ESG Reports Production Deployment Advantages of MapR Distribution for Hadoop

MapR Technologies, Inc. announced that IT research, analysis and strategy firm, Enterprise Strategy Group (ESG), published an ESG Lab Spotlight that documents the enterprise capabilities of the MapR Distribution.

The report titled MapR Distribution including Apache Hadoop discusses and validates the differentiating features of reliability, HA, NoSQL operations, performance, native NFS support and point-in-time snapshots available in the MapR Distribution.

“MapR continues to advance its Hadoop product offering to support the most demanding big data environments,” said Nik Rouda, senior analyst, ESG. “Enterprises with real-time, operational analytic needs will find that the MapR Distribution provides a high-performing distributed data platform which reliably and efficiently stores and processes volumes of data.“

Regarding reliability, the report concluded that the MapR architecture requires fewer resources while enabling higher availability when compared with other Hadoop distributions. Because data is spread across every node with MapR, the resync process to overcome a node failure occurs faster on MapR than on other Hadoop clusters with HDFS. The overall uptime of the cluster is improved by eliminating single points of failure and improving speed of recovery from minutes instead of hours.

The report also evaluated in Hadoop NoSQL operations. It validated the superior NoSQL database performance achievable with the MapR Distribution that includes an integrated database, MapR-DB, which is compatible with HBase applications. When comparing application performance, MapR-DB delivered better operations/sec/node for a wide-ranging set of workloads on both a ten-node HDD cluster and five-node SSD cluster.

ESG concluded that MapR-DB should be considered a top tier solution for NoSQL database online applications and large scale analytics.

In reviewing the performance, ESG reported that MapR-DB produced low, predictable latencies for every test, while the other distribution’s latency fluctuated wildly. With MapR-DB, average read latencies achieved a reduction as much as 6.7x with random reads. For the update latency across the batched put and mix workloads, average update latencies yielded an advantage between 3-15x when comparing different workloads across distributions.

MapR NFS and snapshots were also identified as key differentiators. ESG Lab witnessed MapR NFS in action, which revealed a much faster and more efficient implementation compared to other distributions. Additionally, the report noted the consistency of MapR snapshots, meaning that a MapR snapshot reflects the exact state of the cluster when it is taken, and unlike in other distributions, does not get altered due to subsequent changes to open files.

“The superior results MapR achieved from ESG Labs substantiates the deployment advantages of the MapR Distribution and is an important recognition,” said Jack Norris, CMO, MapR. “This independent analysis reinforces what we hear from customers worldwide that have experienced the difference for themselves by using MapR to address their production Hadoop and real-time requirements.”

To donwload the report (registration needed)