You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 17 Next »

Bluehole (https://www.bluehole.net)

  • We are the developer of 3D MMORPG, TERA.
  • We are using Tajo in TERA client log analytics system. Tajo made our analysis works simple through its direct JSON data format support. For more information on our Tajo adoption story, take a look at our slide : http://www.slideshare.net/zenos2408/aws-tajo

Database Lab., Korea University (http://dbserver.korea.ac.kr)

  • Tajo was originally developed at Database Lab. Korea Univ.
  • They have 32 Tajo cluster nodes.

Encored Technologies (http://encoredtech.com)

  • The company provides real-time and cumulative electric power for residential and business usage, via NILM (Non-Intrusive Load Monitoring).

  • A Tajo cluster is running with 14 nodes, and the system is increasingly expanding.
  • The cluster is constructed on AWS EC2, and on-premise environment; it is becoming more dependent on Tajo.
  • To archive the purpose of research and data analysis for per-device usage, more than one hundred million records are processed by Tajo on a daily basis.

Gruter (http://gruter.com)

  • Gruter is a Hadoop-based infrastructure company which builds big data platforms and provides technical services for the data-driven enterprise market.
  • We provide a social network analysis service.
  • We have about 32 cluster nodes.
  • We use Tajo for ETL and ad-hoc queries on collected social network data sets.
  • We process hundreds of giga bytes per day.

HYUNDAI U&I (http://www.hyundai-uni.com/eng/index.jsp)

  • We use 20 Tajo cluster nodes.
  • We process 4TB (400,000,000 events) per day.

LINEWALKS Inc. (http://www.linewalks.com)

  • LINEWALKS provides visual data discovery services for advanced analysis of large-scale data sets of various fields including Medicine and Finance.
  • We use Tajo to analyze medical data sets provided by  HIRA (Health Insurance Review and Assessment service) .
  • We use a Tajo cluster of 10 nodes on HDFS for performing ETL and complex data analysis tasks.

Locket (http://www.getlocket.com/)

  • They use 10 nodes on Amazon EC2.

LOEN Entertainment (http://www.iloen.com)

  • We provides the online music service called MelOn.
  • We have about 50 Tajo cluster nodes on HDFS.
  • Usually, data analysts use Tajo for ad-hoc queries.
  • It is roughly 1.5x - 3x faster than old Hive, but Tajo only uses some portion of cluster resources.
  • We will adopt Tajo to our batch jobs.

SK telecom (http://sktelecom.com/en/)

  • SK telecom is the largest mobile carrier in South Korea.
  • In overall, Tajo is used to analysis logs collected from our cellular network.
  • We replaced some commercial data warehouse system by Tajo.
  • We use Tajo for ETL and OLAP workloads.
  • We use 2 nodes for Tajo Master and 51 nodes for Tajo workers.

  • For only ETL, we process over 120 SQL queries that process over 4TB per day.

  • We also submit more than 500 different queries through some integrated OLAP tool per day.
  • No labels