Jubatus: Scalable Distributed Computing Framework for Realtime Analysis

Technological fields
Information Sharing Platform Technologies
Keyword
  • Big data
  • Distributed computing
  • Machine learning
Laboratory organization
NTT Information Sharing Platform Laboratories

Download PDF (1,111KB)

Overview

Jubatus* is a fast and scalable distributed realtime analysis framework for big data. It improves availability and reduces communication overhead by using parallel data processing and by loosely sharing intermediate results with component servers. Jubatus is Open Source Software.

Features

  • Realtime processing: on-demand processing without data accumulation
  • High scalability: scales linearly with the number of servers
  • Profound analysis: able to deploy sophisticated algorithms such as machine learning
  • Pluggable architecture: easy to plug in new analysis engines and data storage

Application scenarios

  • Trend analysis of SNS and blog data for marketing
  • Anomaly detection and resource allocation from using sensors and network traffic measurements
  • Making realtime recommendations based on user activities
  • Making market and stock predictions from financial data
  • * Jubatus Website: http://jubat.us/
figure