Whats the best balance of data recovery and availability ?
If you have supported IT systems that are are used for internal processes and external customers you have been in the situation where managers are insisting to access their multiyear reports while accounting is complaining about vendor access to web forms. This is when planning for a scalable and robust application is important to address high-volume (OLTP) vs high-analysis (OLAP) requirements. There are several popular versions of database clustering and replication tools with products from Oracle, MySQL and SQL Server databases with the purpose of allowing organizations to segment their data based on business functions for data warehousing,reporting/business intelligence and sales transactions. To have the best of both worlds I have often used these replication technologies with source and destination concepts like Oracle RAC and Microsoft replicas or MySQL clusters. There can sometimes be duplicate transactions and a “sync gap” between the source and destination sites. I am writing on my experience on these 3 “enterprise” traditional relational model products as they are touted to offer handle CRUD (create, read, update, delete) transactions by isolating and versioning rows better.
Oracle tends to be used in financial, education and some lab/medical systems and therefore in my opinion has historically had more fine grained recovery and audit options. SQL Server has has many newer updates that are comparable with Oracle but has been more “web centric” for web-based applications tightly bound with .NET. I have personally have used Oracle for .NET applications, which is fine but from a licensing standpoint is probably uncommon.
MySQL is a good “hybrid” alternative because of its light weight processes and flexibilty with web-development platforms such as PHP on reliable linux. Using SQL coding for joins and functions across all three I have found is pretty similar and the commands are not that different. However at the administrative level they require different requirements and nannies to monitor them. I will get into the technical particulars in tech section of my blog, refreeds.com
If every data center was wallpapered in money weighing these factors would less relevant. Does anyone have experience with analytics oriented products like PostGres, Informatica and SAP on cloud or virtual platforms I’d like your comments or guidance at Quora too.