Engagement at a Glance
Pactera provides enterprise SmartData professional services and solutions that bring visibility to a treasure trove of data already flowing through the organization. We help clients determine where they should focus to get the most value from SmartData and develop a solution unique to the business. Pactera’s strategies and technologies help clients capture structured and unstructured data at the source and transform it into the kind of insights that drives market opportunities and revenue growth.
The world’s leading online travel company and operates localized websites for travelers in the US, Canada, France, Germany, Italy, Denmark, Austria, Belgium, Ireland, The Netherlands, Norway, Spain, Sweden, UK, Australia, New Zealand, Japan, India, China (through a controlling investment in eLong) and Singapore.
- Structured and non-structured Big Data.
- Requires unified design of ETL, data warehouse, and reporting across all brands to perform global enterprise data integration.
- Lacking well-defined business rules to automate manual processes.
- The SLA time was a must requirement from performance point of view.
- Users were accessing legacy data warehouse every day, requiring that these features get migrated to the new data warehouse smoothly.
- New features and change requests were coming non-stop while redesign and implementation work was on-going.
- Use DB2 as enterprise data warehouse and Hadoop (CDH4) as enterprise big data platform.
- Use the various of ETL tool, such as Informatica, Hive, MapReduce, etc.
- Use Business Object (BO), Hue, Tableau as reporting tool.
- Redesign of new data warehouse tables sourcing from Hadoop.
- Migrate unstructured data to Hadoop.
- Restructured and simplified the ADS tables and type I, II and III slow changing dimension (SCD) to avoid data issues caused by complex.
- Redesign job on the key SLA path.
- Scale : 120 man-year, started in 2009, ongoing.
- Completed 40+ data marts, 800+ ETL tables, and 500+ reports.
- More than 100+ new features were also launched along with the EDW.
- Migrate/Load 6PB data to Hadoop.
- The SLA of EDW reports rending time was reduced by 2 hours and loaded twice the data volume.
- Savings of a minimum of $3M every year for EDW after using Hadoop.
- Set up deep partnership with different departments, especially finance department.