Ryft, a big bata outfit that has products which offer ‘actionable intelligence from complex data’ has rolled out a new analytics platform that claims to be 200X faster than conventional hardware, in analyzing historical and streaming data simultaneously.

Dubbed, Ryft ONE, the new offering promises to provide actionable business insights by simultaneously analyzing up to 48 terabytes of historical and streaming data at an 10 gigabytes/second or faster. It claims to be the only commercial 1U platform capable of doing so.

“The loosely coupled, physically distributed compute-and-memory nodes of today’s high-end clusters turn daunting data analytics problems into IO bottlenecks that can slow solution times. The Ryft ONE platform aims to supercharge performance on challenging analytics workloads with a scalable 1U device designed for easy integration into existing server environments,” explains Steve Conway, IDC research vice president for high-performance computing and data analysis.

“The Ryft ONE is also designed to accelerate both batch and streaming data — a dual requirement more and more user organizations face today,” he added.

Some features of the Ryft ONE :

  • Ryft Analytics Cortex (RAC): Large-scale parallel, hardware-accelerated compute architecture that speeds historical and streaming data analytics, providing intelligence in real time at speeds of 10 gigabytes/second and faster.
  • Ryft Algorithm Primitives (RAP) Library: Expanding collection of prebuilt algorithm components such as exact search, term frequency and fuzzy search that accelerate the development and execution of applications.
  • Linux Front-end and Open API: Compliant with open standards to work with a wide range of visualization, scheduling, performance monitoring and systems management tools. Supports popular programming languages such as C/C++, Java, R, Python, Scala and others.
  • High Performance With High Security: Analyzes SSL encrypted data without added latency to protect sensitive data without sacrificing performance. Data can be stored either encrypted or unencrypted, with no impact to analytics performance.

It has been under development for years now, the company says. The offering uses less power than a hair dryer and can store and “analyze the equivalent of the contents of Wikipedia in 4.5 seconds, without any data indexing, preprocessing, tuning or partitioning.”

According to benchmark tests conducted against the highest-performing in-memory solutions in early 2015, a single Ryft ONE bettered large Spark clusters of 100 to 200 nodes, at a 70 percent operational savings.

(Image credit: Ryft)

Previous post

An Introduction to Recommendation Engines

Next post

Ex-VP of FINRA Appointed Managing Director at Swift