OpenStack Foundation

me@ryanorban.com (Ryan Orban) — Tue, 29 Jan 2013 00:00:00 +0000

Summary

Hortonworks joined the OpenStack Foundation in January 2013, signaling the beginning of a more formal convergence between the Hadoop/big data world and the open-source cloud infrastructure world. Hortonworks was founded in 2011 by former Yahoo engineers who had built much of the original Hadoop codebase; the company commercialized Apache Hadoop and competed with Cloudera for enterprise big data customers.