Slides from my talk at the NYC Cassandra meetup on 4/29/15.
Abstract:
You might think that knowledge of how your database performs writes, reads, and other internal operations is arcane knowledge for ops witches, but no! This knowledge can give you tools to dramatically improve the performance of your application. For example, at Venmo, we halved response times on our slowest endpoint by better understanding the implications of our read and write patterns and tuning our Cassandra cluster accordingly.
In this talk, we'll clarify what is meant by "in Cassandra, writes are fast, but reads are slow." We'll cover how a node performs writes, reads, and compaction, and we'll share some lessons learned about choosing the right compaction strategy for your query patterns.
16. Cassandra reads
• read row(s) from sstables
• get updates to row(s) from memtable
• reduce any multiple values for columns
• respond to coordinator