Expected duration: 1 day or less We have different databases engines (source systems) and a data warehouse (Teradata). All of this are in an internal domain. We plan to have some public webservice (using existing Apigee & Microservices layer) to read from this data.
Hence, in this professional assessment presentation we need to show the detailed pros and cons for the following scenarios, including short term cost and long term cost:
1- All future public webservices to get its data directly from the data warehouse (EDW) 2- add new separate DB for each webservice which gets its data daily from EDW (regardless of the operational cost for DBs). Regardless of long period of EDW backup time during the day time! 3- having a BI cube for each webservice which is built on top of EDW, regardless of the performance and security issue as BI cubes will need to be connected to the public webservice. 4- introducing a new data delivery layer which gets its data replicated from EDW and source systems using HVR (CDC replication tool). This layer can be the contact points for public webservices and future applications. For illustration example, Singlestore Db can be used as a data delivery layer due to its unified distributed SQL engine and its capabilities to handle both transactional and analytical data.
Please not that all of our data are structured for now. Kindly use best practices suggestions and elaborations for this study. More approaches can be added to the above list.