Azul Launches Consulting Services to Help Enterprises Optimize Apache Cassandra Deployments

SUNNYVALE, CALIFORNIA--(Marketwired - July 29, 2015) -

News highlights:

• New consulting services identifies and addresses Java-related barriers in Cassandra deployments
• Customized services target improvements to Cassandra response time, reliability, up-time, and throughput
• Azul field consultants can diagnose and propose remedies for runtime issues that impact Apache Cassandra or supporting components such as Lucene, Solr and Spark

Azul Systems (Azul), the award-winning leader in Java runtime solutions, today announced the availability of Cassandra HealthCheck, a new business consulting service designed to help enterprises eliminate Java-related barriers and improve the runtime consistency, reliability, and up-time of their Cassandra deployments. As an example of the unique capabilities of this new service, Azul recently achieved over 56x improvement in worst-case latency, 25% improvement in throughput, and dramatic reduction in client disconnects due to node timeouts for a large-scale Cassandra deployment with a customer in the financial services industry.

Cassandra HealthCheck (Cassandra HC) consists of a combination of onsite and remotely-delivered runtime analytic services. Leveraging both open source and proprietary tools, Azul consultants have deep Java experience helping IT organizations and DevOps teams optimize Java runtimes and system and application configurations to significantly improve Cassandra deployments, including components leveraging Cassandra data stores such as Apache Lucene, Solr and Spark.

Apache Cassandra is a highly performant, scalable open source distributed database management system written in Java. However, many aspects of Cassandra performance, consistency, and availability depend on the capabilities of the underlying Java runtime platform. Legacy Java Virtual Machines (JVMs) may suffice for some use cases, but enterprises that require highly consistent and reliable low-latency Cassandra deployments are often limited by the JVM itself, specifically the detrimental impact of garbage collection. The Java garbage collection process freezes the application while memory is defragmented and compacted, resulting in random Cassandra pauses that lead to response time inconsistency, increased time to data consistency, and can even trigger serious conditions like cascading node failures, where multiple stalled Cassandra nodes trigger cluster-level outages and crashes.