Unisys Corporation New · NYSE:UIS
Unisys Corporation New — Unisys — ClearPath Future Matters, Birmingham 2019
16 slides from Unisys Corporation New's conference presentation, Sep 11, 2019
Slide 1: Making ClearPath Forward relevant in the digital age
Making ClearPath Forward® relevant in the digital age Rudy Everaert, Enterprise Architect, ING Belgium Introduced by David Howard, Unisys
Slide 2: Unisys Data Exchange
Unisys Data Exchange Digital Applications and Big Data ClearPath Forward Applications Data Exchange Extract, Transform and Load SQL Server / Oracle (copy of the data) KAFKA – enabling publishing ClearPath Forward of ‘events’ Databases The changes happening to DMS II / RDMS / (DMS coming) the CPF dat
Slide 3: Making ClearPath Forward relevant in the digital age
Making ClearPath Forward® relevant in the digital age ClearPath Future Matters Birmingham, September 11, 2019
Slide 4: One ING experience built on one global foundation
One ING experience built on one global foundation 2
Slide 5: We have a plan to unite ING...
We have a plan to unite ING... thinkforward > unite @ Unite ING @) Unite foundations Unite businesses @) Unite customer experience ... and offer one differentiating customer experience G ING 2
Slide 6: MCP Unisys eco-system at ING Belgium
MCP Unisys eco-system at ING Belgium Most of the ING Belgium core banking applications are running on ClearPath platform Our challenge Business Applications International Payment Account Mgt Client Referential Unlock the value of the MCP business-critical Domestic Mortgage applications in the digita
Slide 7: API oriented and event driven architecture enables heritage applications to be relevant
API oriented and event driven architecture enables heritage applications to be relevant in the realization of the thinkforward > unite strategy Enables independent development of Mobile and Web app requests are personalized user experiences and decoupled from enterprise applications customized B2B o
Slide 8: Making ClearPath Forward relevant in the digital age
Making ClearPath Forward® relevant in the digital age Virtualized Steaming Data Mainframe MCP ETL masking Offloading Standardization (ING Private cloud) Near real time Availability & compliance Adapt to DevOps culture integration responsiveness 6
Slide 9: MCP virtualization
MCP virtualization Shared mainframe, Any application Dedicated mainframe, Tribe application clustering ClearPath® MCP Libra ClearPath® MCP Software Standardization (ING Private cloud) Align and support DevOps culture 7
Slide 10: Continuous near real time integration
Continuous near real time integration Streaming ETL with change data transfer mode Interested bounded context Shared model • a domain event contains all data which might be needed by all other bounded contexts • Interested bounded contexts are tolerant readers ignoring all data that is not relevant
Slide 11: Mainframe offloading
Mainframe offloading Client Streaming ETL with change data transfer mode commands queries API Solve the caching consistency data freshness problem with continuous streaming read Service store ▪ Improve Customer Experience DMS II * domain event C ▪ Increase responsiveness store D C local ▪ Increase a
Slide 12: Data masking
Data masking Bulk data transfer mode to mask all data Service * proactively protect ING’s data to avoid security breaches and C ensure compliance. DMS II D system C of record MCP Static data masking creating a structurally similar but inauthentic version Production of the data that can be used for p
Slide 13: DataExchange 5.0 ic2 – performance
DataExchange 5.0 ic2 – performance DataExchange 5.0 ic2 Mode Bulk Data transfer Change Data Transfer DMS II Complex Simple Complex Simple Complex withwork- table around Record mutations 1500 15.000 7.500 1500 750 per second with 2 shards 11
Slide 14: Observations with DataExchange in change data transfer (CDT) mode
Observations with DataExchange in change data transfer (CDT) mode Complexity of DMS II table Single-threading Grouping of record mutations to speed up publishing to Kafka - 50% RMPS* Maximum 1500 RMPS* no atomic events in CDT mode 12 * RMPS : record mutations per second
Slide 15: Observations with DataExchange in change data transfer (CDT) mode
Observations with DataExchange in change data transfer (CDT) mode Performance issues (700 MPS) because of single- Ic2 work-around allows to scale-out threaded implementation on complex tables DMS II excellent expected performance good 5000 acceptable performance … CDC* 3500 average fair 2500 With wo
Slide 16: Key Takeaways
Key Takeaways ➢ Keep the domain event stores simple. Complex DMS II tables have an high impact on performance ➢ DataExchange 5.0 – ic2 work-around allows the improve Kafka provider performance Though complex and at a high cost. Just a set of disconnected DataExchange instances ➢ No atomic event. Gro