Distributed SQL Databases Deconstructed
SQL is a popular database language for modern applications, given its flexibility in modelling workloads and how widely it is understood by developers. However, most modern applications running in the clouds require fault tolerance, the ability to scale out and geographic data distribution of data. These are hard to achieve with traditional SQL databases, which is paving the way for distributed SQL databases. Google Spanner is arguably the world's first truly distributed SQL database. Given its fully decentralized architecture, it delivers higher performance and availability for geo-distributed SQL workloads than other specialized transactional databases such as Amazon Aurora and Google Percolator. Now, there are a number of open source derivatives of Google Spanner such as YugaByte DB, CockroachDB and TiDB. This talk will focus on the common architectural paradigms that these databases are built on (using YugaByte DB as an example). Learn about the concepts these databases leverage, how to evaluate if these will meet your needs and the questions to ask to differentiate amongst these databases.