August 2014 Summaries

The Spark Cassandra Connector has been moved to open-source on GitHub, with an additional Java-API module developed to meet user demand for Java compatibility alongside its primary use in Scala. This blog post explains how to use the new API to access Apache Cassandra via Apache Spark from Java applications, specifically targeting the 1.0.0-rc4 version of the connector compatible with Cassandra 2.0 and Spark 0.9. The example application described demonstrates the process of reading sales data, computing roll-up summaries, and displaying results, while utilizing the connector to create schema, generate random data, and save data to Cassandra using methods from the CassandraJavaUtil class. The tutorial highlights key features of the Spark Cassandra Connector, such as reading and writing RDDs to and from Cassandra, and serves as an introductory guide to integrating Cassandra with Spark applications. Additional resources, including full source code and documentation, are provided for further exploration.

The latest version of the Java driver has been released, compatible with Cassandra 1.2, 2.0, and now supporting Cassandra 2.1. It introduces a new object mapping API that simplifies conversion of query results to custom Java classes. Additionally, it includes support for User Defined Types (UDTs) and tuples in Cassandra 2.1. The driver's perspective on UDT values is as map-like objects with getters and setters. Tuples are essentially anonymous UDTs with unnamed fields and predefined types. To specify target tables and UDTs, Java classes can be decorated with annotations. The object mapper generates boilerplate code for CRUD operations and complex queries. It is available from Maven repositories and as a packaged binary.

At Datastax, interns Daniel Chin and Dolphin built a 32-node DataStax Enterprise cluster running on Raspberry Pi's to demonstrate the fault tolerance of Cassandra. The display features a Big Red Button in their lobby that simulates taking down an entire data center. They designed the hardware using Solidworks, sourced materials from McMaster Carr, and used ethernet cables for network connectivity. Each Raspberry Pi runs at factory settings with vnodes enabled to manage resources efficiently. The cluster is set up as a two-datacenter DSE 4.5 cluster with OpsCenter 5.0 running to show the status of all nodes. Despite its limited hardware capabilities, this unique demonstration aims to excite people about Cassandra's always on capabilities.

Apache Cassandra's release 3.0 introduces User Defined Functions (UDF), allowing users to write code that runs within Cassandra daemons. UDFs are stateless and can be implemented in various languages, including Java and scripting languages with JSR-223 support. They can be used for functional indexes or aggregation. Future enhancements may include additional language support and improved user permissions.

DataStax has released the latest version of its DataStax Sandbox, version 2.0, which features DataStax Enterprise 4.5, OpsCenter 5.0, and DevCenter 1.1. This new release includes a guided tutorial with an updated analytics demo based on Apache Spark™. Users are encouraged to download the software and explore its capabilities.