If you’ve previously worked with databases like MySQL, Postgres, or SQL Server—especially in Scala—you may expect working with storage to involve many hassles and headaches, including these:
Indeed, in many real-world programs, storage becomes both a site of enormous complexity as well as a performance bottleneck for the whole system. That said, a well-configured database provides benefits as well:
Durability. Once a storage operation succeeds, we can rely on it being kept around until we say otherwise.
Transactionality. We can be guaranteed that a set of operations happen all at once, or don’t happen at all.
If we were so inclined, we could use a traditional database system with Scala Native. Just about every database under the sun, including Postgres and MySQL, as well as newer options like Redis, have excellent C client libraries we could use, just like we used libcurl for our asynchronous HTTP client in Chapter 7, Functions and Futures: Patterns for Distributed Services. But if you’ve read the rest of this book, you already know how to do that, and the bulk of the implementation will be tied to the patterns and idiosyncrasies of whichever database you choose, rather than anything particular or special about Scala Native.
Instead, in this chapter, we’ll explore a different strategy, one that’s familiar to C programmers but few others: embedded storage. And you’ll learn how to use a recent and very exciting embedded data store called the Lightning Memory-Mapped Database, or LMDB. Let’s first see what embedded storage is all about.