Advanced FTI Customizations Deepshikhar Tyagi OTM & Business Analytics Consultant Infosys
Agenda Introduction FTI Features FTI As Product General Pain Points Customized FTI Advantages FTI Customization Methodology OTM Design & Customization : Best Practice ODI Customization & Best Practice DB Customization & Best Practice BI Customization & Best Practice 3
Introduction The presentation shares experiences and best practices drawn from various FTI implementation across different projects. 4
FTI Features FTI offers interactive dashboards and reports that offer insight into transportation operations. The application contains prebuilt, role-based dashboards that help organizations to quickly analyze key performance indicators, reports, and metrics to spot trends and determine if business is on track. Quick visual status of performance such as: freight cost per mode, cost per mile, trailer utilization and on-time arrival rates Drives Continuous Improvement 5
6 FTI Architecture Re-Visited
FTI As Product Fusion Transportation Intelligence is a product integrated with OTM : Not mere a module within OTM It has it s own product life cycle, which should be in sync with OTM Resources with strong technical background in OTM, ODI & OBIEE are required 7
General Pain Points Integrating new OTM/FTI Business objects Not everything in OTM is present in FTI Missing Data Heavy Technical Intervention Direct Database Requests Lesser Flexibility Bulky Reports Performance Issues 8
End State (customized FTI) : Advantages & Opportunities Minimum intervention by technical team for creating reports Automated FTI Data purge Framework for Health Check Frequent OTM- FTI Data Reconciliation, Server health check, FTI Data Object Management 9
FTI Customization Methodology Business Requirement Assesment ETL Development Data Mining Infrastructure & Architecture Meta Repository Design Meta Repository Development Project Planning ODI Design Implementation Roll Out & Monitoring Project Requirement Definition FTI Data-Target Design Unit Test User Acceptance Technical Analysis OTM Data-source Design Integration Test Performance Test 10
OTM Design: Best Practice Report Team Involvement : Active participation of reporting team using FTI while designing OTM solution takes care of many issues at first hand. Impact Analysis of OTM Workflows: Table level impact in OTM should be analyzed to check whether pertinent mappings in FTI are getting impacted 11
OTM Customization : Best Practice Refnums : Ideally design should have between max 25-30 refnums per OTM object for high data load. More refnums hit down the performance severely! Flex Fields: Flex fields column available in OTM should be judiciously used. Additional Columns & Tables: Additional column to existing mapped tables is preferred. Going next step additional tables should be kept at minimal level for efficient maintenance. 12
ODI Customizations Managing ETL is the key part of FTI We not only can control what & how data goes inside FTI from OTM but also can check and constantly monitor data reconciliation between two systems 13
ODI Customizations: Best Practice Check Knowledge Module (CKM) : CKM is in charge of checking that records of a data set are consistent with defined constraints. CKM Customizations lead to effective tracking of error during and ETL loads and Improved quality of data load to FTI 14
ODI Customizations: Best Practice For higher data loads customizing user functions can improve the performance and can reduce errors to a great extent. 15
FTI DB Customizations Purging Strategy : Defining purging strategy is important part of FTI DB Management. Usually data for past couple of years is kept for historical analysis. There is no standard purging solution available yet but customized scripts can be written. As a best practice FTI shipment table should be center of the purging solution. 16
Management of Materialized Views: Best Practice MVs are helpful to bring in the FTI objects which are not mapped to Business Intelligence. Customized MVs refresh should be scheduled on time intervals as existing OOB MVs are refreshed via ETL and at instances it is seen that it often results in deadlock 17
BI Customization: Best Practice RPD Customization: Different business objects should not be grouped together inside. Whereas customized MVs having related objects are advised to be mapped together in BMM layer Right usage of System & BI Cache Eliminate Bulky Direct Database Requests Not only affects performance but creates unnecessary clutter at BI Server Optimum number of connections should be defined for a connection pool. As a best practice anywhere between 150-200 connections are kept for good performance 18
FTI References http://docs.oracle.com/cd/e38437_01/otm/ html/docset.html 19
20 Thank You