Software Provides Real-Time Monitoring and Control for Hadoop Clusters Running HBase, MapReduce, YARN, and Spark to Enable New Use Cases and Maximize Big Data Return on Investment
CUPERTINO, California—February 18, 2016 – Pepperdata, the world’s experts in the performance of distributed systems, announced today the certification of Pepperdata Supervisor 4 on Cloudera CDH 5.5 through the Cloudera Certified Technology Program. This latest certification adds support for Spark—in addition to HBase and MapReduce—allowing joint customers to expand use cases across multi-workload, multi-tenant cluster environments while controlling the chaos that often ensues from such complexity on the cluster. As large combinations of different users, applications, and business units compete for resources on the same cluster, they begin to interfere with each other, causing application slowness, downtime, and excessive efforts spent troubleshooting. Pepperdata manages that complexity through a set-it-and-forget-it approach that eliminates manual tuning for IT and guarantees Quality of Service (QoS) levels across the entire Hadoop environment.
Customers are increasing their use of Spark for new applications and more business-critical jobs. With Pepperdata, companies can monitor and control their Hadoop clusters in real time, reducing the complexity of managing the clusters. Pepperdata’s software makes multi-tenant cluster environments predictable and reliable through active monitoring that uncovers granular insights into capacity and usage, and then automatically adjusts the hardware usage of every job—faster than humanly possible—to make it run at peak optimization.
“Spark has rapidly gained support as a faster processing engine for certain applications, and organizations are inadvertently introducing more complexity into their distributed systems,” said Ed Colonna, Vice President of Marketing and Business Development at Pepperdata. “By utilizing Pepperdata on Cloudera, joint customers can unleash powerful insights that come from running all different types of workloads simultaneously, without the worry of non-critical job interference or application downtime, which helps them realize the true promise of their big data strategy.”
As more organizations seek to achieve interactive analytics through high-performance batch computation and real-time stream processing with Spark, predictable performance becomes nearly impossible and users run into resource contention between jobs. Pepperdata actively manages Spark by seamlessly allocating the necessary resources to the highest priority jobs, maximizing the performance of a single multi-tenant cluster to deliver required QoS levels.
“After ten years of growth, Hadoop adoption is entering a new phase closely linked to real-time data processing, making cluster performance a key metric; something Pepperdata is addressing proactively,” said Tim Stevens, vice president of Business and Corporate Development at Cloudera. “By using Pepperdata with Cloudera, users can feel confident that they will be successful in their implementation of Spark for a greater set of business-critical uses cases and in a larger number of production environments.”