Difference between revisions of "IS485 IS Project Experience (Research)- Fintech project examples"
Alanmegargel (talk | contribs) |
Alanmegargel (talk | contribs) |
||
Line 30: | Line 30: | ||
'''Research Question 2:''' What are the performance characteristics (eg; response time, replication time) of an in-memory-data-grid as compare to a relational database, as invoked across microservice instances? <br> | '''Research Question 2:''' What are the performance characteristics (eg; response time, replication time) of an in-memory-data-grid as compare to a relational database, as invoked across microservice instances? <br> | ||
'''Research Question 3:''' How is transaction management done using an in-memory-data-grid as compared to a relational database, in order to avoid concurrency issues across microservice instances? | '''Research Question 3:''' How is transaction management done using an in-memory-data-grid as compared to a relational database, in order to avoid concurrency issues across microservice instances? | ||
− | |width="30%"|<!-- Research Method--> | + | |width="30%"|<!-- Research Method--> '''Research Method:''' Develop a microservice (banking context) which requires transaction management. |
+ | Configure a microservice management tool to replicate instances of the microservice. Load test instances of the microservice across heavy loads, to generate data. Do the load testing for; a) in-memory-data-grid, and b) relational database. Analyse the data generated from load testing. Results and Conclusions.<br> | ||
+ | '''Integrated Report:''' Write up an integrated report covering all 3 research questions. Relate each question within a banking context.<br> | ||
+ | '''Integrated Demo:''' Demonstrate a fully functional microservice for a banking context, using the techniques and tools discovered. Run the demo using; a) in-memory-data-grid, and b) relational database. | ||
|- | |- | ||
|} | |} |
Revision as of 16:24, 3 June 2019
Project deliverable:
- A research report covering 2-3 research questions
- A working prototype
- Presentation slides
Research Problems:
- Approach the faculty advisor
- Project selection should be mutually agreed
Item | Research Problem | Research Questions | Research Method |
---|---|---|---|
1 | Microservices Architecture in Banking - In recent years, there is much written about microservices architecture. Even so, there is very little written about how microservices are used in the banking industry. I study is needed to understand microservices implementation in banking. Faculty Advisor: Alan Megargel |
Research Question 1: How do banks decide on the boundary context and granularity of microservices? Research Question 2: What kind of tool sets do banks use for building, testing, and deploying microservices? |
Research Method: Data collection through surveys and interviews of bank staff, and existing case studies. Analysis of qualitative and quantitative data collected. Results and Conclusions. Integrated Report: Write up an integrated report covering all 3 research questions. |
2 | Data Replication Across Microservice Instances - In a microservices architecture, as the demand (# of invocations) for a microservice increases, management tools elastically create replicate instances of the microservice. A study is needed to understand mechanisms for replicating the data underlying multiple instances of a microservice. Faculty Advisor: Alan Megargel |
Research Question 1: Can an in-memory-data-grid be used instead of a relational database, underlying replicated instances of a microservice? What is the feasibility of each, in a banking context? Research Question 2: What are the performance characteristics (eg; response time, replication time) of an in-memory-data-grid as compare to a relational database, as invoked across microservice instances? |
Research Method: Develop a microservice (banking context) which requires transaction management.
Configure a microservice management tool to replicate instances of the microservice. Load test instances of the microservice across heavy loads, to generate data. Do the load testing for; a) in-memory-data-grid, and b) relational database. Analyse the data generated from load testing. Results and Conclusions. |