Levi Masonde stands up a Redis instance:
Databases play a vital role in software applications—they need to keep updated data or state, which is served by the database acting as the source of truth for the application. How the database performs affects how the entire application performs. Besides obvious factors that affect the performance of a database (hardware, database type, networking infrastructure), there are techniques designed to help you improve the performance of your database and ultimately, your applications. One way to do this is to add caching to your database. But which cache technique works best for which application requirements? This article sheds light on one strategy to implement a cache system.
Click through for one pattern of interaction between cache and database. My preference with the cache-aside pattern is to hide the two data platforms from the calling application as much as possible. In a classic object-oriented language like C#, the actual database + cache calls would be in a separate project and would expose methods on classes that were database-agnostic. With Python, I’d use different .py files in the same project unless I wanted to build a wheel file and deploy it to multiple projects, but the concept would still be the same. I’m not the biggest fan of the way that Levi did it, forcing API developers to have knowledge of both the cache and the database, as that increases the risk of a future developer messing something up.