中國人壽it戰略規劃項目數據架構規劃-資料下載頁

【導讀】Day1(morningonly)?Day2. Lunch. Costeffective. orcollectively.change.orintroduced.canbeimplemented.managementsystems.datamappingtime.plexityandrisk.orcollectively.

　　

【正文】 an involve millions of data items with plex relationships. Its objective is to analyse these relationships and look for patterns, trends, and exceptions Oracle, Visual SQL Designer, Hyperion, BRIO, Crystal, Business Objects, Microstrategy, Essbase, Informatica etc Data Tools 28 Burning Timeline ? December 2020: No vendor support of Informix after then DBMS evaluation criteria ? Performance ? Cost ? Compatibility ? Management ? Skilled resource availability Timing relative to system replacements / upgrades ? Migration from Informix will require at least some data cleansing activities ? Migration from Informix to the new platform should be performed in conjunction with the migration to new applications. Data Architecture – DBMS Migration Considerations for DMBS migration: 29 Controls to ensure Data Quality going forward ? Business Business process control ? OLTP Input: Meta Data, Application Validations, Single Source of input ? ODS Reconciliation of inconsistencies ? Infrastructure Physical control, Infrastructure controlled Controls to ensure Data Quality for History Data ? Legacy Approaches – (When and Where) Data migration should be considered after/before new applications are integrated into the architecture Data quality improvement project should be considered “...productivity will result from focusing on quality, as quality does not e about by pursuing productivity.” Data Architecture – Data Quality 30 Contents – Centralization/Decentralization General Concept Requirements Data Architecture Centralization/Decentralization 31 Data Architecture Centralization or Decentralization Limited Provincial flexibility Performance challenges with scale Consistent data Excessive overall cost No metadata controls Limited ability to aggregate No synergy Functional IT leadership Enterprise perspective Scale economies Control of standards for core data Critical mass of skills Provincial ownership and flexibility Improved performance Responsive to diverse needs Pooled experience Synergy Centralized Model Decentralized Model Federated Model The Federated model emerges from the pros of both the Centralized and Decentralized mode, and provides the right amount of flexibility and the governance 32 Data Architecture Centralisation or Decentralisation Corporate Provincial Branch Individual Provincial Branch Branch Individual Individual Physical DSS Logical access to Provincial data Logical access to Branch data Physical DSS, ODS, and OLTP … … … … Province Groups ? Provinces are grouped into several groups ? The physical location of the groups is centralized ? Each group’s individual requirements are centrally supported by a dedicated IT team ? Groups may physically share infrastructure such as servers, storage… Under the Federated model, the degree of centralization can be adjusted to meet the needs of China Life Group Group Group Provincial Store personal data, logical access to all other data … … … …33 Qamp。A integration simplification standardization modularity + + +