Using probabilistic data types

Presenter: Andre Chang, Solutions Architect, Aerospike

Operations on large data sets are often extremely resource intensive. But there are many situations where a precise answer is not required or even possible, especially in cases where active concurrent processes are updating the data. Operating within limited space and computational complexity, probabilistic data structures can be used to approximate set sizes and provide data on relationships between different sets. We will first explore some probabilistic data structures, then show examples of how to apply one such data type, HyperLogLog, to solve some common problems.