September 2014 Summaries

DataStax Enterprise (DSE) version 4.7 now officially supports Apache Spark™ MLlib integration, allowing users to perform in-memory analytics utilizing integrated Apache Spark™. This update enables powerful real-time analytics capabilities and makes starting a Spark cluster simple. The article demonstrates how to use the Naive Bayes algorithm with Spark and Cassandra to build a classifier for the Iris flower data set, showcasing fundamental techniques that can be used at scale.

The text discusses the improvements made in Apache Cassandra's sstableloader tool, which is used for bulk loading data into a Cassandra cluster. It highlights that while usage of sstableloader has not changed much, it now requires specifying the address(es) of the node using the -d option. Additionally, the text introduces CQLSSTableWriter, a new feature in Apache Cassandra 2.1.0, which simplifies bulk loading data to complex schemas by allowing users to use DDL for table creation and INSERT statements for data insertion. Despite some limitations, such as lack of parallel processing and support for user-defined types, the tool has been improved over past releases, making it easier for users to experiment with generating SSTables and bulk loading data.

The text discusses the importance of performance testing in Cassandra and highlights the challenges faced when doing so manually. It introduces cstar_perf, a tool designed for running performance tests on Cassandra clusters, making it easier to automate testing and manage variables. Open-sourcing this tool is expected to improve the quality of Cassandra releases by encouraging more frequent and earlier testing during development. The documentation for cstar_perf can be found here, with users encouraged to report issues or contribute improvements through GitHub.

The DataStax Node.js driver for Apache Cassandra has been released in its first beta version. Built on top of a previous driver, it offers high-level features similar to other DataStax drivers for Cassandra. It is designed to be tunable through tuning policies and connection options. The public API shares many names with the Java/C# counterparts, and some internal classes have the same behavior as their counterparts. Development will continue on this project while node-cassandra-cql will be discontinued. Exciting new features are planned, and users can join the driver mailing list or follow DataStax developers on Twitter for updates.

DataStax OpsCenter provides a browser-based user interface for managing and monitoring DataStax Enterprise and Cassandra clusters. It also has a full API that can be accessed using Python, allowing developers to perform various operations such as administering users, configuring clusters, nodes, and OpsCenter itself, performing operations on clusters and nodes, managing alerts, best practice rules, backup schedules, accessing performance metrics data, and more. The API is flexible and can be used in any language capable of communicating over HTTP. However, it's crucial to understand the specifics of your needs and test thoroughly before using it in production.