Pulling Data between Servers

Data movement revolves around the E (Extract), T (Transform), and L (Load) in ETL. In a typical implementation, you will deploy the database (either P6 EPPM or Primavera Unifier and data warehouse on separate physical servers. Because of this, you must maximize the bandwidth and minimize latency between servers. Ideally, the servers reside in the same data center with gigabit Ethernet connecting the servers. (Oracle performance tests servers in a central data center with gigabit connections.) You should verify throughput of server communication by performing basic file copy or FTP operations between the servers.

The data movement process is based on the standard SQL, either with INSERT-SELECT or CREATE TABLE AS (CTAS) syntax using an Oracle database link. While the Oracle RDBMS is efficient at moving data through the link, the overall performance depends on the physical network connection.



Legal Notices | Your Privacy Rights
Copyright © 1999, 2020

Last Published Monday, December 14, 2020