Migrating a NorthWind Database to a NoSQL Database

By Basit Anwer

It's now a fact that databases need to have the ability to scale-out to handle data that's produced and accessed at great velocity and in huge volumes. Traditional databases (RDBMS) were never meant to handle such high a flow of information. To keep up, you either incorporate a distributed cache like NCache to overcome the database bottlenecks, or replace the RDBMS with a NoSQL database. NosDB is one such NoSQL database and is made from the ground up in .NET. It easily handles peak loads, and scales out at runtime — and therefore increases throughput on an ‘as-needed’ basis.

It's clear at this point that considering a NoSQL solution will free you of RDBMS bottlenecks. But what's less clear is how to properly migrate from a relational database to NoSQL. NoSQL is just so easy to use that developers often dive in head-on without properly understanding and constructing data models. With the infinite flexibility and features that a NoSQL database provides, it is very easy to mess up your data during migration, which causes a lot of problems later on.

Planning will help to properly migrate your existing data from an RDBMS to a NoSQL database without hurting your existing business processes. To completely demonstrate it, we are going to use Microsoft's sample application, NorthWind, as an example. To begin, let’s understand some of the basic business operations of this fictitious company, ‘NorthWind Traders.’

Why understand business operations? Well, with a traditional RDBMS, you are locked into database-driven development. But with a NoSQL database, your development is always domain-driven, giving you complete control over how you want your data to be stored. Therefore, the easiest approach is to build the database schema around your business use case and your application.