Now to cover some data modeling basics that apply no matter whether your OLTP database is on premises or in the cloud. I do not want to overprovision to be safe I want to right-size this from day one based upon a database structure that does not waste resources due to inefficient design.įigure 1: Amazon Cloud cost calculator Data modeling basics If I can cut down on the CPU and memory needs, I can cut this cost significantly. So that’s $136,000 per year just to run this one data warehouse in the cloud. I chose “OnDemand” for up to 64 virtual CPUs and 448 GB of memory, since this data warehouse wanted to leverage in-memory processing. I was pricing a data warehousing project with just 4 TB of data – small by today’s standards. So you really should get familiar with your cloud provider’s sizing versus cost calculator. Static overprovisioning or dynamic scaling will run up monthly cloud costs very quickly on a bad design. When you make poor database design choices for OLTP applications deployed to the cloud, your company will pay every month for the resulting inefficiencies. The cloud offers infinitely scalable resources – but, at a cost. Data modeling helps you right-size cloud migrations for cost savings No matter where you implement it, you must fully document the business requirements in order to produce an effective final result. I’m not proposing that business logic be held in the database, but it needs to be documented in the data model even if it will be implemented in application code. One must also capture the vast quantity of metadata around the OLTP business requirements that must be reflected. However, the diagram is merely the starting point for an effective and efficient database design. Other types of databases will be covered in subsequent blogs.ĭata modeling is a serious scientific method with many rules and best practices. This blog outlines a solid foundation for data modeling of online transaction processing (OLTP) systems as they move to the cloud. This makes mastering basic data modeling techniques and avoiding common pitfalls imperative. For standard relational database applications, data modeling that incorporates accepted design paradigms, such as normalization, is essential. By Bert Scalzo How to create a solid foundation for data modeling of OLTP systemsĪs you undertake a cloud database migration, a best practice is to perform data modeling as the foundation for well-designed OLTP databases.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |