Sensor & Log Data from Airplanes

What is the business need?

  • Client provides paid Airplane Internet services in both domestic and international routes, for many airlines
  • Current Oracle based design is taking nearly 18 hr. to build custom reports

Sensor & Log Data from Airplanes

Challenges

  • Current SLA reporting is based on Oracle and client is not able to meet SLA reports
  • Current design will not scale out, and scale up requires high investment
  • Current design will not scale out, and scale up requires high investment

Our Solution

  • Used Hadoop HDFS to store the files and MapReduce jobs to process the data. Used Hive, SQOOP and Pig as ETL
  • Hadoop’s exceptional parallel processing using MapReduce framework reduced processing time

Result

  • Reduced Average Claim Cost
  • Reduced processing cost to 10 cents per part
  • Reduced process time from 10 hrs. to 2 hrs. with minimal investment
  • Eliminated Datastage
  • Increased Customer Satisfaction Index
  • More accurate RPS inventory