Mike’s opening of Strata + Hadoop World started with some interesting context. Strata + Hadoop World is now a global event. Since partnering with O’Reilly the event has gone global with sessions in Barcelona, San Jose, London, NYC and Asia. Next month the conference will be in Singapore. The spread in interest is really impressive.
Mike also provided context around how discussion of Hadoop is really disappearing, with the real importance of discussion on the applications on top of the platform. Most discussion today is around how applications on top of the platform solve business challenges. Mike underscored use cases from Digital Reasoning (focusing on use cases in the financial industry), eVariant (focused on use cases in the health care information industry), and CounterTack (focused on use cases in the cyber security domain). In each case, Mike underscored business impacts enabled by Hadoop that do not require deep discussion of the infrastructure, just a focus on the business/results/mission end.
Mike then dove into insights on the Hadoop platform itself, providing updates on Apache Spark (and the One Platform effort supported by the global Apache Spark community), Impala, Search, MapReduce, HBase, HDFS, Yarn, Kafka, Ibis (native Python), and Intel’s trusted analytics platform. Big Data is more valuable now because we can work with it in more ways.
Mike also made two other major announcements. In the security domain, tools like Apache Sentry have been joined by a new way to provide a single lens to let any user look at storage and get a consistent, secure, appropriate view for the user, giving access to the data the user should see. This new project is called Record Service. It enables a single data set to be shared with multiple users in a way that is safe and protects data from unauthorized access.
The video also articulates the very interesting new storage project called Kudu. Kudu deserves a much deeper examination and we will be writing more on it in the future. For now I’ll just underscore that it is designed to provide features and benefits of both HDFS and HBase.