Tachyon In-Memory Option Comes To BlueData Platform for Big Data Analytics
A technology preview of the Tachyon in-memory distributed storage system has been added as an option for the Big Data cloud offered by BlueData Software Inc.
Tachyon can cache working set files in RAM, garnering big performance gains through much faster file access for users of BlueData's EPIC platform.
Those throughput performance increases can be as much as 300x, BlueData said, giving developers another option to EPIC's existing integration support for Apache Spark, yet another fast alternative to the original Hadoop component MapReduce that can also leverage in-memory technology.
BlueData, like many others, is seeking to democratize Big Data by making its tooling and support frameworks more accessible while improving performance for high-speed stream processing at scale. That's something Hadoop -- with its batch-processing-oriented Hadoop Distributed File System (HDFS) -- wasn't originally designed for but which is becoming a growing part of Big Data implementations.
"Big Data is about the combination of speed and scale for analytics," the company quoted Enterprise Strategy Group analyst Nik Rouda as saying. "With the advent of the Internet of Things (IoT) and streaming data, Big Data is helping enterprises make more decisions in real time.
"Spark and Tachyon will be the next generation of building blocks for interactive and instantaneous processing and analytics, much like Hadoop MapReduce and disk-based HDFS were for batch processing," Rouda continued. "By incorporating a shared in-memory distributed storage system in a common platform that runs multiple clusters, BlueData streamlines the development of real-time analytics applications and services."
BlueData pointed out that while the Big Data ecosystem has grown and evolved -- being used for many different use cases such as high-performance trading, fraud detection, cybersecurity and system monitoring -- new solutions such as Spark and Tachyon have been developed for the new applications.
"However, incorporating these technologies with existing Big Data platforms like Hadoop requires point integrations on a cluster-by-cluster basis, which makes it manual and slow," BlueData said. "With this preview, BlueData is streamlining infrastructure by creating a unified platform that incorporates Tachyon. This allows users to focus on building real-time processing applications rather than manually cobbling together infrastructure components."
For the true gearheads, BlueData exec Tom Phelan yesterday went into great technical detail about how Tachyon works with the company's IOBoost technology in a blog post.
BlueData, headquartered in Mountain View, Calif., emerged from stealth mode last September with the EPIC platform for enterprises and received Spark certification from that project's steward, Databricks Inc., in October.
David Ramel is the editor of Visual Studio Magazine.