Google Cloud Spanner

While Cloud SQL provides a fully managed relational database experience, it does so by leveraging MySQL and PostgreSQL, making it subject to the limitations of these databases.

In the context of scalability, this means that Cloud SQL is limited to the upper bounds of these technologies - even more so as Cloud SQL makes no attempt to provide mechanisms for sharding. While technically possible through external tools such as ProxySQL, attempting to horizontally scale Cloud SQL essentially breaks the managed aspect of Cloud SQL.

As we discussed in the previous chapter, these limitations in scalability and ability to operate in a highly-distributed environment are largely what gave rise to NoSQL technologies. By sacrificing strong consistency guarantees (save for special cases that tend to carry poor performance), NoSQL solutions are able to scale far beyond traditional RDBMS solutions. Unfortunately, the NoSQL model simply doesn't work for many applications.

In 2005, Google...