Building A Real Time Trading System Using In Memory Data Grid And D3.Js

Abstract

The result of this thesis is an answer to the real time trade monitoring use case with the implementation of in-memory data grid, then compare the result with a version that implements a traditional database. To complete this, a copy of the data flow of real world systems will be used during the testing period on two versions of the system as mentioned, one with Hazelcast in-memory data grid, and one with PostgreSQL (Relational Database Management System). The results collected from the testing period on thesis systems will be used to evaluate performance and conclusion. A simple user interface will also be designed, in order to display the aggregated data. This design aims to serve the aggregated data as near real time as possible. After these steps, the result of testing on two systems will be compared, to testify my claim. The test result prove that Hazelcast IMDG is outperform PostgreSQL in the same configuration of testing resources. In conclusion, the thesis result is provide a complete small-scale implementation of a trade monitoring system with Hazelcast IMDG can perform many times better than RDBMS version, in use case of big data or heavy load processing.