What do companies like Apple, Netflix, or Instagram have in common? First of all, those are global businesses with millions or even billions of customers around the world, spawning hundreds of billions of requests, downloading petabytes of data, and producing millions of database queries - every single second, worldwide. The second point is that all of the above are using Apache Cassandra™.
Traditional Leader-Follower-designed databases are great, but the strength of leader-based architecture quickly becomes an unbeatable flaw given the need to simultaneously process millions of QpS across multiple continents. Apache Cassandra is an open-source NoSQL natively distributed database known for scalability and highest availability without compromising performance, thanks to unique peer-to-peer leaderless architecture.
In this talk, we explore the strengths and weaknesses of decentralized data solutions, the production experience of the most renowned businesses, the best and worst use cases, and the typical pitfalls of companies starting to upgrade their data infrastructure with Cassandra.Key Takeaways:
- How global companies handle enormously high data pressure.
- What are the design principles behind the most efficient storage engines.
- What are the pros and cons of hyper-performant NoSQL solutions.
- When to consider, when to avoid this type of databases.